Reactor Shielding For Nuclear Engineers

Available as TID25951 for $28.
00 from
National Technical I nformation Service
U. S. Department of Commerce
Springfield, Virginia 22161

International Standard Book Number 0870790048
Library of Congress Catalog Card Number 73-600001
Distribution Category UC80
May 1973; latest printing, September 1981
Preface
Shielding technology has matured considerably in the last decade, and shield
physics must routinely be translated into shield design. Since the publication
in 1959 of Fundamental Aspects of Reactor Shielding, by Herbert Goldstein,
new generations of computers have become available to exploit techniques
heretofore considered too costly, and new measurement techniques have
been devised. The energy and angular distributions of neutrons and gamma
rays can be followed, both in theory and in practice, throughout their
transport histories. Such powerful tools have brought correspondingly large
dividends to the shielding community.
These advances and their underlying fundamentals are recorded in this
volume, which is intended as a text for a two-semester course in reactor
shielding directed at an advanced undergraduate or graduate level. The reader

is assumed to have some familiarity with calculus through partial differential
equations and with nuclear physics through particle interaction theory,
although pertinent aspects of the latter are reviewed in Chap. 3. The material
is arranged to cover fundamental transport considerations in the first
semester; portions of Chap. 4 could be reserved for the second semester. The
second semester could then consist of special topics, such as Monte Carlo
techniques, albedos, ducts, shield-analysis projects, seminars on experimental
shielding, and shield design. Instructors will doubtless follow plans of their
own choosing. Chapters 2 through 6 have problems appended, with solutions
given at the back of the book. Metric units have been used exclusively.
Citations of classified literature have been avoided, and technical reports
have been referenced only where no journal articles could be given.
Although titled Reactor Shielding, this text should be applicable in related
areas where neutron and gamma-ray attenuation are important, as in nuclear
iii
iv
PREFACE
weapons shielding and in isotope source applicati<.Hls. The study of space

radiation and high-energy-accelerator shielding, although closely related to
the present subject, has been considered outside the scope of this book.
Dr. Samuel Glasstone originally conceived the idea for this text; he
concluded that the book was needed and originally proposed to prepare it.
In the preliminary planning of the project, the U. S. Atomic Energy
Commission asked me to collaborate with Dr. Glasstone. Notwithstanding
many plans and discussions for this collaboration, Dr. Glasstone had to
relinquish his role in order to carry out a number of other projects. It is a
pleasure to acknowledge his efforts in the planning of this book and his
useful critiques of early drafts. I sincerely regret that our proposed
association could not be continued.
For their assistance in the preparation of this manuscript, I am greatly
indebted to many people in a number of ways. First, no book on shielding
could be readied for publication at this time without acknowledgment of the
pervasive influence of one man, the late E. P. Blizard. Not the least of his
many contributions to the development of the technology was his
encouragement of the efforts of others, including my own effort in preparing
this manuscript.
The many services and suggestions provided by the staff of the Radiation
Shielding Information Center, Oak Ridge National Laboratory, were
extremely helpful, particularly in scanning the current literature. It is a
distinct pleasure to acknowledge many useful discussions with others at
ORNL: Lorraine Abbott, Clyde Claiborne, Charles Clifford, Paul Stevens,
and Dave Trubey, each of whom supplied references and data in addition to
contributions cited elsewhere. My colleagues Mike Wells and Bob French
have also contributed in this way and in their forbearance.
lowe thanks for reviews and comments on various portions of the
manuscript to Arthur Chilton and his students at the University of Illinois,
Don Dudziak of Los Alamos Scientific Laboratory, Charles Eisenhauer of
National Bureau of Standards, Cliff Horton of Rolls Royce, Ltd.,
Richard Faw of Kansas State University, Norman Francis, David Mesh, and
their associates of General Electric Knolls Atomic Power Laboratory,
Gene Hungerford of Purdue University, John Lamarsh of New York University, Fred Maienschein of Oak Ridge National Laboratory, Ed Profio of
University of California at Santa Barbara, and Leigh Secrest of Texas
Christian University. I am particularly indebted to Lew Spencer of National
Bureau of Standards for his detailed review of the complete manuscript and
PREFACE
for his many useful suggestions. Most of these reviewers provided recommendations based on teaching experience in shielding.
The guidance and counsel of John Inglima during the planning stages and
of Robert Pigeon during the manuscript drafting, both of the U. S. Atomic
Energy Commission, is gratefully acknowledged. For technical editing I am
grateful to Jean Smith and Marian Fox, also of the U. S. Atomic Energy
Commission, and, for typing a difficult manuscript, to Monsita Quave of
Radiation Research Associates, Inc. I am especially grateful to Ceil Schaeffer
for relieving me of many burdensome proofing tasks and, most of all, for her
understanding and encouragement.
N. M. Schaeffer
May 1973
Note to Reader
The occasion of the 1981 reprinting gIVes me the opportunity to
comment on the currency of this text, now eight years after initial
publication. Several observations should be made: (1) The methodology of
shield analysis as described here continues in general use. Many computer
programs are now available which are successors to those mentioned in the
text; the newer versions tend to be more-efficient or more-convenient
versions of the forebears that were originally described. (2) Many of the
tabulated data originally included for the reader's convenience have been
superseded by newer results. We urge the user to check for later references
with the Radiation Shielding Information Center, Oak Ridge National
Laboratory, if the accuracy of his results is important or if he is in doubt
about the latest available computer programs. (3) The shield designs
reviewed in Chap. 10 are obviously dated but should retain some classroom
value for discussion of the principal design considerations.
The general acceptance and use of this text has been particularly
gratifying, and I thank the numerous teachers and students who have offered
suggestions and corrections, which have been incorporated where possible.
N. M. Schaeffer
August 1981
Contributors
H. C. daiborne
Oak Ridge National Laboratory. Oak Ridge, Tennessee
S. T. Friedman
Consultant, Los Angeles, California
c. W. Garrell
Radiation Research Associates, Inc., Fort Worth, Texas
(Now with National Academy of Engineering, Washington, D. C.)
L. G. Mooney
N. M. Schaeffer
W. E. Selph
(Now with Intelcom Radiation Technology, San Diego, California)
P. N. Stevens
Oak Ridge National Laboratory, Oak Ridge, Tennessee, and University of Tennessee,
Knoxville, Tennessee
D. K. Trubey
Oak Ridge National Laboratory, Oak Ridge, Tennessee
vi
Contents
PREFACE . . .
. iii
CONTRIBUTORS
. vi
1 HISTORICAL BACKGROUND
11
2 SOURCES OF RADIATION
2.1 Gamma-Ray and Neutron Sources
12
12
2.1.1 Gamma-Ray Sources

. . . .
2.1.2 Neutron Sources
.....
2.2 Basic Mathematical and Physical Concepts

2.2.1
2.2.2
2.2.3
2.2.4
Differential Distributions .
Average and Most-Probable Values
Solid Angle
. . . .
Measures of Radiation Intensity
2.3 Spatial and Directional Characteristics

2.3.1
2.3.2
Spatial Distributions
Directional Distributions
2.4 Energy Distributions
17
19
19
24
24
26
39
39
42
.....
48
2.4.1 Energy Distributions of Gamma-Ray Sources

2.4.2 Neutron Spectra from Fission
2.4.3 Effect of Medium on Spectra
49
54
57
References
59
Exercises
60
3 INTERACTIONS OF RADIATION WITH MATTER

3.1
Cross Sections
63
63
64
65
66
67
. . . . . .
3.1.1 Microscopic Cross Section

3.1.2 Macroscopic Cross Section
3.1.3 Radiation Reaction Rates
3.2 Radiation Interactions .

3.2.1 Photon Interactions
3.2.2 Neutron Reactions
68
83
3.3 Responses to Radiation
94
95
3.3.1
3.3.2
3.3.3
3.3.4
Absorbed Dose
First-Collision Dose and Kerma
Exposure . . . . . .
RBE Dose; Dose Equivalent
99
103
105
vii
CONTENTS
3.3.5 Maximum Absorbed Dose; Maximum Dose Equivalent
3.3.6 Multicollision Dose
108
112
References
113
Exercises .
114
4 RADIATION TRANSPORT
4.1 Fundamental Considerations
4.2 The Boltzmann Transport Equation
4.3 Spherical Harmonics Method
4.4 Discrete-Ordinates Sn Method
4.4.1
4.4.2
4.4.3
4.4.4
Transport Equation and PhaseSpace Geometry

Derivation of Finite-Difference Equation
Numerical Solution of the Discrete-Ordinates Equation
Advantages and Disadvantages
4.5 Moments Method

.....
4.6 Application of Diffusion Theory
4.7 Invariant Imbedding Method

4.8 Kernel Technique
4.8.1 Gamma-Ray Calculations
4.8.2 Neutron Techniques
4.9 Combination Removal-Diffusion Methods

4.9.1 The Spinney Method . . . .
4.9.2 Variations of the Spinney Method
4.9.3 Differences in Current Methods
119
120
123
129
132
134
136
144
148
149
160
163
168
169
178
188
190
192
199
References
201
Exercises
204
5 MONTE CARLO METHODS FOR RADIATION TRANSPORT

5.1 Sampling from Probability Distribution Functions
5.2 The Evaluation of Integrals . . . . .
5.3 Source Parameters
. . . . .
5.3.1
5.3.2
5.3.3
5.3.4
Selection from an Energy Distribution

Selection of Spatial Point of the Source Particle
Selection of Initial Direction of Source Particle
Source-Biasing Parameters
5.4 Path Length . . . . . . . . .

5.5 Collision Parameters
.....
5.6 Particle Parameters After Collision
5.6.1
5.6.2
5.6.3
5.6.4
5.6.5
5.7
5.8
5.9
5.10
Neutron Elastic Scattering

Neutron Inelastic Scattering
Compton Scattering
Particle Absorptions
Calculation of Emergent-Direction Cosines
Particle Scoring
.
Statistical Variance
. . . .
Demonstration Monte Carlo Program
Programming Suggestions
207
209
216
218
218
219
221
223
225
234
236
236
238
240
240
241
242
247
251
254
CONTENTS
ix
References
257
Exercises
258
6 SIllELD ATTENUATION CALCULATIONS

6.1 Analysis of the Source . . . .
6.2 Direct Solutions . . . . . .
6.3 Application of Parametric Data
6.3.1
6.3.2
6.3.3
6.3.4
Moments-Method Differential Energy Spectra

Monte Carlo
Measured Data
Fitted-Parameter Data
6.4 Simplified Solutions

6.4.1
6.4.2
6.4.3
6.4.4
Applications of Gamma-Ray Buildup Factors

Applications of Neutron~Removal-TheoryKernels
Other Point-Kernel Applications
.
Methods for Estimating Low-Energy Neutron-Flux Density
6.5 Application of Kernel Technique to Calculations of

Secondary Gamma-Ray Dose
6.5.1 Calculation for Slab Shield
6.5.2 Calculation for Semi-Infinite Shield
261
261
263
264
265
270
274
277
283
284
286
288
298
301
304
308
References
310
Exercises
311
7 ALBEDOS, DUCTS, AND VOIDS

7.1 Introduction to Albedos
7.2 Definitions . . . . .
7.2.1 Differential-Dose Albedos
7.2.2 Total-Dose Albedos
7.2.3 Other Albedos
7.3 Neutron Albedos

7.3.1 Fast-Neutron Albedos
7.3.2 Intermediate-Neutron Albedos
7.3.3 Thermal-Neutron Albedos. .
7.4
7.5
7.6
7.7
7.8
Gamma-Ray Albedos , . . .
Secondary-Gamma-Ray Albedos
Applications of Albedos
Ducts
.
Line-of-Sight Component
7.8.1
7.8.2
7.8.3
7.8.4
Rectangular Ducts
Rectangular Slots
Cylindrical Ducts
Cylindrical Annulus
7.9 WallPenetration Component

7.9.1 Application to Cylindrical Ducts
7.9.2 Application to Partially Penetrating Cylindrical Ducts
7.9.3 Comparison with Experiment
7.10 Wall-Scattered Component

7.10.1
Analog Monte Carlo Calculations
313
313
315
316
317
318
318
319
331
332
341
350
355
356
357
360
363
365
366
367
370
372
374
375
376
CONTENTS
7.10.2
7.10.3
7.11
Albedo Methods
Additional Experimental Invcstigations
380
398
404
404
412
Voids
7.11.1 Single Voids

7.11.2 Small Random Voids
References
414
8 SHIELD HEATING, AIR TRANSPORT, SHIELD MATERIALS,

AND SHIELD OPTIMIZATION
8.1 Shield Heating . . .
8.1.] Gamma-Ray Heating
8.1.2 Neutron Heating . .
8.1.3 Chargcd-Partide Heating
8.2 Air Transport
430
8.2.1 Infinite Air Medium. .

8.2.2 Air-Over-Ground Calculations
8.2.3 Air-Ground Interface Effects
8.3 Shield Materials
419
419
420
426
429
.....
432
437
439
8.4 Shield Optimization
443
445
447
450
456
462
References
465
8.3.1
8.3.2
8.3.3
8.3.4
Considerations in Materials Selection

Shield Materials for Stationary Reactor Systems
Shield Materials for Mobile Reactor Systems
Comparison of Attenuation Properties
9 EXPERIMENTAL SHIELDING
9.1. Detectors for Shielding Experiments
9.1.1 Active Neutron Detectors
9.1.2 Passive Neutron Detectors
9.1.3 Active Gamma-Ray Detectors
9.1.4 Passive Gamma-Ray Detectors
9,1.5 Interpretations of Detector Output.
9.2 Shield-Material Measurements

9.2.1 Reactors
9,2.2 Accelerators
9.2.3 Fixed Sources
9.3 Phenomenological Measurements

9.3.1
9.3.2
Air-Transport and Air-Ground Interface Experiments

Duct Penetration.
References
10 SHIELD DESIGN . . . . . . . .
10.1 Iterations in the Shield Design .
10.1.1 Preliminary Conceptual Design
10.1.2 Detailed Conceptual Design.
10.1.3 Final Engineering Design.
10.2 Fast Breeder: Enrico Fermi

10.2.1 The Reactor Plant. .
10.2.2 Shield Design Criteria
471
472
472
474
475
476
476
477
477
493
503
507
507
514
516
519
520
521
522
524
527
528
532
CONTENTS
10.2.3
10.2.4
10.2.5
10.2.6
Reactor-Shield Systems .
Shield Costs . .
Calculational Techniques
Comparison of Measurements and Calculations
10.3 Fast Breeder: Dounreay Fast Reactor

10.3.1 Calculational Model for Bulk Shield
10.3.2 Measurements . . . . . . .
10.3.3 Effect of Streaming . . . . .
lOA
Heavy Water, Natural Uranium: Agesta

10.4.1 Description of Reactor and Shield
10.4.2 Calculational Model
1004.3 Measurements
10.5 Boiling Water: Pathfinder

10.5.1 Calculations . . .
10.5.2 Survey Measurements
10.6 Ship Propulsion: N.S. Savannah

10.6.1
10.6.2
10.6.3
1 0.6.4
10.6.5
10.6.6
10.6.7
Description of Ship, Reactor, and Main Shielding

Shielding for Refueling and Control-Rod Maintenance
Shield Design Criteria
.
Lead~PolyethylcneShield Construction
Attenuation Calculations
. . . . .
Measurements . . . . . . . .
Comparison of Measurements and Calculations
10.7 Space Powe" SNAP-lOA Flight Test
xi
534
548
548
553
559
562
563
565
566
566
567
570
574
578
579
579
580
581
583
585
587
591
592
10.7.1 Shield Analysis. .

10.7.2 Flight-Test Results.
595
595
598
References . . . . . . . .
602
Appendix A:
GAMMA RAYS FROM INELASTIC NEUTRON SCATTERING

AND FISSION . . . . . . . . . . . . . . . . . . . . . 605
Appendix B: NEUTRON FLUENCETOKERMA CONVERSION

FACTORS FOR STANDARD-MAN MODEL
611
Appendix C:
616
CYLINDRICAL TISSUE PHANTOM
Appendix D:
COORDINATE SYSTEMS. VECTOR OPERATIONS,

AND LEGENDRE POLYNOMIALS . .
D.l Coordinate Systems . . . . . .
D.2 Coordinate-System Transformation
D.3 Vector Operators and Functions
D.4 Dirac Delta Function
....
D.S Legendre Polynomials . . . .
D.6 Associated Legendre Polynomials
D.7 Associated Spherical-Harmonic Functions
Appendix E: EXPOSURE BUILDUP FACTORS
....
617
617
617
619
621
622
623
625
626
Appendix F:
628
COEFFICIENTS FOR GAMMA-RAY BUILDUP FACTORS
Appendix G: GRAPHS AND FORMULAS OF EXPONENTIAL

INTEGRAL FUNCTIONS . . . . . . . . . . . . . . . . . 642
xii
CONTENTS
Appendix H: TABLES OF ATTENUATION FUNCTIONS

FOR FINITE SLAB GEOMETRY
....
653
Appendix I: RANDOMNUMBER GENERATORS

. . . .
1.1 Properties of RandomNumber Generators (RNG's)
1.2 Recursion Equations for RNG's
1.3 Testing RNG's
663
664
665
668
669
1.3.1 Global Tests . .

1.3.2 Equidistribution
1.3.3 Independence
669
670
].4 Pathological Numbers

1.5 Useful RNG's
671
672
Appendix J: DEMONSTRATION MONTE CARLO PROGRAM
674
Appendix K: MOMENTSMETHOD RESULTS FOR FISSIONNEUTRON

PENETRATION IN Be, C, CH, CH 2 , H, AND H 2 0 .
687
Appendix L: GAMMARAY DIFFERENTIAL ENERGY SPECTRA

FOR WATER AND LEAD
.
. .
692
Appendix M: GRAPHS FOR NEUTRON ATTENUATION

CALCULATIONS
.....
700
Appendix N: GRAPHS OF THE if; FUNCTION
717
Appendix 0: CONSTANTS FOR EMPIRICAL EXPRESSIONS

OF ALBEDO DATA
726
Appendix P: RADIATION PENETRATION OF CYLINDRICAL DUCTS
734
SOLUTIONS TO EXERCISES
737
AUTHOR INDEX
771
SUBJECT INDEX
775
Historical Background
N. M. SCHAEFFER
Early reactor shields were largely a matter of educated guesses. The complex
of phenomena that had to be considered fo)' an accurate shield analysis was
an imposing obstacle. Microscopic-particle interaction processes were reasonably well understood, but their relative importance depended on largely
unknown physical parameters called cross sections. Bulk attenuation
properties of materials for two principal radiations of interest, neutrons and
gamma rays, were also unknown. Even for an empirical approach, there was
no opportunity under the wartime pressures of the Manhattan Project to
launch a systematic investigation of the attenuation properties of materials.
It was obvious that hydrogenous materials were needed for neutrons and
dense materials for gamma rays. It was also evident that simple exponential
attenuation based on the total cross section was a thoroughly inadequate
concept for determining layer thicknesses. The shield of concrete and
paraffinized wood for the Argonne National Laboratory graphite pile in
1943 was adequate for gamma rays and was overdesigned for neutrons. The
X-I0 reactor at Oak Ridge National Laboratory (ORNL) included a 2.1-m
concrete shield, of which the central 1.5 m contained a special mixture
incorporating the mineral haydite. The large water-of-crystallization content
of haydite made it appear especially useful for neutron attenuation. This
shield was also overdesigned for neutrons and about adequate for gamma
rays, although streaming problems were evident for both radiations around
access holes in the shield. t
The special requirement for a thin shield for the Hanford reactor was
dictated in 1944 by the maximum length of aluminum tubing that could be
tHistorical material for this chapter has been drawn from H. Goldstein, Everitt Pinnel Blizard,
1916-1966, Nuclear Science and Engineering, 27: 145 (1967), the dedication of a special issue
prepared as a memorial to E. P. Blizard. Additional information was graciously provided by Mrs. L. S.
Abbott from the archives of the Neutron Physics Division, Oak Ridge National Laboratory. Mr. C. C.
Horton of Rolls Royce, Ltd., has kindly provided reminiscences of British developments.
REACTOR SHIELDING FOR NUCLEAR ENGINEERS
drawn. E. Fermi and W. Zinn had made some provisional attenuation

measurements in Chicago in 1943. H. Newson and L. Slotin made some
gold-foil measurements for masonite and iron slabs in the core hole at the
rear of the X-10 pile in 1944. A young engineer, C. E. Clifford, was assigned
to help them. The Hanford reactor shield was built of iron slabs sandwiched
between masonite layers. Although initially a good neutron attenuator, the
masonite suffered severe radiation damage and decomposed. The CP-3
(Chicago Pile-3, 1944) shield was composed of ordinary concrete; although
thicker than necessary, it performed satisfactorily.
The early reactor projects clearly demonstrated that the design of a
shield for neutrons and gamma rays that was optimal, efficient, or
economical required answers to a great many questions. In 1946 the Navy
initiated an intensive study program for a nuclear-powered submarine, and
the Air Force, a similar study for a nuclear-powered aircraft. Space and
weight limitations for these nuclear applications added more impetus to the
open questions in shielding. In the spring of 1947, E. P. Blizard, then a Navy
physicist assigned by Capt. H. Rickover to ORNL, was directed to start a
program of shielding measurements. He proposed a program of neutron and
gamma-ray attenuation measurements through several types of concrete
placed in the rear core hole (a 60-cm square aperture) of the X-10 reactor.
C. E. Clifford of the laboratory ,staff was assigned to work with him because
of his experience with measurements for the Hanford shield in 1944. Slabs
of material were placed in the aperture, and detectors were positioned within
and beyond the slabs. This effort marked the first organized research
program in reactor shielding. A spiral-duct mock-u p placed in the hole
demonstrated that proper! y designed passages could penetrate the shield
without transmitting excessive radiation. These studies also led to the ./
recognition that the production of se~ry g'!!!lJTIa r~ by neutron
interactions in the shield was clearly a significant design consideration.
By 1948 shielding studies supporting various reactor projects were in
progress at Hanford, Knolls Atomic Power Laboratory, Bettis Atomic Power
Laboratory, and Massachusetts Institute of Technology (MIT). As additional
results of measurements in the X-10 core hole were made, Blizard became
convinced that too much radiation streamed around the test samples for
accurate measurements and a better facility was needed. He concluded that a
fission plate-a thin disk of enriched uranium covering the core holewould provide a local source of fission neutrons and would be more
accessible for attenuation measurements. Clifford suggested that a tank of
water be placed adjacent to the fission plate so that materials and
HISTORICAL BACKGROUND
instruments could be submerged, which would greatly reduce the radiation

background. These two ideas culminated in the Lid Tank Shielding Facility,
which began operating in 1949.
In the United Kingdom shielding research efforts were started in 1948
and were geared to the British philosophy or reactor development: large
gas-cooled reactors for plutonium production to be rollowed by development of these systems for electricity generation. The research reactor BEPO
had just been completed; it had a 15-cm iron thermal shield followed by a
bulk shield of barytes concrete and had a layout similar to the Oak Ridge
X-10 reactor. The Windscale reactors were under construction in 1948 and
included a thermal shield similar to BEPO but Portland concrete was used
rather than barytes. Eatly design calculations were made by B. T. Price, D. ].
Littler, and F. W. Fenning.
A shielding group was set up under C. C. Horton as part or Fenning's
reactor physics group at Harwell to investigate shielding problems connected
with large concrete shiclds l heating effects, and radiation streaming in the
large ducts that arc integral to gas-cooled systems. In these systems heat
generation in the first 30 em or so of the shields was recognized to be an
important problem, and Hortoll, later with K. Spinney, developed some
models to predict the distribution of heat generation by neutrons and
gamma rays. Horton,
J.
R. Harrison, and D. Halliday of the Harwell group
also initiated a program of duct-streaming measurements in 1952 at the
BEPO facility.
During an intensive working session at ORNL in shielding in the summer
of 1949 with interested participants from a number or installations (one of

many organized by Blizard), T. A. Welton of MIT developed the concept of
the removal cross section for treating neutron attenuation in heavy materials
mixed with hydrogenous materials. Recognizing the importance of the
removal concept, Blizard initiated a new series of measurements in the Lid
Tank to verify applicability and to obtain removal cross sections for many
materials. The removal-cross-section concept quickly came into widespread
use and became the principal method of treating neutron attenuation. Two
decades later it is still regarded as a useful, valid technique for many

applications. So great were the demands on the Lid Tank that a second
fission-plate facility was constructed on the reactor at Brookhaven National
Laboratory, and a program of additional removal-cross-section measurements
was carried out under the direction ofR. Shamberger (Chap. 4).
Blizard proposed an additional test apparatus for complete 41T shields
since they could not be tested in the Lid Tank. Tests for the mock-up for the
Materials Testing Reactor (MTR) indicated that this type of reactor would
make a useful source for shicld tests. Construction was authorized, and the
Bulk Shielding Reactor (BSR) was completed in 1950. The facility was so
versatile that it becam;"the pattern for swimming-pool research reactors
around the world. The BSR group included L. Meem, F. Maienschein, and
R. Peelle. Numerous basic and applied results were forthcoming on materials,
shield mock-ups, and a definitive measurement of the fission gamma-ray
spectrum (Chap. 2).
The British workers also realized the need for a special facility; they
required data to support the design of large shields for power reactors. A
group under the direction of Fenning was set up to design and build this
reactor. Horton was responsible for the physics and general layout of the
facility. The reactor (LIDO) was completed in 1956. Unlike the Oak Ridge
facility, the entire pool was constructed above the ground to allow access to
three caves in the shield wall, in which substantial dry mock-ups could be
placed. The reactor could be traversed through the pool to provide a source
for these mock-ups, and an important design criterion for the pool layout
was that construction of a mock-up in one cave could be carried out while
experiments were continuing in another.
Aircraft shielding required measurements away from the ground; thus
Blizard and Clifford conceived the idea in 1952 of a facility in which a
reactor might be suspended at a sufficient height to eliminate the effects of
ground scattering. They planned an arrangement of four towers in a
rectangle with cable hoists for elevating a BSR-type reactor and crew
compartment 60 m above ground. The Oak Ridge Tower Shielding Facility
began operation under Clifford's direction in 1954, and it proved versatile in
applications far beyond the ill-fated nuclear aircraft program (Chap. 9).
Although destined for cancellation in 1961, the aircraft nuclear
propulsion (ANP) program produced a number of other useful shielding
efforts. The Nuclear Aerospace Research Facility at Convair, Fort Worth,
Tex., included two reactors: the Ground Test Reactor (GTR), a copy of the
BSR, and the Aircraft Shield Test Reactor (ASTR). In 1954 B. Leonard and
N. Schaeffer proposed a program of ground and flight studies with these
reactors to resolve the major shielding uncertainties affecting airframe
design. The GTR was operated in a small water tank suspended from a crane
at a height of 30 m to obtain an early measurement of ground scattering. It
was also placed in a mock-up consisting of the empty fuselage of a retired
aircraft (the XB-36) with a shielded cylinder representing a crew compartment. From these measurements and concurrent air-transport results at the
Oak Ridge towers, the large contribution of secondary gamma rays produced
by neutron radiative capture in air was first observed in 1955. The
importance of these secondary gamma rays was a surprise to both groups;
previous estimates of the probability for gamma-ray production by neutron
capture in nitrogen had been too low, and these measurements were the first
to reveal the discrepancy. The ASTR was carried in the aft bomb bay of a
specially modified B-36 in a series of test flights from 1955 to 1957 at
altitudes from sea level to 11 km. The program provided data on radiation
transfer by air and aircraft structure from reactor to shielded crew
compartment. The program culminated with a joint effort at ORNL in which
the ASTR and the crew compartment were suspended at the towers in the
same relative positions as when installed in the B-36 (Chap. 8).
The decade from 1951 to 1961 is the period when shield technology
came into its own. The major facilities were all in operation from 1954
onward, and large shielding groups at General Electric in Cincinnati, Ohio,
Pratt and Whitney in Hartford, Conn., Convair in Fort Worth, Tex., and
Lockheed in Marietta, Ga., were participating in the ANP program. The
submarine effort was concentrated at the Westinghouse Bettis Laboratory
near Pittsburgh, Pa., and the General Electric Knolls Atomic Power
Laboratory in Schenectady, N. Y. The Oak Ridge group was extremely busy
supporting both efforts. These groups contributed to the technology by
developing design methods, by measuring attenuation through shield
materials (including mock-ups of various shield designs), and by devising new
experimental and analytical approaches. The demise of the ANP program
and the successes of the nuclear submarine are well known. The U.S.S.
Nautilus sailed on nuclear power for the first time in January 1955. This
date is to be compared with 1954, 1956, and 1957, the years in which
nuclear-fueled electric plants first went on line in Russia, Great Britain, and
the United States, respectively.
The nuclear-applications programs gave impetus to the development of
shield-analysis methods as well as to large-scale experimental programs. By
the early 1950s an intensive program in radiation physics was under way at
the National Bureau of Standards (NBS) under the direction of U. Fiillo.
G. W. Gro~n published a definitive set of X-ray attenuation coefficients,
and L. V. SP~'s meth~nts solution of the Boltzmann tr~ns
port equa~ was first described. Shortlyillerward a group at Nuclear
Development Associates, Inc., under the direction of H. Goldstein joined
with Spencer and Fano in an intensive program of moments-method
calculations, which culminated in 1954 with publication of the Goldstein
and Wilkins report on gamma-ray buildup factors. R. Aronson, J. Certaine,

M. Kalas, and P. Mittelman, with Goldstein, applied the method to neutrons.
Fano, Spencer, and M. J. Berger published a definitive exposition of
gamma-ray penetration in 1959, which included a summary of the moments
method as well as other techniques (Chap. 4).
Work on neutron attenuation in the United Kingdom followed a
somewhat different path from the efforts in the United States. Horton and
J. D. Jones devised the removal-diffusion method, the first results of which
were repotted at the second Atoms for Peace conference at Geneva in 1958.
Since 1956, A. Avery and J. Butler have further developed these techniques
at Harwell (Chap. 4).
Although H. Kahn of Rand Corporation published two papers on the
application of Monte Carlo techniques to shielding in 1950, in which he
identified virtually all the principal concepts, widespread use of the
technique and its subsequent development had to await the improvement of
the digital computer. Prominent among early contributors in delineating the
techniques and concepts were E. Cashwell and C. Everett of Los Alamos
Scientific Laboratory (LASL) and G. Goertzel and M. Kalas of Nuclear
Development Associates, Inc. Kalas and F. Clark of ORNL reported on the
theory of importance sampling and finite variance estimators. At NBS,
E. Hayward and J. Hubbell reported photon albedo calculations in 1953;
M. Berger and J. Doggett extended their results in 1955. The first successful
Monte Carlo applications in air scattering were reported in 1957-1958 by
Berger, C. Zerby of ORNL, and M. Wells of Convair. The completion of the
0 5 ~ of Monte Carlo programs by R. Coveyou at ORNL in 1958
must be regarded as~igfiificant advance in shield technology. The 05R
system required a great deal of its users, but it was extremely flexible and
widely used. At the Geneva (Atoms for Peace) conference in 1964, Blizard
and Mittelman reported on eight major Monte Carlo programs in use in the
United States. The MORSE MQ..nte Carlo code of E. Straker, P. Stevens,
D. Irving, and V. Cain ";as comPleted in 1969; it has produced results in
excellent agreement with analytic solutions (Chap. 5).
Monte Carlo has been regarded as one of the sophisticated techniques,
but the workhorse method of shield design has been the point-kernel
approach. Blizard, J. Miller, D. Trubey, and G. Chapman of ORNL and
J. MacDonald, W. Edwards, and J. Moteff of General Electric (GE) made
notable contributions to the use of removal cross sections. K. Shure of
Westinghouse Electric Corporation developed an analysis tech~ for
metal-hydrogenous shields based on a combination of point kernel and a
numerical method using spherical harmonics called a PI multigroup (later a

) solution of the one-dimensional transport equation. In the development
and application of gamma-ray buildup factors to kernel techniques, the work
of J. Taylor of Westinghouse, M. Capo of GE, M. Berger and J. Hubbell of
NBS, R. French of Convair, M. Grotenhuis of Argonne, F. Clark and
D. Trubey of ORNL, and A. Chilton of the University of Illinois should be
listed as principals in devising empirical representations of the data and
simplified schemes for its application (Chaps. 4, 6, and 8).
From the outset many investigators in shield analysis sought manageable
numerical techniques for achieving analytical solutions of the Boltzmann
equation. Of all the efforts in this direction, such as the method of moments,
spherical harmonics, numerical integration, and invariant imbedding, perhaps
the most significant in terms of present usage is the discrete-ordinates
method. B. Carlson of LASL had developed a discrete-ordinate method
for reactivity calculations in 1955 which became known as Sn and which has
been successfully applied to a variety of transport problems. F. Mynatt and
W. Engle of ORNL developed ANISN in 1965, which incorporated improved
differencing and convergence techniques and made the method more suitable
for shielding applications. A two-dimensional version of ANISN called DOT
was described a year later by F. Mynatt, F. Muckenthaler, and P. Stevens
(Chap. 4).
The Sn programs, although not without problems in some geometries,
have been used with a great deal of success in obtaining detailed radiation
distributions in complicated two-dimensional geometries. Several laboratories have recently studied the utility and applicability of coupling Monte
Carlo and discrete-ordinate calculational links. Thus the latter is used for
those portions of a geometry reducible to two dimensions and the former
where the description requires three dimensions.
A more complete historical survey would include the developments and
researchers in nuclear instrumentation for shielding. As shield analysis has
been paced by the development of the digital computer, so shielding
experimentation has been gaited to innovations in particle detectors and fast
electronics. A survey of neutron and gamma-ray detectors is given in
Chap. 9.
In the foregoing chronology we have been limited to an outline of
United States shielding research with only brief insertions of corresponding
British activities. Significant and occasionally large shielding efforts have also
been maintained elsewhere, notably, Belgium, Canada, France, Italy, Japan,
the Netherlands, Norway, Russia, Sweden, and West Germany. These
P3
....
programs principally support national power reactor developments, although

the literature also contains many reports of maritime and space reactor
shielding studies from Europe and Asia.
We have used some terms that will be meaningless to the uninitiated.
However, the chapter references will aid the curious in locating the
appropriate explanations; the objective here has been to trace the early
developments and to introduce some of the literature.
The newcomer will find the following earlier books on this subject to be
useful references: The first handbook in reactor shielding was published in
1956 and was edited by T. Rockwelj,1 who had been in the original shielding
group at Oak Ridge. B. Price, C. Horton, and K. Spi nney 2 of the British
group active in reactor shielding wrote the first text to appear (in 1957) on
the subject. The text by Goldstein 3 was published in report form in 1957
and appeared in hard cover in 1959. The Shielding volume of the Reactor
Handbook, edited by Blizard and Abbott,' was published in 1962.
T. JaegerS wrote a text on Principles of Radiation Protection Engineering,
which was published in 1960 in German and translated by L. Dresner of
ORNL for publication in English in 1965. From the standpoint of
dissemination of shielding information, probably the most important event
was not a publication date but the founding in 1962 of the Radiation
Shielding Information Center (RSIC) at Oak Ridge. Originally organized by
K. Penny, D. Trubey, and B. Maskewitz, RSIC has performed a remarkable
job of serving the needs of the shielding community. The specialized needs
of civil defense have lead to a separate technology of fallout shielding, which
is available in a 1962 monograph by Spencer 6 and a 1966 collection edited
by Kimel. 7
From 1966 to 1970, ORNL published Chaps. 2, 3, 4, and 5 of the
Weapons Radiation Shielding Handbook, edited by Abbott, Claiborne, and
Clifford." Authors for this handbook contributed revised material from the
earlier pllblication to the present text. Recenrly the Engineering Compendium on Radiation Shielding,9 R. G. Jaeger, editor-in-chief, was published,
Vol. I in 1968, Vol. III in 1970, and Vol. II in press. This compendium is
sponsored by the International Atomic Energy Agency, Vienna, and is an
excellent source for the international shielding literature.
The extensive Russian shielding literature deserves further mention here
since it is referenced in only a few instances elsewhere in this work. A
guide to the Soviet literature was published by J. Lewin, J. Gurney, and
D. Trubey' 0 for RSIC in 1968. A recent computer scan of Russian
entries in the RSIC bibliography produced over 200 entries. Most of these
articles are available in English rranslation 1 to 2 years after their original

Russian publication. Current reports will be found in Soviet Atomic Energy
(Atomnaya Energiya).l' Collections that are useful reports of current
progress are found in a series entitled Problems in the Physics of Reactor
Shielding; Vol. 4 is the most recent volume available in English.' 2
REFERENCES
1. T. Rockwell III (Ed.), Reactor Shielding Design Manual, USAEC Repoct TID-7004,
1956.
2. B. T. Price, C. C. Horton, and K. T. Spinney, Radiation Shielding, The Macmillan Co.,
New York, 1957.
3. H. Goldstein, Fundamental Aspects of Reactor Shielding, Addison-Wesley Publishing
Company, Inc., Reading, Mass., 1959.
4. E. P. Blizard and L. S. Abbott (Eds.), Reactor Handbook, 2nd. ed., Vol. 3, Part B,
Shielding, Interscience Publishers, a division of John Wiley & Sons, Inc., New York,
1962.
5. T. Jaeger, Principles of Radiation Protection Engineering (English translation by
L. Dresner). McGraw-Hill Book Company, Inc., New York, 1965.
6. L. V. Spencer, Strncture Shielding Against FaUout Radiation from Nuclear Weapons,
National Bureau of Standards, Monograph 42, Superintendent of Documents, U. S.
Government Printing Office, Washington, D. c. , 1962.
7. W. R. Kimel (Ed.), Radiation Shielding Analysis and Design Principles as Applied to
Nuclear Defense Planning, Report TR-40, Office of Civil Defense and Kansas State
University, 1966. See also C. H. Koontz (Ed.), Shelter Design and Analysis, Vol. 1,
Fallout Radiation Shielding, Report TR-20, Office of Civil Defense, Revised 1970.
8. L. S. Abbott, H. C. Claiborne, and C. E. Clifford (Eds.), Weapons Radiation Shielding
Handbook, USAEC Report DASA-1892, Oak Ridge National Laboratory, 1970.
9. R. G. Jaeger (Editor-in-Chief), Engineering Compendium on Radiation Shielding,
Springer-Verlag, New York, Vol. I, 1968; Vol. III, 1970; Vol. II, in press.
10.J. Lewin, J. Gurney, and D. K. Trubey, A Survey of Recent Soviet Radiation Shielding Work, USAEC Report ORNLRSIC-23, Oak Ridge National Laboratory, 1968.
11. Soviet Atomic Energy, Consultants Bureau, New York, monthly.
12. D. Broder, A. Veselkin, Yu. Yegorov, A. Suvorov, and S. Tsypin (Eds.), Problems
in the Physics of Reactor Shielding, Report JPRS~51083, Joint Publication Research
Service, 1970.
Sources of Radiation
W. E. SELPH and
c. W. GARRETT
Shield analysis is usually subdivided into three or four phases; separate

calculations are performed for the radiation components that are significant
at each phase. Occasionally two or more of the phases may be combined, but
each represents a distinct consideration in a shield analysis. The first phase
involves defining the source in sufficient detail to provide the parameters for
input to the succeeding phases. The second p~ase involves calculating the
intensity and distribution of the radiation that penetrates the shield about
the source; it is the principal phase in the analysis of a reactor shield. The
third phase consists in determining the intensity and distribution of radiation
traveling from the shield to the receiver. "Receiver" describes the points or
region for which the radiation intensity is desired, e.g., a radiation detector
(real or hypothetical), a human for whom biological exposure is desired, or
equipment for which the exposure is desired to assess possible radiation
damage. This thi~e typically involves a calculation of radiation
interactions in air. A fou.rth-. phase may be defined if the receiver has a
separate shield, e.g., for some applications of nuclear propulsion it is
advantageous to divide the shielding and use separate shields at the reactor
(the source) and the crew (the receiver).
Thus the first task in a shield analysis or a shield design is to define and
characterize the radiation sources. This chapter contains a classification of
the various types of sources encountered in reactor shielding and a brief
description of each in a form convenient for shield analyses; the characteristics important to the shield designer are emphasized.
Discussions of the nuclear and atomic processes that give rise to the
emission of radiation are given in any nuclear physics text. Also, the
shield designer is primarily concerned with the characteristics of the emitted
radiation, so only a cursory description of the physics of radiation
production is given here. In addition, some mathematical and fundamental
11
12
radiation
physics concepts are reviewed which are necessary for the
development of the material in this and succeeding chapters.

Subsequent chapters are devoted to the other phases of shield analysis
and to topics relevant to their application.
2.1
GAMMA-RAY AND NEUTRON SOURCES
The primary concern in reactor shield design is the specification of

suitable barriers around sources of neutrons and gamma rays to limit the
radiation exposure to biological systems or equipment that. must function in
proximity to these sources. Both types of radiation are sufficiently
penetrating to be difficult to attenuate; yet they are sufficiently interacting
to be damaging to tissue and other materials.
Several other types of radiation arise from the fission event or from the
interaction of fission neutrons with nuclei. These include charged particles
and neutrinos.
Neutrinos, which possess no charge, mass, or magnetic moment, cannot
interact with matter except through the very weak, purely nuclear forces.
Thus, despite the fact that they carry away 5% of the power of a reactor,
they do not pose a shielding problem because they are incapable of causing
damage.
Conversely, charged particles are of little concern because they are so
highly interacting that relatively small amounts of material provide an
adequate barrier. The absorption of energy associated with charged particles
may, however, be an important consideration in the thermal design of a
system.
The sources of radiations of primary interest in a reactor, neutrons and
gamma rays, are discussed in the following sections.

2.1.1 Gamma-Ray Sources
V
A variety of sources contribute to the gamma radiation produced by a

fission reactor. The relative importance of these sources to the total
gamma-ray intensity depends primarily on the reactor design; within a given
reactor the importance of the components may vary with position in the
reactor and with reactor operating historY'O:8):..!xample, at a point near the
reactor core, the prompt fission gamma rays may be most important during
reactor operation, the fission-product gamma rays may predominate for the
first few hours after shutdown, and subsequently the gamma rays from
activated materials in the vicinity may be more importantJ
SOURCES OF RADIATION
13
In power reactors the most intense gamma-ray component penetrating
the shield is frequently that from neutron interactions in the thermal shield,
pressure vessel, or biological shield. However, in special situations any of the
sources discussed in the following paragraphs may be of importance, and
each must be considered as a potentially significant source by the shield
designer.
.i
I
(a) Fission Gamma Rays. The discussion on gamma rays from fission
and fission products is limited to those from 235 U. Spectral distributions
and energy partitions are not known as well for other fissionable materials;
similarity to 235 U is frequently assumed, but some differences have been
observed. Strictly speaking, gamma-ray energy released in fission is divided
into four time ranges, the first and last contributing more than 90% of the
total energy released as gamma rays. These time (t) ranges are:
Prompt, t ~ 0.05 Jisec (7.25 MeV).
Short-life, 0.05 < t ~ 1.0 Jisec (0.43 MeV).
Intermediate-life, 1.0 Jisec < t ~ 1.0 sec (0.55 MeV).
Delayed, t> 1.0 sec (6.65 MeV).
The values in parentheses are for the gamma energy released per 235 U fission
and were taken from an evaluation by Holden, Mendelson, and Dudleyl
except the value for prompt fission, which is quoted from a recent note by
Peelle and Maienschein. 2 Prompt fission gamma rays have energies from
10 keV to 10 MeV. An average of 8.1 0.3 photons are released per fission,
and these photons carry off 7.25 0.26 MeV per fission event. Prompt
fission gamma rays can contribute a significant amount to the total
gamma-ray field at points near the reactor core during reactor operation and
should be included in the core-shield analysis. The energy distribution of
prompt gamma rays is discussed in Sec. 2.4.1.
The short-life interval is similar to the prompt in energy distribution and
accounts for 5.9% (0.43 MeV) of the prompt gamma-ray energy release.
The intermediate-life interval is also usually assumed to have the same
energy distribution as the prompt interval with an energy release of about
0.55 MeV.
(b) Fission-Product-Decay Gamma Rays. The two middleweight nuclei
resulting from a fission event are called fission fragments or fission products.
Because of an excess of neutrons, most of approximately 80 possihle initial
fission-product isotopes are radioactive, initially decaying by beta emission.
The beta decay is followed by gamma emission whenever the beta decay
14
results in an excited state of the daughter nucleus. In a few special cases,

only gamma radiation is emitted.
The vast majority of daughter products of the initial fission fragments
are also radioactive, decaying with the emission of beta particles and gamma
rays. Having, on the average, about three stages of radioactive decay before
they become stable, fission products form a complex mixture of gamma-ray
emitters with half-lives ranging from less than a second to millions of years.
Of the approximately 21.5 MeV per fission emitted by fission products, the
greatest fraction is carried by beta particles and neutrinos. About 6.65 MeV
per 235 U fission is emitted by the fission product as delayed gamma rays;
over three-quarters of this energy is released with 10 3 sec following fission.
Fission-product activity will not dictate the thickness of the shield
surrounding an operating reactor. There are a few shield-design situations,
however, where protection from fission-product activity is the primary
criterion. Included in this category are primary loops in circulating-fuel
reactors and enclosures for the shipment or dismantling of used fuel rods or
assemblies. Knowledge of the dose rates from fission products is essential to
the formulation of procedures to be followed in any maintenance involving
compromise of the primary shield.
Delayed gamma rays can be classified into two groups, depending on the
time of their emission following fission. Early fission-product gamma rays
are those emitted within a few minutes after fission; they contribute to the
total core source during the operation of a reactor. They can also be of
prime importance in regions external to the core during the operation of a
circulating-fuel reactor. Late fission-product gamma rays (those which are
emitted several minutes or longer after fission) are not of much importance
during reactor operation, but they can be a very significant source following
reactor shutdown.
Because most of the fission-product energy is contained in the early
fission products, fission-product intensity reaches a state of quasiequilibrium in a reactor core that has operated at steady state for only a few
hours. Thus the early decay gammas are sometimes lumped with the prompt
fission gamma rays and are considered a part of the prompt source. Energy
spectra for both the early and late fission products are discussed in
Sec. 2.4.1.
(c) Capture Gamma Rays. Radiative capture of neutrons by nuclei at
thermal and epithermal energies produces secondary gamma rays, commonly
called capture gamma rays. They are emitted promptly after neutron capture
15
to relieve an unstable energy situation generated within the compound

nucleus.
The total energy available for gamma rays from capture is the sum of the
kinetic energy of the incident neutron and its binding energy in the
compound nucleus. Since the probability of capture decreases rapidly with
increasing kinetic energy, capture reactions generally are of importance only
for neutrons with kinetic energies below 25 keV. Typically, neutron binding
energies are in the region of 6 to 8 MeV, although they can range from 2.2 to
about 11 MeV. Thus binding energy is the greatest component of the total
gamma energy emitted.
Although this energy may be carried off by only one photon, it most
often is shared by two or more photons as the compound nucleus decays
through several excited states. The energy distribution of capture gamma
rays can range from sharp discrete line spectra to almost continuous spectra
produced by cascading decay; the capture interaction process is further
discussed in Chap. 3.
Since investigators studying capture spectra are primarily interested in
nuclear processes, many times only discrete spectral lines with significant
intensities are reported. In using such data, we must be careful to ascertain
whether or not the entire neutron binding energy is contained in the
reported gamma spectrum. Where it is not (differences can be as much as
30%), adjustments should be made by increasing the number of photons
emitted to account for the total amount since this energy is usually an
important contributor to the radiation that penetrates the shield.

Cap!ure gamma rays are a significant source and occasionally constitute
the most important consideration in shield design because of their high
energy and the fact that they are generated throughout the shield.
(d) Inelastic-Scattering Gamma Rays. In neutron inelastic scattering,
part of the energy of the incident neutron is carried off by the scattered
neutron and part is absorbed by the target nucleus. The latter is left in an
excited state and subsequently decays by gamma-ray emission. An alternative process that may be involved is the formation of a neutron-heavy
compound nucleus by union of the target nucleus and the incident neutron.
The extra neutron is then emitted, and the target nucleus is left in an excited
state. The time between neutron interaction and gamma-ray emission is
negligible (10- 14 sec).

As with capture gamma rays, the excess energy may be carried off by
one or more photons. However, gamma rays from this source are generally
16
less penettating than those from neutton capture; their energies range from a
fraction of 1 MeV to several MeV.[Since the energy of the incident neutron
must exceed the energy of the first excited level of the target nucleus,
inelastic scattering becomes more important as the neutron energy increases. J
In general, neutrons must be in the MeV region to generate significant
gamma rays by this process.
(e) Reaction-Product Gamma Rays. The reaction-product source results
from a process resembling that of inelastic scattering except that some
particle other than a neutron is ejected from the nucleus. The nucleus is left
unstable and emits a gamma ray. An example is the 10 B(n,a) 7 Li interaction,
which is accompanied by the emission of a gamma ray of approximately
0.5 MeV. Sources of this type are significant only in materials containing
isotopes which have a reasonable probability of undergoing a particular
reaction and which are located in strong neutron fields. In reactors
boron-containing materials generally arc the only materials that fit these
criteria.
(f) Activation-Product Gamma Rays. Capture and inelastic-scattering

gamma rays are emitted simultaneously with the neutron-target nucleus
interaction. However, the nucleus formed by a neutron interaction may be
radioactive and decay with a half-life that can range from seconds to years,
emitting photons and other radiations in the process.1..These activationproduct gamma rays may be of importance to the shield analysis and are of
particular concern after reactor shutdown.IThey may also be emitted in
significant quantities from materials whichhave been exposed to the high
neutron flux of the core and which are subsequently circulated in regions
external to the reactor shield. Irradiated samples and reactor coolants both
fit this description.
For example, the 160(n,p)1 6N reaction produced by fast-neutron
activation of water emits gamma rays with energies of 6.1 and 7.1 MeV. The
half-life of 16 N is 7.13 sec, short enough to produce high activities in
irradiated-water-coolant streams. Liquid-metal fuels and coolants must also
be considered as a source of activation-product gamma rays. In sodium
mixtures, 23Na(n(y)24Na produces 1.38- and 2.76-MeV photons with a
half-life of 14.8 hr.
(g) Annihilation Radiation. A few activated materials decay by the
emission of positrons, which are annihilated by subsequent combination
with electrons. Because of the relatively short distance that positrons travel
before annihilation, the process can be considered to occur at the time of
17
decay and at the site of the activated nucleus. 'Two O.Sl1-MeV photons are
emitted from each positron-electron reaction. High-energy gamma rays also
can react in a converse process called pair production (discussed in Chap. 3)
to produce electron-positron pairs. These positrons also are annihilated near
their source in an identical manner and contribute further to the source of
O.Sl1-MeV gamma rays.
(h) Bremsstrahlung. The acceleration and deceleration of electrons in
the atomic electric field produces electromagnetic radiation called br!1!'.sstrahlung. JThe process is identical to that occurring in the X-ray tube and is
an important consideration only where high-energy beta particles (or
accelerated electrons) interact with materials of high atomic number. An
example is found in the use of lithium as a coolant. Neutron absorption in
7 Li produces 8 Li. The latter undergoes decay to 8 Be by emission of beta
particles with energies as high as 13 MeV. Bremsstrahlung produced by these
high-energy electrons as they slow down in piping or containment materials
requires evaluation as a gamma~ray source.
2.1. 2 Neutron Sources
By far the greatest neutron source in an operating reactor is that created
by the fission process itself, in which, along with the prompt gamma rays,
free neutrons are released as part of the fission event. However, other
reactions can also produce neutron sources of importance to the shield

designer, and all sources discussed in this section should be considered in a
shield analysis.
(a) Fission Neutrons.. Approximately 2.S neutrons are emitted per
fission event in 235 U by thermal neutrons (more in other cases), and they
carry away a total energy of approximately S MeV. Although energies can
range from the eV region to beyond 18 MeV, the average energy of a 235U
fission neutron is about 2 MeV, and an upper limit is often taken to be
14 MeV. In fact, less than 1% of the total energy of fission neutrons is shared
by neutrons whose energies exceed 10 MeV. However, these high-energy
neutrons are very penetrating, and in some cases they can be of overriding
importance. Prompt fission-neutron spectra are discussed more fully in
Sec. 2.4.2.
For shielding purposes, fission neutrons may be assumed to be evolved
simultaneously with the fission event. The very small fraction 1%) with
delayed emissions requires consideration as a separate source only in the case
18
of a circulating-fuel reactor where the fuel loop extends beyond the core
shield.
"(b) Activation Neutrons. Under certain circumstances the decay of a
radioactive nucleus can be followed by the emission of a neutron. This
occurs when the energy of excitation of the daughter nucleus is in excess of
the binding energy of the last neutron in the nucleus. In fact, when such an
energy imbalance exists, neutron emission is the preferred mode of release.
An example is the beta decay of 17 N with a 4.14-sec half-life, which leaves
an 17 0 nucleus with more than enough excitation energy to eject a neutron.
Neutrons from this source have a most probable energy around 1.0 MeV.
Nitrogen-17 is formed by the I 60(n,p jl 7 N reaction and can be important in
fast-neutron bombardment of water.
"(c) Photoneutrons. A photon whose energy is greater than the neutron

binding energy of a nucleus can impart enough energy to the nucleus to
cause neutron emission. The photon energy required to make such a reaction
possible exceeds 7 MeV for all but a few nuclei, and the probability for the
photo neutron reaction is quite low until photon energies above 10 MeV are
reached. Thus photoneutrons do not contribute a significant source
component in the vast majority of reactor shielding problems. The few
nuclei whose neutron binding energies are low enough to create a possible
problem in reactor shielding include 2 D, 9 Be, 13 C, and 6 Li. The threshold
photon energies for these isotopes are 2.23, 1.67, 4.9, and 5.3 MeV,
respectively. Since all four occur in moderator materials, these exceptions
are occasionally important.
V(d) Particle-Reaction Neutrons. The interaction of alpha particles with
nuclei of lithium, beryllium, oxygen, boron, and fluorine produces neutrons.

Thus these elements are often combined with alpha-active isotopes, such as
polonium or plutonium, to form neutron sources for use in experimentation
or reactor start-up. Neutrons from this source may be important to shielding
and safety during assembly or in the pre-start-up environment of reactors

containing beryllium in the fuel-element material. Similarly, neutrons from
(<>,n) reactions in oxygen may be dominant in oxide fnel elements.
The energy distribution of these neutrons is broad since the neutron
kinetic energy depends on its angle of emission relative to the direction of
the incident alpha particle as well as on the kinetic energy of the alpha
particle. In polonium-beryllium and plutonium-beryllium sources, for
example, emergent-neutron energies range from <] to> 10 MeV.
19
Neutrons may be generated by other charged-particle interactions, but

they have not been important in reactor designs to date. Incident neutrons
with energies above 10 MeV can excite a compound nucleus sufficiently to
emit two or more neutrons. Such reactions are rarely of importance in
reactor shield design.
2.2
BASIC MATHEMATICAL AND PHYSICAL CONCEPTS
The physical properties of radiation sources and fields arc discussed in

the remaining sections of this chapter and in subsequent chapters. The
manner in which these properties arc characterized requires an understanding
of certain basic concepts, including differential and integral distributions,
normalization and averaging techniques, solid-angle concepts, and definitions
of flux density and current density. This section contains a review of these
necessary basic considerations.
2.2.1
Differential Distributions
Most properties that characterize radiation fields and sources arc

functions of one Of, morc often, several independent variables. Mathematical
functions may be used to describe the dependence of such a property on its
variables; functions of this type arc called difj,,,ential distributions. Errors in
the treatment of these distribution functions are among the more common
mistakes in shielding practice. Because of the need for precision in these
concepts, a careful explanation of the methods of treating such distributions
is in order.
Consider, for example, a radiation source nonuniformly distributed in a
three-dimensional slab whose emission rate varies with both time and one
spatial dimension. Such a source can be described by a function of four
variables, P(x,t;1l VAt). Its values give the number of source particles
emitted in time interval J1t centered about time t from a spatial volume
element of size Il V centered at point x in the slab. The dependent variables
Il V and Ilt arc functions of the independent variables x and t, respectively.
Such a function is sufficient to characterize completely the spatial and
temporal properties of the source.
The concept can be generalized to any number of variables. The general
form can be written as
P(x,y,z, ... ; Ilx,lly,llz, ... J = D(x,y,z, ... J Ilx Ily Ilz. ,.
(2.2-1)
20
The function D(x,y,z, ...) is called a differential distribution function, t and

its values give the number of events per unit measure of each of the
independent variables, x,y,z, .. " involved.
In the radiation-source example, the differential distribution
D( x,t )
= P(x,t;Li V,At)
Li V Lit
(2.2-2)
gives the number of source emissions that occur per unit volume and per unit
time in volume element AV about x and in time interval Lit about t. It has
units of l-3e l , for example, particles per cubic centimeter per second.
Differential distributions are often used to describe energy spectra of
sources and fields. For example, if D(E) is a differential distribution
describing the continuous energy spectrum of a photon field, the values of
D(E) give the number of photons per unit energy interval about E, and
D(E) AE is the number of photons whose energies are in the energy interval
liE containing energy E.
Although differential distributions are useful in conjunction with finite
intervals (i.e., LiV,Lit,LiE), which need not be infinitesimals, they are
frequently applied to differentials. Thus D(x,y,z, ...) dx dy dz ... specifies,
for a distribution of events, the amount that lies in the differential volume
element dx dy dz ... containing the point (x,y,z, .. .). iThe integral of a
differential distribution function over the entire domain of all variables gives
N, the total number of events or quantity contained in the distribution,
N=fxfy D(x,y, ... ) dxdy ...
(2.2-3)
In our two examples, Ix It D(x,t) dt dx is the total number of particles

emitted from the radiation source and IE D(E) dE is the total number of
photons of all energies contained in the energy distribution. Differential
distributions may be displayed in graphical form, as shown in Fig. 2.1.
[It is important to recognize that the value of D(x)+ at a selected value of
x, say XI' does not give the number of events occurring at point Xt. Rather,
D(Xl) is the number of events occurring in a unit interval of x that contains
the point
Thus, although the unit selectd to measure the independent
x)
fMathematicians reserve the term distribution function for inlcgral distributions [i.e., 1--'1 {x} and
F, (x), Eqs. 1.2-4 and 2.2-5 J; however, our use of the term here is consistent with common usage in
the radiation shielding field.
:j:Th~ discussion that follows is easily extended to many independent variables.
21
x/
Fig. 2.1-Differential distribution.
variable x is arbitrary (i.e., if x is a length, centimeters, feet, yards, etc., can

be used), it is essential that this unit be specified in order for the distribution
function and its plot to have meaning.
Distributions are sometimes expressed in other forms. The cumulative
distribution function for a differential distribution of one variable is defined
by~
F,(x)=
I_.D(x)dx
(2.2-4)
where P, (x) gives the amount of the distribution that lies below x.t
Alternatively, the function
P 2 (x)
= Ix
D(x') dx'
(2.2-5 )
gives the amount of the distribution above x. It IS sometimes called the

survivorship function.
Often, distributions are normalized to the total distribution, an
operation that is valuable in comparing different distributions of the same
variables:
[(x) = D(x)
N
x)
f1(
tin Eqs. 2.2-4 and 2.2-5, _00 and

alters the range of the distribution.
D(x)
J_:D(x) dx
=P, (x) =e.

N
00
D(x') dx'
J: D(x) dx
can be replaced by any a,b such that a < x
(2.2-6)
(2.2-7)
< b; this
merely
22
f,(x) =
(X) = f~ D(x') dx'

N
f:D(x) dx
['2
(2.2-8)
where the function f(x) is the fraction of the distribution per unit inrerval
about x, f, (x) is the fraction of the distriburion lying below x, and f2 (x) is
the fraction lying above x. Nore that f(x).!, (x), and f2 (x) are constrained to
the interval 10,1], and f, (x) + f2 (x) = 1. If Xb and Xu denote the lower and
upper bounds of the distribution,
f, (xb)
=f,(x u ) =0
(2.2-9)
f, (xu)
=f2 (xb) = 1
(2.2-10)
and
{u
xbf(x)dx=
(2.2-11)
Differential distribution functions can be used to obtain the portion of

the distribution or the number of events that occur over a specified region.
In the photon-field energy-spectrum example,
f:.' D(E) dE
gives the number of photons with energies between E, and E 2 . Such integral
distributions may be plotted in histogram form (Fig. 2.2). The ordinate
D(Llx;) is given by
D(Llxi) =
r, +'
i
(2.2-12)
D(I<) dx
and gives the total number of events within LlXi, where LlXi = Xi+1 - Xi'
Observe that the shape of a histogram depends on the selection of the
increments, LlXi. The distribution function that is plotted in Fig. 2.2 with
equal increments could give a histogram of the form shown in Fig. 2.3 if
different increment sizes were chosen. However, if a differential histogram is
constructed in which
D'(Llxi) =
l
A
~xi
x,
i 1
+ D(x)
dx
(2.2-13)
23
I-~
r-x
XI
xi + /
Fig. 2.2-Histogram for constant intervals.
Fig. 2.3~H!i.stogram for variable intervals.
the shape of the histogram always approximates the distribution function

D(x), and the approximation improves as the number of intervals is
increased. Note that
N; ~ D(,hj) ; ~ D'(dxj) dXj
(2.2-14)
If a differential distribution function has more than one independent

variable, new differential distributions are obtained when an integration is
performed over some, but not all, of the variables. Consider the following
differential distribution function, which can be derived for the radiationsource example:
D'(t) ;
fall 'pace D(x,t) dV
(2.2-15)
24
Most of the time, differential data presented in shielding analyses have been
integrated (either mathematically or through the inherent characteristics of
an experimental device) over one or more of the variables involved. Such
data arc commonly denoted singly differentiated, dOIf!J[y dljfercntiated, etc ..
depending on the number of independent variables that remain after
integration of the fundamental multidimensional distribution.
2.2.2 Average and Most-Probable Values

Several parameters derived from differential distributions arc of value in
characterizing the distribution. One is the integral value N (Eq. 2.2-3). wbich
gives the total number of events in the distribution and is often llsed for
normalizing purposes.
The average value of an independent variable within a distribution is
obtained by computing the integral

1
x = N I, Iv
=!,Iy
x . D(x,y
x j\x,y,
J dx dy ...
)dxdy ...
(2.2-16)
which is called the first moment of the differential distribution.

The most-probable value of an independent variable of a differential
distribution is its value where the distribution function has its maximum
value. Note that the most-probable value may not be unique; for example. if
D(x) is constant over all x, every value of x is a most-probable value.
At times, higher moments of differential distributions are useful (c.g., in
the computation of standard deviations and variances). The nth moment
about the variable x is defined by
x(n)
1
N
I, Iv x n . J)(x,y ... J dxdy ...
(2.2-17)
Examples of calculations of average and most-probable values are given

the exercises.
III
2.2.3 Solid Angle

Directional characteristics arc essential to a complete description of
radiation fields and sources. Since the concept of solid angle is used in
specifying directional properties, some elementary definitions arc reviewed.
25
The solid angle subtended at a point by a surface is the projection of the

surface on a sphere of unit radius surrounding the point. Consider the solid
angle subtended at the point D by K, in Fig. 2.4. (We could consider K 2 or a
surface of any other shape.) The area of the projection of K, on the sphere
of unit radius is a solid angle. A unit of solid angle is called a steradian. Since
the area of a sphere is 47Tr 2 and that of a unit sphere is 47T there arc 47T
1
Fig. 2.4-Solid angle.
stcradians about a point. If the unit vector normal to surface K 1 is n and the
unit vector along r through K, is Sl, then the solid angle is the scalar product
n Sl (K, I r2 ) , t when r is the length of the radius vector to K I
Radiation-source and -field angle distributions are defined by differential
functions of the form D(Sl), where D(Sl) gives the distribution per /lnit solid
angle along the direction of vector n. In polar coordinates, where measures
polar and measures azimuthal angles,:!: the differential area on a unit
sphere is given by
dA
sin e de d
(2.218)
d.
and is numerically equal to the differential solid angle, dSl, subtended by

dA. Thus. if D(Sl) is the differential angle distribution of a point radiation
t Strictly speaking, this (ormul,l is accurate only when K 1 <~ r 2 .

+Polar and
az_imuthal angles
radi,ltion sourCl'S are discussed,
:lfe
defined in Sec. 2.3.2 .. where directional characteristics of
26
source located at the center of the sphere, D(n) sin IJ dIJ dl/> particles are
emitted which pass through dA, and
f:: f:" D(n) sin IJ dIJ dl/>

gives the number of source particles having directions bounded by polar
angles IJ 1 and IJ 2 and by azimuthal angles 1/>1 and 1/>2' Note that in the
preceding discussion, the angles IJ and I/> define the directional vector, n,
which has a unit magnitude.
2.2.4 Measures of Radiation Intensity
Essential to a discussion of intensity is the concept of a receiver, or a
detector, a mathematical concept used in defining the necessary units and
describing the fields. As will be seen, theoretical detectors take the form of
points, surfaces, and volumes, depending on the quantity being measured.
Measures of radiation intensity commonly used in shield analysis include
particle density, flux density, current density, absorbed dose rate, kerma
rate, and exposure rate. The first three aTC used to characterize a radiation
field and are defined in this section. The last three are used ro characterize
radiation interactions with matter and are discussed in Chap. 3. The time
integral of the last three quantities over an operating cycle or exposure
duration provides a basis for estimating the damaging effect of the radiation
involved.
The quantities discussed in this section are obtained by taking a limit of
measured radiation-particle densities t in a finite volume or passing through a
finite area as the volume or area approaches zero in size. This is a very special
limiting process that requires careful definition to avoid a contradiction
between the mathematical concept of a differntial and the physical reality
of the random nature of particle tracks. In moving to such a conceptual
limit, we can reach a point where the detector size is so small that for further
reductions in volume the probability of finding a particle within it also
approaches zero, and thus the limit of the particle density is zero. In the
special limiting process implied in the definitions that follow, this is not
tThe terms particle and density arc used in a general senslt. Particle refers both to particles having
rest mass and to photons. Density refers to a differential function of one or more variables, i.e.,
partides/cm 2 , particles/em 3 , MeV cm- 3 sec-I, etc.
27
allowed to occur. Rather, in setting the limit, we stop shrinking the detector
volume or area as soon as a further reduction in size would not change the
value of the density quantity being measured. At the same time, we must
keep the detecting volume large enough to contain a statistically significant
number of particles or particle interactions. Tn radiation fields with very
steep gradients, this special limiting process raises difficulties because the
two conflicting reguircmcnts on detector size may be mutually exclusive. In
such cases we introduce an averaging process to overcome the conceptual
problem,
A complete description of a radiation field includes the numher of

particles of a given type and energy which at a given time exist at a given
position and travel in a given direction, all particle types, energies, positions,
and directions being considered at all times. Such complete descriptions arc
seldom required, however, except in the most sophisticated partielctransport calculations, and a radiation field cannot be measured in such
detail. Therefore particle populations arc usually described in terms of a
lesser number of parameters.
Instead of the indefinite number of variables used in the previous
discussion on distributions in general (x1y,z, .. .), we now limit the
discussion to the seven dimensions needed to define the kinematics of a
particle. These are the three spatial coordinates: the particle kinetic energy
(velocity is used alternatively), two angles defining direction, and time. The
three spatial coordinates can be specified as the radius vector, r, and the two
direction angles reduced to a unit direction vector, n. These variables define
phase space, which is indicated functionally asf(r,i:,n,I).
The following discussion of particle, flux, and current densities is
adapted from a recent report by Stevens and Cl a iborne. 3 They have followed
the recommendations of the International Commission on Radiation Units
and Measurcments,4,5 which this text has also attempted to follow.
(a) Particle Densities, Knowledge of the particle density over all phase
space is equivalent to a complete solution of a particle-transport problem
and comprises more information than is available from most calculational
schemes now in usc. When given in seven-dimensional phase space, particle
density is defined by
,,(r,E,n,l) dii dn
_ 1. . .
28
the number of particles per unit volume at space point r and time t having
energies in dE about energy E and directions in dn about the unit direction
vector n.
The particle density so defined is doubly differential, in energy and in
direction, and less detailed forms will often suffice. For example, we may
use the steady-statet particle-density differential in energy only, commonly
called the differential particle density and defined by
n(r,E) dE
the number of particles per unit volume at space point r having energies in
dE about E.
n(r,E) dE = 14. n(r,E,n) dn dE
(2.2-19)
Or one may use the steady-state total particle density, defined as the number
of particles per unit volume at space point r and given by
n(r) = L.
10- n(r,E,n) dE dn = 10- n(r,E) dE
(2.2-20)
(b) Flux Densities. Even though the concept of particle density is

basically simple and has a unique interpretation, experience has shown that
the flux density,4 or, as it is commonly called, theflux,:t serves better as the
dependent variable in solutions of the transport equation (discussed in
Chap. 4). The flux density is related to the particle density through the
particle's speed, and, when described in terms of seven-dimensional phase
space, the flux density is given by
(r,E,n,t) = v n(r,E,n,t)
(2.2-21)
where v is the particle's speed and corresponds to the energy E. (The speed is
the scalar magnitude of the particle's velocity vector, v.)
tThesteady-state,ortime-independent, condition is denoted in phase-space notation by dropping

the time symbol t; seven-dimensional phase space becomes six dimensional and in phase-space notation
is given by (r,E,n).
:t:Although this quantity is truly a density and the International Commission on Radiation Units
and Measurements 4 recommends the use of the term flux density, the simpler term flux is ingrained ill
shielding terminology and is used extensively elsewhere. We have used flux density exclusively in this
text.
29
The flux density defined in the preceding paragraph is doubly differential and is usually referred to as the angular flux density. Greater insight into
the use of the angular flux density as the cfependent variable in mathematical
descriptions of particle transport is provided by its interpretation either as
the track lengths traversed per unit volume and time or as the flow of
particles per unit area and time.
Flux density is a measure of a radiation field in terms of its potential for
interaction with the material through which it is passing and is measured
with volume detectors. In the theory of radiation interactions, nuclear forces
(or coulomb forces) are additive; i.e., no shadowing of nuclei or electrons by
others occurs. This implies that the probability of a radiation particle
interacting with matter is directly proportional to the number of nuclei (or
electrons) in whose vicinity it passes. This number, in turn, is proportional to
the distance traveled by the radiation particle. Thus the total interaction
probability of a radiation field with matter is proportional to the sum of the
distances, or total track length, traveled by all the radiation particles
traversing the medium, and the interaction rate is proportional to the total
track length generated per unit of time.
The track length interpretation of angular flux density follows from the
observation that the speed of an individual particle can be considered as its
scalar track length per unit time. The product of particle density and speed is
then the sum of the track lengths traced by all the particles within a unit
volume per unit time, t in which case the definition of the angular flux
density would be
(r,E,n,t) dE
dn
the total track lengths traversed per unit volume and time at space point r
and time t by particles having energies in dE about energy E and directions
in dn about n.
The interpretation of the angular flux density as a flow of particles per
unit area and time is closely related to the concept of angular current (to be
discussed in the next section). Itwill be shown in the discussion on current
that the angular flux density is identical to the magnitude of the current
vector J and th us can be interpreted as
tNote that the per-unit-time units of the flux density are associated with the particle's speed,
which is a function only of the energy. However, the time dependence of the flux density is a
consequence of the time behavior of the particle density, which does not have time units even though
the time symbol, t, is included in the phase-space notation to denote a dependence on time.
30
(r,E,n,t) dE dn
the flow per unit area and time at space point r and time t of particles having
energies in dE about E and directions in dn about n.
When the transport and the deposition of the particle kinetic energy are
of interest, the energy-flux-density differential in energy and angle is often
used. This quantity, called the angular energy-flux density, is defined by
f(r,E,n,t) dE dn
the energy flow per unit area and time at space point r and time t due to
particles having energies in dE about E and directions in dn about n, and is
related to the angular flux density by
f(r,E,n,t) = E (r,E,n,t)
(2.2-22 )
Many calculational models employ less detailed descriptions of the flux

density; for example, they may use a description in which steady state is
assumed and the angular dependencies are removed by the appropriate
solid-angle integration. This reduces the description to four-dimensional
phase space and results in a dependent variable representing the particleflux-density differential in energy. Commonly known as the differential flux
density, this quantity is given by
(r,E) = J4n (r,E,n) dn
(2.2-23)
Like the angular flux density, the differential flux density can be
interpreted in terms of track length per unit volume and time or in terms of
the number of particles that enter a unit sphere per unit time. tn the latter
case, the solid-angle integration can be regarded as a summing of particles
that enter a sphere of unit cross section regardless of their directions of
motion. The sphere is, in effect, generated by the rotation of a circular unit
area during the integration over a 41T solid angle (sec Fig. 2.5). [n this
context the definition for the differential flux density can be restated as
(r,E) dE
number of particles having energies in dE about E which enter a sphere of

unit cross section per unit time at space point r.
"~.
31
12,
'\
~A
(Project! .re.i
Fig. 2.5--Incremental sphere concept of flux density.
Although this definition of the differential flux density is descriptive, it

is not exact. The mathematically rigorous definition is
"'(
I') dE" == I'1m
... r,"
.6.A
-+
N(E)
AAdE
(2.2-24)
L.l
which implies the limit process LlA .... 0, with N(E) dE denoting the number
of particles having energies in dE about E which enter an incremental sphere
of cross section LlA per unit time.
The concept of the incremental sphere is the best way to visualize the
energy-flux-density differential in energy. Referred to as the differential
energy-flux density, this quantity may be defined by
I(r,E) dE
the energy flow per unit time into a sphere of unit cross section at space
poin t r due to particles having energies in dE about E. The differen tial
energy-flux density is given by
[(r,E) dE
= J""
[(r,E,n)
dn dE =E "'(r,E) dE
(2.2-25)
Other quantities used are the total flux density, the total energy-flux
density, the group flux density, and the group angular flux densityfrhe total
flux density, defined alternatively as the total particle track lengrh per unit
vOTume and time at space pOjnt [ or as the number of particles that enter a
32
sphere of unit cross section per unit time at space point r, is obtained by
integrating the differential flux over all energies:
Similarly, the total energy-flux density, defined as the total energy flow
per unit time into a sphere of unit cross section at space point r, is obtained
by an integration of the differential cncrgy~flux density over all energies:
I(r) = ~)~ I(r,1':) dL
(2.227)
The total flux density has only limited application to practical shielding
problems because of the strongly energy-dependent nature of the particle
behavior. A morc useful approach is to divide the total energy range into L
energy intervals, called energy groups, t
C=1,2, ...,L
and to define the grollp flllx density as the integral of the differential flux
density over the corresponding energy group,
I
c(r) '"
~R:)
<P(r,E) dE
(2.2-28)
with the constraint that
<P(r) = I: <Pc(r)
I.
(2.229)
The group angular flux density (group flux density differential in angle)
has a similar definition and is obtained by integrating the Jngular flux
density over a specific energy group:
<Pc (r,[2) '"
I::
f. g
L g +l
<P(r,li,.Q) dE
(2.2-30)
tThe s\lbscript~ ~ and g + 1 ref{'r to the upper and lower lilllits, respectively, of the Gth energy
group, and C c= 1 eorn:spoJlds to tll<: highest energy group. An altern'lte convention would ,ls$oeiate
G'-' 1 with the lowest energy group; the subscripts g + 1 and g would then correspond to the upper
and low\er eJlergy limits. L is the number of groups.
43
33
For problems that involve directional symmetry, the group angular flux
density can be rewritten in terms of a new angulor variable f.l = n . rllrl, the
direction cosine:
(2.2-31)
The group angular flux density can then be defined as the total particle track
length per unit volume and time at space point r of particles with energies
within energy group I:iE c and directions defined by direction cD.'iincs that lie
in <ifl about fl.
JI1 neutron physics, another variable used to characterize a neutron speed

is lcth"rJ:Y. The lethargy, lA, of a particle with energy Ii is given by
(2.232)
where 1'0 is an upper limit of n, often taken to be 1() MeV in fission
reactors. Appropriately, lethargy increases as energy decreases and
<\)(r,u)
= I' <\)(r,E)
(2,2-33)
where <P(r,l1) is tbe differential flux density per unit lethargy at point r.
A useful measure of total exposure to a flux density for applications
involving energy deposition is the integral quantity called j7ltcflce. Flucncc is
defined by the International Commission on Radiation Units and Measurements 4 ,5 as the quotient of t1.N divided by t1.A, where t1./\/ is the number of
particles that enter a sphere of cross-sectional area Ll.A and the Ll.'s imply the
special limiting process described at the beginning of Sec. 2,2.4. This
definition is equivalent to regarding fluence as a time-integrated flux density
over some specified time interval. As such, the fluence can be written
(2.234 )
where ~t corresponds to some specified time interval and <P(t) can be any
one of the several kinds of flux density described in the preceding text. For
example, the f'nergy flHcnce is
J-'(r)
{,""M'
t,
I(r,t) <it
(2,235)
34
(c)\current Densities. The characteristic property of the current variable is its close relation to the convective (leakage) effects in the theoretical
description of particle transport. The most general form of the current
variable is differential both in energy (or speed) and in angle. Called the
arlgular currerlt density, or, more frequently, the arlgular currerlt, this
quantity is symbolized by ](r,v,n,t) and is defined as the directed flow per
unit area (normal to the n direction) and time at the space point r and time
t of particles having speeds in dv about v and directions in dn about nJ
[The relation between current and particle density can be established by
considering that (1) the product of v and the particle density can be regarded
as a vector sum of the individual codirectional velocity vectors (v), yielding
the resultant vector] = rI V or (2) that the (v x dt x dA x N) particles
contained within the volume element shown in Fig. 2.6 will all exit through
Fig. 2.6-Particle flow concept of current. The particles contained within the volume
element will all exit through the differential area dA 'Within dt if v = il.
the differential area dA within differential time dt if v = vn. These models

can be expressed mathematically in terms of the angular current as
~
](r,v,n,t) dv dn = nv rI(r,v,n,t) dv dn
(2.2-36)
and, since the particle's kinetic energy is a function of its speed, Eq. 2.2-36
can be rewritten
J(r,E,n,t) dE dn = nv rI(r,E,n,t) dE dn
(2.2-37)
where ](r,E,n,t) dE = ](r,v,n,t) dv and rI(r,E,n,t) dE = n(r,v,n,t) dv. Then,

when v rI(r,E,n,t) is identified as the angular flux density, the relation
35
between the angular flux density and the angular current noted in Sec. (b),
above, is obtained:
J(r,[;,D,I) dE dD ~ D <P(r,E,D,t) dE dD
(2.2-38)
Other less detailed forms of the angular current are the group angular
current, Je(r,D), and the total angular current, 1(r,D), which are obtained
by integrations over an energy group and all energies, respectively:
Jc (r,D) '"
t'"
l::R+l
1(r,D) '"
fa"
J(r,[;,D) dE ~ D<Pc(r,D)
J(r,1:,D) dE ~ D<P(r,D)
(2.2-39)
(2.240)
It is apparent that the angular current variables all have essentially the same
simple relation with the corresponding angular flux density because the
energy integrations arc performed directly on the flux density. For example,
in the case of the group angular current,
Je (r,D) ~
J.)Eg
-g+l
D<P(r,E,D) dE
(2.241)
The integral of the angular current over all directions (4" solid angle)
constitutes a vector summation, and the resultant vector is regarded as the
net current, often called simply current.
The net current differential in energy only, referred to as the differential

net current, is defined as the nct flow per unit area and time at space point r
and time t of particles having energies in dE about E, where the unit area is
normal to the direction of the resultant vector J(r,E,t), or J(r,E,t) dE,
J(r,E,t) dE ~ /4" J(r,E,D,t) dD dE

~
/4" D<P(r,E,D,t) dD dE
(2.242)
which is the angular flux-weighted vector summation of the unit vectors n

over a 4" solid angle. The group >let current, Je(r), and total net current,
J(r), are 4" solid-angle integrations of the group angular current and total
angular current, respectively:
36
REACTOR SHIELDING FOR NUCLISAR ENGINEERS
(2.2-43)
(2.2-44)
The flow of n-directed particles across an arbitrarily oriented differential
area is a necessary concept in the description of the directed flow of particles
in terms of a specific coordinate system and can be related to the angular
current, J(n), t by consideration of Fig. 2.7, where the direction vector, n, is
At-,.L..-(JA (Differential area

Aormal to n I
J(U)
~n
Fig. 2.7 -Schematic diagram of particle flow across an arbitrarily oriented surface.
normal to the differential area. The number of n-directed particles crossing

the differential area dA per unit time is equal to J(n) (n dA). A scalar
current, In(n), that describes the flow of the n-directed particles per unit
area normal to the direction n is defined as
(2.2-45)
It follows that
.\
(2.2-46)
where n = the unit vector corresponding to an arbitrary direction

n = a coordinate-identifying subscript, for example, n == x when n == i
In(n} = the flow of n-directed particles per unit area (normal to the
direction n) and time
tin this further discussion of current, the notationJ(n) will be used to denote any of the angular
currents and J ttoe corresponding net current: for example, J(O) may represent ](r,E,n, t), Jc{r,n,t).
or ](r,nj, and J may represent ](r,/:',t), JC(r), or J(r).
..1...
37
(Q) ~ the angular flux variable corresponding to the angular current

J(Q)
Note that In(Q) is a scalar quantity but is uniquely related to the
coordinate system through the direction unit vector n. The corresponding
vector current is given by
(2.247)
where the vector In(Q) is the component of the vector l(Q) with respect to
the n direction and In(Q) is the projection of the vector l(Q) on the n axis
and also the magnitude of the component In(Q).
The three components of J(fl) in cartesian coordinates arc given by
1 x (Q) ~ iJ x (Q) ~ i (Q) (i' Q)

ly(Q) ~ jJy(Q) ~j (Q)
Q)
(2.248)
lz(Q) ~ kJz(Q) ~ k (Q) (k' Q)

andl(Q) is equal to the vector sum of the three components:
l(Q) ~ lx(Q) + ly(Q) + lz(Q)

~ iJx(Q)
+ jJy(Q) + kJz(Q)
(2.249)
It is important to recognize that, although they have the same units, the
scalar quantity flux density is not equivalent to the vector quantity current
density. The latter includes directional properties of the radiation field but
the former docs not. Current density should be considered as a measure of
the passage of particles through a surface (a density per unit area), and flux
density should be considered as a measure of track length in a volume (a
density per unit volume). Only in rare instances arc the two numerically
equal at a point in a radiation field. The following examples illustrate this
point.
A plane source S I (Fig. 2.8) emits monodirectional and monocnergetic
particles at the rate of] 0 1 0 particles cm- 2 sec- 1 in a direction normal to thc
surface. Let us compute the current density and flux density:
(a) 'At a point in a plane A whose normal nA is parallel to that of S,.
(b) At a point in a plane R whose normal nlJ is at an angle

to nA'
e with respect
38
--- ---------
~-----------~
---
5,
52
Fig. 2.B-Plane sources.
We proceed as follows:
(a) The flow of particles per unit time and area in the positive direction
through plane A is 10 10 particles cm- 2 sec-I. The flow rate in the
opposite direction is zero. Therefore the net current density, J, is in
the direction nA and is 10 10 particles cm- 2 sec-I. To obtain the flux

density, <1>, consider a unit cube (1 by 1 by 1 em) whose center lies in
A with two faces parallel to A; 10 10 particles/sec enter the cube, and
each generates a track length of 1 em in crossing the cube. Thus the
total track length generated in the cube per unit time is 10 10 em/sec,
and, since the cube volume is 1 em 3 , the flux density is 10 10
cm/sec -;- 1 em 3 := 10 10 particles cm- 2 sec-I.
(b) The net flow rate through a unit area of plane B is cos times that
through a unit area of plane A; i.e., the component of J in the
direction nB is J . nB, and the current density in the direction of nB
is cos X 10 10 particles em -2 sec -1. The flux density is not a
function of the direction nB and is 10 10 particlescm- 2 sec-I.
Suppose now that we add a second plane source, 52, also shown in
Fig. 2.8, which emits monodirectional particles with the same energy but in
a direction opposite to that of 51 at a rate of 6 X 10 9 particles cm- 2 sec-I.
We compute the current density in the direction nA and the flux density at a
point in plane A when both sources are emitting particles simultaneously.
The particle flow rate is 6 X 10 9 particles cm- 2 sec- I in the negative
10
direction and 10
particles cm- 2 sec- I in the positive direction. Thus the
net current density is 4 X 10 9 particles cm- 2 sec-I. The track length
contribution from 5, is 6 X 10 9 em sec- I cm- 3 The flux density is thus
1.6 X 10 10 particles cm- 2 sec-I.
39
2.3 SPATIAL AND DIRECTIONAL CHARACTERISTICS

A number of parameters are needed to characterize a sQurce. It may be
distributed in space and emit a variety of radiation types, its intensity may
vary with time, and the radiation it emits may be distributed in hoth energy
and direction of emission. Spatial and directional properties of typical
radiation sources are discussed in this section; important energy and time
distributions are considered in the following section.
2.3.1 Spatial Distributions

Radiation sources arc classified into four spatial categories, point, line,
surface, and volume, depending on the number of dimensions they
encompass. Often, however, a source of two or more dimensions is
considered as a configuration of sources of lower dimension. For example, a
volume source may be viewed as an aggregate of point sources distrihuted
throughout the volume.
The spatial variables describing sources arc independent of each other
and of other source parameters. Thus sources of any number of dimensions
and any spatial distribution within those dimensions can have independent
distributions in radiation type, direction of emission, energy, and time.
Direction, energy, and time distributions may vary from point to point in
the spatial distribution.
(aJ Point Source. A source emitting radiation from a single point 1Il
space ma y be considered the fundamental source configuration since all
other sources can be constructed from numbers of such point sources. Total
point-source strengths are measured in units of particles or MeV per second.
When point sources have distributions in direction, energy, and time,
differential distributions characterizing such dependencies have typical units
of particles sec-I MeV- 1 steradian-I. Differential point sources are of great
value in developing analytic functions of radiation-source and -field
properties. For example, at point r let S(r) be the emissions cm- 3 sec- 1 of a
volume-distributed soorce. A differential point source is then given by S(r)
dV particles/sec. If f(r; r') is a function relating a radiation-field property at
r' to a point source of unit strength at r, the field property generated at r' by
the entire volume source is given by
qr') ~
JJJ
source
volume
f(r; r'J S(r) dv
(2.31 )
40
In (act, functions relating two spatial positions. such as f above. play an

illlportant role in shield analysis. They arc called point kernc!s and afC
discussed more fully in Chap. 6.
A source need not be infinitesimal in size to be classified as a point
source. When a radiation field at a detector point is being evaluated. a
point-source approximation may be used to represent a volume or area
source if the source dimensions arc small compared to the source--dctcctor
distance. In making such an approximation for volume-distributed source
configurations, we must be careful to consider radiation interactions that
occur within the source volume (self-shielding). and adjustments should he
made if required.
(b) Line Source. A source with cmiSSlOll confined to a line is a line
source; total line-source strengths have typical units of particles or MeV per
second per unit length of the source. Like all sources, line sources can have
differential energy, angle, and time distrihutions that may vary along their
length. Typical units of differential distribution describing such line sources
are particles sec-] MeV- 1 steradian- 1 cm-\ or MeV scc- 1 MeV- 1 stcradian- 1
em-I. Examples of approximate line sources encountered in reactor shielding
problems include pipes carrying radioactive material and long, thin fuel pins.
(c) Surface Source. A surface source is one in which radiation emanates
from a plane or other two-dimensional surface. The units of source strength
are particles or MeV per unit time per unit source area (i.e., particles sec- 1
2
cm- ). Since surface source strengths have the same units as flux and current
density, care must be taken not to confuse the three LJ.uantities.
Surface-source strength is a scalar, specified in terms of a particle emission
rate per unit area of source surface. The need for such care is illustrated in
Sec. 2.3.2, where it is shown that a differential surface-source strength
function, although sharing the same units, may not be the same as the
differential function defining the flux density generated by that surface
source.
Surfaces that can be described mathematically are of high value in
specifying surface sources since many times functions describing such
surfaces admit to analytic integration and are convenient in hand and
computer numerical computations. Also, volume sources arc often
characterized by the radiation passing through their exterior surfaces. Such a
representation often allows the shielding analysis to be divided into two
phases: (]) determination of the radiation transport within the source,
resulting in the definition of an equivalent surface source. and (2) analysis of
41
the radiation transport external to the source. In such an approach the

calculations involved in phase (1) need not be repeated for shield-design
parameter studies, which arc a part of phase (2). Reactor-core and
pressure-vessel surfaces arc examples of often-used c<-]uivalcnt surface
sources.
(d) Volume Source. A radiation source distributed throughout a closed
surface constitutes a volume source. Typical volume sources encountered
Inlet
11
'0
'0
It
I
I
I
._---T--_ . . .
z = Zo
~,
/_-----1---I
pro,
z)
Outlet
=l~z
Zo
Fig. 2.9-Cylindrical volume source.
include cylinders, cubes, slabs, spheres, and other regular geometries in

which total source strengths arc expressed in particles or energy per unit
volume per second.
As an example of a common volume source, consider a cylindrical
reactor core whose power distribution is symmetrical about its axis and
whose radial and axial power distributions arc separab1et (sce Fig. 2.9). Let
p(r,O) and p(O,z) be the differential power distributions (watts/em 3 ) along
the radius at ::::: 0 and along the reactor-core axis, respectively. If K
particles/watt-sec are emitted in the fission process, the differential particle
source density, S(r,z), at (r,z) is given by
K
5(r,z) =--p(r,O) p(O,z) particles cm- 3 sec- 1

p(O,O)
tI( a function p(r,z) can be written PI (r) P2 (z), the VJl'i:lbles arc separable.
--l
(2.3-2)
42
and the total reactor source is
s = J;:o
1::0 S (r,z) 21fr dz dr particles/ sec
(2.3-3)
In morc complex systems, power-distribution functions describing

volume sources may not be separable. In such cases, if analytic expressions
are unavailable or unwieldy, a power map can be constructed, and
integrations can be performed numerically.
2.3.2 Directional Distributions

The directions of travel of radiation particles are another important
characteristic of radiation fields and sources. As suggested in Sec. 2.2.3 on
solid angle, particle directions are usually defined relative to the origin of a
conveniently chosen coordinate system. Once such a coordinate system is
established, a direction is specified by a vector, n, parallel to the direction of
travel and having unit magnitude.
Frequently, a radiation field or source is symmetrical in direction about
one or morc lines or planes in space, and it is usually convenient to choose
the coordinate system such that one or more axes lie in the plane or line of
symmetry. In such cases directions can then be specified by a single angle.
For example, Fig. 2.10 shows that two angles, 0 and <P, are required to
specify direction without symmetry. The angle 0 is called the polar angle and
the angle <P the azimuthal angle in such a system. If a field is symmetrical
about a line, aligning the z-axis with that line enables one to specify
direction by use of the polar angle alone, and the field is said to have
azimuthal symmetry.
Although the direction of n is always taken to be parallel with the
(
direction of travel of the radiation particles, there is no universal convention
\ that specifies whether n points toward or opposite to that direction of
travel. Thus care must be taken to specify accurately which is the case when
direction distributions are defined. For example, when the directional
properties of a point source are given, the coordinate system is usually
\
chosen to have its origin at the point, and n points in the direction of travel
of the emitted particles. However, when directional properties of a field at a
, point detector are given, the origin is again situated at the point, and thus
\ the direction of n is opposite to the direction of travel of the particles
\arriving at the detector point.
As with the other properties of fields and sources, directional properties
are specified by differential distribution functions. Direction functions are
40
43
sometimes separable from other distribution functions describing the source

or field, in which case, for example, the differential flux density, <1>, may be
written
(r,n,E,t) ~ 1'1 (r,E,t) 1'2 (n)
(2.34)
and direction attributes can be examined independently of other parameters.

In cases where the directional dependence is not separable and analytic
manipulation of the total differential function is not possible,
multidimensional maps may be constructed and numerical manipulations
performed to obtain directional characteristics.
z
In
I
I
I
I
o
~
-;Ir---Y
I
"
x
'..J
Fig. 2.10-Direction vector
n.
The following sections discuss some of the more common direction

distributions encountered in shielding problems. The exercises at the end of
this chapter provide additional insights into the definition and manipulation
of direction distribution functions.
(a) Isotropic Distributions. An isotropic direction distribution is one in
which all directions of travel are equally likely. Since there are 41T steradians
of solid angle surrounding a point, the normalized isotropic differential
distribution function is a constant function, 1/411' per steradian, independent
of angle or direction.
Many nuclear reactions that cause the emission of radiations are
considered to be isotropic in nature. Thus the neutrons and gamma rays
emitted from a fissioning nucleus, fission products, activated nudei, and
44
electron positron annihilations may be assumed to be isotropic. However,

certain reactions, nlOst notably those which involve a scattering process, arc
not isotropic, and assuming so can lead to significant errors (sec Chap. 3).
Point, surface, and volume-distributed sources can be isotropic. When
analyzing isotropic surface sources, we must he careful to ascertain if the
quuted source strength applies to one or to both sides of its surface. Suppose
c cos 0
-
rc
Fig. 2.ll-Point isotropic sources in a plane.
we arc g1Ven a plane source that isotropically units Sa particles scc- l cm- 2 .
In some texts, Sa applies only to the 21f space ;lbovc the plane; in others, Sa
applies to 47T space. The two interpretations of Sa differ by <l factor of 2. We
will choose the latter alternative and assume, unless stated otherwise, that 47T
emission is intended.
Isotropic point sources distributed in a finite volume do not necessarily
generate equivalent isotropic surface sources on the boundaries of the
volume since particle interactions within the source medium may change the
directional distribution of the emerging particles. Also, isotropic surface
sources do not generate isotropic flux densities at detectors removed from
the source plane, as shown by the following example.
Consider a plane isotropic surface source approximated by eClually
spaced isotropic point sources spaced c centimeters apart, as shown in
Fig. 2.11.
As the polar angle
is increased, tracks of particles from the surface
move closer together and consequently have a flux density that is higher by a
45
factor of l/eos () than the particles emitted nonnal to the surface. The
angular distribution of the flux density above such an isotropic surface
source is given by t
1
<1>(8)::: 4
'IT C
S
cos 0
particles scc- 1 stcradian- 1
Cill
-2
(2.35)
where S/c 2 is the total surLlcc source strength (in particles scc- 1 crn- 2 ).
(h) Cosine Distribution. Radiation emerging froln the surface of a
volume-distributed source often depends on the cosine of the ;111glc between
the normal to the surface and the direction of emergence. In many cases the
dependence closely approximates or is exactly a cosine distribution of that
e
dS
Fig. 2.12--Cosine distribution.
angle. For such a source, if Sa particles scc- 1 cm- 2 is the source strength, the
diFFerential source anglc distribution function Idemonstrated in Sec.
2.3.2(d) I is (1/27fJSa cos 0 particles sec-I steradian-I cm- 2 emitted along a
direction inclined at all angle 0 to thc normal (Fig. 2.12). As beFore, the
tracks in direction 0 are closer together than those which are normal to the
surface (0 = OJ, and we must divide by cos 0 to obtain tile flux density. The
flux density at r is then
(cJ Other Distributions. Various other analytical expressions have been

derived to fit observcd or expected angular emittance patterns. SurFace~angle
functions of higher powers of cos
have been used to approximate the
tProuf is given ill l'roblelll 2.(,.
46
leakage characteristics of a volume-distributed self-absorbing source whose

activity per unit volume increases with depth into the source. If the
differential angle surface-source function of such a source is proportional to
cos n (), the differential flux density above the surface is proportional to
cos n - I (). However, in some cases the complexity of the leakage pattern may
prohibit use of an analytical approximation, and numerical techniques must
be employed.
In the treatment of radiation reflection from a medium, the intensity
varies with the azimuthal angle cP between the incident and reflected
directions as well as with the polar angle cPo In such cases the angular
distribution will be a function of the variables () and cP, as shown in Fig. 2.13.
n
Reflected particle
Incident particle
Fig. 2.13-Reflection angles.
The intensity in a given direction is still reported per steradian (or other
designated increment), but both angles are required to designate the
directional properties of the scattered radiation from a beam incident at a
specified angle.
(d) Normalization and Monodirectional Approximation. When an
analytic expression has been used to represent the variation of emittance rate
with angle, the expression may be normalized to the total source strength,
So, by the relation
So
= In
KG (n) dn
(2.3.-6 )
in which G( n) is the functional variation of the source with angle and K is a

normalization constant. Thus, for a cosine distribution from a plane surface
source, G (n) = cos (), and
47
(20
(0/2
50 = J=o Jo=o K cos O(sm 0 dO
d
q
(2.3-7)
50 = 2TrK particles per unit time per unit area, and K = 5 0 /2Tr. The source
differential angular distribution function is then given by

5(0) = K G(n)
cos 0 partie
. Ies (unit, t1tne
. )-1 stera d'Ian- I (
' arear I
= 50 ~
umt
(2.3-8)
Occasionally other powers of the cosine are useful in describing an angular

variation. For G(n) = cos n 0 (where n is an integer), the normalization is
(2.3-9)
and
K= (n
+ 1)5 0
4Tr
5(0)=
---- -
(n + 1)5
cosnO
4Tr
--
(2.3-10)
In a like manner we can show for an isotropic distribution from a plane

su~so~e emit~ partie III pe~it time and_~ tha.u.;(nL:: 1,
lS.::....5 0HTr ...ansl
5(0) =
;~ particles (unit timer l steradian-I
(unit arearl
(2.3-11)
For a calculation of the transport of radiation with variable angular

distribution, it is sometimes convenient to assume monodirectional sources
with representative directions. The results are combined and normalized to
the incident distribution. Such an approximation assumes that the
differential distribution shown in Fig. 2.14(a) may be approximated by the
monodirectional beams shown in Fig.2.14(b). In normalizing these two
sources, we must assign representative solid-angle increments (L\n j ) to each
of the discrete sources and weight the monodirectional intensity by use of
the relation
N(nj) =
fAn., K G(n) dn
(2.3-12)
48
(al
Fig. 2.14-Angular distribution represented by array of monodirectional sources.
where IV(Q i ) is the monodircctional source strength of the ith source

pointing in direction rl j . The total source strength is given by
S
"
~ i~'
N([2,)
~ f",nKG(Q)
dQ
(2.3-13)
where n is the number of representative rnonodircctional sources used.
2.4
ENERGY DISTRIBUTIONS
Monoenergetic radiation fields arc rarely encountered. A few

monoenergetic gamnla-ray sources aTC available in the form of separated
isotopes with no daughter products. and monoenergetic neutrons arc
obtained from charged-particle reactions in accelerators. Even though only
one radiation energy is emitted by these sources, interaction with
environmental materials or within the source itself will lead to the presence
of radiation of various degraded energies in the radiation field.
Some is()tope sources emit two or three discrete energies: other sources,
such as fission, provide a continuum of evolved radiation energies within
certain energy ranges. In analytic treatments an energy-distributed radiation
field is often approximated as a series of monoenergetic sources. All the
radiation within an energy increment is assumed to have the same energy,
which may be chosen as the midpoint of the increment, the average energy
within the increment. or some weighted average based on a particular
interaction probability. This is analogous to approximations used in the
other distributions, such as the usc of a fixed direction to represellt an
angular increment, a point source to approximate a source-volume
increment, etc. Methods that are used to select the representative energy of
49
an energy increment are discussed in Sec. 2.2.2. In particular, the average
energy is often used, in which case the integral of Eq. 2.2-16 is carried out
over the range of the energy increment.
For many purposes an average energy will not suffice, and the actual
energy distribution must be carried through the analysis from source to
detector. The tetm spectrum IS used interchangeably

distrihution, following the practice of the spectroscopist.
with
energy
2.4.1 Energy Distributions of Gamma-Ray Sources

A few relevant characteristics of the spectra of gamma-ray sources
encountered in reactors were included in the discussion of the various source
types (Sec. 2.1.1). The two most important sources arc the prompt fission
reactions and fission products; their energy spectra arc discllssed in the
following two sections. Capture and inelastic neutron-scattering gamma-ray
spectral data ore given in Appendix A.

(al Prompt Fission Gamma-Ray Spectra. The spectrum of gamma rays
given off simultaneously with the fission of 2 3 5 U has been rather extensively
studied. The measurements of Pcel1e and Maienschein 2 are the most accurate
published data; they contain uncertainties of at most ] 5(X), and, in most
energy regions, the uncertainty is less. Figure 2.15 presents the differential
energy distribution (photons fission- 1 MeV-I) measured by Peelle and

Maienschein 6 Kirkbride 7 found that the spectra for 233 U and 234 Pu were
not significontly different from that for 235 U. Even 252 Cf. which fissions
spontaneously, exhibits a prompt gamma-ray spectrum very similar to that
of 235U (Ref. 8). The spectrum of Fig. 2.15 may be approximated by the
segmented fit:
f(li)
6.6
0.1 < Ii < 0.6 MeV
20.2 exp (-1.781')
0.6<1'< 1.5MeV
7.2 exp (-1.091')
1.5 < 1'< 10.5 MeV
(2.4-1)
The equation agrees with the experimental spectrum plus uncertainty to
within 10% from 0.6 to 7.5 MeV except at 1.2 and ~ 5.0 MeV, where it
deviates ~16%. The constant yield agrees with experiment from 0.1 to
0.6 MeV to within -20% except at 0.26 MeV. An approximation for the
energy region from 0.01 to 0.6 MeV, which may be adequate for shielding
50

10'
5
2
10'
5
>
~
::;;
10
'0
0
]
~
"'"
10- 1
5
2
10- 2
5
2
10- 3
0
23456
GAMMA-RAY ENERGY, MeV
Fig. 2.IS-The energy spectrum of gamma rays emitted within 69 nsee after fission of
U by thermal neutrons. The two lines, which represent the random (largely
propagated from counting statistics) 2/3 confidence limits on the spectrum, are drawn as
straight lines between adjacent mean window energies. The nearly Gaussian shapes shown
at the lower left and upper right indicate the energy resolution (From Peelle and
Maienschein. 6 )
235
calculations, is the emission of 3.75 I?hotpns/fission of average energy

0.324 MeV.
Although the approximation of Eq. 2.4-1 is useful, for most applications
accuracy demands the use of a numerical representation of the spectrum
IV-ther than these less-precise analytical functions. photon yields per fission
in fine and broad energy groups are given in App.'ndix A.
-J~
51
10 '0 ,.....---,_-,-_-,-_,---._.-_,.-_
Decay by isomeric transition
f-
:!'
X
a:
""-
10-11 L.,;--l"'7-'-7"""-,-:,-J..".-l.l-:;-LLf--';:---l
1~1~1~41~21~1021041~1~
TIME AFTER FISSION, sec
Fig.2.16 Gamma-ray energy release from fission of 235U as a function of time after
fission. The so-called prompt fission gamma rays afC emitted at times much shorter th,lTl
those shown on the graph. (From F. C. Maicnschein, lirlJ{ineerin)? Compendium on
Radiation Shieldinft, Vol. 1, p. 76, Springer-Verlag, New York, 1968.)
(b) Fission-Product Gamma Rays. Analysis of fission-product decay is

often complicated by the time history of fission-product generation. Data on
the decay of products of simu ltaneous fissions arc of interest in the analysis
of nuclear-weapon fallout or of a short reactor power burst. of more general
interest in reactor shielding is the decay of fission products built up during
sustained operation or over complex operating cycles.
Two types of events contribute gamma rays to fission-product decay.
8
Between 10- and 10- 3 sec aftcr fission of 2 3 5 U, decay of isomers in excited
states to the ground states contributes most of the decay gamma rays. After
1
about 10- sec, beta decays of the unstable nuclei contribute most of the
decay gammas. Figure 2.16 illustrates the relationship of these processes.
In an operating reactor the primary importance of the gamma rays from
early decays is their contribution to the steady-state cnvironment. Decay
gamma rays reach saturation very soon after start-up. Gamma rays from
these short-lived isotopes constitute the majority of all fission-product
gamma rays; about 75% arc released within 10 3 sec after fission. If the
52
fission-product gammas are integrated over time, the magnitude and shape of
the energy distribution is close to that for prompt fission gammas.
Maienschein 9 gives
N(E)
= 7.4 e-I.l DE photons fission-!
MeV- l
(2.4-2)
as a simple, though approximate, expression for the fission-product

spectrum. Equation 2.4-2 corresponds to an estimated total fission-product
gamma-ray energy release of 6.8 1.0 MeV/fission, which is composed of
5.9 0.7 MeV/fission for 1 sec < t < 10' sec and E > 0.28 MeV;
0.6 0.6 MeV/fission, E < 0.28 MeV; and 0.3 0.2 MeV/fission for t < 1 sec.
As noted in Sec. 2.1.1, this component is often combined with the prompt
fission gamma rays.
Measurements of the gamma radiation evolved by decay of the fission
products between 0.2 and 45 sec after fission were made by Engle and
Fisher.! D Listed in Table 2.1 are the values of the total radiation integrated
over energy and over the time period from 0.2 to 45 sec for four fissionable
isotopes, Comparison of these values with the totals from the long-lived
fission products shows that a significant fraction of the fission-product
energy is evolved within the first 45 sec following fission.
Table 2.1-RESULTS OF DELAYED GAMMA SPECTRA
INTEGRATED OVER ENERGY AND TIME FROM
0.20 TO 45.0 SEC AFTER FISSION' 0
Isotope
Photons/fission
MeVlfission
MeV/photon
233U
2.02
3.31
5.50
3.26
1.97
3.18
5.08
2.86
0.975
0.961
0.924
0.877
23S
238
239pu
The energy spectra for five time intervals are shown in Fig. 2.17 for
fission of 23 S u. Engle and Fisher l D give similar data for other fissionable
isotopes, 2 3 U, 2 3 3 U, 23 2 u, and 2 3 9 Pu. It is worth noting that the spectra
for these isotopes vary considerably from that of 2 3 S U (Y2 to 2 times) in this
early time interval.
The long-lived (late) fission products are unimportant during operation
) of the reactor, but they represent a significant source of radiation after
shutdown, particularly for a reactor core with many hours of operating
\'history. From a knowledge of the direct yield and half-life of each fission
53
L~~-'-l 0.20 - 0.50 sec
1.0 -2.0 sec
4.0 - 5.5 sec

10.0 - 13.0 sec
35.0 - 45.0 sec
105':-_....l._ _-!-__-!-_-..l__....L_ _~-___,':_---...J

o
234567
PHOTON ENERGY. MeV
Fig. 2.17-Photon spectra of gammas for different time intervals following

(From Engle and Fisher, I 0)
235U
fission.
fragment and the intensity and energy of each emitted gamma ray. along
with similar data for each daughter isotope in each decay chain, it is possible
to calculate the energy and time distributions of fission-product gamma rays
as a. function of reactor operating time. In this manner the important gamma
e~s ~h time an>! energy interval can be identified, and more
importantly a source term for the shutdown reactor core may be obtained
for any~operating history. Although the task is onerous and exacting. a
number of investigators have tackled it with results that compare increas-
54
ingly better with measurements as the number of identified decay chains

increases. Perhaps the most authoritative work was published by Perkins and
King' I and later updated by Perkins. 1 2 The results of the later work arc
given in Appendix A. For instantaneous fission and 1,10,100, and 1000 hr
of reactor operation, these data give disintegration rate, beta-ray energy
release, and total gamma-ray energy release and further subdivide the gamma
rays into seven energy intervals. Unfortunately, the uppermost energy
in terval contains everything above 2.6 MeV. Results calculated by Scoles I 3
avoid this shortcoming since his energy intervals arc 1 MeV wide, except the
topmost, which contains only one linc at 5.4 MeV. These results were used
considerably for some years since they were uniquely suited to shielding
problems for times to 10 6 hr of reactor operation and to 10 4 hr after
shutdown. Perhaps a reader of this text willupdatc this work.
For approximate calculations the total fission-product decay rate is given
within 20% from 10 min to 30 days after fission by Goldstein 14
(2.43)
where t is the time (in seconds) after fission.
2.4.2 Neutron Spectra from Fission

The energy distribution of neutrons from fission is obviously onc of the
principal inputs in the preparation of the source term for a reactor. The
energy range of importance in shielding for thermal fission in 2 J 5 U, 3 to

17 MeV, was measured by Watt, 15 who also reviewed measurements by
others in lower ranges and proposed an empirical expression to fit the data
from 0.075 to] 7 MeV:
N(E) = 0.484e -E sinh (2El'''' neutrons MeV- l fission- l
(2.44 )
where N(l;) is the fraction of neutrons per unit energy interval emitted per
fission and E is neutron energy in MeV. The Watt fission spectrum was
widely used until Cranbcrg, Frye, Nercson, and Rosen 16 rcported new
measurements from 0.18 to 12 MeV.
These results were based on time-of-flight measurements to about 8 MeV
and photographic emulsion exposures to 12 MeV. Cranberg et a!. reported
that
55
v' N(E) = 0.453e- E/o.965 sinh (2.29E)\I neutrons MeV-I fission- I (2.4-5)
was a morc accurate fit over the entire range.
Note, however, that uncertainties in the measurements, 15% or less to
8 MeV, were 30% or more at 12 MeV and above. Equation 2.4-5 is plotted in
Fig. 2.18 and tabulated for numerical use in Ref. 14. The tabulation
10- 2
!!! 10.3
Z
10'"
10. 5
10.6
0
12
15
18
E, MeV
Fig. 2.18-Fraction of neutrons per MeV interval emitted at energy E from the thermal
fission of 235 U . (From Herbert Goldstein, Reactor Handbook, Second Edition, Vol. III,
Part B, Shielding, E. P. Blizard (Ed.), p. 19, Interscience Publishers, a division of John
Wiley & Sons, Inc., New York, 1962.)
also gives the fraction of neutrons above E and the energy per fission carried
by neutrons above E.
An even simpler expression that is within 15% of Eq. 2.4-5 over its range
of validity (and within 7%, 5 < E < 13 MeV) is due to Goldstein: 14
N(E)
= 1.75e-O.766E
neutrons MeV-I fission-I
(4<E< 14 MeV)
(2.4-6 )
This form is very convenient in analytical manipulations and adequately

covers the energy range of greatest interest in many reactor shielding
problems.
Both the spectrum and the total number of neutrons evolved in fission
vary with the energy of the incident neutron and with the species of
fissionable material. In experiments at Los Alamos, Grundl and Neuer l 7 and
56
Leachman 18 compared the spectra from the fission of 235 U, 233 U, and
239PU. Using the average energy, E, as a parameter, they represented the
spectra from all three isotopes by the formula
N(E)
IX
E'" e- li
-=EE
(2.4-7)
From correlations of these and other data, TerreiP 8 found that E may be
expressed in terms of V; the average number of neutrons given off per fission,
by
E'" 0.78 + 0.621
(V + I)'" MeV
(2.4-8)
Values ofl' for fissions resulting from thermal-neutron absorption in various

isotopes are listed in Table 2.2.
Table 2.2-NUMBER OF NEUTRONS PRODUCED BY FISSION

Fissionable isotope
2.54 0.04
2.46 0.03
2.88 0.04
The value for j) increases almost linearly with energy as the incidentneutron energy rises above thermal. For 23 S U
v(E) "" vT + 0.15(E - E T )
(2.4-9)
where "T is the value at thermal energy, E T , and E is the energy of the
neutron-producing fission. From Eqs. 2.4-8 and 2.4-9, it is seen that the
average emitted neutron energy, E, increases at a rate of about 4% of the
increase in incident-neutron energy. For most practical shielding problems,
thi~ effect may be safely ignored. Variations between fissionable isotopes are
.~more significant, however, and should be considered. The values ofv given in
Table 2.2 indicate that a breeder reactor whose power is. evolved primarily
from 23 9pU fissions will be a 20% stronger neutron source than an
equivalent power reactor employing 23 S U.
57
2.4.3 Effect of Medium on Spectra

Ot is evident that the shape of the neutron or gamma-ray energy
distribution depends on the type of source. However, it is not so obvious
that th_~'!ttenJlation.ErQJl~rti~e medium through ",hi"h th".radi'!!.ion is
tr'!!.'smj tted...also h.aJLe. a .-S!J"ong...JJJ.flu~ on the spectrllm.IIn fact, the
spectrum of the neutrons or gamma rays emerging from a reactor shield is
determined by the shield moderator and structure layers and not by the type
of fission. The detailed reasons for this will become apparent from the
discussion on interactions in Chap. 3 and on transport in Chap. 4. For the
present it will suffice to Ulustrate the importance of the medium for two
familiar media, water and lead.
Figure 2.19 shows the measured and calculated spectra from a
fission-neutron source after the neutrons have penetrated various thicknesses
10',--,----..,--,---,---..,---,
~10
o
.......__-N~ CALCULATION
em
100_L---7----7---~--_=_--_,':,__-_:'.
10
12
NEUTRON ENERGY. MeV
Fig. 2.19-Neutron spectrum vs. penetration through water measured in a direction

normal to the slab with a collimated detector. [From V. V. Verbinski, M. S. Bokhari, J. C.
Courtney. and G. E. Whitesides, Nuclear Science and Engineering, 27: 283 (1967).]
58
Source
energy
>
12.3 em
...J
W
0:
....>0;
....Zw
>-
0:
J:
::;
::;
"
10
11
12
E. MeV
Fig. 2.20-Differential energy spectrum for a IO-MeV plane monodirectional gamma-ray

source vs. penetration depth through a lead slab. (From Goldstein and Wilkins. 2o )
of a water shield. As the thickness increases, the lower energy neutrons

become less important, and the higher energy neutrons become more
important; Le., the higher energy neutrons comprise a higher percentage of
the total neutron-flux density.
59
Both the measurements and the calculations demonstrate the effect of

the water medium. In this case the calculation is performed by direct
integration of th~tzmann equation, a technique discussed in Chap. 4.
The effect of attenuation on photons from a plane monoenergetic source
through various thicknesses of lead is shown in Fig. 2.20. In this illustration
the initial source spectrum is a vertical bar at 10 MeV. Despite this the
distriburion is almost horizontal at 1.76 em. Note that a pronounced peak
develops between 2 and 3 MeV as the depth increases.
REFERENCES
1. N. E. Holden, M. R. Mendelson, and T. E. Dudley, Total Gamma Energy Release Due
to Thermal-Neutron Fission in 235 U, Nucl. Sci. Eng., 30: 461 (1967).
2. R. W. Peelle and F. C. Maienschein, Prompt Gamma Rays from Thermal Neutron
Fission of Uranium-235, Nucl. Sci. Eng., 40"485 (1970).

3. P. N. Stevens and H. C. Claiborne, Basic Concepts of Radiation Shielding Analysis, in
Weapons Radiation Shielding Handbook, Chap. 2, USAEC Report DASA-1892-5, Oak
Ridge National Laboratory, June 1970.
4. International Commission on Radiation Units and Measurements, Radiation Quantities and Units, Report ICRU-ll, 1968.
5. International Commission on Radiation Units and Measurements, Neutron Fluence,
Neutron Spectra and Kerma, Report ICRU-13, 1969.
6. R. W. Peelle and F. C. Maienschein, The Absolute Spectrum of Photons Emitted in
Coincidence with Thermal-Neutron Fission of Uranium-235, USAEC Report ORNL4457, Oak Ridge National Laboratory, 1970.
7.J. Kirkbride, The Prompt Gamma photons from 235U, 233U, and 239pu Due to the
Absorption of a GLEEP Neutron Beam, British Report NRDC-58, Feb. 13, 1955.
8. A. B. Smith, P. R. Fields, and A. M. Friedman, Prompt Gamma Rays Accompanying
the Spontaoeous Fission of 2"Cf, Phys. Rev., 104: 699 (1956).
9. F. C. Maienschein, Engineering Compendium on Radiation Shielding, Vol. 1, p.76,
Sptinger-Verlag, New York, 1968.
10. L. B. Engle and P. C. Fisher, Energy and Time Dependence of Delayed Gammas from
Fission, USAEC Report LAMS~2642, Los Alamos Scientific Laboratory, July 1962.
I1.J. F. Perkins and R. W. King, Energy Release from the Decay of Fission Products,
Nucl. Sci. Eng., 3: 726 (1958).
12.J. F. Perkins, Decay of 235U Fission Products, Report RR-TR-63-11, U. S. Army
Missile Command, Redstone Arsenal, July 1963.
13.J. F. Scoles, Calculated Gamma-Ray Spectra from 235U Fission Products, Report
NARF-58-37T (FZK-9-132), Convair, Aug. 29, 1958.
14. Herbert Goldstein, Sources of Neutrons and Gamma Rays, in Reactor Handbook,
2nd cd., Vol. Ill, Part B, Shielding, E. P. Blizard (Ed.), Chap. 8, Interscience
Publishers, a division of John Wiley & Sons, Inc., New York, 1962.
60
15. B. E. Watt, Energy Spectrum of Neutrons From Thermal Fission of 2 3 5 U, Phys. Rev.,
87: 1037 (1952).
16. L. Cranberg, G. Frye, N. Nereson, and L. Rosen, Fission Neutron Spectrum Of 235 U,
Phys. Rev., 103: 662 (1956).
17. J. A. Grundl and J. R. Neuer, Comparison of 235 U and 239 Pu Fission Neutron
Spectra, Bull. Amer. Phys. Soc., 1: 95 (Abstract N4) (1956).
18. R. B. Leachman, Neutrons and Radiations from Fission, in Proceedings of the Second
United Nations International Conference on the Peaceful Uses of A fornie Energy,
Geneva, 1958, Vol. 15, p. 331, United Nations, New York, 1958.
19. V. V. Verbinski, M. S. Bakhari,]. C. Courtney, and G. E. Whitesides, Measurements
and Calculations of the Spectral and Spatial Details of the Fast-Neutron Flux in Water
Shields, Nucl. Sci. Eng., 27: 283 (1967).
20. H. Goldstein and J. E. Wilkins, Calculations of the Penetration of Gamma Rays,
USAEC Report NYO-3075, Nuclear Development Associates, Inc.,]une 30, 1954.
EXERCISES
2.1 The survivorship function for the time dependence of radioactive decay is given by
2.2
2.3
2.4
2.5
where N(t) is the number of radioactive atoms existing at time t, No = N(O), and Ais
a constant for a given isotope. (a) Find the differential time distribution function
p(t), where p(t) dt gives the number of atoms disintegrating in interval dt. (b) Using
the definitions given in Sec. 2.2.1., obtain the cumulative distribution F I (t)
(Eq. 2.2-4) and the fractional distributions I, (t) (Eq. 2.2-7) and I, (t) (Eq. 2.2-8).
(c) Sketch these distributions. (d) Compute the number of atoms disintegrating
between times T I and T 2
(a) Compute the solid angle
subtended by a circular disk of radius a at a point P
that is located a distance R from the disk, where P lies on the normal n passing
through the center of the disk. (b) As R -,l> 0 or a -,l> 00, n -,l> ?
An isotropic surface source of Sa particles sec~1 em -2 is located on the surface of a
sphere of radius R. Assuming a nonabsorbing medium within the sphere:
(a) compute the flux density at the center of the sphere; (b) compute the net
current density at the center of the sphere through a midplane.
Compute (a) the flux density and (b) the current density for the geometry of
Problem 2.3 when the surface source is located only on the hemisphere above the
midplane.
In the geometry shown by the following sketch, the plane P is parallel to the
isotropic disk source 5 I, which emits Sa particles sec -1 cm""2. Point source
52 is isotropic and emits 5 particles/sec. Compute (a) the flux density and (b) the
net current density through plane P at point D.
.n
D,
61
...: : ; ; ; - - _......_ _ n
Exercise 2.S-Disk plus point source.
2.6 Derive Eq.2.3-5 given in Sec. 2.3.2 for the flux density above a surface array of
isotropic point sources; Le., show
(8)
2.7
2S
1fC
cos
() particles em-2 sec-1 steradian-I
A slab of thickness X that is infinite in the other two dimensions (see sketch for this
exercise) is nonabsorbing and contains a uniformly distributed volume source of Sv
particles cm- 3 sec-I. At a point P on the surface of the slab, compute the following:
(a) Compute the surface source differential distribution G(O), where G(e) dD. is the
flux density emerging in solid-angle element dU ahout n. (Q is inclined at an angle ()
to the normal of the slab.) Compare the result with that of Problem 2.6
(b) Compute the differential flux~density disrribution <1>(8), where <1>(8) dS is

particles sec-I steradian-I emitted into dn about n from the slab surface element
dS. (c) Compute the total surface flux (particles sec- t cm-2 ) emerging from the slab.
(d) What is an equivalent isotropic flux? (e) Compute the differential angle flux
density H(8), where H(8) d8 is the flux density of particles whose directions lie in
angle element dO about n, integrated over all 21f azimuthal angles. (f) Compute the
relative total flux density in the following angle intervals of 8: 0 to 10, 40 to 50
75 to 85, 80 0 to 90.
t----x
.>t---1-f--.n
Exercise 2.7-Nonabsorbing slab source.
62
2.8 A differential energy distribution for the fission spectrum is often chosen to be
~
because it is simpler than Eq. 2.4-7 and accurate within 12% for E <9 MeV. For this
distribution, compute: (a) the most probable energy, E; (b) the average energy, E,
and its associated speed, vE; and (c) the average speed. v, and its associated energy
Ejj. Recall that for E in MeV and v in cm 2 /sec,
E = 0.525 X 1O- 18 V 2
Interactions of Radiation
with Matter
C. W. GARRETT, W. E. SELPH, P. N. STEVENS,and H.C. CLAIBORNE
Once the source of radiation has been properly defined for a particular shield
analysis, the second phase of the analysis can be performed. This phase
involves the calculation of the intensity and distribution of the radiation that
penetrates the shield or, alternatively, the attenuation afforded by the shield.
Although radiation attenuation is simply the macroscopic ramification of
the microscopic interaction processes between the particle and the atoms of
the medium, the transition from the individual processes to the gross
behavior of the radiation field is a complex one. Not only must some
consideration be given to the statistical nature of these processes but also the
relative importance of the various interactions must be considered for the
type of radiation and the appropriate range of energy(Thus an understanding of the energy-transfer mechanisms involved in the various interactions
and of the relative probabilities among the several possible interactions is
fundamental to shield analysis)
To begin a discussion of interactions, we will review the concept of cross
section, the basic measure of a radiation interaction. Then we will review the
atomic and nuclear processes important to photon and neutron transport and
to attenuation calculations. We conclude the chapter with a discussion
of some important interaction rates, namely, absorbed dose rate, kerma
rate, and exposure rate, which are often used to characterize radiation
fields. As in Chap. 2, for convenience we use the word particle to include
photons as well as particles with mass.
3.1 CROSS SECTIONS
Consider a particle traversing a medium; numerous factors influence the
probability of interaction with the nuclei and electrons of the medium.
Some of the more important (but not necessarily independent) factors
63
64
include (1) nuclides in the medium, (2) partial densities of each nuclide,
(3) electron density within the medium, and (4) energy and direction
distributions of the incident radiation.
Other factors must sometimes be considered as well. For low-energy
particles, particularly thermal neutrons, the physical state of the medium
may be important; for example, a nucleus bound in a crystal lattice may
have a different interaction probability associated with it than an identical
but unbound nucleus. Further, although we usually assume in shield analyses
that target nuclei are at rest, thermal motion of the nuclei can sometimes be
significant. For example, when the shield design involves the analysis of
thermal-neutron transport through high-temperature regions of a reactor, the
shift in value of the thermal-neutron energy must be considered.
3.1.1 Microscopic Cross Section
The basic measure of the probability that a neutron or photon will
interact with a nucleus, an electron, or an atom is a quantity called the
microscopic cross section. Microscupic cross sections are functions of all the
variables listed above and traditionally are denoted by the symbol a.
Subscripts are used to denote specific interactions to which a cross section
refers; e.g., (1a refers to an absorption cross section, ain to an inelasticscattering cross section, (1s to scattering, and at to a total cross section.
Functional dependencies are often shown; e.g., aa(E) represents the value of
an absorption microscopic cross section as a function of the energy, E, of the
incident radiation.
A microscopic cross section has units of area, hence the term cross
section. It may be visualized as the effective projected cross-sectional area of
a sphere centered about the target particle through which an incident
radiation particle must pass if an interaction is to occur. In general, this
effective area represents the range of the interaction force between incident
particle and target nuclei and is not directly related to the size of the
nucleus. The effective area for absorption (aa) may be very different from
the effective area for scattering (as) for the same target and incident
radiation.
Calculation and measurement of microscopic cross sections have been
active fields of endeavor for many years, and extensive files of evaluated data
on computer tape serve as a primary source for input to radiation shielding
analysis. (jor a given target, radiation type, and interaction process,
microscopic cross sections are usually tabulated as functions of incident
radiation energy1..0metimes, however, cross sections are reported as average
INTERACTIONS OF RAD1ATION W1TH MATTER
65
values, the average being taken over a distribution in energy of the incident
particle. For example, thermal-neutron cross sections are reported in this
manner where the energy distribution of the neutrons is assumed to
approximate a Maxwell-Boltzmann distribution.

The body of information on nelltron cross sections has grown so rapidly
in the past ] 0 years that it is no longer feasible to revise and update
comptlations; they have become so large as to be practicable only through
the usc of a cornputer. The National Neutron Cross Section Center at
Brookhaven National Laboratory has concentrated more on developing

cross-section data files on computer tape than on updating publications (for
example, those cited in the section on neutron interaction later in this
chapter). Evaluated nuclear data files, abbreviated ENDF/B, arc available
from National Neutron Cross Section Center, Brookhaven National Laboratory, Upton, New York 11973.
If very current data arc not needed, the compilations of neutron cross
sections cited in the following sections should be helpful. Sources of
gamma-ray cross-section data are also cited.
Microscopic cross sections arc usually measured in units of barns (b),
milJibarns (mb), and microbarns (Jlb). A barn is ]0-'4 em'; a milliharn,
10-27 cm 2 ;andamicrobarn, ]0- 30 cm 2 .
3.1.2
Macroscopic Cross Section
Although microscopic cross sections deal with probabilities of the

interaction of radiation with individual targets (nuclei and electrons),
macroscopic cross sections are related to probabilities for interaction with
the aggregate of targets that compose the medium through which the
radiation is passing. Almost invariahly shield analysis recJuires macroscopic
cross sections that, as shown in the following text, are computed from a
knowledge of microscopic cross sections and the constituents of the
medium. Two symbols universally used to denote macroscopk cross sections
are Jl and 1:; the former is most ofren applied to photon cross sections and
the latter to neutron cross sections. As with microscopic cross sections,
subscripts and functional dependencies are often useo in conjunction with
these symbols.
If N is the target particle density (nuclei/em 3 ), the relation between the
macroscopic cross section, r:, ano the microscopic cross section, u, for a
specified interaction as a function of incident-radiation energy, n, is
2:(E)
N a(E)
(3.1-1)
66
Macroscopic cross sections as defined by Eg.3.1-1 thus have units of

reciprocal length (usually cm- 1 ) and are often called linear cross sections or
coefficients. (The term macroscopic cross section is used principally for
neutrons; following historical precedent in attenuation of light, the term
coefficient, with various modifiers to distinguish types, is used for gamma
rays.) In fact, such a macroscopic cross section is the total apparent
cross-sectional area of interaction presented to a radiation
fidel per unit
volume of material (i.e., cm 2 Icm 3 ).

For mixtures of target nuclides, the macroscopic cross section for a
specified interaction process is given by
"Z(E)=f.Nioi(E)
;
(3.1-2)
where N i is the volumetric density of the ith nuclide and 0i(Li) is the
microscopic cross section for the ith nuclide. It is easily shown that
where Au is Avogadro's number (6.023 X 10' 3), Ai is the atomic weight of

the ith nuclide, and Pi is the partial density (glcm') of the ith nuclide in the
mixture.
Macroscopic cross sections can be added. Thus, if Ls(l:) and La(Ij) are
scattering and absorption macroscopic cross sections, respectively, the cross
section for both processes taken together is L s (1i) + La(E).

If Eg. 3.1-1, 3.1-2, or 3.1-3 is divided by the density, p, of the medium,
the apparent cross-sectional area of interaction per unit mass of material
(cm 2 /g) is obtained. Such mass attenuation coefficients arc most often
encountered in photon shielding calculations and are usually denoted by /lIp,

J1 being the linear macroscopic cross section.
3.1.3
Radiation Reaction Rates
It was noted in
the discussion of measures of radiation intensity
(Sec. 2.2.4, Chap. 2) that interaction rates of a radiation field with its
environment are important and often-used characterizations of radiation
intensity. However, further discussion of the concept was deferred to this

chapter so that the various microscopic processes of interaction could first
be reviewed.
INTERACTIONS OF RADIATION WITH MATTER
67
Many interaction processes can be used to measure a radiation field. With

a photon field, for example, the total photon collision rate (number of
photons interacting by all processes per unit volume per unit time) can be
used. The rate of energy removal (MeV cm' 3 sec") from the primary photon
beam can also be used. A third example might be the rate of local energy
deposition (MeV cm' 3 sec"), which has the same dimensions but may not
equal the energy removal in the local region being considered because of the
energy carried away from andlor delivered to the local site by secondary
processes. Rates of these types are regularly used in the calculation and
evaluation of reactor shields; they are discussed later in this chapter. First,
however, some fundamental considerations arc in order.
The macroscopic cross section may be viewed as the reaction rate per
unit number flux density of a radiation field. Thus, for a linear macroscopic
cross section, the reaction-rate density is given by
RR
:=
/lcfl reactions em -3 sec- 1
(3.1-4)
if the cross section, fl. is in units of cm- 1 and the flux density, 4>, is in units
of cm-1 sec-1. of course, the reaction or reactions whose rates are calculated
by Eg.3.3-] are those associated with the particular macroscopic cross
section used.
In a similar manner, the reaction rate per unit mass density is given
RR ~
f.l. reactions g"

p
sec"
by
(3.]-5)
where is again the flux density and III p is the masS macroscopic cross
section. The total reaction rate over energy flux spectrum and volume, V, is
given by the integral

RR ~
5v 51' Il(F) (r,F) dE dv reactions/sec
(3.]-6)
which the differential flux density, <t>(r,E), is defined by Eg. 2.2-23 of

Sec. 2.2.4 and dV is the differential volume element.
In
3.2 RADIATION INTERACTIONS

The physics of importanr neutron and photon interaction processes for
radiarions having energies found in fission reactors is well known, and
68
thorough discussions can be found in nuclear science texts as well as in

specialized treatments such as those of Foderaro,' Blizard,2 and Goldstein. 3
We will summarize the basic interactions here and make reference to sources
of cross-section data for them.
3.2.1 Photon Interactions

[The energy range of interest for gamma rays associated with fission or
from fission-neutron interactions extends from the lQ..w...keV range to about
1Q-MeV. There are many possible mechanisms for the interaction of gamma
rays in this energy range, but only three interactions contribute in a
significant way to photon attenuation. These are the photoelectric effect,
pair ~ction, and Compton scatEering] Of the numerous possible
references to photon-interaction data, we will mention only two: Hubbell,'
a recent tabulation with summary discussions of the processes of interest,
and Hubbell and Berger,S an article prepared a year earlier and appearing in
the Engineering Compendium on Radiation Shielding.
(a) Photoelectric Effect. In the photoelectric process, a photon is
absorbed and an orbital electron is emitted. The entire energy of the photon
is transferred to the electron, which is ejected from its shell and emerges
from the atom as a photoelectron. Momentum is conserved by recoil of the
whole atom, and the more tightly bound electrons have the greatest
probability of being ejected. The photoelectron is ejected from the atom
with kinetic energy T = E y - Be' where Be is the binding energy of the
orbital electron. The orbital vacancy is then filled by transition of an outer
electron; this transition is~m~njyd by emission of characteristic X rays
called j1uores~radiation. Neither the photoelectron nor its fluorescent
radiation is sufficiently energetic to be of further interest in shield analysis.
Extensive tabulations of photoelectric cross sections are given by
Hubbell' and Hubbell and Berger. S Qualitatively the photoelectric cross
section depends on atomic number, Z, and incident gamma-ray energy, E-y,
(3.2-1)
where n varies from 3 to 5. We see that the probability of interaction is
proportional to the -3 power of the photon energy and to the third to fifth
power of the atomic number of the target atom{Jhus the importance of the
--------
_._._----~
69
photo~ic effect is greater fat _!he higher Z elements and diminishes

rapidly with insreasing photon energ): for any element.")
Since the K-shell electrons are the most tightly -o'ound, they have the
highest probability of a photoelectric absorption. For incident-photon
energy below the K-shell ionization energy, only L-shell or higher electrons
are available (with a lower probability for interaction). trhus discontinuities
in the cross sections for the photoelectric effect are found at the ionization
energies of the various electron shellsJFigure 3.1 illustrates this phenomenon
for lead; both the K-edge and the L-edge are shown.
10 5
"t~- edge
.D
,.:
zw
4
C3 10
u::
u.
w
K-edge
;:
0.
10'
II:
'"
III
120
PHOTON ENERGY. keV
Fig. 3.1-Photoelectric effect in lead.
The ionization energy for a given electron shell increases with increasing
atomic number; the energy of the K-edge as a function of Z is shown in
Table 3.1. Also shown in Table 3.1 is the photon energy at which the
photoelectric effect accounts for half the total photon interaction probability. At energies above the Ell value, the photoelectric effect diminishes in
importance. At energies below Ell' it becomes increasingly more important
(see discussion on total attenuation coefficients at the end of this section).
As shown in Table 3.1, the energies at which the photoelectric effect
predominates lie well below the region of 0.5 to 10 MeV, which is of greatest
concern in reactor shielding. Although not predominant, the effect is still
significant within that range. [However, the greatest significance of the
photoelectric effect is that it establishes a lower limit to the photon energy
requiring consideration in a shield analysis)Becaus~! the rapid increase in
70

Table 3.1-ENERGY AT WHICH THE PHOTOELECTRIC EFFECT PROVIDES ONE-HALF
THE TOTAL ABSORPTION COEFFICIENT
(E% l AND ENERGY OF K-SHELL
IONIZATION (EKlt
Z
I
4
6
8
Element
Hydrogen
Beryllium
Carbon
Oxygen
I3
Aluminum
20
26
42
50
74
Calcium
Iron
82
Lead
92
Uranium
Molybdenum
Tin
Tungsten
EK,MeV
E%,MeV
1.4
2.2
2.8
5.2
1.5
X 10- 5
X 10'
X 10'
X 10-'
X 10.3
10"
0.011
0.016
0.025
0.046
4.0
6.9
2.0
2.9
6.06
X 10- 3
X 10-3
X 10-2
X 10.2
X 10-2
0.079
0.11
0.195
0.25
0.42
8.8 X 10-2
11.6 X 10.2
0.50
0.62
t From Goldstein. 3
the I'hotoel"-llic absorption cross section wjth decreasing photon energy,.:

point in energy is reached b~low which, for all practical purpOSes. 0.0
photons exist.
(b) Compton Effect. (:rhe Compton effect is the scattering of a photon
by a free electronJThe photon imparts energy to the electron and is altered
in direction and energy. A very important feature of this effect is the fact
that, except when the scattering angle is latge, the photon emerges from the
interaction with a significant fraction of the incident-photon energy. This
fact accounts for much of the complexity associated with gamma-ray
transport analysis.1
In the analysis of the Compton effect, the electrons are assumed to be
free, neither interacting among themselves nor bound within the atom. The
Compton effect for an atom is therefore the additive effect of all its
electrons, and the ~acroscopic cross section for Compton scattering is
determined by the electron density)rhus the dependence of this process on
atomic number is merely a linear dependence on the number of electrons per
atom. Figure 3.2 illustrates the process. The photon is scattered through an
angle () and carries a portion of its incident energy. The remainder of the
energy is taken up as kinetic energy by the recoiling electron, which scatters
through an angle ljJ.
71
/Pho,on
Scattered
Incident
...
O"l~--
......
photon
~ReCOil
ek!ctron
Fig. 3.2-Compton scattering.
The equations describing the angle and energy relations of the Compton
process are most easily expressed when photon energies are measured either
in units of the rest mass energy of an electron (m e c 2 ) or in units of Compton
wavelengths. 111 these units, if E* is energy in units of MeV, the energy E in
units of m e c 2 is given by
E*
E = 0.511
(3.2-2)
and the Compton wavelength is given by the reciprocal:
x = .!.=O.511
E
E*
(3.2-3)
One unit of Compton wavelength is equal to hlmec, or 0.02426 A, where h

is Planck's constant (6.625 X 10'27 erg-sec). Subsequent equations in this
section on Compton scattering are written in terms of these units of energy
and wavelength.
From the conservation of energy and momentum, it may be shown that
the relation between the incident-photon energy, E, scattered-photon
energy, E', and scattering angle, 0, for a Compton collision is
E' =
1 +E(l- cosO)
(3.2-4)
In terms of Compton wavelengths, the relation seems even simpler
x' -
A = 1 - cos ()
where X' refers to the scattered photon.
(3.2-5)
72
From these equations it is seen that straight-ahead scattering (0=0)

results in no energy loss for the photon; the electron recoil angle, 1/1, is 90 in
this casc. A lower limit to the scattered photon energy occurs in backward
scattering (0 - 180), where
E
=-=1+2E
(3.26 )
A~ax::::A+2
(3.27)
F'
.
~mlO
and
The electron recoil angle, If; , is zero in this casco Thus the photon can scatter
in any direction, but electron recoil is linlited to forward (relative to the
initial photon's track) directions.

The energy of the scattered photon varies from the initial energy down
to the lower limit. For initial energies much greater than 1 rest mass
(>0.511 MeV), the lower limit is approximately E/2.
The essence of the process is contained in the Klein-Nishina formula,
which describes, to a very good approximation, the probability per electron
that a photon will Compton scatter into a unit solid angle about a scattering
angle, (J. This differential function takes on a convenient form when the
microscopic electron cross section is measured in Thomson units (T.U.). One
T.U. = 8"/3(e 2 /m e c 2 )2 = 0.665 b. In these units the Klein-Nishina relation, expressed in terms of energy or Compton wavelength, is
3 (Ii')2 (E--c;y+-Ii' sin' 0) T.U. clectron-
a(O) = -6- --":1 "E
Ie
steradian- t
(3.2-8 )
or
(,,)2 (A'"
); +?::' - sin' 0) T. U. e1ectron-
3 ,,'
a(O) = 16"
steradian-
(3.2-9)
lSince a(O) is a differential microscopic cross section (unit area per unit solid
angle), some authors prefer to denote these functions by da/dr2 to
emphasize the differential.] It should be noted that the variables E', E, and
sin 0 (or ,,', ", and sin 0) are not independent in the preceding equations
but arc related through Eq. 3.2-4 (or 3.2-5). Thus a(O) is actually a function
73
only of initial and scatteted photon energy or initial photon energy and
scattering angle. This is explored in the exercises.
The total cross section for Compton scattering by an electron may be
obtained by integration of Eq. 3.2-8 over all scattering angles.
o (Ii)
e
3 {1-+E [2Ii (1 + E)
li 3
1 + 2
= -
In(l + 2E) ]
1 + 3l! }
+ 2E In (1 + 2E) - (1 + 2E)2 . T.U./electron
(3.2-10)
where oc(E) is the microscopic cross section for Compton scattering per
electron for a photon of energy E. For very low energies (E <{; 1), oe(E)
approaches 1 T.U. [n the other extreme (Ii ~1),
0c(li) '"
8~
(In 2E
+~) T.U.Jelectron
(3.2-11)
As noted in the preceding text, the energy of the incident photon is

divided into two parts in a Compton collision: that contained in the kinetic
energy of the recoiling electron and thus deposited very close to the site of
the collision, and that carried away from the collision site by the scattered
photon. The fraction, /(8), of energy deposited locally (i.e., transferred to
the electton) as a function of scattering angle is given by
(3.2-12)
The average fractional energy loss per Compton collision, j~, is given by
1
Ie
= - f 1(8) 0(8) dn
c
(3.2-13)
where the integration is carried out over all solid angle. The quantity
(3.2-14)
a cross section reflecting the probability of local energy deposition in a
Compton collision and is called the Compto" e"ergy absorption crOSS
section.
IS
_ _l!
74
The relations between initial- and scattered-photon energies, scattering
angle, Compton cross section, and energy transferred to the electron are
sufficiently complex that they are not intuitively obvious. Figure 3.3
illustrates the relation between initial and scattered photon energies as a
function of scattering angle. For initial photon energies of 0.1 MeV or less,
there is little energy degradation. For larger initial energies, degradation is
large for large scattering angles. Figure 3.4 shows the dependence of electron
recoil energy on initial photon energy and scattering angle. Figure 3.5 shows
the differential Compton cross section, a(a), for several representative initial
photon energies. This graph shows that in the 1- to 10-MeV range the cross
section is peaked in the forward direction and demonstrates that Compton
scattering is highly forward. Figure 3.4 is taken from Nelms,' who presents
over 80 graphs showing the interrelations of these parameters with scales
sufficiently expanded to allow reasonable accuracy. Finally, Table 3.2 lists
the total Compton cross section per electron, 0c. as a function of initial
photon energy. Also given in this table are values of a ca and a c - a ca ; the
former cross section relates to the energy deposited by the incident photon,
and the latter is a cross section related to the energy carried off by the
scattered photon.
Some gamma-ray analysis methods require that the differential crosssection data be put into some format other than that just given. For
example, in many Monte Carlo calculations the integral probability
distribution for scattering at an angle ex ,;;; a is expressed as a function of a.
Such a function has the limits of 0 probability at ~ 0 and 1 at ~ 180.
All the more sophisticated methods require differential angle cross-section
data in some form.
(cJ Pair Production. In the process of pair production, a photon

interacts with the electric field of atomic electrons or the nucleus. The
incident photon is completely annihilated, and its energy is converted into
the mass of an electron-positron pair. Clearly, the incident photon must
have at least enough energy to create the mass of the electron-positron pair.
Thus a threshold for the pair-production process is 2m e c', or 1.022 MeV.
Any excess energy of the incident photon is given to the kinetic energies of
the two charged particles produced in the process and to the recoiling
nucleus. As in the photoelectric effect, the nucleus plays an essential part in
pair production since its recoil is necessary to conserve momentum in the
process.
For a given nuclide the nuclear pair-production cross section increases
rapidly from rhe threshold of 1.022 MeV to 10 MeV. This is illustrated in
7S
0.05
-----J0.l0
0.15
0.20
>
~
!!!
'"
0.30
'""
0.40
0.50
0.70
1.00
1.50
2.00
3.00
5.00
7.00
10.00
00
40
80
120
160
SCATTERING ANGLE. dog
Fig.3.3-Reduction in gamma-ray energy by Compton scattering (Eo

E = reduced energy).
= initial energy;
76

1 0 , - -.......=---,---,--...--,--..,..--,,...--...----,---,
1.0
,.0~
Cl
001
a:
w
z
w
Z
t....
w
8w
0.01
a:
0.001
000001 !---:7--:~-_:!;:_-_=-_!;;_-*-~;;--_;;;;-___;t--~
o
10
20
30
40
50
60
70
SO
90
100
ANGLE OF RECOIL ELECTRON, deg
Fig. 3.4-Recoil-electron energy vs. angle of recoil. (From Nelrns. 6 )
Fig. 3.6, which shows the dependence of the cross section on incidentphoton energy in lead. As a function of atomic number, pair production for
interactions with atomic electrons is proportional to Z, and that for nuclear
pair production, to Z2. This latter effect increases the dominance of nuclear
pair production in reactor shield marerials. In fact, below lOMeV the
probability of pair production with atomic electrons ranges from 10 to 30%
of that for nuclear pair production in hydrogen and is negligible in high-Z
materials such as lead.
10.1 ..-
-,-
,-
-,-
77
,----,
0001
0.05
0.10
0.15
0.20
0.30
~
w
en
0.40
l:l
oa:
0.50
'a:z"
0.70
>
:;;
;0
>-
'wa":
::
1.00
"
en
zW
..J
"
1.50
:I:
!!l
2.00
..J
2.50
><
..J
";::
3.00
a:
4.00
w
w
LL
LL
5.00
is
6.00
7.00
8.00
10-3~----f=----_:!::---_:_!:;;--.......::::;:;~--'
10.00
40
80
120
160
SCATTERING ANGLE. deg
Fig. 3.5-Differential Compton scattering cross sections.
"E
z
78

Table 3.2-TOTAL AND ENERGY-ABSORPTION CROSS
SECTIONS FOR THE COMPTON EFFECTt
E,MeV
Gel
b/electron
U ca
' b/elcctron
Uc
- u ca b/electron
0.0
0.01
0.015
0.020
0.030
0.665
0.637
0.627
0.616
0.596
0
0.0077
0.0138
0.0196
0.0295
0.665
0.629
0.613
0.596
0.566
0.040
0.050
0.060
0.080
0.10
0.578
0.561
0.546
0.517
0.4929
0.0380
0.0451
0.0509
0.061
0.0685
0.540
0.516
0.459
0.456
0.4244
0.150
0.200
0.300
0.4
0.5
0.4436
0.4066
0.3535
0.3167
0.2892
0.0812
0.0886
0.0958
0.0982
0.0984
0.3624
0.3180
0.2577
0.2185
0.2866
0.6
0.8
1.0
1.5
2.0
0.2675
0.2350
0.2112
0.1716
0.1464
0.0984
0.0959
0.0929
0.0849
0.0777
0.1691
0.1391
0.1183
0.0867
0.0687
3.0
4.0
5.0
6.0
8.0
0.1151
0.0960
0.0828
0.0732
0.0599
0.0664
0.0582
0.0519
0.0471
0.0399
0.0487
0.0378
0.0309
0.0261
0.0200
10.0
15.0
0.0510
0.0377
0.0349
0.0268
0.0161
0.0109
tFrom Goldstein. 3
As discussed in the next section, the photoelectric process, the Compton

process, and pair production all compete for the absorption of the energy of
a photon field. However, rarely are photoelectric effect and pair production
simultaneously of importance. In low-7 materials, the photoelectric and
Compton processes are dominant. Only in iron and higher Z elements does
pair production account for over one-half the energy absorption for photons
with an energy less than 10 MeV. Table 3.3 lists the photon energy above
which pair production accounts for one-half or more of the total energy
absorptio'n.
79
15
10
9
B
7
E
0
j:
6
5
4
t)
W
'J>
'"Z
t)
ti
:0
0
0
'"'T
'"
;;:
~
o.5o1:----!-L--!2:---3!:--~4---!6~-~6---!7:---!B~--:9--~10
PHOTON ENERGY, MeV
Fig. 3.6-Pair-production cross section in lead. (From Hubbell.4
Hubbell4 and Hubbell and Berger S give discussions of the models used to
calculate pair-production cross sections and tabulations of both nuclear and
electron pair-production cross-section data for a wide range of photon
energies in many elements and several compounds and mixtures.
Although pair production results in the annihilation of the incident
photon and plays an important part in the attenuation of high-energy
photons, it also results in a secondary photon source that should be
considered in some shield analyses. The positron created in the process
combines with an electron very close to the site of the pair production, and
both the positron and electron are, in turn, annihilated. This gives rise to two
O.511-MeV photons that emerge in opposite directions. For most purposes
this secondary source may be considered to originate at the site of the initial
pair-production event. This secondary source can account for as much as 5%
of the total energy deposited.
80
Table 3.3-ENERGY AT WHICH PAIR PRODUCTION

PROVIDES ONE-HALF THE TOTAL ABSORPTION
COEFFICIENTt
Z
Element
E.MeV
Element
E.MeV
I
4
6
8
13
20
Hydrogen
Beryllium
Carbon
Oxygen
Aluminum
Calcium
78
35
28
20
15
12
26
42
50
74
82
92
Iron
Molybdenum
Tin
Tungsten
9.5
7.5
6.5
5.2
5.0
4.8
Lead
Uranium
t From Goldstein. 3
(d) Other Processes. Several other photon inreractions, such as coherent

electron scattering, Thomson nuclear scattering, and De1briick scattering,
eXist;[however, coherent scattering, also called Rayleigh scattering, is the
only one that has any possible significance in reactor shielding situations]
These other interactions are summarized in Refs. 4 and 5.
Coherent scattering is similar to Compton scattering in that the incident
photon is scattered from an electron. However, in contrast to the Compton
process, the binding forces of the orbital electrons are significant in coherent
photon scattering. In this process the recoil momentum is assumed by the
whole atom with the result that the scattered photon emerges with an energy
that is, for all practical purposes, equal to that of the incident photon.
Further, in the photon energy range of interest for fission sources, the
photon scattering angles, f), are small, for example, less than ISo at O. I MeV
and 20 at 1.0 MeV in aluminum. The practical effect of coherent scattering is
to scatter a photon in the forward direction with no reduction in energy, and
thus it may be treated as though no interaction took place.
However, if coherent scattering is ignored, care must be taken with the
cross sections used in photon-attenuation calculations. Often the cross
section for coherent scattering is included in tabulations of total photonattenuation coefficients. Since such use would treat the coherent scatter as
an attenuating event when, in fact, it has essentially no effect on photon
attenuation, the coherent portion should be subtracted from the total cross
section. The tabulations in Refs. 4 and 5 list such data both with and
without the coherent portion.
(e) Total Photon-Attenuation Coefficients. The three processes that
contribute to photon attenuation in reactor shields are the photoelectric
81
effect, Compton scattering, and pair production. Each process causes some
or all of the energy of the incident photon to be deposited at the site of the
interaction, and the last two cause photons of reduced energy to be emitted
in new directions. Thus the total microsc9pic cross section per atom of
atomic number Z for photon attenuation is given by
(3.2-15)
where ope is the cross section per atom for the photoelectric effect,
Oze ~ ZOe' and opp is the cross section per atom for pair production. If we
include coherent scattering as well as Compton scattering, we replace ozc by
aeon' The term 0zc is sometimes written 0incoh to explicitly show that
coherent scattering is not included. As functions of photon energy, ope and

6r---.-,r--,------r-------,
5
4
.c
b
I
1
1
E
3
I
I
I
,, Photoelectric
\
Total
effect
\
\
,,
'-
Pair production .
o~=-----?':;:__-----,l::_-~=-"'--"'-:.:.-~-
0.01
0.10
1.0
10.0
E, MeV
Fig. 3.7-Total gamma-ray cross sections fOT'oxygen. (From Goldstein?)
decrease with increasing energy; 0p , however, increases as the

incident-photon energy increases. Thus for air elements a minimum occurs in
the total cross section at some energy. This behavior is illustrated in Figs. 3.7
and 3.8, where all three components and the total cross section are plotted
for oxygen and lead, respectively. The Compton effect is predominant for
intermediate energies, 1 to 5 Me V for lead and other high-Z materials and a
wider range for low-z materials. In hydrogen the Compton effect accounts
for the total cross section over the entire energy range of interest. As
suggested by these two figures, the point in energy where the minimum in
0zc
82
Fig. 3.8-Total gamma-ray cross sections for lead. (From Goldstein?)
100
50
>0
:2'
."E
10
5
3
25
50
75
100
Fig. 3.9-Energy of minimum total photon cross section. (From Goldstein. 3 )
total cross section occurs decreases with increasing atomic number.

Figure 3.9 is a plot of the location of the minimum as a function of z.
It should be noted that a mactascopic CtaSS section computed ftam the

total microscopic cross section of Eq. 3.2-15 is a total interaction coefficient
and includes mote than energy deposition by virtue of the fact that the
entire Compton cross section is included. Thus these macroscopic cross
sections are sometimes called narrow-beam absorption coejJtcients since all
ptacesses, including Compton events, effectively remove photons ftom a
83
collimated beam. However, they are more commonly called attenuation

coefficients. Thus J.l.t(= Nat) is called a linear attenuation coefj,cient, and
J.1t!p, a mass attenuation coefficient. The total interaction rate is sometimes
called the reaction rate. The reaction rate per unit volume is simply the
product <PJ.l.t.
Another macroscopic cross section, which actually reflects the local
energy removed from a photon field by an element of atomic number Z, is
given by
(3.2-16)
where N is the atomic density (atoms/cm 3 ) of the element and a ca ' the
Compton energy absorption cross section per electron, is defined by
Eq.3.2-14. The quantities J.l. a and J.l.a/p are known as linear and mass
energy-deposition coefficients, respectively.
The fact that electrons are reaction products of each of the three
gamma-ray interactions important to transport through shields deserves one
further comment. Bremsstrahlung (Sec. 2.1.1) is produced by the decelerating electrons and represents an additional form of the initial gamma-ray
energy. In this book we will usually assume that this radiation is absorbed at
the point of formation, with one important exception, which is discussed in
Sec. 3.3, Responses to Radiation.
3.2.2 Neutron Reactions
In several respects the processes associated with neutron attenuation
contrast sharply with those associated with photons. For example, only
three photon processes are significant at photon energies found in fission
reactors, but many neutron reactions must be considered. Also, although all
photon reactions of interest involve electrons, no neutron reactions are with
electrons; they are all with nuclei. Most cross-section data for photons are
accurately obtained by calculation (well verified by experiment); exactly the
opposite is true for neutrons. The processes associated with nuclear
interactions are far more complicated and, in many cases, are not well
understood. Thus most neutron cross sections must be measured experimen-
tally. Active experimental programs have been under way for many years;
however, important gaps in certain data still exist. With the exception of the
K-edge, L-edge, etc., in the photoelectric cross section, photon-interaction
probabilities are smooth functions of both energy and atomic number. In
many neutron cross sections, however l resonance and threshold phenomena
84
REACTOR SHIELDING FOR NUCLIlAR ENGINEERS
cause abrupt excursi.ons'LFurther, neutron cross sections are the ramification
of complex interactions of nuclear forces, which depend on several factors in

addition to simple atomic number and weight. Thus interpolations are in
many instances impossible, and neutron cross sections must be determined
nucleus by nucleus;IIt should be noted, however, that for certain neutron

reactions theoretical considerations are effectively used in making interpolations. Foderaro' discusses in detail the application of nuclear models to
calculate neutron cross sections.
[The neutron has a mass of 1.00898 atomic mass units, slightly greater
than the proton mass of 1.00759 amu)Since it does not possess electrical
charge, a neutron is not affected by the atomic electric field. Consequently it
travels through matter unhindered until it passes close enough to a nucleus
to interact with the short-range nuclear forces. A nuclear interaction may
alter the energy and direction of the neutron or may result in its absorption
into the nucleus. Scattering may be either elastic, in which the kinetic energy
of the system is conserved, or inelastic, in which some of the kinetic energy
is transformed into excitation energy of the nucleus. As discussed in Chap. 2,
a neutron absorption can result in the emission of one or more gamma rays,
charged particles, and, at times, one or more neutrons. Inelastic scattering
also produces secondary emissions, and, of course, neutron absorption can
cause fission in certain isotopes. Thus all neutron interactions, except elastic
scattering, produce a secondary source of radiation.
LAs with photons, the total neutron cross section for an interaction with a
specified nucleus is the sum of the absorption and scattering cross sections
V-
at = aa + as b/nucleus
(3.2-17)
Most often the scattering probabilities are subdivided into elastic and
inelastic portions, as are the various types of absorption reactions. Care must
be used in the interpretation of the term absorption. In the present context
absorption implies the disappearance of the incident neutron as a separate
entity, the neutron becoming a part of the new nucleus created by the
interaction. However, as mentioned in the section on inelastic reactions,
certain scattering processes are also assumed to involve the absorption of a
neutron by a nucleus followed by the emission of one or more secondary
neutrons.
Many nuclear processes have certain characteristics that are relatively
invariant over broad bands of neutron energy. It is thus convenient to divide
the energy spectrum of fission-produced neutrons into four energy groups.
tion
rs in
e in
ned
ron
ola-
85
Neutrons in the lowest group, ranging from 0 to about 0.4 e V, are called
thermal, or slow, neutrons. Their energy distribution approximates the
Maxwell-Boltzmann distribution given by
2rr E~ -E/kT
j ',.tE ) -_ (rrkT)%
e
(3.2-18)
, to
lter
leal
y it
where fn (E) is the fraction of neutrons per unit energy about energy E, k is
the Boltzmann constant (8.6 x 10-5 eV/oK), and T is the absolute
temperature of the medium. t In the thermal range neutrons are assumed to
be in equilibrium with the thermal agitation of the nuclei of the medium.
~Us
Because neutron absorption often varies with the reciprocal of neutron speed
lay
in the thermal region, this region is also known as the l/v region.
The next higher energy region ranges from 0.4 eV to about 1 keV and is
called the resonance, or epidrerrnal, region because many cross sections
exhibit one, a few, or many resonance peaks in this region. The cross
sections at these peaks can have values several orders of magnitude above the
base level; theit effect on the avetage cross section across the resonance
region depends on both their number and their widths, These larter
quantities vary greatly among nuclei.
Depending on the nucleus, the resonance region stops somewhere from 1
to 50 keV. For convenience we call neutrons in the range from a few keY to
500 keY intermediate neutrons.
The highest energy region ranges from 0.5 MeV to the upper limit of
energy of fission neutrons, about 18 MeV. In this fast-neutron region, some
on
gy
gy
2,
!S,
rrg
ill
1C
,)
t
t
resonances are found, most often in the low~energy portion, but, for the
most part, cross sections are relatively smooth and, in general, have lower
values than those of the lower energy regions.

Consistent with our treatment of photon interaction, we will briefly
describe neutron reactions in this section and refer the reader to detailed
discussions elsewhere. In particular, a recent text on neutron interactions
by
Foderaro 1 should be noted here since it is well-suited to the needs of

shielding specialists.
(a) Elastic Neutron Scattering. In elastic neutron scattering both kinetic
energy and momentum are conserved. Thus the simple billiard-ball collision
tStrictly speaking, the average neutron temperature equals the temperature of the mediulll only
fur nonabsorbing media. For absmbing media the average neutron temperature will be somewhat
higher, and the diSTribution will depart from that of Maxwell-Boltzmann.
86

",.1>
'OC;:'<<f'
de
I
~~n~cl~n~t
neutnon
\1"
~~:_D~
,,,'-
LaboRltol'"Y system
Recoil
nucleus
Center of mass
at collision",
Incident
~~~ _ _"""''-'-.,...-{
~~~s~~:~,.Of_~"
nucleus
system
Fig. 3.10-Collision coordinate systems.
model can be used to obtain the direction of motion and the speed of the
scattered neutron and recoiling nucleus relative to the direction and speed
(or energy) of the incident neutron.
Actually two processes are assumed to produce elastic scattering of
neutrons. The first is known as resonance, or capture, scattering. In this
process the neutron is assumed to be absorbed by the target nucleus and
reemitted in (possibly) another direction. The second process is called
potential, or diffraction, scattering. In this process the neutron is assumed
not to enter the target nucleus but rather to be elastically scattered by
interaction with the potential well created around the nucleus by the
short-range nuclear forces. The probability for elastic scatter is the sum of
the probabilities for both processes; the cross section for elastic scattering is
based on this sum and is often denoted by u es ' In adding these probabilities,
we must take account of the spin states of the particles since coherent
effects are involved.
Resonance elastic scattering is most important at low or intermediate
neutron energies, where it causes an oscillating behavior of the elastic-
87
scattering cross section. At higher energies (i.e., the fast region), almost all
resonance scattering is inelastic since the compound nucleus formed by
resonance scattering results in an excited nucleus following neutron
reemission. Kinetic energy is therefore not conserved, and some of the
kinetic energy reappears as a photon. Potential scattering, which is generally
a smooth function of neutron energy, is the predominant component of
elastic scattering in the fast region.
The scattering angle-energy relations for elastic interactions are given
most simply in the center-of-mass coordinate system in which the total
momentum is constant zero, and particle speeds after collision are equal to
those before collision. Figure 3.10 illustrates the collision geometry in both
the laboratory and the center-<>f-mass systems.
The energy, E, of the elastically scattered neutron is related to the
energy, Eo, of the incident neutron, the atomic mass, A, of the target
nucleus, and the scattering angle, q, (in the center-of-mass system), by the
equation
A 2 + 2A cos q, + 1
(A + 1)2
E
Eo
(3.2-19)
The scattering angle, IJ, in the laboratory system is related to q, by the

equation
cos IJ =
Acosq,+1
(A2+2Acosq,+I)~
(3.2-20)
It is seen that for hydrogen scattering (A = 1) the neutron can lose all its
energy in just one collision. Maximum energy loss occurs where q, = 1r, and
Eq. 3.2-19 becomes
Em;n
= (A -1)
Eo
A +1
(3.2-21)
For low-energy neutrons elastic scattering is approximately isotropic in

the center-of-mass system, and the average scattering angle in the laboratory
system is given by
cos IJ
2
3A
=-
(3.2-22)
88
Equation 3.2-22 illustrates the point that scattering (in the laboratory
system) is peaked in the forward direction [1> cos 0> 0) and that, as A
increases, (j approaches rr/2. Thus, for low energies and heavy nuclei,
scattering approaches isotropy in the laboratory system as well.
However, the approximation of isotropic scattering in the center-of-mass
system deteriorates as the incident-neutron energy increases, and above

about 0.1 MeV scattering is anisotropic for all except the lightest nuclei. In
this energy region Goldstein3 quotes an approximate equation for {j (where
Eo is in units of MeV)
200
8 = (E )IIA II degrees
o
(3.2-23)
which indicates that for incident energies in the fast region anisotropy
increases with increasing A.
As with photon scattering, differential angle scattering probabilities for
elastic neutron scatteting play an important role in neutron-transport
calculations. The compilation of experimentdly determined angIe-distribution data by Goldberg, May, and Stehn 7 is widely used.
(b) Inelastic Neutron Scattering. In inelastic neutron scattering the
inelastic collision differs from an elastic event primarily in that a portion of
the incident-neutron energy appears as excitation of the target nucleus. The
excited nucleus subsequently decays by photon emission; thus inelastic
neutron scattering was first introduced in Chap. 2 as a secondary gamma-ray
source. The inelastically scattered neutron leaves the collision site usually
altered in direction and generally with much-reduced energy as well. In fact,
inelastic scattering is an important means of reducing fast (>1 MeV) neutron
energies in reactor shields not only because a large amount of energy can be
transferred to the nucleus in one inelastic collision but also because of the
increasing importance of inelastic scattering with energy in the 1- to 14-MeV
range. For most elements of interest, elastic-scattering cross sections are
slowly oscillating, generally decreasing in the fast-energy region. Since
inelastic scattering is a threshold reaction, the corresponding cross section
usually increases with increasing energy. (An exception occurs if particle
reactions are present; the inelastic cross section will decrease with energy if
there are competing particle reactions.)
Inelastic scattering cannot occur unless the incident neutron has a kinetic
energy somewhat greater than the first exci~ed state of the target nucleus.
Depending on the nucleus, this threshold vades from 0.1 to 4 or 5 MeV. As
89
the neutron kinetic energy ex~eeds energy states of the nucleus above the
first, these higher states may also become excited, and thus the neutron may
excite the nucleus to any level up to the limit of the incident-neutron
energy, As the excitatiob energy is increased, spacing between levels
decreases until a continuum Is reached, For a given nuclear species and
neutron energy, there is a fixed probability for exciting each energy leveL
For ali but the first level, there are alternate routes by which the nucleus
Table 3.4-TYPICAL CROSS SECTIONS FOR NEUTRON
INELASTIC SCATTERING AT FOUR ENERGIESt
Approximate
threshold
Cross section, b
energy,
Element
MeV
14MeV
5,16 MeV
2,0 MeV
1.0 MeV
Tungsten
at/ tI ' (L)
ann' (C)
0.0115
0.99
0
2.49
0.095
2.47
123
135
2.34
0.005
LL.d
ann' (L)
ann' (C)
0.57
3.13
0
2.52
0
2.06
7.7
0
0.34
0
Nickel
ann'(L)
ann' (C)
1.4
4.02
0
1.13
0
1.17
0.6
0
0
0
4.91
9.4
0.268
0.216
0.0436
0
0
0
0
0
Carbon
on,,' (L)
0 111/
(C)
t From Troubetzkoy et al. B
may release the excitation energy: one photon may be given off, dropping
the nucleus to the ground state; or a cascade of photons may be emitted as a
result of the presence of the intermediate levels. There is, therefore, one set
of probabilities for exciting the various levels and another set of probabilities
for alternate decay routes from each of the levels. Cross-section data that
account for ali these alternatives are necessarily somewhat complex,
The total cross section for all inelastic events is often split into two
components: Gnn'(L), which is the probability of excitation to any level
from which discrete gamma rays are emitted, and unn'(C), which is the
probability for excitation to the region in which levels are so closely spaced
that the emission is essentially continuous. Table 3.4 shows for four
elements the neutron-energy threshold and the values of unn'(C) and unn'(L)
REACTOR SHIELDING FOR NUCLEiAR ENGINEERS
90
for four energies. The general trend is for the inelastic cross section to
increase as either the atomic weight of the t"'1get nucleus or the energy of
the neutron increases. The threshold for inelastic scattering decreases with
increasing Z.
Inelastic-scattering cross sections useful for detailed radiation-transport
calculations are often expressed in terms of the gamma-ray spectrum emitted
T~b1c 3.5~NUMBEROi"
GAMMA RAYS EMITTED PER NEUTRON,PRODUCING REACfION IN LEADt

"l>
MeV
Ii". IIMV
0.25
0.75
1.25
1.75
2.25
2.75
3.25
3.75
4.25
4.75
5.25
6.50
1.802 X 10 1
0.2730
0.293
0.301
0.305
0.305
0.168
0.182
0.188
0.188
0.186
0.119
0.059
0.130
0.134
0.132
0.123
0.095
0.099
0.099
0.095
0.087
0.069
1.700
1.600
1.500
1.450
0.539
0.567
0.579
0.587
0.587
0.075
0.071
0.063
0.057
0.057
0.041
0.039
0.047
0.043
0.035
0.031
0.023
0.035
0.033
0.025
0.017
0.011
0.031
0.023
O.ol5
0.009
0.005
0.023
0.01 5
0.009
0.005
0.003
0.062
0.029
0.009
0.0026
0.0019
1.400
1.350
1.300
1.250
1.200
0.587
0.583
0.571
0.559
0.540
0.292
0.269
0.277
0.261
0.252
0.180
0.174
0.164
0.1547
0.142
0.123
0.119
0.103
0.087
0.072
0.079
0.069
0.059
0.045
0.032
0.047
0.035
0.027
0.023
0.020
0.021
0.019
0.011
0.009
0.00$
0.Q15
0.011
0.009
0.007
0.010
0.007
0.005
0.003
0.003
0.006
0.003
0.003
0.001
0.003
0.004
0.001
0.001
0.001
0.001
0.004
0.0015
0.0019
0.0036
0.0047
0.Q138
0.522
00404
0.285
0.205
0.248
0.254
0.259
0.277
0.292
0.134
0.125
0.117
0.112
0.111
0.056
0.044
0.040
0.053
0.065
0.020
0.014
0.Q12
O.OUI
0.026
0.020
0.032
0.052
0.081
0.111
0.010
0.014
0.020
0.04$
0.014
0.022
0.034
0.048
0.069
0.010
0.Q18
0.028
0.040
0.059
0.008
0.012
0.020
0.032
0.045
0.006
0.Q12
0.018
0.026
0.037
0.021
0.038
0.058
0.080
0.107
9.000
8.500
8.000
7.500
7.000
0.103
0.054
0.034
0.025
0.021
0.319
0.351
0.391
0.429
0.467
0.109
0.110
0.112
0.113
0.117
0.087
0.112
0.140
0.158
0.183
0.039
0.050
0.063
0.072
0.078
0.153
0.200
0.246
0.291
0.331
0.05$
o.06i
0.06$
0.064
0.061
0.089
0.106
0.119
0.124
0.131
0.Q78
0.098
0.110
0.113
0.113
0.058
0.Q75
0.089
0.098
0.100
0.045
0.054
0.057
0.053
0.043
0.133
0.141
0.136
0.091
0.041
6.500
'.000
5.500
5.000
4.500
0.022
0.027
0.029
0.030
0.026
0.491
0.523
0.553
0.575
0.623
0.120
0.\27
0.133
0.135
0.137
0.196
0.208
0.213
0.200
0.182
0.073
0.062
0.046
0.030
0.Q18
0.370
0.393
00402
0.395
0.376
0.G5$
0.050
0.111
0.101
0.087
0.060
0.021
0.093
0.069
0.038
0.010
0.026
0.007
0.015
0.03$
O.Q2li
0.133
0.134
0.131
0.112
0.069
4.000
3.500
3.250
3.000
2.750
0.018
0.013
0.006
0.692
0.745
0.761
0.870
1.060
0.135
0.120
0.108
0.092
0.080
0.162
0.130
0.115
0.104
0.100
0.006
0.337
0.281
0.231
0.185
0.080
O.Q1i
0.Q18
1.126
1.125
1.095
1.1)(,2
1.037
1.000
0.057
0.043
0.027
0.015
0.068
0.037
0.013
1.150
1.100
1.050
1.000
9.500 x 10
2.500
2.250
2.000
1.750
1.500
1.250
00463
O.O5~
O.Q3l
0.04~
tFrum Troub<'rzkoy ,., "I."
by the nucleus for specific incident-neutron energies. These data may include
gamma rays from other nonelastic events, Le., reactions other than elastic
scattering including inelastic scattering, (n,2n), (n,p), etc. Data of this type
are shown in Table 3.5, which gives the numb~r of gamma rays given off at
various gamma-ray energies per neutron-prodl/cing reaction in lead, (n,n')
and (n,2.. ). These values are not interaction cro~s sections but rather give the
distribution of gamma rays resulting from an interaction. The data were
91
obtained from the report of Troubetzkoy et al.,8 which is an evaluation of

calculations, measurements, and extrapolation. Multiplication of these values
by the total cross section for neutron-producing events would give the
gamma rays produced per unit neutron-flux density.
In most stationary power reactors with concrete shields, gamma rays
from inelastic scattering are usually not a determining factor in the shield
design. In cases such as mobile reactors with lead or tungsten as gamma-ray
shields, however, a careful analysis of inelastic scattering is required. Also, in
fast reactors with high-Z materials in their structure, it may be necessary to
include inelastic gamma rays from the core as a part of the source definition.
As with elastic scattering, theoretical concepts have been used to
describe the nature of inelastic neutron cross sections. For the most part,
however, such treatments rely on parameters that must be determined by
measurement. Reference 7 corttains a compilation of measured neutron
inelastic-scattering angle distributions. A summary of gamma rays from
inelastic neutron scattering based on both calculation and measurement is
given in Appendix A.
(c) Neutron Radiative Capture. For neutron energies in the lower
ranges, radiative capture competes in importance with the elastic-scattering
process. In radiative capture the incident neutron is absorbed by the target
nucleus; the compound nucleus is almost invariably left in an excited state
and decays in a variety of ways, emitting one or more gamma rays. The total
excitation energy is equal to the energy of the incident neutron plus the
neutron binding energy.
The binding energy of neutrons varies from 3.2 MeV in hydrogen to 11
MeV in nitrogen, averaging about 7 MeV. Thus energetic gamma rays can be
emitted. The probability for the emission of such a photon is highest in light
and magiet nuclei. If, however, the compound nucleus formed by the
absorption has energy levels lower than the combined binding and kinetic
energies of the incident neutrOn, a gamma-ray cascade can result. Recent
theoretical work in the prediction of the gamma-ray spectra from capture
reactions is discussed by Foderaro,' Troubetzkoy et al." Yost! and
Howerton and Plechaty. 1 0
Experimental work on gamma-ray yields from neutron capture is a
continuing effort of a number of laboratories (notably in Canada and the
USSR). A compilation of data, to 1960 is given in the Shielding volume of
tThese are relatively more stable nuclei that contain 2, 8, 20, 52, 82, or 126 neutrons or protons.
92
the Reactor Handbook. 2 More recent data can be found in the journal
Nuclear Data. 1 lOne to five gamma rays, ranging in energy from less than 1
to 12 MeV, are emitted per capture. Typical mittance spectra taken from a
compilation in the Reactor Handbook' are shown in Fig. 3.11.
The thermal-neutron cross sections fot radiative capture in most
elements and isotopes are given in the famoIUS "barn book."12 Very few
measurements have been made on the a(n~'Y) cross sections at neutron
energies other than thermal. Generally, the apture cross section becomes
quite small for neutron energies above 10 to 20 keV. Therefore the total
~::I~tJ
00
9 10
E, MeV
~3:INJJ: ~
Silicon
0123456
E, MeV
Chromium
3.0 .
GJ 2.0 -
"
1.0
00
9 10
E, MeV
9 10
Manganese
2.0 Phosphorus
iii
~ 1.0
5.0
4.0
GJ 3.0
"
2.0
1.0
00 1
E, MeV
10
~::[U~"
o
2 3
9 10
E, MeV
Fig. 3011-Typical gamma-ray spectra from neutron capture, V (E) (photons per MeV per
capture) vs. gamma-ray energy. E. [From E. P. Blizard: (Ed.), Reactor Handbook, 2nd ed.,
Vol. III, Part B, Shielding, pp. 50-51, Intersciencc Publishers, a division of John
Wiley & Sons, Inc., New York, 1962.J
93
energy emitted as gamma rays, is almost equal to the neutron binding energy
because the energy of the incident neutron is generally quite small by
comparison. The gamma-ray $pectra resulting from neutrons captured at
epithermal energies may differ significantly from the thermal capture
spectra. However, since the data are frequently lacking, gamma-ray yield
data for thermal neutrons are often used for all capture energies. Yost and
Solomito' 3 show that this assumption leads to considerable error in
gamma-ray transport for many designs. Yost 9 has developed a method for
calculating the capture gammaray spectrum as a function of neutron energy.
(d) Other Nonelastic Reactions. There are reaction,; other than inelastic
scattering in which the kine tid energy of the system is not conserved. These
include (n,2n), (n,fission), and (n,charged particle) reactions. The nonelastic
neutron cross section, one' is defined by
(3.2-24)
in which 0t is the total neutron cross section and CJ es is the cross section for
elastic scatter. Since (Jne is easier to measure than the inelastic.scattering
cross section, 0jn' it is often quoted. It should be recognized that Gjn :< one'
Charged-particle reactions generally are produced by neutrons with

energies greater than 1 MeV. At lower energy these reactions are generally
inhibited either by the energetics of the process or by the Coulomb barrier,
which must be penetrated by the charged reaction products. However, some
light-particle thermal-neutron reactions do occur. One of these of importance to shielding because <:>f its high thermal-neutron absorption cross
section is the 10 B(n,a)7 Li reaction, which is accompanied by the emission of
a relatively soft D.S-MeV gallllma ray. Borated materials are widely used in
reactor shields because lOB cdmpetes for the absorption of thermal neutrons
and thus inhibits the productibn of more-energetic secondary gamma rays by
reducing radiative capture in other isotopes.
(e) Activation Cross Sections. Many neutron interactions result in the
production of excited nuclei, which decay by a variety of means. Cross
sections that describe the probability of the formation of a specific
radioactive nucleus are often dalled activation cross sections and are specified
in units of barns per nucleus. Thus the total radiative capture cross section of
a nucleus may be the sum of several activation cross sections. On the other
hand, the total activation cross section may include processes other than
radiative capture.
94
Activation cross sections are useful in ffi 4king estimates of secondary

sources in reactor shields and materials, in radioisotope-production calculations, and in the relatively new and rapidly growing field of activation
analysis. The details of decay schemes, emitted-particle energies, branching
ratios, etc., can be found in several sources. Perhaps the best current source
for such data is Nuclear Data Sheets. 1 4 Acthration cross sections are also
found in the barn book.' 2
3.3 RESPONSES TO RADIATION

It was noted in the discussion of mea~ures of radiation intensity
(Sec. 2.2.4) that interaction rates of a radiatiqn field with its environment
are important and are often-used characterizeltions of radiation intensity.
However, further discussion of the concept waS deferred to this chapter so
that the various microscopic processes of interaction could first be reviewed.
Calculations of radiation transport are fUl1damentally in terms of flux
densities and currents (described in Chap. 2). However, the quantity most
often measured and the quantity by which the shield design criteria are
specified is a property of the radiation field called dose. We use the term in a
generic sense to include any quantity that rel~tes the energy deposited by
the radiation field to biological damage. Therein lies the rub. Biological
damage is intrinsically difficult to measure and even more difficult to
predict. Fortunately we limit our shielding interest in biological effects to
defining a proper interface between the shield designer and the radiobiologist.
In an ideal situation a given dose woul~ always produce the same
biological effect irrespective of the nature and energy of the radiation or of
the body organ being irradiated. However, studies by radiobiologists in
which doses and biological damage to specifiq organs have been correlated
for known radiation fields have established thai: nature is not that simple. In
addition to being a function of absorbed energy, biological responses are
functions of the irradiated organ and of the type, rate, and energy of the
radiation. From these studies the concept of r~lative biological effectiveness
(RBE) has evolved. The RBE is a weighting f"ctor that is used to compare
the biological effects produced by the same pllysical dose (same amount of
energy deposited) of a standard radiation with radiation of a different type
and/or energy. When the physical dose is multiplied by the RBE, it becomes
a biological dose.
INTERACTIONS
OF RADIATION WITH MATTER
95
The foregoing gives some indication of the problems associated with

establishing doses for shield deslgn that are related to biological hazards. The
shield designer's task is furthh complicated by the perturbation of the
radiation field by the human body, an effect not usually included in a
typical shielding calculation. 0bviously the dose received at a particular
location within the body is nOt the same as the dose in a small detector at
the same location in space with the body absent. It is apparent that shielding
studies should include anthropomorphic phantoms as part of the shield
configuration, a theoretically possible but usually impractical consideration.
As an alternative, slab Qr cylindrical phantoms of a composition
resembling that of the human body have been used, and the doses have been
calculated or measured as a function of depth in the phantom for a given
incident radiation field. The usults are then used as response functions to
relate an unperturbed radiation field (usually called a free field) to the dose
in a human body had it been present.
The following discussion on the various quantities used to define
physical and biological doses reveals the confusion that has developed in the
terminology and in the definition of units, both because of the burgeoning
nuclear science and technolqgy and because radiobiologists and shield
designers basically differ in thelt viewpoints. The International Commission
on Radiation Units and Measutements (ICRU) has as its principal objective
the development of internatiollally acceptable recommendations regarding
quantities and units of radiati~n and radioactivity, procedures suitable for
the measurement and application of these quantities, and physical data
needed in the application of these procedures. The ICRU recognized the
confusion that existed, and in an effort to mitigate it recommended a
consistent set of definitions and units. 15 Although shield designers have
accepted most of the ICRU recommendations, they have continued to use
terms not included in the recommendations because these terms are so
ingrained in the shielding field and because the ICRU did not include all the
concepts needed in shield design. In the following descriptions the ICRU
recommendations and the traditional viewpoints are contrasted.
Trubey' 6 describes the impact of the ICRU recommendations on the

shielding community.
3.3.1 Absorbed Dose
A bsorbed dose is the energy imparted by radiation to a unit mass of

matter and as such is a physical quantity as opposed to a biological effect. A
96
REACTOR SHIELDING FOR NUCL.lfAR ENGINEERS
formal definition for the energy imparted, E~, is: The energy imparted by
ionizing radiationt to the matter in a volume is the difference between the
sum of the energies of all the directly and indirectly ionizing particles which
have entered the volume and the sum of the e/lergies of all those which have
left it plus the energy equivalent (Q) of any 4ecrease in rest mass that took
place in nuclear or elementary-particle reactipns with the volume. For all
practical purposes the energy imparted is usuailly equal to the heating effect,
but in some cases part of the energy deposited may result in changes in
interatomic-bond energies.
As proposed by the ICRU, 1 s the energy imparted can be expressed as
(3.3-1)
where
1: E in = sum
of the energies (excludi1)g rest energies) of all those

directly and indirectly ionizing particles which have entered
the volume
1:Eex = sum of the energies (excludi!J.g rest energies) of all those
directly and indirectly ionizing particles which have left the
volume
1: Q = sum of all the energies released minus the sum of all the
energies expended in any nuclttar reactions, transformations,
and elementary-particle processes that have occurred within
the volume
The absorbed dose (D) is

flED
D=-tl.M
(3.3-2)
where flED:j: is the energy imparted by ionizirg radiation to the matter in a

volume element and tl.M is the mass of the matter in that volume element.
tIonizing radiation is a radiation consisting of directly or indirectly ionizing particles or a mixture

of both. Directly ionizing particles are charged particles (~lectrons, protons. alpha particles, etc.)
having sufficient kinetic energy to produce ionization by cbllision. Indirectly ionizing particles are
uncharged particles (neutrons, gamma rays, etc.) which can ~ibcratc directly ionizing particles or can
initiate a nuclear transformation.
:j:The notation flED implies that the volume element AVassociatcd with the element of mass AM
be of an appropriate size such that the limiting process D := ~E D/AM yields a meaningful estimate of
the absorbed dose, and we invoke the special limiting process,pf Sec. 2.2.4.
INTERACTION$ OF RADIATION WITH MATTER
97
The special unit of absorbed dose is the rad:

1 rad; 100 ergs/g
(3.3-3)
The, absorbed dose rate is

ill)
Absorbed dose rate; -
.::It
(3.3-4)
where ill) is the increment in absorbed dose in time, .::It. A special unit of
absorbed dose rate is any quotient of the rad or its multiple or submultiple
divided by a suitable unit of tiIhe (rads/sec, mrads/hr, etc.).
Energy can be imparted to f' volume of matter by many different particle
reactions. Particles having the same initial energy do not necessarily deposit
the same amount of energy b!ecause, of course, energy deposition depends
not only on the initial energy but also on the type of radiation and the kinds
of interactions that occur. Scbematic representations of energy imparted to
the matter contained within a volume element .::l V for two particles having
initial energies E~ and E~ arle shown in Fig. 3.12 (superscripts 1 and 2
identify the reactions produced by particles 1 and 2, respectively). The
corresponding equations for the energy imparted are
(3.3-5 )
(3.3-6)
where .::lED is the energy imparted to the matter within the volume element
.::l V, Q is the energy equivalen~ of any change in the rest mass due to nuclear
or elementary-particle reactions within the volume, and subscripts {3 and 'Y
refer to the type of particle 1eaving the volume. In reaction 1 an incident
gamma ray undergoes Comptdn scattering within volume element .::l V. The
term Ql indicates the bindint energy of the Compton electron, which is
usually of negligible magnitu~e. The Compton electron loses some of its
energy through ionization within the volume and then departs with the
energy E p. The scattered (degraded) gamma ray also leaves the volume. In
the second reaction a neutron 'undergoes radiative capture and a gamma ray
is produced which leaves the I volume. The term Q2 indicates the binding
energy associated with neutron capture, which appears as excitation energy.
Secondary collision.s within tile volume element were purposely omitted in
98
REACfOR SHIELDING FOR NUC~EAR ENGINEERS
E'
t.v
Fig.3.12-Schematic representation of relation beltween absorbed dose (energy deposi.

tion) and kerma.
this schematic representation to illustrate that the volume element t. V

should be large enough for many prim~ry interactions to occur but
sufficiently small so that a primary particle and/or secondary neutrons or
gamma rays usually would not suffer coll,sions subsequent to the initial
collision by the incident particle within the volume element L1 V.
The average energy imparted to the m~tter within the volume element
L1V by all reactions of particles of a paruicular type and energy can be
written as
-
1.: Ri(E) I.iiE)

I
ED(E)=~1.:~.~R~i(E~)---
(3.3-7)
where ED(E) = average energy imparted in e~gs or MeV

Ri(E) = rate of the ith type of reaFtion for particles of energy E
within the volume element L1 V
Ei(E) = energy imparted, in ergs or MeV, associated with a particle of
energy E undergoing reactiort i within the volume element
L1V
INTERACTION$ OF RADIATION WITH MATTER
99
3.3.2 First-Collision Dose and Kerma

The quantity absorbed cilose discussed in the previous section is a
physical variable that is closely related to the biological effect of a radiation
field but for a real situatibn is sometimes very difficult to calculate
accurately or to relate to the ttesponse in a human being. Consequently other
concepts have been devised anli used.
One of these concepts is the first-collision dose (also called the
single-collision dose). This quantity has been subject to several interpretations that differ from one another in subtle ways. We will discuss the two
that are pertinent to the present discussion:
1. The first-collision dose is the absorbed dose contributed by all
particles incident on an isolated small mass (unless the term is explicitly
modified to include only one component, e.g., neutrons) in which
charged-particle equilibriumt exists.
2. The first-collision dose is the energy transferred from uncharged
incident radiation, regardless If origin, to the charged particles formed in a
limitingly small probe that may be located anywhere. Charged-particle
equilibrium in the probe is nOtt a requirement.
The small mass specified in the first interpretation means that the probe
is small enough to leave the radiation field unperturbed and the probability
is negligible that the incident !l'articles will interact with the probe mote than
once or that secondary gamtna rays:j: produced within the mass will be
absorbed within it. The small probe specified in the second interpretation
has the same requirements, btiIt it can be smaller than that required for the
first interpretation because charged-particle equilibrium is not necessary.
The first interpretation is widely used, particularly by experimentalists,
and is often called the free-field dose, the air dose, or the free-air dose.
However, in an attempt to alleviate confusion, the ICRU chose the second
interpretation to use as a basis for an exact definition and called the quantity
so defined the kerma (kinetic energy released in material). In fact, the
ICRU 15 encourages the exclusive use of the term kerma with the implication
that the term first-collision do.e and its other interpretations be avoided.
fCharged-particle equilibrium may ibe viewed as that condition when, on the average, as many
charged particles, such as electrons, ent~r the volume element
a v as leak out, thereby resulting in an
essentially zero net transfer of energy h}! the flow of electrons.
:j:As used here secondary gamma r~ys refcr to those gamma rays produced by interaction of the
incident neutrons with the medium (e,g'i, capture gamma rays and inelasticscattering gamma rays).
REACTOR SHIELDING FOR NUCLl'AR ENGINEERS
lOa
The definition of kerma, K, as recommended by rhe ICRU 15 is
(3.3-8)
where i:J.E K is the sum of the initial kinetic ellergies (ergs) of all the charged
particles liberated by indirectly ionizing pal1ticles in a volume element of
material i:J. V and liM is the mass (grams) o! the material contained in the
volume element i:J. V.
The kerma rate is
i:J.K
Kerma rate = ~
(3.3-9)
where i:J.K is the increment in kerma in time ~t.

Since i:J.EK is the sum of the initial winetic energies of all charged
particles liberated by indirectly ionizing par~icles, it includes not only the
kinetic energy these charged particles exp~nd in collisions but also the
bremsstrahlung they radiate and the energy '1f any charged particles that are
produced in secondary processes. Note also that no restriction is placed on
the volume into which these secondary radiatjons penetrate.
The kerma or kerma rate for a specifie4 material in free space or at a
point inside a different material will be tha~ which would be obtained if a
small quantity of the specified material werei placed at the point of interest.
With the definition of kenna in mind, w have interpreted the reactions
shown in Fig. 3.12 in terms of the initial kintic energy imparted to charged
particles
(3.3-10a)
(3.3-10b)
We can see that the expression for i:J.Ek and tl>e expression given in Eq. 3.3-6
for
are the same. However, the expres$ion for
differs from that
given in Eq. 3.3-5 for i:J.Eb because the Q:>mpton electron produced by
reaction 1 did not deposit all its energy in th~ volume element and there was
no compensating inleakage of an electron; i.~., charged-particle equilibrium
did not exist, and i:J.Ek > i:J.Eb. If charged-p~rticle equilibrium did exist, we
i:J.Eb
i:J.Ek
!OI
Eb
could assume that, on the av~rage, a recoil electron with energy

would
enter .::iV of Fig. 3.12 from lln adjacent region. This would then increase
.::iEb and make it equal.::iEk (latter is unaltered).
The choice of the second interpretation was not a popular one for
experimentalists. Auxier' 7 eJ<lpressed the issue clearly. He argues that the
first interpretation should be formally recognized as first-collision dose,
principally because the qu;).ntity measured with ionization chambers,
proportional counters, and other devices is absorbed energy and not
transferred kinetic energy.
Therefore, since measured doses cannot be interpreted as kerma, except
by deduction, some dosimetrists are insisting on retaining the term
first-collision dose. On the otfu.er hand, if charged-particle equilibrium exists,
which it mayor may not, and if bremsstrahlung radiation is ignored, then
first-collision dose and kerm~ are equivalent. In bulk shields or large tissue
masses, first-collision dose and' kerma have nearly the same magnitude, but in
thin layers, such as clothing or skin, they can be quite different. It is possible
to find the factors by which fluence must be modified to convert to kerma.
Such calculations have been made and are indispensable for comparing
analytical results with measurement.
The early calculations of first-collision dose per unit fluence were
performed by Snyder' 8 for monoenergetic neutrons incident on a fourelement tissue model. Althlough the term kerma had not yet been
introduced, the doses obtaine41 by Snyder are those described by the second
interpretation and thus are kerma per unit fluence. His calculations
considered only neutron cal/tures and elastic scatterings; the latter was
assumed to take place isotrdpically in the center-of-mass system. Later,
Henderson [9 made similar cal4ulations for neutrons for a four-element tissue
model and for ethylene. He rdported his results as absorbed dose (in rads per
hour per unit flux), but they ate actually kenna rate per unit flux.
The most recent and mos~ comprehensive calculations of neutron kerma
factors (kerma per unit fluence) were made by Ritts, Solomito, and
Stevens. 20 They included eS$entially all reactions, considered anisotropic
scattering in those reactions for which angular cross-section data were
available, and in all cases used the most recent cross-section data. One of the
models they used was an 11-dlement standard-man model in which all tissue
organs and bone were homollenized; other models were for specific body
organs. The elemental compqsitions used in the calculations, the reactions
considered, and the neutron-lfluence-to-kerma conversion factors for standard man are given in Appendix B. The kerma factor for a particular
REACTOR SHIELDING FOR NUCL!l:AR ENGINEERS
102
t:
10-10
::::>
a:
w
a.. 10. 11
:;:
a:
w
"" 10-12~---':-:!:.,L...L.,.L,;-L-.L...!,-t.---L~,-'----'-:S-'---'+:':::f--...L:::"L-~::t----L--!-;:;'--~
10-2 10-'
10"
10'
10'
103
10'
105
10
10' 10"
NEUTRON ENERGY. eV
Fig.3.13-Neutron-fluence-to-kerma factors for sta""dard-man model and for elements

contributing to standard-man model. (From Ritts et at 20)
irradiated material can be found by summil1g the averages of the kinetic

energies imparted to the struck nuclei and the energies associated with
charged particles that are emitted.
The kerma per unit luence can be expressed as
RK =
1:1:j C N.(E)
a(E)
E
t
I) ,
'}
j
(3.3-11)
where N i = number density of element i (atoms/g)

aij = microscopic cross section of reactipn j for element i (cm 2 /atom)
E ij = average energy deposited by reaction j for element i (MeV)
C =1.602 X 10-6 (ergs/MeV)
Since the kerma values are given per un,t luence, they are luence-tokerma conversion factors (usually referred tOlsimply as kerma factors). Such
kerma factors for the standard-man model, are plotted in Fig. 3.13 and
tabulated in Appendix B along with ker'1'" factors for the individual
elements making the greatest contributions to.the total factor.
Kerma-rate factors for monoenergetic g'lmma rays incident on carbon,
air, and a four-element tissue model were calc\llated by Henderson I 9 and are
INTERACTIONSIOF RADIATION WITH MATTER
103
Carbon
ENERGY, MeV
Fig.3.14-Gamma-ray kerma-rate factors for air, carbon, and tissue. (From Hender500.
19
plotted in Fig. 3.14. (Hendersdn's gamma-ray results, like his neutron results,
were actually reported as absbrbed dose, the term kerma not having been
introduced yet.)
3.3.3 Exposure
Exposure is a term that should be used only for gamma rays. Formerly
called exposure dose, it carj1e into common use after the problem of
specifying biological dose associated with X rays was first encountered. As
recommended by the ICRU, eKposure (X) describes the deposition of energy
in air by electromagnetic radiabon and is defined by
(3.3-12)
104
REACTOR SHIELDING FOR NUClEAR ENGINEERS
where .a.Q is the sum of the electrical char$es of all the ions of one sign
produced in air when all electrons liberated by photons in a volume of air
.a. V are completely stopped t and liM is the mass of the material contained in
the volume element .a.v. Exposure rate is gi\1en by .a.X/.a.t, where.a.X is the
increment of exposure during the increment df time .a.t.
The unit of exposure is the roentgen (R). It is the quantity of X or
gamma radiation that produces, in air, ions c'lrrying 1 electrostatic unit (esu)
of charge per 0.001293 g of dry air (or 2.58x 10-4 coulomb/kg). In terms
of ergs, the roentgen is equal to 87.7 ergs per gram of air or 96 ergs per gram
of tissue. For gamma rays above 3 MeV in energy, the range of secondary
electrons in low-atomic-number materials It>ecomes comparable with the
relaxation length of the gamma rays. Consequently the ionization produced
in a small volume is no longer a sole measure ~f the intensity of the radiation
at that point. The ICRU does not recommen~ the use of the roentgen above
3 MeV; in practice, however, it is still used above 3 MeV with instruments
calibrated in terms of energy absorption in air.
Since 1 rad is equal to 100 ergs per gram of irradiated material, the
roentgen and the rad are frequently interchanged when tissue exposure is
referred to, the difference of 4% being no real consequence in shielding
calculations. Strictly speaking, however, the rad should be reserved for
describing an absorbed dose.
Away from boundaries the distinction betj.veen kerma and exposure loses
relevance if only the three principal processes (Compton scattering,
photoelectric effect, and pair production) are considered and if the
photoelectric effect and pair production are assumed to be absorptions. For
this simple and widely used model,:j: the lluence-to-exposure conversion
factor is proportional to the product of the photon energy and the
energy-deposition coefficient, in which c~se exposure, absorbed dose,
first-collision dose, and kerma in air are all eqt/al in magnitude.
In some calculations, however, a slightly more complex model is used in
which pair production contributes to scattering as well as absorption. As in
the preceding model the total kinetic energy of the electron pair created in
the pair-production process is assumed to be absorbed at the point of its
emission. However, the two 0.51-MeV g~mma rays produced by the
tThc ionization that would be produced by the brfmsstrahlung associated with secondary
electrons is not included in .6.Q. Except for this small dif~erence, which is insignificant for photon
energies less than 15 MeV, exposure is equivalent to kcrma.
:t:For this and the subsequent model, the bremsstrahlun8j radiation that would be produced by the
secondary electrons (and positrons) is assumed to be immedi*ely absorbed.
105
annihilation of the positron a~e treated as scattered gamma rays, and pair
production is also viewed asa scattering process. Absorption from pair
production is described in terrns of a modified cross section, which is given
by
1.022)
(3.3-13)
(for E > 1.022 MeV)
Il
; Ilpp ( 1 - - E PPm
where E is the energy of the incident photon (MeV).
3.3.4 RBE Dose; Dose Equivalent
Whenever the biological resronse of a human organ to radiation exposure
is of concern, merely knowing the absorbed dose (energy deposited) is not
sufficient to predict the biological consequence. Biological responses vary
both with the nature and ener~y of the radiation and with the composition
and function of the irradiated brgans. Thus, when an organ is exposed to a
mixed radiation field or to a 'field of radiation comprised of one type of
particle with various energies, the energy deposited by each type of particle
of a given energy must be weighted by some factor before the total
biological dose received by the brgan can be determined by a summing of the
individual contributions.
For a specific biological effect in a particular mass, the weighting factor
is mainly a function of the linear rate of energy transfer (LET) to the system
by charged particles set in mo~ion by the interactions of incident radiation.
The LET of a material is related to the linear stopping power, but the
concepts are somewhat differ~nt. The linear stopping power is the average
energy loss per unit path length by a charged particle in traversing a medium
regardless of where the energ~ is absorbed. The LET, however, refers to
energy imparted within a limited volume. These two quantities are equal
only in the special case wh4n the LET includes the absorption of all
secondary particles, in which case it is called LETw These quantities are
described more completely by Trubey.l 6
The LET-dependent weighting factor is called the relative biological
effectiveness (RBE) and is defined as
RBE = (rads of standard radiation producing a given
biological effectJV(rads of another type of radiation
prodUcing the sa\l1e effect)
j
(3.3-14)
The standard radiation referred to is X rays of 250 keV energy, and thus by
definition the RBE for 250-keV X rays is 1.
106
REACfOR SHIELDING FOR NUCLEAR ENGINEERS
When a dose given in rads is weighted by an RBE value, the resulting

dose, sometimes referred to as the RBE dose, is given in rems, a unit derived
from the term roentgen equivalent mall:
RBE dose in rems '" dose in rads x RBE
(3.3-15)
It follows that for 250-keV X rays the rad dose and the rem dose are
numerically equal.
Table 3.6-QUALITY FACTOR AS A FUNCTION
OF LINEAR ENERGY TRAN$FER AND
ION-PAIR FORMATION? 1.22
QF
I
2
5
10
20
Average specific
ionization,
ion pairs/lJ of water
Average LET,
kV/I.l of watet
<100
200
650
1500
2000
<3.5
7.0
23
53
175
The rem is quantitatively defined as the absorbed dose due to any

ionizing radiation that has the same biologic"l effectiveness as 1 rad of X
rays with an average specific ionization of 1Olll ion pairs per micron of water
in terms of its air eqUivalent. It turns out that X rays and gamma rays
generally do not exceed this specific ionizatiort (LET < 3.5 keV/IJ), which is
considered to be the boundary between IOWl-LET and high-LET radiation
and thus to be the limiting condition for RBE 1= 1.
In the past RBE has been used in the field$ of radiobiology and radiation
protection, but this generated concern because of the differences in
application and, to some extent, in concep~. Consequently ICRU recommends that RBE be used only for correiatiI1g radiobiological experiments
and that a new term, the quality factor (~F), be used in the field of
radiation protection. Quality factors are actu4lly those values of RBE which
are intended to embrace all effects that are hazardous to human beings. In
other words, QF values are not related to specific organs of the body as are
some of the RBE values. When a QF value is used to weight absorbed dose,
107
the resulting dose, in rems, i$ identified as the dose equivalent (DE). Dose
equivalents, like RBE doses, ar~ assumed to be additive.
Although the QF is defined in purely physical terms (that is, as a
function of LET only, which, in turn, is a function of the number of ion
pairs produced per centimeter ,of travel by charged particles), the basis of the
legislated value is biological. '!fable 3.6 shows the recommendations of the
International Commission on Radiological Protection (ICRP)21,2 2 for the
relation between QF and both LET and ion-pair formation.
The lCRP recommends a QF of 1 for X rays, gamma rays, electrons, and
positrons of any specific ioni~ation;2 1 QF values for neutrons, protons, and
heavy recoil nuclei vary with energy. The lCRP recommendations for QF
values for neutrons between 0.01 and 1000 MeV are plotted in Fig. 3.15.
Tabulated values for lower energy neutrons are also shown. These values are
11 r-,--rrrmrr--r-n'TTTTrr--,-rTTTmr-,...,..,...,=,--,--rrTTTTTl
10
Qtlality
fa~or
lL-.LJ...L.L.llllL-.LJ...L.LJ.J.U'-:-...J-LJ..J.llU.L.-...L.l...LJlJ.JJ."-:-...L.I...l..w'w
10-2
10- 1
10
10'
ENERGY, MeV
Fig. 3.15-Quality factors for neutrons.
102
103
108
REACTOR SHIELDING FOR NUCLIIAR ENGINEERS
based on the calculated results of Snyder ani:! Neufeld. 23 As a rough and

conservative approximation, a QF of lOis applicable to protons and a QF of
20 can be used for heavy recoil nuclei.
The preceding QF values are assumed robe applicable for whole-body
irradiation. When specific organs are being cOtllsidered, additional modifying
factors may be needed. For example, ICRP specifies that, when the lens of
the eye is being irradiated, a modifying fact~r of 3 should be used if the
QF;;;' 10 and a modifying factor of 1 should be used when the QF = l.
Strangely enough, no recommendations are 'given for 1 < QF < 10. Presumably, linear interpolation is to be used between these limits.
When the dose in an organ is nonuniform, a' dose distribution factor (OF)
should also be applied. These factors cannot be established with certainty,
and only one, the relative damage factor applied in calculating the dose
equivalent in bone from internal radiation, is in current use.
In general, the shielding community has aqcepted the use of QF and DE,
as is reflected by the latest shielding literature. From a practical viewpoint
the change merely amounts to replacing RBE by QF and calling the product
of the quality factor and absorbed dose the dbse equivalent (DE) instead of
the RBE dose. Consequently no confusion should exist when the older
literature is consulted. It should be kept in mind, however, that QF values
are chosen by such groups as the ICRP and thus are more subject to change
than would be an RBE value based on experimental data.
All quality factors and other modifying actors are intended solely for
chronic exposure te low-level radiation fields, with genetic damage being the
hazard considered. High accidental exposures must be assessed on the basis
of particular circumstances. The ICRP has no recommendations for such
exposures, but in many cases the absorbed dOise will give a better indication
of the biological risks than will the dose equivalent. In other words, for acute
effects due to massive exposures, the QF shQuld be taken as unity, which
partially results because energy may be wasted from the standpoint of the
production of biological effects. Consequently, for military applications
during war, when early death or incapacitation is the effect to be considered,
the dose, in rems, may be taken as numerica'lly equal to the absorbed or
physical dose until better information becomes available.
3.3.5 Maximum Absorbed Dose; Maximum Dpse Equivalent
As was pointed out in the introduction to this section, the usual
shielding calculation provides a detailed de~cription of the unperturbed

radiation field (free field); that is, perturbations of the field by a human
109
body are not considered. This greatly simplifies the analysis without
seriously compromising the overall accuracy. However, the results must
somehow be related to the ha~ards to a human being if he were exposed to
such a radiation field.
A correlation between the unperturbed radiation field and the dose in a
human body has been accomplished through the use of slab and cylindrical
phantoms, which have dimensions and compositions resembling those of the
human body. Calculations (or measurements) of the absorbed dose or the
biological dose were made as a, function of depth in the phantom for given
monoenergetic neutrons or g~mma rays incident on the phantom. The
incident angular distribution Was a monodirectional beam or an isotropic
lux. (Calculations for neutrons include the contribution to the dose by
secondary gamma rays.) The maximum values in the depth-dose distributions were identified as the maximum absorbed dose and maximum dose
equivalent for the absorbed (physical) and biological doses, respectively.
Since these values were obtained and reported on the basis of a unit particle
entering the phantom, they can be used as response functions to relate an
unperturbed field of mixed radiation to the maximum dose that would be
received by some part of the body. Use of these maximum values is dictated
by a conservative design philosophy that does not allow the permissible dose
to be exceeded at any point in ~he body.
(a) Neutron Doses in Phllntoms. The first dose calculations in a
phantom were performed by Snyder and Neufeld 23 for a beam of
monoenergetic neutrons norm~lly incident on a phantom represented by a
slab of tissue. The slab, infinite in the transverse directions, was assumed to
be 30 cm thick. Snyder and Neufeld determined the distribution of the
absorbed dose through the slab and found that the maximum dose occurred
at nearly the surface or within a few centimeters of the surface. In these
calculations only neutron captlUres and elastic scatterings were considered,
which limited the contribution by secondary gamma rays to capture gamma
rays. The resulting maximum absorbed doses and biological doses as a
function of the incident-neutron energy are shown in Table 3.7. These values
have been used widely in reactor-shield design to convert neutron luences to
doses and have become virtual standards since these dose equivalents have
been stipulated for use by the Federal Register. 2 4
In similar calculations, perormed later by Auxier, Snyder, and Jones,> 5
the infinite slab was replaced, by a finite cylindrical phantom rhat more
nearly represented a human body. The phantom was 60 cm high and 30 cm
in diameter. It was divided into annular shells cut into 60 angular sectors
110
and five layers 15 em high. This division yielded 150 volume elements. The
maximum doses (averaged for a volume element) occurred in the outermost
volume element at the midplane on the side facing the beam. These
maximum doses are given in Appendix C, Table c.l.
Table 3.7-MAXIMUM ABSORBED DOSE AND MAXIMUM DOSE

EQUIVALENT FOR MONOENERGETIC NEUTRONS
INCIDENT ON A SLAB TISSUE I'HANTOMt
Maximum absorbed
Neutron energy,
dose,:!:
Maximum dose
equivalent,
-I
-2
Effective
QF
MeV
mrad neutron-1 em-2
Thermal
0.0001
0.005
0.02
0.1
3.2 (-7)
6.9 (-7)
5.7 (-7)
5.7 (-7)
1.10(-6)
1.04(-6)
1.39(-6)
1.22( -6)
2.35(-6)
8.3 (-6)
3.25
2.01
2.14
4.12
7.55
2.4
3.8
4.3
5.8
7.1
7.0
2.30(-5)
3.80(-5)
3.41(-5)
3.80(-5)
4.16(-5)
4.16(-5)
9.61
10.0
7.93
6.55
5.85
5.94
0.5
1.0
2.5
5.0
7.5
10
(-6)
(-6)
(-6)
(-6)
(-6)
(-6)
mrelil neutron
em
tFrom Snyder and Neufeld."

*Multiply by 3600 to convert to (mrad!hr)!(neutrnn em"
(mrem!hr)!(neutron!cm" sec").
7
Read: 3.2 X 10- , etc.
sec") or to
The effective quality factors shown in Tables 3.7 and C.l are the ratios
of the maximum dose equivalent to the maximum absorbed dose for a given
incident-neutron energy and phantom config\.1ration. The magnitude of the
effective quality factor is very close to but usually slightly less than that
recommended by the ICRP (see Fig. 3.15). The differences in the values
would be expected since the values recommended by the ICRP consider only
the initial collision of the neutron whereas th values of the effective quality
factors represent an average of all collision$ experienced by the neutron
within the phantom.
INTERACTlON$ OF RADIATION WITH MATTER
III
(b) Gamma-Ray Doses in Phantoms. Claiborne and Trub ey 2. recently

calculated the dose delivered to a phantom by monoenergetic gamma rays.
Using the discrete-ordinates method, with some checks by the Monte Carlo
method (both of these transprt methods are described in the next chapter),
they calculated the dose distributed in a 3D-em-thick infinite-slab phantom
having the standard-man comppsition shown in Appendix B, Table B.2. Their
maximum dose rates, which occurred at nearly the surface or within the first
2 em of the surface, are comp"red in Fig. 3.16 with the experimental results
of Jones."
1~
lO-S
1~ 4
F----;---,--,---'r-rrr,r---+--,r,-r-,r1rT;
10-
5
-
Phantom, discrete-ordinates calculations
o Phantom, Monte Carlo calculations
Experimental (Jones)
1O-7L..::
10-2
.J-_--'_-J.._lo-.L-W...l..L,-_ _....L.._ _lo-.....l._.L-..l-..l-L-J..J 10-"

2
5
10- 1
2
5
10
PHOTON ENERGY, MeV
Fig.3.16-Maximum gamma-ray d.ose rate in slab phantom of standard-man composition. Comparison with kerrna calculations and with phantom measurements. (From
Claiborne and Trubey.')
112
3.3.6 Multicollision Dose

The term multicollision dose, which until recently has been used solely
for neutron exposure, is not recognized by the ICRU. The multicollisiondose concept was developed by health physiists in an effort to relate the
neutron flux or fluence incident upon a configuration of tissue representing
the human body to the absorbed dose or to the dose equivalent. The
multicollision dose could, in turn, be related to a maximum permissible dose
(MPD) established by shield-design criteria. The term has tacitly been
defined by common usage in shield design as nhe sum of the maximum dose
equivalents (or maximum absorbed doses) delivered by each neutron-energy
group to a slab of tissue as defmed by Snyder and Neufeld' 3 This includes
the dose due to secondary gamma rays produced by the neutrons in the slab.
This dose is calculated by
Multicollision dose;
I: Vtg
g
D max (Eg )
(3.3-16)
where Vtg is the group flue nee and D max (E ) is the maximum dose
equivalent (or maximum absorbed dose) in th srab for an incident neutron
of the energy corresponding to the gth neutron group.
Gamma-ray multicollision doses are similarly calculated with Eq. 3.3-16,
with, of course, Vtg and D max (Eg ) being the appropriate values for gamma
rays.
The total dose for a mixed-radiation field incident on a tissue slab is
determined by adding the neutron and gamma-ray multicollision doses
obtained separately for the slab.
Although the concept of multicollision doses, or dose rate, is not
recognized by the ICRU, it seems clear that, td fully implement the trend to
preciseness in the fields of radiobiology and radiation protection, the dose
calculated in shield design should be given reco~nition by some official body.
Perhaps the term multicollision dose can be retained for the neutron-induced
contribution, and the term total multicollision dose can be used when the
incident gamma-ray contribution is also includ~d. Or perhaps an entirely new
term could be introduced; several have been suggested but none have been
accepted as yet.
113
REFERENCES
1. A. Foderaro, The Elements of Neutron Interaction Theory, The MIT Press,
Cambridge, Mass., 1971.
2. E. P. Blizard (Ed.), Reactor Handbook, 2nd ed., Vol. III, Part B, Shielding,
Interscience Publishers, a divisiOln of John Wiley & Sons, [nc., New York, 1962.
3. H. Goldstein, Fundamental Aspects of Reactor Shielding, Addison.Wesley Publishing
Company. Inc., Reading, Mass", 1959 (out of print); now available from Johnson
Reprint Corporation, New York.
4.]. H. Hubbell, Photon Crasis Sections, Attenuation Coefficients, and Energy
Absorption Coefficients from 10 keY to 100 GeV, Report NSRDS-NBS-29, National
Bureau of Standards, 1969.
5.]. H. Hubbell and M. J. Berger" in Engineering Compendium on Radiation Shielding,
R. G. Jaeger (Ed.), Vol. I, p. 1617, Springer.Verlag, New York, 1968.
6. A. T. Nelms, Graphs of the Compton Energy.Angle Relationship and the KleinNishina Formula /rom 10 keVito 500 MeV, National Bureau of Standards, Circular
542, Superintendent of Documents, U. S. Government Printing Office, Washington,
D. C., 1953.
7. M. D. Goldberg, V. M. May, and J. R. Stehn, Angular Distributions in NeutronInduced Reactions, USAEC Report BNL-400, Vol. 2, Brookhaven National Laboratory, October 1962.
8. E. S. Troubetzkoy, M. C. Berteh, J. Celnik, H. Flesher, G. Grochowski, M. H. Kalos,
and J. H. Ray, Neutron Cross Sections of 238U, 235U, 237U, 239U, 234U. 236U,
239pu, 240pU, W, Ph, Ni, Cr, C, 8Li, 7 Li, and T. ReportUNC-5099, United Nuclear
Corporation, 1964.
9. K. J. Yost, A Method for the alculation of Neutron Capture Gamma-Ray Spectra,
Nucl. Sci. Eng., 32: 62 (1968).
10. R. J. Howerton and E. F. Plechaty, A Formalism for Calculation of Neutron-Induced
Gamma-Ray Production Cross Sections and Spectra, Nucl. Sci. Eng., 32: 178 (1968).
11. Nuclear Data, A3: 367650 (1967);A5: 1242 (1968);A5: 243-431 (1969).
12. D. J. Hughes and R. B. Schwar~z, Neutron Cross Sections, USAEC Report BNL-325
and Supplements, 2nd ed., Brookhaven National Laboratory, 1958-1966.
13. K. J. Yost and M. Solomito, Sensitivity of Gamma-Ray Dose Calculations to the
Energy Dependence of Gammi-Ray Production Cross Sections, in Neutron Cross
Sections and Technology, Proceedings of a Conference, Washington, D. C., Mar. 4-7,
1968, D. T. Goldman (Ed.), NB~.SPEC.PUBL..299 (Vol. I), p. 53, Superintendent of
Documents, U. S. Government Brinting Office, Washington, D. C., 1968.
14. Nuclear Data Sheets, originally ~ppearing as the Annual Cumulation of Nuclear Data
in Nuclear Science Abstracts, now published as Nuclear Data, Section B, by Academic
Press, Inc., New York.
15. International Commission on R/adiation Units and Measurements, Radiation Quantities and Units, Report ICRU-11, Sept. 1, 1968.
16. D. K. Trubey, Use of ICRU'Defined Units in Shielding, USAEC Report ORNLRSIC16, Oak Ridge National Laboratory, 1968.
114
17. J. A. Auxier, Kerma Versus First Collision Dose: The Other Side of the Controversy,
Health Phys., 17: 342 (1969).
18. W. s. Snyder, Protection Against Neutron Radi~.tion up to 30 MeV, Report of the
National Council on Radiation Protection and Measurements, National Bureau of
Standards, Handbook 63, p. 6, Superintendent of Documents, U. S. Government
Printing Office, Washington, D. C., 1967.
19. B. J. Henderson, Conversion of Neutron or Gam"1a Ray Flux to Absorbed Dose Rate,
USAEC Report XDC-59-8-179, General Electric qompany, Aug. 14, 1959.
20. J. J. Ritts, E. Solomito, and P. N. Stevens, Calct4ation of Neutron Fluence-to-Kerma
Factors for the Human Body, USAEC Report ORNL-TM-2079, Oak Ridge National
Laboratory, 1968.
21. Recommendations of the International Commisswn on Radiological Protection, ICRP
Publication 4, Report of Committee IV (1953-1959) on "Protection Against
Electromagnetic Radiation Above 3 MeV and EIctrons, Neutrons and Protons," The
Macmillan Company, New York, 1964.
22. Report of the RBE Committee on the International Commission on Radiological
Protection and on Radiological Units and Meas~rements, Health Phys., 9: 357-386
(1963).
23. W. S. Snyder and J. Neufeld, On the Passage of Heavy Particles Through Tissue,
Radiat. Res., 6: 67 (1957); also reprinted in NBS Handbook 63.
24. Standards for Protection Against Radiation, Federal Register, Title 10, Part 20, 1966.
25. J. A. Auxier, W. S. Snyder, and T. D. Jones, Ne~tron Interactions and Penetration in
Tissue, in Radiation Dosimetry, Vol. I, Chap. 6, f. H. Attix and W. C. Roesch (Eds.),
Academic Press, Inc., New York, 1968.
26. H. C. Claiborne and D. K. Trubey, Dose Rates in a Slab Phantom from Monoenergetic
Gamma Rays, NucL Appl. Technol., 8(5): 450-495 (1970).
27. A. R. Jones, Proposed Calibration Factors for Various Dosimeters at Different
Energies, Health Phys., 12: 663-671 (1966).
EXERCISES
3.1 Assume that air has the following composition by volume: nitrogen, 78.5%; oxygen,
20.99%; and argon, 0.96%. Calculate, at standarti temperature and pressure (273K,
76 cm Hg): (a) The partial density in air of each constituent. (b) The weight percent
of each constituent. (c) The atomic density in air of each constituent. (d) The
electron density of air. (e) The average atomic number of air. (Densities at STP are as
follows: nitrogen, 0.00125 g/cm'; oxygen, 0.0(\)143 glcm'; argon, 0.00178 g/cm';
and air, 0.00129 g/cm 3 Atomic numbers and weights, respectively, are nitrogen, 7
and 14.0; oxygen, 8 and 16.0; and argon, 18 and 39.9.)
3.2 It is often convenient to treat mixtures such as air and compounds such as water as
homogeneous media made up of one pseudQelement. For water, compute the
following properties of the pseudoelement: (a) Atomic density. (b) Electron density.
(c) Atomic number. (Assume the density of watet to be 1.0 g/cm 3 .)
"5
3.3 At 0.10 and 3.0 MeV, Hubbe1l4 lists the photon microscopic cross sections shown in
the table for Exercise 3.3.
TAB~E
FOR EXERCISE 3.3,

PHOtON CROSS SECTIONS
(1s. b/atom
E:= 0.1 MeV

Nitrogen
Oxygen
Argon
Hydrogen
coh~rent
\\fith
Without
coherent
ape.
b/atom
3.54
4.06
9.85
0.493
3.45
3.94
8.87
0.493
0.041
0.071
3.6
E= 3.0 MeV
Nitrogen
Oxygen
Argon
Hydrogen
0.806
0.921
2.07
0.1151
(Jpp'
b/atom
0.025
0.032
0.17
0.00052
(aJ Explain the blanks in the tahle.

(b) For nitrogen at STP, compUte the total mass attenuation coefficient (cm'jg) and
the linear attenuation coefficient (em-I) without coherent scattering at 0.1 MeV.
(c) Using the results of Proble$ 3.1, .compute the total mass and linear attenuation
coefficients for air at 3.0 MeV.
(d) What is the mean free path in air for 3.0 MeV photons? Express your answer in
meters.
(e) Compute the two scattering, the photoelectric, and the pair~production microscopic cross sections at 0.1 and 3.0 MeV for the water pseudoelement. Express the
results in terms of barns per molecule.
3.4 A thin (0.04 g/cm2 ) aluminum disk 2 cm in radius is subjected on one side to a
monodirectional photon flux density of 10 10 photons cm-2 seC -1, normal to the disk.
The photon energy is 0.10 MeV. Hubbe1l4 lists the following microscopic cross
sections for aluminum whose density is 2.7 g/cm 3 :
Ucoh = 6.79 b/atom
ujncoh
= 6.41 b/atom
ape = 0.78 blatom
Compute: (a) The total reaction rate (photons per second) in the disk. (b) The rate at
which photons are absorbed on their first collision in the disk. (c) The filSt-collision
116
REACTOR SHIELDING FOR NUCljEAR ENGINEERS
scattering rate in the disk. (d) The exit flux density of O.I-MeV photons.
(e) Approximately what fraction of the disk's altoms are reacting with the photon
beam per unit time?
3.5 For scattering angle () (0:<;; () ~ 1T/4), the diffl1rential scattering cross section for
Compton scattering of 1.25-MeV photons is closely approximated by
G(fJ) = (29.5 - 78.3 cos () + 56.6 cos 2 8) X 10""26 cm 2 electron-I steradian-1

For the disk of Problem 3.4, compute the flux density of singly scattered photons at
a point P located 2 em from the center and on the opposite side of the disk from a
normally incident 1.25-MeV photon flux density of 10 1 0 photons cm""2 sec-I.
3.6 The differential speed distribution of thermal neUitrons is given by
. dV about v
n (v ) d v = A v 2 e -mv' 12kT d v neutrons I em J w hase spee d'15 in
where A and k are constants, T is the neutron temperature, and m is the neutron
mass. Show that, for a material whose cross settion varies as ltv [a(v) = C/v]. the
average cross section for the reaction rate of neutrons with this speed distribution is
given by
(m,,)~
a = C 8kT
(C is a qanstaDt)
and thus the energy for which a(v) = ois 4kT/rr.

3.7 The accompanying table gives the densities and mass energy transfer coefficients /lk/P
for four materials. Compute the kerma rate (erg$ g~1 sec-I) in a small volume of each
material (a, b, c, d) located in a vacuum 100 em from a l-Ci isotropic 60 CO point
source.
TABLE FOR EXERCIISE 3.7
(a)
(b)
(c)
(d)
Air
Water
Iron
Lead
0.0268
0.0300
0.0253
0.0350
0.00129
1.00
7.90
11.35
(e) Compute the absorbed dose rate in tissue, where Mk/P = 0.029 cm2 /g. Assume that
charged-particle equilibrium exists and bremsstrahlung is negligible. Give the answer
in rad~ per hour.
117
3.8 In Sec. 3.2.1, the energy-angle relation, expressed in units of Compton wavelength,
was noted to be
X'-A=l-cosO
and the differential angle scattering cross section was given as
steradian-}
Show that, when reduced to its one independent variable, A', the differential angle
scattering distribution of photons of initial wavelength A is given by
(a) a(8) =
1~1T (trrl~ + ~ + 2(A -A') + (A -
A')2] T.V. electron-' steradian-'
and thus that
(b) a(A')
=~= H~),
X
[~+ ~ + 2(A -
A') + (A - A')2] T.V. electron- 1 unit wavelength-'
where a(X') dX'/oc gives the probability that a photon of energy A which undergoes a
Compton scatter has a scattered energy in wavelength interval dA' about A'.
Radiation Transport
P. N. STEVENS, D. K. TRUBEY, C. W. GARRETT, and W. E. SELPH
Radiation transport is to neUtron or photon interactions what dynamics of

rigid bodies is to particle kinematics. In each case the former represents the
macroscopic manifestation of the latter. The principal task in shield analysis
is to solve the transport equation. In the following sections we will discuss
some simple concepts and then develop the Boltzmann transport equation,
which is not so simple. Subsequent sections are devoted to descriptions of
several techniques for solving this equation by direct analysis and by
alternate approaches. The Monte Carlo method of solution has been so
successful in reactor shielding that we have devoted Chap. 5 exclusively to it.
In many instances solutions are tabulated. Thus a shield designer must, at
most, choose one method from several methods of solving the transport
problem or, frequently, apply a predecessor's result. The selection of one of
these techniques for application to a particular problem depends on the
nature of the problem, the accuracy required, and, mundanely, the time and
funds available. of course, the most economical technique for handling the
problem and obtaining the necessary accuracy is the optimum choice.
However, there is no simple' way in which to make this choice. We have
attempted to describe the theory involved in the various methods in a
manner that will allow the shield designer to utilize his best judgment in
making a good selection. Slilbsequent chapters discuss the application of
these methods in various practical situations.
The radiation field at a point in space removed from the source can be
divided into two components. The fltst is the uncollided, or, as it is
sometimes called, the unscattered radiation (particles that arrive at the point
without having undergone any interactions with the transporting medium).
The second component is composed of the collided, or scattered, particles
(particles that have undergone one or more interactions that have caused a
change in direction or energy or both). Primary radiation particles and those
generated in secondary reactions both may be so classified, although
119
120
REACTOR SHlELDlNG FOR NUCLEAR ENG1NEERS
secondary particles are sometimes treated as a portion of the collided

component of the primary beam.
Solution of the generalized Boltzmann equation can result in the
determination of both components simultaneously; however, the calculation
is often divided into two parts, which are performed separately. The first is a
computation of the uncollided component, and the second, a calculation of
the scattered component. Analysis of the uncollided component is rather
straightforward even though complexities of energy spectra and geometry
sometimes require use of numerical techniques. Analysis of the scattered
component, on the other hand, may become quite complicated, and it is
toward this task that most of the efforts in shield analysis are devoted.
Following scatter, a particle or photon will have altered interaction
probabilities owing to a shift in energy, and subsequent scatters will further
alter these probabilities. The distribution of absorptions within the shield is
also complicated by scattering since abs<)rptions can occur following
scattering.
As the radiation penetrates deeper into the shield, a higher percentage of
the total flux density will be scattered radiation. The ratio of scattered to
uncollided radiation may increase to some maximum value, which is a
complex function of the incident spectrum arjd material cross sections, or (as
is usually the case) the ratio may continue to increase indefinitely. Thus the
thicker the shield, the more important the scattered radiation becomes.
As will be seen in subsequent sections, most practical situations involve
complicated geometries and source characteristics, and calculation of the
scattered component requires the use of simplifying assumptions.
There has been an active effort to develop and refine methods for
predicting scattered radiation. It is important to remain in close touch with
the technical literature because new developments appear frequently.

4.1
FUNDAMENTAL CONSIDERATIONS
The flux density at a detector point is influenced by two factors: the

geometric relation between the source and the detector (including the source
angular distribution) and the character of the material between the source
and the detector. These two factors combine in a multiplicative fashion in
simple geometries and are illustrated in the following example.
Consider the uncollided flux density at the detector point in Fig. 4.1.
Even if the two slabs of materials are identical and the source strengths
(particles cm-2 sec" incident on the surface of the slab) are equal, the
121
RADIATION TRANSPORT
Source
Detector
Monodirectional source
Isotropic source
Fig. 4.1-Geometric differences due to sources.
dA
(x)
--------.
-------_.-
I,
I,
_Il.......-dx'
,"
,,
"
o
"
x
Fig. 4.2-S1ab attenuation, monodirectional flux density.
radiation intensity will be different at the two detectors owing to the

different source angular distributions, i.e., different geometries. On the other
hand, if the material in one slab were altered, the detector response behind
that slab would change with no change in source-detector geometry. This
illustrates the attenuation (sometimes called barrier) factor.
In the absence of external forces, particles travel in straight lines. From
this elementary statement it follows that the flux density in the beam of a
plane monodirectional source does not vary with distance from the source
unless, of course, it is attenuated by some material. This fact is useful in the
analysis of individual components of a radiation field, and we apply it here
to develop another elementary result.
Consider the slab geometry of Fig. 4.2, in which a monodirectional
source of <1>0 particles cm-2 sec-\ is normally incident on a slab. Every
i22
REACTOR SHiELDiNG FOR NUCLEAR ENGiNEERS
interaction within the slab removes a particle from the uncollided beam. We
shall compute the flux density, (x) , of uncollided particles that emerge
from the slab. Incident on the differential slab volume dV of cross-sectional
area dA and thickness dx' (dV=dA dx') are (x') particles cm-> sec-I, and
emerging from dV are (x' + dx') particles cm-2 sec-I. As shown in Chap. 3,
the interaction rate per unit volume occurring in dV is given by ~t (x'),
where ~t is the total linear attenuation coefficient (or total macroscopic
cross section). Thus particle conservation in volume dV demands that
(x' + dx') dA - (x') dA = - ~t (x') dV
(4.1-1)
where the minus sign on the right side is introduced because collisions
remove particles from the uncollided beam. After both sides have been
divided by dA, Eq. 4.1-1 may be written
(4.1-2)
or
(4.1-3)
This familiar differential equation has the solution

(4.1-4)
where c is a constant. By applying the condition that <1>(0) = <1>0' we obtain
(4.1-5)
Thus the exit flux density of the uncollided component is given by
(4.1-6)
This exponential attenuation law is basic to the calculation of radiation
transport through matter.
ConSIder a point isotropic source at the center of a spherical shell of
material, as shown in Fig. 4.3. With no shell material present, the flux
density of uncollided particles at the detector would be given by
RADIATION TRANSPORT
123
(R) =S/41fR 2 , where S is the source strength. However, the flux density
passing through the shell is reduced by a factor of e-~ ,x. Thus
(4.1-7)
the product of a geometric factor, 1/41fR 2 , and an attenuation factor, e-~'x.
Detector
R
Fig. 4.3-Spherical shell attenuation.
Successively complex radiation-transport calculations occur if (1) the

radiation is allowed to scatter with no change in energy, (2) scattering with
energy degradation is allowed, and (3) secondary radiations are taken into
account. An accurate analysis of neutrons and gamma rays from a reactor
source requires that energy degradation and secondaries be considered.

Consider a material in which only scattering and absorption occur. The
total macroscopic cross section, l:tl consists of two components: 1;$' the
scattering cross section, and ~a, the absorption cross section. For a parallel
monoenergetic beam incident to a slab, one component of the penetrating
radiation will be that which has not interacted with the shield, given by
<l>oe-~,x. The other penetrating component will be the radiation that has
scattered one or more times within the shield. This scattered component will
exit the shield in all directions and over a continuous energy distribution
even though the incident radiation is of only one energy and one direction.
4.2 THE BOLTZMANN TRANSPORT EQUATION

The complex form of radialion transport in which scallering and energy
degradation occur was originally described by the differential equation due
124
to Boltzmann. In a book on transport theory, Kuscer 1 notes that Boltzmann
formulated this equation in 1872 to calculate the coefficient of self-diffusion

for a gas in which the molecules wete assumed to scatter as elastic spheres.
The gas problem is equivalent to the radiation problem of interest here.
Kuscer states that a solution of the equation was not achieved until early this
century and contrasts the ease with which transport problems are formulated
with the difficulties encountered in obtaining solutions. Basically this
equation is a "bookkeeping" statement that accounts for additions to and
subtractions from the radiation in a given increment of space, enetgy, and
direction. Particles may be interpreted as either neutrons or gamma rays in
the derivation.
The differential flux density is a function of seven independent variables:
three that define the spatial position of the detector point, two that define
direction of particle motion, one that measures particle energy, and one that
measures time. In vector notation these can be reduced to three quantities: a
vector defining spatial position, a vector defining momentum (specifying
both direction and energy), and one scalar time variable. A radiation field is
completely defined (i.e., the radiation-transport problem of interest is
solved) when the particle or energy-flux density at every spatial point within
the region of interest is known.
For shield analysis a steady state is ordinarily assumed, and the time
variable is not required. It is also customary to use a unit vector for direction
and an independent scalar variable for energy (or its equivalents, speed or
wavelength). Thus we define
(r,n,E)
dn dE
particles unit arel sec- 1
as the flux density (or track length per unit volume per second) of particles
at spatial point r whose directions of travel lie in the solid angle dn about n
and whose energies lie in the increment dE about E. We derive the
Boltzmann equation using particle conservation in a differential volume
d V dn dE of six scalar dime nsions, in which d V is a differential spatial
volume element, as illustrated in Fig. 4.4. The six-dimensional system, of
which d V dn dE is a differential element, is sometimes called a six-dimensional phase space, and the differential element, a differential phase-space
cell. Care must be taken when using spherical coordinates for the spatial
coordinate system; the polar (li r ) and azimuthal (!/Ir) angles defining the
direction of r must not be confused with the polar (Ii) and azimuthal (!/I)
angles defining unit vector n, which meaS'lres particle direction.
RADIATION TRANSPORT
125
r---.
, o
I
Fig. 4.4-Volume element in phase space.
Particles of the appropriate energy (dE about E) and direction (dn about
0) can be introduced into dv dn dE by the following processes:
1. They can be born in dV at the appropriate energy and direction by a
source located there.
2. They can flow into dV with the proper energy and direction from
adjacent spatial regions.
3. Particles bearing other directions or energies or both can undergo an
interaction within dv such that the scattered particle is in dv dn dE (the
inscatter source).
In a similar fashion, particles within dv dn dE can be removed from that

differential element by the following processes:
1. They can undergo an interaction that causes them to be absorbed or
changed in direction or energy or both.
2. They can flow out of dV into adjacent spatial regions.
In a steady state the losses from dV dn dE must equal the gains. Further,
the leakage gains and losses into and out of dV can be combined into one
term: net leakage. Thus
Net leakage + interactions = inscattering + source
where, by convention, net leakage is defined as being positive in the outward

direction (a loss).
The net leakage term, NL, is equal to V'O(r,O,E) dv dn dE particles/
sec, a fact that we wUl demonstrate. Consider a small spatial volume element
126
.::l.V enclosed by surface.::l.S (Fig. 4.5); dS is a surface differential area on.::l.V

with outward normal n, and dA is the projection of dS normal to O. The
particle flow through dS and dA in direction 0 1S given by
(r,O,E) dn dE dA particles/sec, which may be written
n
Fig. 4.5-Differential surface element.
(r,O,E) dn dE (dA/dS) ds. However, since dA!dS = (n'O), the net leakage,
d(NL), through dS is
d(NL) = (r,O,E) dn dE (n'O) dS
(4.2-1)
[Proof that Eq. 4.2-1 gives net rather than outward leakage through dS is
given as one of the problems in the exercises. J The vector property
(u'v) = v'u enables us to write
d(NL) = [O<l>(r,O,E) dn dE] n ds
(4.2-2)
The total net leakage from AV is then given by

ANL
las
d(NL)
!as
[O<l>(r,O,E) dn dE]'n dS
(4.2-3)
But AV, AS, and O<l>(r,O,E) dn dE meet the conditions for application of
the divergence theorem (stated in Appendix D), and the net leakage from
AVis thus
ANL = av V'O<l>(r,O,E) dn dE dv
Now consider the limit as AV approaches dv. The function V'O<l>(r,O,E)
dn dE becomes a constant with respect to r, and

lim
av~v
ANL
[V'O<l>(r,O,E) dn dE]
lim
av~v
Jav dV
127
RADIATION TRANSPORT
The limit of the integral is simply dV. Thus the net leakage, NL, from dV is
given by
NL = Vn<l>(r,n,E) dV drJ. dE particles/sec
(4.2-4)
Every interaction within dV (except for coherent scattering without a

change in direction) removes particles initially in d V drJ. dE from that
differential element. Thus the interaction loss term, IL, is given by
IL = ~t(r,E) (r,n,E) dv drJ. dE particles/sec
(4.2-5 )
The macroscopic loss coefficient is taken to be the total attenuation

coefficient, ~t.
The source term is handled formally. If S(r,n,E) is defined to be the
differential source distribution (particles per second per unit volume, solid
angle, and energy), the total source in dV drJ. dE is
S = S(r,n,E) dV drJ. dE particles/sec
(4.2-6)
The remaining term expresses the inscattering. Let p(E.....E, n.....n)

drJ. dE be the probability that a particle of energy E' and direction n'
scatters into dE about E and drJ. about n. The rate at which scattering
events in dv drJ. dE' occur is given by ~s(r ,E') (r,n'E') dV drJ. dE'
particles/sec, in which ~s(r ,E') is the macroscopic scattering coefficient.
Inscattering from dV drJ. dE' into dv drJ. dE is then
~s(r ,E') p(E'.....E,n.....n) drJ. dE (r,n' ,E')dV drJ.' dE'
The total inscattering rate, IS, is given by

IS =
IE' In'
~s(r,E') p(E.....E,n.....rJ.) drJ. dE (r,n' ,E')
dv drJ. dE' particles/sec
Conservation requires that losses equal gains: NL + IL
(4.2-7)
=IS +S or
Vn<l>(r,n,E) + ~t(r,E) (r,n,E)
=IE' In' ~s(r,E') p(E.....E,n.....n)(r,n',E) drJ.' dE' + S(r,n,E)
(4.2-8)
128
REACTOR SHlELDlNG FOR NUCLEAR ENGINEERS
in which the common factor dV dn dE has been cancelled from each term.
Equation 4.2-8 is the integra-differential form of the Boltzmann equation
and is that most often used as a starting point in the solution of the radiation
transport problem, although other forms (I.e., purely integral) are sometimes
encountered. Equation 4.2-8 is applicable to both neutron and photon
transport. Simple transformations (some are given in the exercises) allow the
Boltzmann equation to be expressed in terms of wavelength A (A = hV/E)
rather than energy and differential energy-flux densities f(r,n,E) and
f(r,n,A) rather than number-flux densities. In photon transport, because of
the relatively simple expressions that result, the form involving f(r,n,A) =
A4>(r ,n ,A) (A in units of Compton wavelength) is convenient.
The appearance of the Boltzmann equation varies considerably with the
coordinates involved, the type of radiation being considered, and whether
energy degradation is allowed. For this discussion the vector notation is
used, and various symbols representing interaction parameters are defined.
Appendix D lists the transformation between the three common coordinate
systems and defines various vector operators, functions, and theorems
helpful in deriving and manipulating the Boltzmann equation.
A direct, analytic solution of the Boltzmann equation can be found only
in a few very simple and highly idealized cases. For the most part,
approxima te solutions are all that are possible in practical situations.
Approximate solutions can be obtained in two ways: simplifying assumptions can be made which alter the Boltzmann equation so that an analytic
solution is obtainable or numerical methods can be used. An example of the
first type is the elementary one-speed diffusion equation of reactor physics
which, although usually derived differently, can be shown to be a form of
Eq. 4.2-8 with the proper assumptions invoked.
Successful numerical solutions to the Boltzmann equation have been
obtained with difference methods and iterative procedures. These methods
evolved in parallel with, and were influenced by, the development of
high-speed computing machines, which are a necessary tool in their
application.
The approach generally used involves the derivation of a difference
approximation to the Boltzmann equation for each point of a mesh filling
the shield volume. Steps in such a calculation include:
1. Choosing the division points between energy groups.
2. Cnoosing a method of representing the differential scattering cross

section and the angular dependence of the flux density.
RADIATION TRANSPORT
129
3. Integrating the Boltzmann equation over each energy group (cross

sections must be suitably averaged over each group).
4. Approximating the equations relating the spatial derivatives of the
differential (angle and energy) flux density as functions of r by a finite
difference system at each mesh point.
5. Solving the resulting system of equations by an iterative method.
The various difference methods used in solving the Boltzmann equation
differ primarily in their method of representing the angular dependence,
although they may also differ in the iteration scheme indicated in step 5.
Before proceeding to a review of various solution methods, we should
point out that Eq. 4.2-8 includes some undefined functions, in particular
those contained in the functions 5(r,n,E) and p(E' .... E,n' .... n). To solve
the Boltzmann equation, we must replace these expressions with functions
that describe the processes involved in the problem being considered.
4.3 SPHERICAL HARMONICS METHOD

Although the method of spherical harmonics is not currently in
widespread use, it was one of the first developed and embodies concepts
used in other techniques in this chapter. The method as applied to the
solution of the Boltzmann transport equation consists in representing the
various angle-dependent terms as expansions in the spherical harmonics

polynomials. These polynomials, commonly called associated spherical
harmonics, t vary in definition from one reference to another. We use the
polynomials described by Weinberg and Wigner. 2
Applying the spherical harmonics technique to the general transport
problem is inherently complex. However, a simplified and lucid illustration
of the method can be shown for a steady-state (no time dependence),
one-speed (no energy dependence), one-dimensional (slab geometry with
azimuthal symmetry), homogeneous (constant system parameters) neutron
transport problem. Consistent with these simplifications, the general
Boltzmann transport equation can be written:j:
P.
3;
34>(x p.)
+ ~t 4>(x,p.) = 5(x,p.) +
f+1
-1
(4.3-1)
tDefinitions and theorems of associated spherical harmonics and the related Legendre functions
are given in Appendix D.
tThe derivation of Eq. 4.3-1 is given as a problem in the exercises.
130
REACTOR SHlELDING FOR NUCLEAR ENGINEERS
where "'(x,lt) =angular flux density (neutron-flux density per unit It)
x = spatial variable in slab geometry, the direction of which is
specified by the unit vector i
J1 == direction cosine with respect to the x-axis
= nj = cos IJ
~t = total macroscopic cross section
'rs(It,It') dlt = a scattering cross section that describes the probability that a
neutron with an incident direction cosine It' will be scattered
so that its emergent direction has a direction cosine in dlt
about It
'rs(It,It') =f~ff 'rs(n,n') dl/i
'rs(n,n') = the macroscopic differential scattering cross section
S(x,lt) = source particles per unit It, volume, and time
The angle-dependent terms of Eq. 4.3-1 can be represented as a series of
spherical harmonics of the first kind, the Legendre polynomials Pj(It).
Expanding the angular flux density and source term in terms of these
polynomials yields
"'(x,lt) =
j;O "'j(x) Pj(lt)
(4.3-2)
S(x,lt) =
j;O Sj(x) Pj(lt)
(4.3-3)
where "'j(x) = position-dependent Legendre coefficients corresponding to

the flux density:
...
( ) _ 2j + 1
....j x - - 2 -
1+1...
d
1 .... (x,lt) Pj(lt) It
(j = 0,1,2,... )
(4.3-4)
and Sj(x) = position-dependent Legendre coefficients corresponding to the

source term:
(j = 0,1,2, ... )
(4.3-5)
Since for most practical situations the differential scattering cross section
depends only on the change in direction given by Ito = nn', the series
expansion for 'rs(n,n') is made in terms of the Legendre polynomials
Pi(lto) :
RADIATION TRANSPORT
131
(4.3-6)
where the values of lJi are the Legendre coefficients (i; 0.1,2.....00 )
corresponding to the differential scattering cross section
2i + 1
lJi; - 2 -
f+1
-1
~s(n.n) Pi(jlo) djlo
(4.3-7)
The spherical harmonics form of the Boltzmann equation is obtained by

introducing the preceding series representations for (x,jl), S(x,jl). and
~s(n,n') into Eq. 4.3-1. multiplying each term by the Legendre polynomial
Pn(jl). and integrating over all jl (-1 to +1). When Eqs.4.3-2. 4.3-3. and
4.3-6 are substituted into Eq.4.3-1 and the orthogonality property of
Legendre polynomials is used along with the addition theorem and a
recursion relation. t the following set of coupled differential equations is
obtained:
(for n; 0.1.2, ... ,00) (4.3-8)

This set of equations, which no longer involves the directional variables and
therefore is more amenable to solution than Eq. 4.3-1, is called the second
(or spherical-harmonics component) form of the Boltzmann equation by
Weinberg and Wigner2 and others.
Practical methods of solution require that the series representations of
(x,jl) be limited to a finite number of terms, say (n + 1) terms; n is
commonly called the truncation number, and the corresponding calculation
is called the Pn approximation. The PI approximation is equivalent to
diffusion theory (see Sec. 4.6) and involves only a linear representation of
the flux density. which restricts its application to situations wherein the
neutron-flux density is nearly isotropic, a severe limitation for deep-penetration problems.
The accuracy of the spherical-harmonics calculation is also influenced by
the number of terms used to represent the differential scattering cross
section. Only a few terms are necessary for nearly isotropic scattering, but a
tThese properties are given in Appendix D. The use of the addition theorem makes possible the
evaluation of the inscattering-integral term containing I:s(O,o'). which is necessarily expanded in
terms of Pj(JJo) rather than Pj(Il).
132
large number of terms are required for adequate treatment of anisotropic

scattering; in the past this has limited the use of the spherical harmonics
treatment. However, recent advances in cross-section technology and
increased computer capacity have for all practical considerations removed
this limitation.
Shure' found that a multigroup P, approach in one dimension for
calculating spatial and spectral neutron distributions in metal hydrogenous
reactor shields yielded satisfactory estimates of neutron attenuation for
reasonable amounts of computer time. Further, Lanning" recognized that
for some design problems the low-order approximations were sufficiently
accurate. He successfully calculated the spatial distribution of the gamma-ray
energy-flux density in one-dimensional slab geometry.
4.4 DISCRETE-ORDINATES Sn METHODt
The discrete-ordinates Sn method is a means of effecting a numerical
solution of the Boltzmann transport equation. The most recent versions of
the method permit anisotropic scattering to be included and thus make it
suitable for both neutron and gamma-ray deep-penetration calculations in a
wide variety of shielding problems. Since the method is fundamentally
formulated as a finite-difference equation, a minimum number of limiting
assumptions are required, and the solutions apparently approach the exact
solution of the Boltzmann equation as the space, angle, and energy mesh
approach differential size. The method can be applied without significant
restrictions to the problem of calculating criticality, and it can be used for
both homogeneous and laminated shields with a variety of source configurations, including surface- and volume-distributed sources.
We should note that a finite-difference equation is defined as an algebraic
statement relating values of the variable (in this instance, the discrete
ordinate flux density) from point to point in phase space. To obtain the
difference equations, we section phase space into a finite number of discrete
points and relate the flux density at each point to the flux densities at
adjacent points. As will be shown, this is accomplished by integrating the
conservative form of the Boltzmann equation over a finite-difference cell in
phase space. This procedure replaces the Boltzmann integro-differential
equation with a system of simultaneous difference equations. The latter may
then be solved numerically by an iterative technique.
tThis section is primarily the work of F. R. Mynatt.
RADIATION TRANSPORT
133
Early applications of discrete ordinates, such as the Wicks -Chandrasekhar 6 method, were limited to simple problems, such as the transport
of monoenergetic neutrons isotropically scattered in one-dimensional slabs.
The fundame ntal assumption in the Wick-Chandrasekhar method was that
the integral in the Boltzmann equation could be approximated by a Gaussian
quadrature formula; consequenrly functions involved in the integral had to
be evaluated only at the angles corresponding to the Gaussian zeros.
Although this original discrete-ordinates method could be extended to
anisotropic scattering, it was limited to slab geometry.
A discrete-ordinates technique that could be extended to spherical and
cylindrical geometries was introduced by Carlson;7 this method is commonly
called the discrete-ordinates Sn method.
Other approaches that can be classified as discrete-ordinates methods are
the direct numerical integration techniques employed by the NIOBE" and
other codes, but these techniques have not been so widely used for shielding
problems.
Early versions of the Sn method assumed that the angular flux density
varies with angle as connected line segments in an even number of equally
spaced angular increments. This representation, although reasonably accurate
for homogeneous one-dimensional systems, was found to be unsuitable for
the general problelll- Recursions involving many terms are required, and an
extension of the method to two-dimensional geometries is most difficult.
These shortcomings are largely alleviated by the use of the diamond
difference technique described by Carlson, Lee, and Worlton; which relates
the angular flux density within each particular angular increment in a general
way to the end-point values of the increment. With the diamond difference
method, the Boltzmann equation can be integrated over an angular
increment, yielding, for the derivative terms, a two-point difference equation
involving the angular flux density evaluated at the increment end points.
The linear Boltzmann equation is a flow balance for a differential
phase-space cell, treating the events causing an increase or a decrease in the
number of particles contained in the cell. The discrete-ordinates difference
equations can be formulated in an equivalent manner considering a
finite-difference cell (it is presented this way in most references). For some
time it was not clear that the difference equations would, in general,
approach the analytic form of the Boltzmann equation as the finite-difference phase-space cell approached differential size. Lathrop! 0 showed that
they would for the one-dimensional geometries, and this is established
implicirly in the following paragraphs in which the difference equations for
134
spherical geometry are derived directly from the analytic Boltzmann

equation. Spherical geometry, although simple, serves to illustrate all the
characteristics of the discrete-ordinates equations except for discrete ray
streaming, which occurs only in two- or three-dimensional geometry.
4.4.1 Transport Equation and Phase-Space Geometry
The derivation of the discrete-ordinates equations is given for the special
case of spherical symmetry in spherical (position) coordinates. With only
slight modifications these results can be made to apply to slab and
infinite-cylinder geometries. This discussion also embodies the central
features of two-dimensional derivations while avoiding much of the
complexity. The differential phase-space cell is defined by three variables:
the scalar value of the radius, r (Fig. 4.4), the cosine of the angle of the
particle direction relative to the radius, J1 ; (1/r)rn, and the energy of the
particle, B; that is,
Differential phase-space cell; dv dll dE
; 41fr2 dr dll dE
(4.4-1 )
The finite-difference cell is obtained by integrating Eq.4.4-1 over

selected finite intervals of radius, angle, and energy; it is given by
Finite-difference cell; VI 1!l.IlD 1!l.EG
3 - r 3)(Ild+1 - J1d )( E +1
-_41f
3 (r;+1
g
i
- Eg )
(4.4-2)
The following subscript notation is used throughout this section:

subscripts I, D, and G denote functions whose values are associated with the
Ith space interval, Dth angular interval, and Gth energy group, respectively; i
and i + 1 refer to a function evaluated at the lower and upper limits of the
Ith space interval, d and d + 1 refer to a function evaluated at the lower and
upper limits of the Dth angular interval, and g and g + 1 refer to a function
evaluated at the lower and upper limits of the Gth energy group.
For this problem (one-dimensional spherical geometry), the following
two analytic forms t of the Boltzmann transport equation can be considered:
fDerivation of these equations is given as a problem in the exercises.
RADIATION TRANSPORT
I'
-ar (r,I',E) +
135
(1_1'2)
-a
(r,I',E) +:Et<l>(r,I',E)
r
I'
= S(r,I',E) +
f-1+1 J.-0 :Es(r,E'--+E,l'o) (r,E' ,1") dE' dl"
(4.4-3)
and
I'a
1a
2" -a [r 2 (r,I',E) J + - -a
r
= S(r,I',E)
where
[( 1 -
J1
,
1'2 )(r,I',E) 1 + :Et(r,E)(r,I',E)
/ , d 'dI''
0 :Es(r,E --+ E,l'o )(r,E ,1') E
f.1+1 J."",
(4.4-4)
(r,I',B) = particle track length per unit volume (flux density)

about T, per unit time, per unit energy about E, and
per unit direction cosine about I'
:Et(r,E) = position- and energy-dependent macroscopic total
cross section
:Es(r,E' --+ E,l'o) dE' dl" = differential scattering cross section describing the
probability that a particle with an initial energy E'
and direction cosine 11' undergoes a collision at r,
resulting in a change of flight direction described
by the cosine of the scattering angle 1'0, which
places it in a new direction that lies in dl' about I'
with a new energy in dE about E
1'0 = cosine of the scattering angle = n n'
n,n' = final and initial flight direction unit vectors,
respectively
S(r,I'') = source particles per unit volume about r, per unit
time, per unit energy about E, and per unit
direction cosine about I'
Equation 4.4-4 is called the conservative form of the transport equation,

and its integration over any phase-space volume results in interface terms,
which maybe identified as leakage terms, that satisfy the divergence
theorem exactly. As a consequence the conservative equation is the preferred
formal basis for numerical analyses.
For convenience we number the terms in Eq. 4.4-4 consecutively so that
it becomes
(4.4-4a)
136
4.4.2 Derivation of Finite-Difference Equation

The discrete-ordinates difference equation is obtained by applying the
following integral operator to the transport equation (Eq. 4.4-4) in a manner
consistent with the classical technique for obtaining difference equations:
(4.4-5 )
This operator integrates each term of the transport equation over the
difference cell. [The integration limits are expressed symbolically by xeX,
which implies a definite integral with respect to the variable x over the
interval Xl Application of the operator to the first term of Eq. 4.4-4 gives
(4.4-6 )
which when rearranged becomes
(4.4-7)
The integral of the flux density over the energy group G may be identified as
the group angular flux
(4.4-8)
in which case Eq. 4.4-7 becomes
(4.4-9)
The volume integral in Eq. 4.4-9 can be modified and evaluated in the
following manner:
r aar [r
V
Jre I
cI>c(r,J.l)] dr=
JreVI
d[r 2 cI>c(r,J.l)]
(4.4-10)
where cI>C,i+l (Il) = cI>c(ri+l ,J.l) and cI>C,i(J.l) = cI>c(ri,J.l).
RADIATION TRANSPORT
137
Substitution of Eq. 4.4-10 into Eq. 4.4-9 yields the following expression for
the first term:
(4.4-11 )
It follows from the mean-value theorem that any integral can be approximated by
x,
i x,
x f(x) dx
2;
x f(x) Ax
(for
Xl
< X < X2, and Ax; X2
xd
(4.4-12)
The parameter. x may be adjusted to give the equality; for well-behaved

functions, the closer
x is to the true mean, the better the approximation.
Applying the mean-value theorem to Eq. 4.4-11 to evaluate the solid-angle

intervals results in
T l ; 41T(PD rl+l ct>G,i+I.D AP.D -1iD
rr
ct>G,i,D Ap.D)
(4.4-13)
where ct>G ,i,D == ct>G ,i(liD) and liD is a mean-value approximation for the
direction cosine over the direction increment AP.D' Identifying the surface
areas of the volume incremenr by
Ai
=:
41T r
l
(4.4-14)
Ai+} = 41rTl+t
yields the final form for the first term:

(4.4-15)
The integral operator (Eq.4.4-5) is next applied to the second term in
Eq. 4.4-4, and the result is rearranged:
T 2 ; 41T
r J~ea~D
r r ~a
JreVI
up
[(1 _p.2)
JEe4EC
ct>(r,p.,E)
dE) dp. dr
(4.4-16)
If we introduce the group angular flux as before, Eq. 4.4-16 becomes
. T 2 ; 41T
JreV[
r dr
~a
J~eJ1IJ.D uIJ
[(1 _p.2) ct>G(r,p.)] dp.
(4.4-17)
138
The integration over 11 is accomplished according to the procedure suggested

by Eq. 4.4-10 with the following result
T. =41T[f"v/(1-1lZ/+ 1 ) 'l>G,d+,(r) rdr
-J..v/(1-1lZ/)'l>G,d(r) rdrj
(4.4-18)
The remaining integration over the radius variable is performed using the
mean-value approximation (Eq. 4.4-12)
(4.4-19)
Equation 4.4-19 reduces to a two-point difference in the angle index
T. = (B[,d+l 'l>G,d+l,[ - B[,d 'l>G,d,[)
(4.4-20)
if a curoature coefficient, B[,d, is defined by the expression

(4.4-21)
Consistent with the conservation property of the technique, Eq.4.4-19
or Eq. 4.4-20 gives an overall neutron balance. This is apparent since a
summation of Eq.4.4-19 over M direction increments tlllD yields
[(1 -IlAt+,) 'l>[,G,M+l - (1 -Ili) 'l>[,G,I]' which is identically zero since
IlM+1 = -1 and III = +1 (M is the total number of angle increments),
Equation 4.4-21, which defines the curvature coefficients, can be recast in
the form of a recursion relation that involves the coefficients B[,d+ 1 and
B[ d, First, B[ d is subtracted from B[ d+l (where B[ d and B[ d+l are given
"
" ,
by Eq. 4.4-21):
(4.4-22)
We assume that T[ in Eq. 4.4-22 is the arithmetic mean; that is,
__ ri+ I
r[ -
+ ri
(4.4-23)
Then it follows that
(4.4-24 )
RADIATION TRANSPORT
139
Following similar arguments, the factor (Ila+t - Ila) can be expressed as

(4.4-25 )
Introducing Eqs. 4.4-24 and 4.4-25 into Eq. 4.4-22 yields the following
recursion relation:
(4.4-26)
The final form for the recursion relation is obtained by introducing the cell
areas Ai+. and Ai (from Eq. 4.4-14) and rearranging:
BI,d+. = BI,d - TiD tJ.IlD (Ai+ 1
Ail
(4.4-27)
where BI,M+. = O. Equation 4.4-27 is the form of the curvature coefficient

found in the literature. The only approximation made in the preceding
derivation is in the application of mean values.
When the integral operator, Eq. 4.4-5, is applied to the fifth term of
Eq. 4.4-4 (the inscattering integral), the result is
The differential scattering cross section can be approximated by a truncated

Legendre polynomial expansion in the cosine of the scattering angle:
(4.4-29)
where the ~n values are Legendre coefficients of the expansion. The
Legendre polynomial, P n (llo), of the scattering angle cosine, Ilo, is related to
the initial and final angle direction cosines, 1l,1l', by the addition theorem for
Legendre polynomials, t which for spherically symmetric geometry is simply
(4.4-30)
tThe addition theorem is quoted in Appendix D.
140
In adapting Eq. 4.4-28 to a multigroup calculation, the integrals over all

incident energies and all incident' angles are replaced by sums of integrals
over the primed phase-space cell. Symbolically this is denoted by
(4.4-31 )
where Land M signify the number of energy groups and the number of
angular increments, respectively.
Combining Eqs. 4.4-29 and 4.4-30 with Eq. 4.4-28, expressing the
incident energy and angle integrals by Eq.4.4-31, and evaluating all
remaining integrals by the mean-value theorem yields (after considerable
rearrangement of terms) the following forms for the inscattering integral:
T - V1t.IlD
5 2
C'=1
(4.4-32)
where T,'l;!... c is the nth Legendre moment of the multigroup scattering
cross section (multigroup macroscopic transfer coefficient), defined by
T,'C/... c == [41T tVI
fEaEc fE,.aEc' T,n(T,E' ... E)
X;I (T,E',Il') Pn (Il') dll' dE' dE T2 dT)
(4.4-33)
and ll,c' is the nth Legendre coefficient of the angular dependence of the
group flux density, calculated from
M
J7 c'= D'=l
E <1>1 ' C' D' PnUID ') t.IlD'
J
(4.4-34 )
Application of the integral operator (Eq.4.4-5) to the removal term

(third term) ofEq. 4.4-4 gives
T 3 = f"vI
f~.a~D fEaEc T,t(T,E) (r.Il,E)
41TT2 dT dll dE
(4.4-35 )
141
RADIATION TRANSPORT
The evaluation of Eq. 4.4-35 requires some effort to avoid the

assumption of angle-energy separability in the weighting of the multigroup
cross sections. As the first step in evaluating Eq. 4.4-35 in terms of a cross
section that is independent of angle, the energy integral of Eq. 4.4-35 is
written as
yIj
= fE<AEc };k,E) (r,/.1,E) dE
(4.4-36 )
= };t(r) <l>c(r,/.1) - R
where R is the correction factor that is to be determined and

flux-density weighted group-G total cross section defined by
=
c(r) -
};t
IE,AEC };,(r,E)
j" (r,E) dE
I jO (r,E)
dE
};t (r)
is the
(4.4-37)
in which jO (r,E) is the zeroth Legendre coefficient of the angular dependence

of the flux density; jO(r,E) is identical to the differential flux density,
(r,E).
Rearrangement of Eq. 4.4-36 provides an explicit expression for R:
(4.4-38)
The correction factor, R, is determined by expressing the angular flux
densities as truncated Legendre series and then combining the two terms that
comprise Eq.4.4-38. The truncated Legendre series representation of the
flux density is
(4.4-39)
When Eq. 4.4-39 is substituted into Eq. 4.4-38, the result is
(4.4-40)
142
which can be written as

N
=L.J
2n+ 1
- 2 - [~b(T) - ~g'(T)] j~(T) Pn(IJ)
(4.4-41 )
n=O
where
and the energy moments of the cross section are defined by

(4.4-42)
Substitution of Eq. 4.4-41 into Eq. 4.4-36 yields the final form for the
energy integral:
(4.4-43)
With this form for the energy integral in Eq. 4.4-35, the remaining
integrals are evaluated by the mean-value approximation; the result is
T 3 = VI Ll.IJD
[~b,Icf>G,I,D
N
~ 2n; 1 (~b,I - ~g:I)j~,I Pn (IID )]
(4.4-44)
The series in Eq. 4.4-44 is very similar in form to the inscattering integral
term (Ts ) and may be included there by replacing ~(/->G in Eq. 4.4-32 with
(4.4-45)
where 0G G' = 1 if c'
has the d;sired form
=C and 0 if c' 0/= c. The modified removal term then

(4.4-46 )
RADIATION TRANSPORT
143
Application of the integral operator to the source term of Eq. 4.4-4 is

straightforward since, with the exception of multigroup constants, the
mean-value approximation is used for all variables. The final result for a
general fixed source is
(4.4-47)
If multiplication (fission) is present, the source term is represented by
5(r,p.,E) = If-X(E) fo- vT,pr,E') j"(r,E'l dE'
eff
(4.4-48)
which gives
Xc
~f
0
T 4 = VI !:;.P.D -k1: v"'I
,C' he'
,
eff G'=!
(4.4-49)
where keff = effective multiplication constant of the system

T,f C = macroscopic fission cross section at energy C'
, v =number of neutrons per fission by neutrons of energy C'
Xc = fission spectrum defined by
Xc '"
f E'"EG
X(E) dE
fE',"Ec' vT,t(E') iJ(E')
vT,I C' -
."
h,c'
The discrete-ordinates difference equation is obtained by substituting the

derived expressions for each of the five terms into Eq. 4.4-4 and then
dividing through by !:;.P.D' The result is
ILD (A i+ 1 <l>c,i+ 1 ,D ~ Ai<l>C J'i D)
+ -.---J.J.IlD (Bd+ 1 <l>C I D+ 1

I
Bd<l>C , I , d)
144
Although derived for spherical geometry, that is, for Ai = 41rrt and VI =
(41r13)(rf+l - rf), Eq.4.4-50 is the general discrete-ordinates difference
equation for one-dimensional geometries. The equations for the other
geometries can be obtained from Eq. 4.4-50, with Ai = 1.0 and VI = t:..rI for
a slab and Ai = 21rri and VI =1r(rt+l - rt) for a cylinder.
4.4.3 Numerical Solution of the Discrete-Ordinates Equation
Equation 4.4-50 contains discrete flux-density variables having both
centered and end-point subscripts. This in effect increases the number of
unknowns such that an insufficient number of determining relations are
available for their solution. This difficulty can be resolved by relating the
centered and end-point flux densities in some consistent fashion. The
diamond difference technique is the most widely used method for this
purpose and includes two relations for the spatial variable,
+ (l-A) <PG "iD
<PG I D =A<PG i+l D
"
, ,
(forll> 0)
(4.4-51 )
and
<PG,I,D = (1 - A) <PG,i+l,D + A<PG,i,D
(forll < 0)
(4.4-52)
and a single expression for the angular variable

<PG,I,D
=B<PG,I,d+l
+ (1 - B) <PG,I,d
(4.4-53 )
where A and B are constants that can be assigned values of the interval
(1/2,1). When A =B = 112, Eqs. 4.4-51 and 4.4-52 are the same for all values
of J.l and, together with Eq. 4.4-53, are known as the ordinary diamond
difference equations, which we rewrite:
<PG ,I'+ 1, D
= 2<PG I D
I
- <PG ,t,'D
(for TiD> 0)
(4.4-54 )
< 0)
(4.4-55)
or
<PG J'
,'D = 2<PG , I , D - <PG "i+ 1 D
(for TiD
and
<PG,I,d+l = 2<pG,I,D - <PG,I,d
(4.4-56)
RADIATION TRANSPORT
145
These equations form rhe basis for most current computer solutions. For a
spatial sweep when /lD > 0, Eqs. 4.4-54 through 4.4-56 are combined to
provide the following explicit expression for the centered discrete-ordinate
flux, <l>C ,I ,D:
<l>C,I,D = [IlD(A i+1 + Ai) <l>C,i,D + (1/il/l D )(BI ,d+1 + BI,d) <l>C,I,d
+ VISC,I,D] [ilD(2A i+ l
+ (2BI ,d+l/il /lD) + VI ~b,I]-1
(4.4-57)
where the source term SC,I,D includes the fixed source and all inscattering
sources. For a typical spatial mesh sweep (ilD> 0), Eq.4.4-57 is used to
solve for the centered flux <l>C I D' Then the end-point fluxes <l>C i+1 D and
<l>C I d+1 are calculated by Egs: 4.4-54 and 4.4-55, respectively: The next
cen~~red flux <l>C HI D is then calculated again with Eq. 4.4-57 and so on. If
the flux is decre~sing so rapidly that the centered flux <l>C ,I ,D is less than
one-half of either previous end-point flux, <l>C i D or <l>C I d' then the newly
calculated end-point flux, <l>C,i+I,D or <l>C,~:d+I' will 'be negative. This
phenomenon is called diamond difference breakdown and will result in a
meaningless positive-negative oscillation of the calculated fluxes. The
calculation rna y be modified by refining the space or angle mesh or both to
remedy this, but this would necessarily increase the computational time.
However, most problems are reasonably well behaved except possibly for a
few points. A technique called negative flux fix-up has been used where
negative values occur. The troublesome fluxes are immediately recalculated
with the step difference equations, which always yield positive fluxes.
If A = B = 1 in Eqs. 4.4-51 through 4.4-53, the step-function relation is
obtained which equates the centered fluxes to the appropriate end-point
fluxes. The step difference equations are:
C,i+t/n = cI>C,I1D
(for ilD > 0)
(4.4-58)
or
4>C,i,D = 4>C,I,D
(for ilD
< 0)
(4.4-59)
and
G,I,d+l -
- G,I,D
(4.4-60)
146
An explicit expression for the centered flux <l>C I D is obtained for the
/lD > 0 spatial mesh sweep by substituting Eqs.'4.4-58 and 4.4-60 into
Eq. 4.4-50. The result is
(4.4-61)
These difference equations are solved by the same calculational sequence

described earlier for the ordinary diamond difference equations. The
step-function relation is less accurate than the ordinary diamond difference
scheme for the same mesh; however, it has the advantage of always giving
positive flux densities for positive sources.
The choice of the discrete directions plays an important role in the
discrete-ordinates Sn method. It does not appear that a most accurate (or
best) quadrature scheme for a specific problem can be selected in advance.
The efficiency of a given set of discrete directions (quadrature set) depends
on problem parameters, such as geometry, optical thickness, energy-group
structure, spatial mesh size, etc., and a generalization of these dependencies
is not possible.
The discrete directions and associated weights (which represent solid

angle) define the quadrature used in the inscattering integral; the directions
also define the mean values for the angles, such as liD' and thus affect the
approximations in the convection term.
In all Sn codes the discrete directions are represented as points on the
surface of a unit sphere located at the point in space for which the flux
density is to be defined and oriented in a fixed manner with respect to the
coordinate system. The points or directions are located on the sphere
symmetrically with respect to the three planes defining an octant such that
the point description of one octant defines the whole sphere. This is not an
absolute necessity but is usually required because of reflecting boundaries.
The more recent Sn codes allow specification of direction weights as well
as the directions themselves. Lee 9 developed an elegant method of areas
which computes directions and direction weights that are symmetric with
respect to rotational interchange of the axes of the unit sphere. Although the
directions and weights in this method are somewhat adjustable, the best
results occur with the recommended values, which satisfy various approximate moment conditions and asymptotic theories. The area method has the
advantage of rotation symmetry and the important advantage of all positive
weights for any order of Sw
RADIATION TRANSPORT
147
Although rotation-reflection symmetry has desirable qualities, only

three-dime nsional calculations would benefit from full symmetry; twodimensional problems thus require twofold symmetry, and one-dimensional
problems require no symmetry conditions within the octant. Thus for oneand two-dimensional geometries, and especially for problems where other
conditions outweigh the symmetry considerations, some liberty in choosing
directions maybe exercised.
The selection of a set of directions and the weights (biasing scheme) is
particularly important and is geometry dependent. The recommendation of
the programmer should be followed when his machine program is used. A
bad direction set will lead to implausible flux densities in particular
directions.
A typical computer solution of the discrete-ordinates problem is a
procedure for iterating the solution to some prescribed degree of convergence. The sweep of the mesh points (i.e., the sequence for moving through
the discrete points) is carefully ordered to follow the neutrons (or gamma
rays). For shielding problems all particles undergoing scattering will always
be degraded in energy. Therefore the calculation will begin with the highest
energy group (G = 1) and progress sequentially through the lower energy
groups. The angular sweep is performed in the direction of increasing D
(decreasing liD) beginning with D = 1, which for most penetration problems
corresponds to the most important direction. The spatial sweep begins at a
boundary along which the inwardly directed flux values are specified, and
the sweep is made to the other boundary and then repeated for the next
angle. The spatial sweep for negative liD begins at the other boundary, at
which point the reentrant fluxes are usually specified as zero, and proceeds
to the source boundary. After the spatial sweep has progressed through all
angles at one energy, the next lower energy group is treated in a similar
manner and so on.
In the solution of Eq. 4.4-57 or Eq. 4.4-61 for the centered flux <l>G,I,D'
some of the required discrete-ordinate fluxes have not yet been calculated;
e.g., the within group scattering involves some as yet undetermined fluxes,
those which correspond to angles liD" where D' ;;;. D. Therefore the solution
is obtained through the process of inner iteration, whereby values for the
unknown fluxes are taken as their previous iterate estimates. Details of the
various iteration schemes and of the related convergence problem are
omitted here. Several iteration schemes are described by Mynatt!! as part of
a detailed development and experimental evaluation of the two-dimensional
equation in his dissertation. An iterative technique called the synthetic
148
method was developed by KOppl2 to calculate neutron transport and has

been adapted by Gelbard and Hageman 13 to accelerate convergence of Sn
calculations.
4.4.4 Advantages and Disadvantages
From the results of calculations made with Sn codes, the discreteordinates methods appear to have the following advantages for shielding
applications:
1. Depending somewhat on the sophistication desired, the Sn calculations are easy to prepare.
2. The method is not stochastic, and flux-density errors at deep
penetration are systematic rather than statistical.
3. A series of problems having similar characteristics benefit from
knowledge of flux densities calculated in a similar case.
4. Secondary gamma rays can be calculated by the same method, either
as a second calculation or simultaneously with neutrons. The gamma-ray
yield distribution may also be made a function of the energy of the captured
neutron.
S. The range of neutron energies from highest fission eGergies to thermal,
including upscattering; in energy, can be calculated by the same method.
6. The one-dimensional calculations are much faster (in computer time)
than similar Monte Carlo calculations (see Chap. 5). In two dimensions the
type of problem and the desired answers determine whether Sn or Monte
Carlo is better.
The following disadvantages are evident, but proponents believe that
additional development can alleviate or eliminate them:
1. Convergence of an iteration method is not always uniform and
well-defined. The best method currently used is to determine from each
iterate the maximum deviation in the scalar flux density at any point in
space relative to the previous iterate value. Iterations proceed until the
maximum deviation falls below a specified limit.
2. Flux-density aberrations are frequently observed in two dimensions
owing to localized sources and the propagation of neutrons in discrete
directions (this is sometimes called the ray effect).
3. No basic ground rules exist to define for a particular problem the best
direction set, space mesh, multigroup structure, and polynomial expansion
limit.
RADIATION TRANSPORT
149
4.5 MOMENTS METHOD
The method of moments was formulated by Spencer and Fano!4 and

was the first technique to be successfully applied to the .Boltzmann equation
for solutions useful to reactor shielding. [n addition to its historical
prominence, this technique has some important advantages not shared by
other methods, one being that foreknowledge about the behavior of the
solution can be incorporated analytically in a very natural way, thereby
often reducing the effort required to achieve a specific result or a desired
accuracy or both. Another is that the type of recursio!l relation developed
precludes a truncation at a crucial part of the calculation; that is, a finite
number of moments can be calculated exactly (ignoring errors due to the
numerical solution) without considering the influence of higher moments.
In the moments method ol1e considers first the formal definition for the
moments and the manner in which they relate to some system parameter of
interest, say f(x). If f(x) is defined for all x within the interval A";x ";B,
then the nth moment of !(x) is
(4.5-1)
provided the integral exists. Only nonnegative integer values of n are

considered in practical applications.
Definite interpretations may be associated with the various moments.
For example, the zeroth moment is a normalizing number, and the first,
second, third, and fourth moments are closely related to the mean value,
variance, skewness, and kurtosis, respectively. In the physics of statics and
dynamics, the first moment of the mass is the center of gravity and the
second is the moment of inertia.
No such particular meanings are given to the moments as they are used in
the solution of radiation-transport problems (although the second moment
of the flux density is proportional to the Fermi age). Rather, they are
regarded as a transform, much the same as Laplace, Fourier, or finite
trigonometric transforms. The major portion of the calculation is performed
in terms of the transform (moments) space; then, by an appropriate
inversion, the desired answer is reconstructed.
The application of the moments method to the solution of the
.Boltzmann transport equation is limited with respect to the source-shield
configuration. It is usually applied only to infinite homogeneous media with
150
REACTOR SHIELDlNG FOR NUCLllAR ENGINEERS
plane, line, or point sources. The method as applied to gamma rays has been
described by Fano, Spencer, and Berger'S and by Goldstein and WUkins,' 6
and, as applied to neutrons, by Goldstein. 1 7 The technique is basically the
same for both neutt'Ons and gamma rays, and a description for one should
suffice for the other. The most significant differences lie in the treatment of
the scattering integral and in the more complex nature of the neutron cross
sections. The following description is for shib geometry in terms of the
simpler gamma-ray problem in which the dependent variable is the angular
energy-flux density, I(x,A,p.) and the Compton wavelength is taken as the
energy variable.
Consider the following specialized form of the Boltzmann equation: t
p.
~I
+ ~(A) I(x,A,p.) = f' r I(X,A',P.')Ek(A',A)
ux
Jo J4tr
X 5(1 + A' -;; - 0-0') dn' dA' + S(A,P.) 5(x)
(4.5-2)
where I(x,A,p.) dE dp. = energy-flux density (MeV per unit area and time)
due to gamma rays wlth energies in dE about E and
direction cosines that lie in dp. about p.
x =spatial coordinate in slab geometry
A = gamma-ray energy aft~r scattering expressed in terms
of its Compton wavelength
A.' = gamma-ray wavelength prior to scattering
p. = direction cosine with nespect to the x-axis
~(A) = total macroscopic cross section evaluated at the
energy corresponding to the gamma-ray wavelength,
A
0-0' = cos 8 = cosine of the scattering angle between initial
and final gamma-ray directions, where 0' and 0 are
the initial and final unit direction vectors, respectively
E = electron density
k(A',A) = 271' (A/A') 0(A',8)
0(A',8) = microscopic cross section per electron for Compton
scattering given by the Klein-Nishina formula

tThe derivation is given as a problem in the exercises.
RAIDIATION TRANSPORT
lSI
6 (x) =Dirac I delta function t that locates the plane at x =0

6(1 + X' - X - O 0') = Dirac delta functiont that prescribes that the
angular change (0'0') be consistent with the change
in wavelength (X - X') as given by the Compton
scattering equation. X - X' = 1 -- no'
S(X.Il) dE dll =plane source of gamma rays (energy emission per
unit area and time of gamma rays with energies in
dE about E and direction cosines that lie in dll about
Il).
Solving Eq. 4.5-2 by t~e moments method is similar to using the
spherical-harmonics treatmem (see Sec. 4.3) in that the angular energy-flux
density is first expanded as a Legendre polynomial series:
I(x,X.Il) =
L
-
2j+ 1
-2- Ij(x,X) Pj(ll)
(4.5-3)
j=o
where the Legendre coefficiellts are given by

(4.5-4)
It can be shown that I o (x,X) is the energy-flux density and that I I (x,X) is the
energy-current density.
With this series representation for the angular energy-flux density. the
integra-differential equation (Eq. 4.5-2) with the dependent variable I(x,X.Il)
and three continuous independent variables can be transformed into a
sequence of integra-differential equations for the variables Ij(x,X). which are
dependent on only two independent variables. This desired result is obtained
by multiplying Eq. 4.5-2 by the Legendre polynomial Pj(ll) and integrating
over all solid angle. The result (after some manipulation) is the following
sequence of equations:
j+ 1 alj+l
~,
alj_l
2j + 1 ax + 2j + 1 ---ax + ~(X) Ij(x,X) = Sj(X) 6(x) + f o Pj(1 + X - X)

X ek(X'.X) Ij(x,X') dx'
tProperties of the Dirac delta function are given in Appendix D.
Ij = 0.1.2.....00)
(4.5-5)
152
REACTOR SHIELDING FOR NUCL.EiAR ENGiNEERS
The spatial variable in this sequence of equations is eliminated by

applying the moments of the Legendte coefflcients of the angular energyflux density, which are defined as
,<,n+1
"'0
bn )'(>") =' - n!
f+~
.~
I)'(x,A) x n dx
(4.5-6)
where ~ 0 is the total macroscopic cross setion evaluated ar the source

energy. The equations satisfied by bnj(>") are obrained by multiplying
Eq. 4.5-6 by :rZ+ 1 (xnjn!) and integrating wid, respect to x from _00 to +00.
Because of this integration over all space, th application of the moments
method to the transport problem becomes restricted to the infinite-medium
geometry. With the use of Eq. 4.5-6, the original Boltzmann equation can be
reduced to the following doubly indexed sequence of linear integral
equations (the variable upper limit in the integral classifies it as the Volterra
type) :
~(>..) bnj{>") = fo"
k{X',>")Pj(1+>"'->") bnj(>"') d>"'+
2t:
1 [(j+ 1) bn_l,j+'
+ jb n" ,j_I] + J.l.oSj{>") lino
(4.5-7)
where j = 0, 1, 2, ..., 00, and n = 0, 1, 2, ... ,00. The Kronecker delta function
ono = 1 if n = 0, Ono = 0 if n O.
The moments for a given problem can be evaluated by a straightforward
numerical solution of Eq. 4.5-7. The ease of numerical calculation depends
on the form of the source function Sj{~)' Many problems involve
monoenergetic sources, and the Sj(X) are given by >"oSjO(>" - Xo ), where >"0 is
the Compton wavelength of the source energy. Since the presence of the
delta function is undesirable for machine calbulation, the following transformation is made:
*'
(4.5-8)
Introducing the transformation defined by E'i!' 4.5-8 into Eq. 4.5-7 yields
the defining equations for Bnj and Cnj'
~Bnj = f:,
ek{>..' ,>..) Pj( 1+>..'->..) Bnj{>"') d>'" + 2j:0 [(j + 1)
X B n _1 ,j+ I + jB n _1 ,j_ d + >"ok{>"o ,X)
Bj( 1+>"0 -X) Cnj (4.5-9)
RADiATION TRANSPORT
.+ 1
~Cnj = ~o ( ~+ 1 Cn_l,j+l
2j~ 1 Cn-I,j-I
153
The equations that define Cnj ate similar in form to the equations that
define bnj and Bnj except that the inscattering integral does not appear. It
follows that Ao 5(A-Ao) Cnj are the moments for the unscattered-energy-flux
density. Therefore the transformation given by Eq. 4.5-8 separates the
unscattered-energy-flux density (energy-flux density corresponding to Cnj)
from the total energy-flux density (energy-flux density corresponding to
bnj)' The solution to Eq.4.5-9 requires values of Cnj as input, and the
calculated moments Bnj are associated only with the scattered-energy-flux
density. This is convenient since the uncollided angular energy-flux density,
10 (x,A,Il), is easily calculated, and values of Cnj are then uniquely
determined.
For a typical calculation, the quantity of greatest interest is usually the
total, or scalar, energy-flux density, 10 (x, A). Therefore only the moments
B no (n = 0,1,2, ... ,N) are required. However, the calculation of a given Bnj
requires the prior calculation of B n - l ,j+! and B n - l ,j-l ; therefore moments
other than the Bno momenos must be calculated. In general, the moments
Boj (j = 0,1,2,... ,]) can be calculated directly, but a Bnj moment cannot be
calculated until calculations have been made of all the Bn'j"s for which
(n + j) - (n' + j') is a nonnegative even integer (including zero) and n' < n.
Table 4.1 illustrates a typical calculation sequence (for n = 5). As noted, all
Table 4.1-SEQUENCE OF MOMENTS CALCULATION FOR n = 5
o
1
2
3
4
j= 0
j=l
j=2
j =3
j= 4
j =5
154
REACTOR SHIELDING FOR NUCL!lAR ENGINEERS
the moments shown in the table must be calulated to determine B no for

n';; 5.
We rewrite the Eq. 4.5-9 integral in the follGwing form:
(4.5-11)
where An =AO + n.1.A
.1.A = arbitrary increment of wavelength
V(A n ) =Bnj(A n )
H(AnX) = ek(A',A n )Pj(1+A' -An)
T(A n ) = [~o/(2j+ 1)] [(j+ 1)B n _l ,j+1 +jBn_l,j+l] + AoH(An,AO) Cnj
This type of integral equation (Volterra tYWe) is characterized by (1) the

limi t of integration being the independent variable An; (2) the value of the
dependent variable V(A n ) depending on the values of V(A') if A' < An but not
if A' > An; and (3) T(An) involving only knGwn or previously calculated
quantities. A numerical evaluation of the intgral is required, and several
schemes are available. Regardless of which scheme is used, there are
coefficients Mnk such that Eq. 4.5-11 can be rewritten as
~(An) V(A n ) =
n-I
k=O
H(An-Ak) v(Ak) Mnk
+ H(An,An) V(A n ) M nn + T(A n )
(4.5-12)
where
(4.5-13)
Trapezoidal rule is used for the (n = 1) interval for which M nn
the following explicit expression for V(AI) is ob~ained:
V(A ) = T(A.) + H(AI ,AO) ~(>,o) (.1.A/2)

I
~(A.)- (3.1.A(8)
For n
= .1.A/2, and
(4.5-14)
> 1, M nn = .1.A/3, and Eq. 4.5-12 is rewrit~en

(4.5-15)
RADIATION TRANSPORT
155
The coefficients Mnk for n even are determined by Simpson's rule. For
n = 3, the fonowing coefficients are used: M 30 = 5/16, M31 = 1, and
M 32 = 5/4; for n odd but greater than 3, the integral from Xo to X3 is
evaluated using the special (n '" 3) coefficients, and the integral from X3 to
Xn is evaluated by Simpson's rule. The gamma-ray scattering process is such
that
(when Xn > Xk + 2)
(4.5-16)
Therefore the sums on the index k involve only a fixed number of terms.
Also
(when
Xn > Xo + 2)
(4.5-17)
and the second term in T(X tI > AO + 2) also vanishes; H(An,Xk) can be
calculated directly since only the Klein--Nishina formula and the Legendre
polynomials are involved.
At this point we assume chat the moments for a given problem can be
calculated and consider the problem of reconstructing the flux densities. It
should be emphasized that the calculation to this point can be performed
with very few approximations, t excluding the approximations involved in
the numerical procedures. 'I1he major source of error will lie in the
subsequent reconstruction process since only a finite number of moments
are available. In fact, for a finite number of moments, there is an infinite
number of allowable functions. The problem is basically one of choice: the
selection of a functional foem that win come as close as possible to
describing the spatial dependence of Ij(x,A).
Two methods have been used to reconstruct the flux densities: the
polynomial expansion method and the method of undetermined parameters.
Both were developed by Spencer and Fano." A more recent description is
given by Spencer. IS The polynomial expansion method assumes that Ij(p,A)
behaves roughly as some trial function f(p), where p is measured in mean
free paths at the initial energy; i.e., p == ~ox. Then
(4.5-18)
tThese few assumptions can be ve.. y limiting, however I since they include the assumptions of
homogeneity and infinite extent of the ttansporting medium.
156
REACTOR SHlELDlNG FOR NUCLlJAR ENGINEERS
where gj(p,A) contains the A dependence of the jth coefficient and provides a
correction for the p dependence. If a reasonable choice of f(p) can be made,
then g.(p,A). need be only a gently varying smooth function of p; for
example, a polynomial of degree N in p when (N + 1) moments are available.
The g/p,A) could be represented as an infinite series with respect to a set of
orthogonal polynomials of degree n:
(4.5-19)
The orthogonality relation is given by
(4.5-20)
where f(x) is a weighting function as well as the trial function f(p). In
practice this representation of gj(p,A) is limitd to a finite number of terms
since, given (N + 1) moments, only (N + 1) values of Anj can be obtained.
The approximation for If(p,A) can then be written
N
IJ(p,A) = f(p)
n;o Anj(AI Pn(p)
(4.5-21)
where IJ(p,A) is the jth Legendre coefficient of the scattered component of

the angular flux density. This assumes that ~alues of (N + 1) moments Bnj
are available for the reconstruction of IJ(p,A), which is accomplished by
evaluating the (N + 1) coefficient Anj in terms of the known (N + 1)
moments Bnj for a given value of j. To this end IJ(p,A) is multiplied by
pm (p) dp, and the product is integrated from _00 to +00:
(4.5-22)
The polynomial Pn(P) can be written
n
Pn(P) =
1; api
i=O 1
(4.5-23)
where the a/s are known parameters for a given type of polynomial. The
expression for Anj(A) then becomes
(4.5-24 )
tThe polynomial set U'n) is any orthogonal set, one of which is the Legendre polynomial set.
RADIATION TRANSPORT
157
The moments Bjj are defined !'s

1
..
f+~
BiP,) "'"" -~ IJ(p,A) pI dp
(4.5-25)
Elimination of the integrals between the preceding expressions for AnjCA)

and Bjj( n) provides the desired relation
n
A"j(A) =
E (il)
;=0
ajB il'
(4.5-26 )
Practical considerations Will usually restrict accurate calculation to the

differential energy-flux density If/..p,A), and then only the Ano(A)'S are
required; that is,
N
n(p,A) = f(p)
E Ano(A) Pn(p)
n=O
(4.5-27)
where
n
Ano(A)
= ;=0
E (i!) AjBio
In principle, Anj(A) for j > 0 can be calculated. However, since the angular
flux density I(x,A,p.) is usu"Uy highly peakec: in the forward direction, the
series
(4.5-28)
converges slowly, thereby requmng a large number of values of IJ(x,A),

which, in turn, would require a large number of moments. Finally,
unscattered-energy-flux densities are easily obtained for most simple
geometries and combined with the scattered energy-flux densities:
10 (X,A) =
rg (x, A) + n (X,A)
(4.5-29)
The polynomial expansion method described in the preceding text is

most often used for reconsttucting the energy-flux density of the gamma-ray
problem. This is partly for historical reasons and partly due to the ability of
the method to make full use of large numbers of moments within the same
systematic framework of analysis. For the neutron problem the selection of
158
a suitable weighting (or test) function is not obvious, and the method loses
much of its flexibility.
When the method of undetermined parameters is used to reconstruct the
flux densities, I/p,X) is represented as
(4.5-30)
where hij(p) is a function having the general expected behavior of Ij(p,X) but
containing one or more undetermined parameters, and the aij are undetermined parameters that include the X dependence. In particular, let j = 0 and
assume that (N + 1) values of the Bno(X) momemts are known; then
(4.5-31)
The moments corresponding to j = 0 can be written as
(n = 0,1,2,oo.,N)
(4.5-32)
Substituting Eq. 4.5-31 into 4.5-32 yields the following set of (N + 1)

equations:
(n = 0,1,2,oo.,N) (4.5-33)
Values of hio(P) should be selected so that the above integration can be

evaluated either analytically or numerically, and, if (N + 1) moments are
available, then a total of (N + 1) undetermined parameters are allowed.
Problems not amenable to other methods can sometimes be solved by
the method of undetermined parameters because of the much greater choice
that can be made in the hij(p) values. As a result this method has been more
widely applied to the neutron-penetration problem. A characteristic of the
method is when it fails, it fails catastrophically, leaving no doubt about its
applicability. Usually not all the moments available are needed to obtain a
satisfactory solution. The surplus moments can be used to check the
accuracy by constructing moments correspondimg to the unused moments, a
feature not so easily accomplished by the other methods.
The application of the moments method to neutron-transport problems
is similar in many respects to the gamma-ray formulation. The primary
differences are the use of differential neutron-scattering cross sections rather
RADIATION TRANSPORT
159
than Compton scattering crQss sections, particle-flux density, (r,E, (l),

rather than energy-flux density, and the use of neutron lethargy in
performing integrals over energy.
The spatial moments in the neutron case are defined as
1 f~
Inl(u) = -,
_~ Il(u,x) x n dx
n.
(4.5-34)
and the set of interlocked integral equations has the form

Ot(u) Inl(u)
= 21T "'", fl-1 Inl(U ,) On,i(U ,) fi(u ,,Ili) Pl(CXi) dlli

1
+ 2/+1 [(/+I)In_1 ,1+1 +lIn - I ,I-d +Snl(u)
(4.5-35)
where U is a variable based on the neutron energy, U = In (E/E o ); Il is the

cosine of the scattering angle in the center-of-mass system; and cx is the
cosine of the scattering angle in the laboratory system.
Differences also appear in the means of reconstructing the flux density
from the mo ments. For neutrons the results are more sensitive to the choice
of f(p}.
Cross sections change more rapidly with energy for neutrons, and the
scattering integral limits vary from element to element. For hydrogen the
scattering integral may be integrated directly by numerical means as was
done for gamma rays. For all other elements, three optional approaches are
available, depending on whether energy degradation in scattering is treated
exactly or approximately or ignored entirely.
For sufficiently large nuclei, the energy degradation may be ignored. The
minimum mass for which this approximation can be made will depend on
the other constituents of the material. For example, energy degradation by
oxygen may be ignored in the presence of hydrogen but not in a thick shield
consisting primarily of heavy elements.
The approximate treatment consists in ignoring degradation for heavy
elements but accounting for it with hydrogen scattering. This treatment has
been found to give results within 20% of the exact treatment for a mass of 6
or 7 amu where hydrogen accounts for most of the slowing down. Similar
accuracy is obtained at 10 to 20 mean free paths for atomic masses above 20
in the absence of lower weight elements. A method for a more exact
treatment of degradation was developed by Certaine.' 9
160
REACTOR SHIELDING FOR NUCr;EAR ENGINEERS
4.6 APPLICATION OF DIFFUSION THEO~Y

An approach to the particle-transport problem which neglects the
detailed directional aspects of particle motion comes from diffusion theory.
If we consider the neutron balance in a medi\llm, the equation of continuity
is a statement that the neutron gains are equal to the losses in a given volume
element. We assume no energy degradation and that all neutrons have one
velocity.
rate of gain
[ per unit volume
J _ [ source per ]
-
unit volume
absorption
leakage
]
- [ per unit volume - [ per unit volume
In four-dimensional phase space, (r,t), this statement becomes
an(r,t)
---at" = S(r,t)
~a
(r,t) - V' ](r,t)
(4.6-1)
where n(r,t) = neutron density (neutrons/em')

~a = macroscopic absorption cross section (cm- r )
S(r,t) = general source term (neutrons cm-' sec- r )
(r,t) = total neutron-flux density (neuttons C11l-2 sec- r )
](r,t) = net neutron-current density (ne\lltrons cm- 2 sec- r )
an/at = time rate of change of the neutron density (neutrons cm-'
sec-I)
~a (r,t) =loss of neutrons due to absorption (neutrons cm-' sec-I)
V '](r,t) = loss of neutrons due to convection (neutrons cm-' sec-I)
Equation 4.6-1 can be regarded as a precise relation that can be applied
without restriction to the general problem of particle transport. However, a
basic limitation in its use is that, except for cettain very restricted situations,
a tractable form for the net neutron curren~, ](r,t), does not exist. If we
assume that absorption is small compared with scattering and confine
ourselves to steady-state (time-independent) conditions, the net current
density depends on the change in neutron flux with position
](r) = - D V(r}
(4.6-2)
where D is the position-independent diffusion coefficient (em) and V(r) is

the gradient of the total neutron flux. It is noted that with the steady-state
...
161
RADIATION TRANSPORT
assump tion, phase space has been reduced to three posltlon variables
denoted by the vector (r). Equation 4.6-2 is known as Fick's law, which
simply states that the net diffusion of particles (or molecules) in liquids and
gases will be from regions of high particle density to regions of low particle
density, with the gradient of the particle-flux density as the driving
potential.
Substitution of Eq. 4.6-2 into the steady-state form of Eq. 4.6-1 leads to
the diffusion equation
D 'I7 2 4>(r) - ~a 4>(r) + S(r)
=0
(4.6-3)
Equation 4.6-3 has the same form as the steady-state form of the PI
approximation to the spherical-harmonics treatment of the Boltzmann
equation (see Sec. 4.3).
Certain limitations are inherent to diffusion theory: (I) the scattering
process is assumed to be isotropic in the laboratory frame of reference (the
use of a "transport-corrected" diffusion coefficient, D = 1/3~tr, removes
this limitation); (2) the directional distribution of the particle-flux density is
nearly isotropic; (3) the diffusing medium must be a poor absorber, i.e.,
~a ~ ~s; and (4) the results ate invalid for regions within two to three mean
free paths of boundaries, strong sources, and strong sinks. The existence of
these limitations is a clear indication of the approximate nature of diffusion
theory insofar as the physical situation is concerned. In reality, the preceding
conditions of applicability for diffusion theory are seldom satisfied.
However, with the judicious selection of system parameters, the diffusiontheory solutions of certain problemst compare favorably with solutions
obtained with more exact thedries or with the physical situation itself.
A neutron shielding problem would involve a continuous energy
spectrum over a wide energy range (typically from a low keV region to
10 MeV); thus a group apprbach is required to adequately describe the
diffusion process. The energy range is divided into G energy groups with the
gth group corresponding to the energy width Eg + I - E g .
The group-diffusion equation for the gth group is given by
(4.6-4)
tFor example, diffusion theory is used in fast reactor shielding design since the leakage spectrum
peaks below 0.5 MeV and the materials !nvolved are nonhydrogenous (sodium and graphite). Also, the
small source (reactor core) requires tiwo*dimensional calculations, which are much simpler with
diffusion theory.
162
where D g = group-averaged diffusion coefficient
JEEg +!
D(E) (r,E) dE
g
= ----''------<l>g(r)
<l>g(r)
= total
flux density corresponding to the gth energy group (g =

1,2, ... ,G)
=f
Eg l
+ (r,E) dE
Eg
= group-averaged macroscopic absorption cross section

g l
fE +
Eg
~a(E) (r,E) dE
=---"----:----:-:--<l>g(r)
Sg(r) = general source termt for thegth group
=
Eg
S(r,E) dE
However, the typical neutron shielding problem is not amenable to solution

by the straightforward application of diffusion theory, because the neutrons
are on the average very energetic and posse$S a strong forward directional
bias. The limitations of diffusion theory under these conditions are clearly
violated, and results thus obtained would be meaningless. However, when
applied to certain special problems in combination with other methods,
diffusion theory has proved useful. Applicatipns of diffusion theory to the
neutron shielding problem are discussed further in Sec. 4.8.
The use of diffusion theory to predict g'l.mma-ray energy-flux densities
seems to be unjustified on superficial examination of the gamma-ray-transport phenomenon. Certainly deep penetration by gamma radiation cannot be
described by diffusion theory, because the resultant gamma-ray-flux density
is due to photons that have maintained a strong directional correlation.
However, diffusion theory seems to be adequate for small-to-moderate
penetrations relatively near the source under conditions where the lowenergy end of the spectrum predominates and the scattering is more nearly
isotropic. These restricted conditions exist, for example, for most gamma-ray
heating calculations.
tObserve that the source term includes both particles b(~rn in energy group (EgEg+l) and particles
that have scattered into that energy interval from other ;groups. Thus the set of equations 4.6-4
(g = 1,2 ....,G) is coupled through the source term.
RADIATION TRANSPORT
163
4.7 INVARIANT IMBEDDING METHOD

The method of invariant imbedding is not another method for solving
the Boltzmann transport e<l)uation; rather, it is a different fundamental
approach to the mathematical description of particle transport. The method
has for its historical basis the early works of a Russian astrophysicist,
Ambarzumian,2 0 whose interest was in the transport problems of astrophysics. Recent investigations of Bellman, Kalaba, and Wing21 and Wing22
have shown that the invariant imbedding approach can be applied with
high-speed computers to a much broader class of problems, including the
neutron- and gamma-ray-transport problems encountered in radiation shielding.
The dependent variables of the invariant imbedding formulation are the
reflection and transmission functions, with the region dimensions (shield
thickness) and the energy and direction of the particle comprising a
six-dimensional phase space. In this context a particular shielding problem is
viewed as being imbedded in a more general class of shields having different
dimensions. Characteristically, and in contrast with solutions of the
Boltzmann transport equation, the invariant imbedding method provides
transmission and reflection information for a large variety of shields as well
as for the specific problem of interest. However, the detailed behavior of the
radiation during transport through the shield is not explicit during the
analysis and for that reason is unavailable, a not too serious shortcoming, if
not a real advantage, for the typical shielding problem. The pertinent
equations can be derived by applying the usual conservation principles of
radiation transport to a shield system, the dimensions of which are allowed
to vary by differential amounts. For simplicity and greater clarity, the
derivations are performed in slab geometry with azimuthal symmetry. Phase
space becomes three dimensional: the shield thickness, X, the energy
variable, E, and the direction variable, 11 (= cos Ii). A schematic representation of this configuration is shown in Fig. 4.6.
The reflection function R{X;Il,E;llo,E o ) dll dE is defined as the number of
particles reflected from a slah of thickness X with energies in dE about E and
directions that lie in dll about 11 per incident particle with energy Eo and
direction 110; the function can be regarded as an angular flux density within
the differential slab thickness dx. The reflection equation describes the
change in the reflection function due to changes in the shield thickness and
is formulated without involving the transmission function. The derivation is
accomplished by equating the difference in the reflection functions for slabs
164
REACTOR SHlELDlNG FOR NUCLEAR ENGINEERS
Incident particles
---::>t-..,.---t~---,,,,-- X
Transmitted
particles
X=o
+ dX
Reflected
particles
Fig. 4.6-Geometry for invariant imbedding technique.
of thicknesses X + dx and X with the net change in the reflection function

which results from collisions suffered by the particles within the differential
slab dX:
R(X+dX;p.,E;p.o,E o ) dp. dE - R(X;p.,E;p.o,E o ) dp. dE
dx
= - ~t(X,Eo) R(X;p.,E;p.o,Eo)-dp.dE
.
P.o
dx
dX
- ~t(X,E) R(X;p.,E;p.o,E o ) - dp. dE + - ~s(X;p.,E;p.o,Eo)
P.
p..
+
J 'l-'
o dp.
-I
dE
"
~s(X;p. ,E
,,
;p.o,E o ) R(X;p., E;p. ,E) -dx d p. d E

P.o
r d~' roo dE' R(X;p.',E';p.o,Eo ) ~s(X;II,E;p.',E') dX dp. dE

Jo p. Jo
ld~' JO dp." roo dE' roo dE" R(X;p.',E';p.o,E o )

r
Jo p. -I
Jo
Jo
X ~s(X;p.",E";p.',E') R(X;p.,E;p.",E") dX
lip. dE
(4.7-1)
where ~t(X,Eo) is the position-dependent total macroscopic cross section

evaluated at the particle energy Eo and ~s(X;IA',E';p.o,Eo) dlA' dE' represents
the position-dependent differential scattering cross section which describes
the probabili~y that a particle with an initial energy Eo and an initial
direction lAo undergoes a scattering collision that places it into a direction
which lies in dp.' about p.' with a new energy in dE' about E'.
RAQIATION TRANSPORT
165
The first and second terms on the right-hand side of Eq. 4.7-1 represent
the particle losses due to colUsions within dxt (any collision is presumed to
alter the particle's energy and direction). The first term is the loss of incident
particles scattered within dX such that they do not enter the slab of
thickness x, and the second term is the loss of particles that are scattered
within dx such that they are prevented from emerging from the slab of
thickness X + dX. The third. fourth, fifth, and sixth terms represent the
inscattering gains due to scattering collisions within dX. The fourth term is
the gain from particles that scntter from dx into the slab of thickness X with
energies in dE' about E' and directions dp.' about p.' and then are reflected
from the slab of thickness X with the proper emergent angle and direction.
The fifth term is the gain from particles that scatter from the slab of
thickness X into dX with energies in dE' about E' and directions dp.' about p.'
and then are scattered within dx with the proper emergent energy and
direction. The sixth term is the gain from particles that scatter from the slab
of thickness X into dX with energies in dE' about E' and direction dp.' about
p.', are scattered back into the slab of thickness X with energies in dE" about
E" and direction dp." about 1/', and are finally reflected from the slab of
thickness X with the proper emergent energy and direction. A rearrangement
of terms leads to the usual form of the reflection equation
d
1
[~t(X,Eo) + ~t(X,E)]
R(X;p.,E;p.o,Eo) = - ~$ (X;p.,E;p.o,E o ) dX
~
~
p.
, ,) - 1 +
X R(X;p.,E;p.,E
P.o
X ~ (X; E ' E') +

$
p."p. ,
r
Jo
1
10oP.o
dP.'L~d
,
" ;}lo,E )
E R(X;p.,E
o
~fo
d "r~ dE' r~ dE"
P. -1 P. Jo
J
o
(4.7-2)
with the initial condition that
(4.7-3)
tThe flight paths within the volUlme element
terms, respectively.
ax
are dX!lJo and dX//J for the fIrst and second
166
REACTOR SH1ELDlNG FOR NUCLEAR ENGINEERS
The transmission function T(X;IJ.,E;lJ.o,B o ) dlJ. dE is defined as the

number of particles that are transmitted through a slab of thickness X and
emerge with energies in dE about E and directions in dlJ. about IJ. per incident
particle with energy Eo and direction Ilo.
The transmission equation is derived in a manner similar to that used to
derive the reflection equation, making use of the fact that the difference in
the transmission functions for slabs of thickness X + dx and S is due to
collisions suffered by the particles within the differential slab dx. A familiar
form of the transmission equation is
d
~t(X,Eo)
dx T(X;Il,E;llo,E o ) = Ilo
T(X;Il,E;llo,E o )
fO
dll'
-1
r~
Jo
dE'
x T(X;Il,E;Il',E') 2
~s(XIIl',E';)lo,E o )
Ilo
r 1 d~' r10 dll" r~ dE' r~ dE"
Jo
Il
J-1
Jo
Jo
'<' ( X;Il,
"E'" ;)l, E')
X R(X;1l ",E ;llo,E o ) '"s
(4.7-4)
with the initial condition that
(4.7-5)
The first term on the right-hand side of Eq. 4.7-4 represents the loss of
incident particles due to collisions suffered within dX (any collision is
assumed to alter particle energy and direction). The second and third terms
represent the inscattering gains due to scattering collisions within dx. The
second term is the gain from particles that scatter from dX into the slab of
thickness X with energies in dE' about E' and directions in dll' about Il' and
finally emerge with energies in dE about E and directions in dll about Il. The
third term is the gain from particles that are reflected from the slab of
thickness X into dX and are then scattered bak into the slab of thickness X,
finally emerging with energies in dE about E and directions in dll about Il.
The reflection equation (Eq. 4.7-2) and the transmission equation
(Eq.4.7-4) are both nonlinear integra-differential equations that for the
radiation-transport problems of nuclear engineering form problems of the
initial-value type. The reflection equation involves only the reflection
RADIATION TRANSPORT
167
function as the dependent' variable; thus it can be solved without

consideration of the transmission equation. The transmission equation
appears simp ler in form (fewer terms) but contains the reflection function,
which must be known before a solution can be effected. Therefore a typical
shielding transmission problem (initial value) would involve the solution of a
coupled pair of nonlinear integro-differential equations. This is in contrast to
the Boltzmann equation, which is a linear integro-differential equation and
for the same application forms a boundary-value-type problem.
Analytical solutions of the reflection and transmission equations for
practical problems are not possible because of their complexity. Consequently all useful solutions are numerical in nature and are accomplished
with digital computers. The numerical techniques are similar to those used to
solve the Boltzmann equation by the discrete-ordinates technique, in which a
specific combination of the independent variables defines discrete values of
the neutron flux <l>C,I,D (see Sec. 4.4 for a more complete description).
In invariant imbedding specific combinations of the energy and direction
of the particle define the irh particle state. In this context the discrete
reflection variable Rij(X) is the number of particles in state i reflected by a
slab of thickness X due to a unit source of particles in state j that are
incident on the slab. The discrete transmission variable Tij(X) is the number
of particles in state i that penetrate a slab of thickness X due to a unit source
of particles in state j that are incident on the slab. The reflection and
transmission equations in discrete-variable notation, along with a general
description of the numerical techniques used in their solution, are -given by
Mathews, Hansen, and Mason. 2 3
The paper of Mathews et al. also describes the application of invariant
imbedding to practical energy.dependent neutron shielding problems, such as
for a thick water shield and a thinner heterogeneous iron-polyethyleneiron shield. A very detailed set of reflection and transmission equations in
particle-state notation for the monoenergetic-neutron-rransport problem in
slab geometry is given by Mingle, 2 4 who includes applications of the method
of escape probabilities, blackness coefficients, and critical-size determinations. Solutions for the gamma-ray-transport problem in slab geometry,
including results for slabs of iron, water, lead, and concrete, are given by
Shimizu and Mizuta. 2 s
The advantages and disadvantages of the invariant imbedding method
relative to other techniques should strongly influence the extent and
direction of future applications. The advantages of the method are that it
yields very detailed solutions (gives energy and angular distributions), is
168
REACTOR SHIELDlNG FOR NUCLEAR ENGINEERS
efficient for deep penetrations with reasonably short computer times, is well
suited for heterogeneous shield configurations, the effects of boundaries are
implicitly and exactly included in the solution, and it has the computational
advantages of being an initial-value problem. The disadvantages of the
method are that it is inefficient for thin shields (the method is very slow
during initial phases of solution), difficult to apply to other than slab
geometries, and does not generate detailed particle-state information within
the shield (actually an advantage from a computational point of view).
Furthermore, the basic equations are nonlinear (not too serious if the
solution is obtained numerically), and the calculational techniques and
computer programs are not as advanced as those for the solutions of the
Boltzmann equation.
4.8 KERNEL TECHNIQUE

The kernel technique, which in the language of mathematical physics is
known as the method of Green's functions, is one of the more widely used
methods for the solution of both gamma-ray and neutron shielding
problems. The point kernel used in shielding, K(lr - r'I), is formally the
solution to the unit point-source problem and is defined as the desired
response of a detector (particle-flux density, energy-flux density, dose, or
energy absorption) at the space point r due to a unit point source of
radiation at the space point r'. This kernel provides the means for solving a
variety of problems that involve distributed sources. As an illustration of the
procedure, consider the surface-source problem. In terms of the point kernel,
the detector response at a distance Ir - r'l '" R away from a differential
source area dA(R) of intensity SA (R) (particles cm-2 sec-I) is
d<l> = [SA (R) dA(R)] K(R)
(4.8-1)
The differential area dA(R) is selected so that the term [SA (R) dA(R)] can
be considered as a point source located at a distance R from the detector.
The total detector response (the desired answer to the surface-source
problem) is obtained by integrating over the entire source surface:
<l>
=f
d<l>
= fA
SA (R) K(R) dA(R)
(4.8-2)
The utility of the method is considerably enhanced if the integral can be

evaluated analytically.
RADIk'\TION TRANSPORT
169
Kernels that are used in practical shielding calculations almost invariably

result from solutions (either analytical or numerical) for infinite homogeneous media. Consequently applications of these kernels to finite-geometry
configurations require corrections for boundary effects.
Kernel techniques for gamma-ray-penetration analysis have evolved
somewhat differently than those for neutron-penetration analysis. For
photon transport the parameters within the kernels have been developed
primarily by numerical methods (predominantly Monte Carlo and moments
calculations) and were validated by experimental measurement. On the other
hand, most parameters used in neutron-attenuation kernels historically have
been empirical quantities obtained by measurement. For certain cases,
calculations have been made to verify the experimental results, and more
recently Monte Carlo and moments calculations have also been effectively
applied to the neutron problem.
It is convenient to divide the discussion of kernel techniques into two
parts. The following two sections discuss the basic concepts of kernel
methods for photons and neutrons, respectively. Applications of these
techniques are further discussed in Chaps. 6 and 8.
4.8.1 Gamma-Ray Calculationst
In the analysis of gamma-ray-transport problems, the uncollided-flux
density (i.e., the flux density due to source gamma rays that arrive at the
point of interest without suffering an interaction) is usually easily calculated.
For a monoenergetic point isotropic source in an infinite medium, the
uncollided-flux density (gamma rays cm -2 sec-I) is given by
(4.8-3)
where S =source strength (gamma rayslsec)
/.1(E) = macroscopic total cross section evaluated at the initial
gamma-ray energy, E (cm- I )
e-Jl(E)R = material attenuation factor, which is the probability that a
gamma ray of energy E travels a distance R (cm) without
suffering a collision
2
1
(41TR
= geometric attenuation for a point source (cm -2 )
tIn this section we will not need direction cosines, and we will return to the usual notation II- for
gamma-ray macroscopic cross section.
170
REACTOR SHiELDING FOR NUCLEAR ENGINEERS
Calculation of the scattered-flux density is, in general, much more

complex. The scattered component can be handled by introducing a buildup
factor, which accounts for the increase (Le., b\lildup) in the flux density at
some point r that is due to the scattered gamma rays. This buildup factor,
defined as
B
== [some desired property (particle-flux density, energy-flux density,

dose, etc.) of the total gamma-ray-flux density at R] I [same property
due to the uncollided-flux density at R 1
serves as the basis for formulating the point kernels required for gamma-ray
shield analysis. For the calculation of dose, the kernel is given by
Ilat (E) E .po (R) B
Pt S
_Ilat (E)Ee-Il(E)R B
4nR 2Pt
(4.8-4)
r
where ll at(E) is the macroscopic energy-absorption cross section for tissue

evaluated at the initial gamma-ray energy E; Pt is the density of tissue
(usually taken as 1.0 g/cm 3 ); and B r is the exposure buildup factor, t which
is the ratio of the actual dose at R to the uncollided dose at R. Similarly, if
the desired property is the energy absorbed per unit mass, the kernel is given
by
11 (E) Ee-Il (E)R

a
B
4nR2p
(4.8-5)
where lla(E) is the macroscopic energy-absorption cross section evaluated at

the initial gamma-ray energy E for the material in which the energy is
absorbed, and Ba is the energy-absorption (or energy-deposition) buildup
factor, which is the ratio of the actual energy absorbed at R to the
uncollided energy absorbed at R.
tFormerly called dose buildup factor. Away from boundaries for energies pertinent to reactor
shields, gamma-ray absorbed dose, kenna, and exposure are numerically equal (see Chap. 2).
RADIATION TRANSPORT
171
The results of many accurate gamma-ray-attenuation calculations are

reported in terms of buildup factors, which in combination with the kernel
technique provide a relatively simple and, in many cases, accurate calculational method. The primary source of buildup data is the collection of
moments-method data reported by Goldstein and Wilkins'6 for point
isotropic and plane monodirectional sources. The point-source data are
perhaps the most widely used because of their application to point-kernel
solutions. The infinite-plane data also find use in approximations such as
predicting the gamma-ray penetration through the side of a large cylindrical
or rectangular shield surrounding a large, diffuse source.
Tables of the exposure buildup factor vs. atomic number for various
penetration depths and initial energies are shown in Appendix E for a point
isotropic source and infinite-medium geometry.
A variety of functions have been derived to fit the buildup data. Some of
the more commonly used functions are:
1. The linear form
B(E,/lR)
= 1 + A dE)
/lR
(4.8-6 )
with only one constant, AI, to be evaluated as a function of energy. This

form has the advantage of simplicity but does not provide a consistently
accurate
fit.
2. The quadratic form

B(E,/lR) = 1 + A, (E) /lR + b(E) (/lR)'
(4.8-7)
should be applied to deep-penetration calculations.

3. The form due to Taylor' 6 is
B(E,/lR) = Ae-" 1 (E)/l
+ (1 _ A)e-"> (E )/lR
(4.8-8)
4. The polynomial form is
(4.8-9)
Capo' 7 published a set of the coefficient f3n(E) for a wide range of
materials and gamma-ray energies.
5. The form due to Berger' 8 is
B(E,/lR) = 1 + C(E) /lReD(E)/lR
(4.8-10)
172
This two-parameter formula is only slightly more complicated than the linear
form but provides better accuracy.
Trub ey29 published a review and evaluation of these and other empirical
equations which contains values of the fitted parameters and comparisons of
the expressions with the data of Goldstein and Wilkins. 6 Some of the
conclusions of his study are:
1. The linear form is generally not adequate.
2. The best fit is provided by the four-term polynomial fit by Capo.
3. For most purposes the function proposed by Berger is recommended
because it is simpler to apply and provides accuracy approaching that of the
more comp licated forms.
Values of the coefficients for all these forms of the buildup factor are
given for a variety of materials, photon energies, and source-detector
distances in Appendix F. Data in all these tables are for a point isotropic
source geometry. Applications of buildup factors to gamma-ray transport
problems are discussed in Sec. 6.4.1 and to gamma-ray heating in Sec. 8.1.1.
There are many other possible forms of buildup factors, such as those
cited by Hubbell, 3 0 but they are generally more complicated than the forms
given here. Hubbell's power-series form, for example, converges adequately
at short distances only and thus usually requires many terms, but it has the
advantage of allowing separation of the variables dependent on the medium
properties, geometry, and thickness.
Sometimes solutions to photon shielding problems can be reasonably
approximated from attenuation data for a disk or rectangular source and a
slab shield. These simple plane-source problems are amenable to analytic or
numerical solution by integrating a point kernel of the form of Eq. 4.8-3
over the source area. The results when tabulated or plotted are directly
usable in practical applications.
Consider, for example, a detector shielded from a plane disk source that
is uniformly emitting S monoenergetic photons cm-2 sec-' isotropically in
41T steradians (see Fig. 4.7). When the point kernel as given by Eq. 4.8-3 is
applied, the unscattered dose rate along the disk axis is
r(JJt,ro,z) - S G(E)
r"
Jo
e-p.t sec 8 (21Tr) dr

41TR 2
(4.8-11)
in which G(E) is a particle-flux-density to dose-rate conversion function. Its

units will determine the units of the dose calculated.
173
RADIATION TRANSPORT
Source
PIa.o
d'
--.LI =
-r~
Off"8xis
position
R~
rX
~:x --1--1--- Z~AXiS

-
z~t
~~~
'0 x
~
Fig. 4.7-Schematic diagram of disk source.
Since R' = " + z', Eq. 4.8-11 can be transformed to
f(lJt,zlro) =
where
S G(E)
i'.J1+(r"z)'
2.'
e-Y
-
dy
(4.8-12)
y == IJt sec e. When integrated, Eq. 4.8-12 becomes

f(lJt,zlro) =
S G(E)
2
{E, (IJt) ~ E, [lJt,j1 + (rolz)'])
(4.8-13)
where E, is the exponential integral function t of the first order and is

defined by
Edx)==
l
x
e-Y
-dy
y
(4.8-14a)
In general,
(4.8-14b)
[Equation 4.8-13, as well as the following equations for computing
uncollided doses, can be used to determine the total dose (uncollided plus
scattered) by using the Taylor form of the buildup factor.]
For an isotropic flux density <1>(0) at the source plane, which is
equivalent to a current density with a cos e angular distribution (as shown in
Chap. 2), the unscattered dose rate is
tGraphs of this function (as well as eX, E2 and 3) are given in Appendix G.
174
e-/J.t sec 6 cos 0 (21Tr) dr

41TR 2
(4.8-15)
wh ich integrates to
r(llt,z!ro) = <1>(0) ~(E) {E 2 (Ilt)
-.11 + 1(ro!z)2 E 2 [1lfV'1 + (ro!Z)2
J}
(4.8-16)
where E 2 is the exponential function of the second order.

The angular current density J(O) equivalent to <1>(0) in the forward
hemisphere is [(0)!41T] cos 0, defined as the number of photons per unit
solid angle crossing a unit area on the source plane in the directions within
the interval -1" cos 0 "1 as measured normal to the source plane.
Therefore
J(O) =
r <1>(0) cos 0 dn = Jor <1>(0)41Tcos 0 21T

In/2 41T
1
x d(cos 0)
= <1>(0)
(4.8-17)
If the current density in the forward direction is used, 4J(0) must be

substituted for <1>(0) in Eq. 4.8-16.
In general, for the cos n 0 angular distribution in the forward direction,
<1>(0)
=(n + 1) <1>'(0) cosn 0
(4.8-18)
21T
where <1>'(0) is the total or scalar flux density in the forward direction only,
and
r(llt,z!ro)
=(n + 1) ~(O) G(E)
{En+2 (Ilt)
_ E n+2 [1lfV'1 + (ro!Z)2
[vf1 + (ro!z)2] n+t
J}
(4.8-19)
175
RADIATION TRANSPORT
For the off-axis position at a distance p measured perpendicularly to the

disk axis (see Fig. 4.7), integrations must be done numetically. Hubbell,
Bach, and Herbold 31 integrated an expression similar to Eq. 4.8-11, the
isotropic source case, for off-axis positions and tabulated the results in terms
of the parameters p.t, zlr o , and plTo' These results are shown in Appendix H.
The quantity tabulated is 411" r(p.t,zITo ,plTo)1 [S G(E) J, which is the same as
411" <1>0 (pITO )IS, where <1>0 (pITo) is the uncollided-Iux density at plTo'
. Similarly, Trub ey 32 determined the data for an isotropic flux denSity.
The results are given in Appendix Has 2r(p.t,zITo,pITo)1 [<1>(0) G(E)] , which
is the same as 2<1>0/<1>(0).
Certain circular aperture and disk-source configurations to which these
results might be applied are shown in Fig. 4.8.
Semi-infinite medium
Plane isotropic
Plane isotropic extended source
circular disk
r7)!'??;o/-''m'7?l
source
~t-+Plane isotropic
circular disk source
t
Z
Infinite
medium
Plane
isotropic
extended
~-.........
source~:
T:1"
Circular'
aperture
J.l
..r:::::,.
rO !:
::
//
I
----.1..
0 . . . . . . .-
r-~~-r
~. I
ti
Slab
barrier
Fig. 4.8-Some circular aperture or disk-source shield configurations to which point

keTnels ate applicable (From Hubbell et al. 31 )
176
A solution was developed by Hubbell, Bach, and Lamkin 33 for the

uncollided-flux density at a distance z from a plane isotropic rectangular
source. Expressed as the product of separable source and geometry
functions, the uncollided-flux density is given by
o
~ 2n+ 1
 (a,b) = LJ - 2 - gn p(a,b)
(4.8-20)
n=O
where gn and Pn(a,b) are Legendre coefficients of the source and geometry
functions, respectively.
If a = H/z and b = W/z, where Hand Ware the height and width,
respectively, of the source plane (see Fig. 4.9), then Eq. 4.8-20 gives the flux
density at the corner position, that is, the flux density at a distance z along
the normal to the corner of the rectangular source. It follows that the use of
aH
//
3
/1
I
I
Fig.4.9-Schernatic dIagram demonstrating the use of the corner position of a

rectangular source to calculate dose at an arbitrary position by point-kernel techniques.
RADIATION TRANSPORT
177
the half-height and half-width gives one-fourth of the flux density at z along
a normal to the center of the source plane. The source function is
gn =
J: g(cos 0) Pn(cos 0) d(cosO)
(4.8-21 )
where g( cos 0) represents the angular flux density at the source plane for the
case of a slab shield of thickness t (z :;;. t) located between the source and the
detector at a distance z from the source; that is,
5 e-/lt!cos9
g(llt,cos 0) = 4
0
1r cos
(4.8-22)
Substituting Eq. 4.8-22 into Eq. 4.8-21, Hubbell et al.," evaluated gn and Pn
numerically and solved Eq.4.8-20. The results for a corner position
(a = H!z, b = W/z) are given in Appendix H as 411"f/[5 G(E)J, or 4<1>/5, in
terms of the parameters Ilt, a, and b.
Similarly, Trubey" numerically evaluated the equivalent of Eq. 4.8-20
for an isotropic flux density (cosine distribution of the angular current
density), that is, for
g(llt,cos 0) =
<1>(0) e-/lt!cos9
411"
(4.8-23)
These results for acorner position are given in Appendix Has 2f/ [<1>(0) G(E) J ,
or 2<1> /<1>(0), which is the same quantity tabulated for the disk source in
Appendix H. For a square, the dose will be slightly greater than that for a
disk of radius W.
Although these results relate directly to the response of a detector in a
corner position, they are also applicable to any arbitrary position lying
within the projection of the source plane. It is obvious from Fig. 4.9 that the
dose rate at the detector is
OIH (3W)
[( 1 -zOl)H,( 1 -z(3) W]
+ f z ( 7' --;- + f,
(4.8-24)
178
4.8.2 Neutron Techniques

As noted previously, the use of buildup factors in the attenuation
function, or kernel, for neutrons has not developed to a large extent,
primarily because neutron interactions are more complex than gamma-ray
interactions. However, a simple kernel developed by Albert and Welton,"
which uses an energy-dependent hydrogen cross section, has been widely
applied to hydrogenous shields. This section contains a description of the
Albert-Welton kernel and a discussion of removal cross sections that are
required in the kernel when other shield materials are used in conjunction
with the hydrogenous medium.
Other kernels that can be used to calculate differential energy spectra of
neutrons in hydrogenous media can be developed from the moments method
or Monte Carlo calculations. For example, the moments-method code
RENUPAK was used 32 to calculate the differential energy-flux density as a
function of distance from a point fission source in an infinite medium of
lithium hydride, and an empirical attenuation function based on the results
was then incorporated in the point-kernel code QAD by Solomito and
Stockton 3 5 for use in space-reactor shield designs. Since these kernels are for
an infinite medium, some caution should be exercised in their use. However,
simple neutron-dose attenuation functions determined from such calculations can be used for preliminary shield-design applications, and some of
these functions for concrete and several other materials are included in this
section.
(a) Removal Cross Sections. Measurements in the Oak Ridge National
Laboratory (ORNL) Lid Tank Shielding Facility 36 showed that the insertion
of relatively thin slabs of material between a fission source and a thick water
shield gives an effect that can be correlated by a simple exponential
attenuation factor that is characteristic of absorption processes alone. This
behavior might not be expected since nonabsorption effects predominate in
fast-neutron attenuation. However, the large thickness of water (at least
6 g/cm 2 ) filters out the neutrons deflected by the sample and thereby effects
their complete removal. Therefore the effect of slabs of shield materials
when followed by large thicknesses of hydrogenous material can be
described by an equivalent absorption cross section averaged over energy,
called the removal cross section.
An ideal way to experimentally determine the validity of the concept
would be to use a plane monodirectional source of fission neutrons incident
on a tank of water. For such a configuration the removal-cross-section
RADIATION TRANSPORT
179
concept would be valid if the doses measured at the source distance z in

water could be correlated by
(4.8-25)
where D, (z) is th'e observed neutron dose attenuated through a distance z of
water, D 2 (z) is the observed neutron dose attenuated through a slab of
material of thickness t (inserted between source and water) plus water of
thickness z, and ~ R is the macroscopic removal cross section.
In the experimental shielding facility where this concept was originally
tested, the source was a finite isotropic disk rather than a plane
monodirectional source. However, with a few simple assumptions about the
behavior of neutron penetration, an analog to Eq. 4.8-25 was derived and
used in obtaining removal cross sections from experimental data. 3 7
Values of microscopic removal cross sections (OR) determined from
measurements at the Lid Tank Shielding Facility for several elements and
compounds are shown in Table 4.2. Empirical functions useful for interpolation in the experimental data have been derived by Zoller: 38
~R
= 0.19Z-0.743 cm 2 /g
=0.125Z- 0 .565
~R =0.206A-\!o
(for Z';;; 8)
cm 2 /g
(forZ>8)
(4.8-26 )
Z-0.294 ""
0.206 (AZj"\!o
(4.8-27)
where Z is the atomic number. Most of the macroscopic removal cross

sections given in Table 4.3 were obtained with Eq. 4.8-26.
It should be emphasized that most of the removal cross sections
determined by experiment were obtained for a slab-type configuration with
water follOwing the shield material and may not be applicable to other
configurations. Trubey and Chapman 3 reported that in a homogeneous
medium the removal cross section for oxygen is 0.75 0.05 b rather than
0.99 0.10 b as shown in Table 4.2. They also note that the removal cross
section may vary with sample thickness (the value for oxygen obtained from
the homogeneous-medium measurements increased from 0.72 b at a distance
90 cm from the source to 0.79 b at a distance 140 cm from the source).
180
Table 4.2-MICROSCOPIC REMOVAL CROSS SECTIONS OF VARIOUS

ELEMENTS AND COMPOUNDS MEASURED AT THE ORNL
LID TANK SHIELDING FACILITYt
Material
Aluminum
Beryllium
Bismuth
Boron
Carbon
Chlorine
Copper
Fluorine
Iron
Lead
Lithium
Nickel
Oxygen
oR' blatom
Material
oR' b/atom
Tungsten
Zirconium
Uranium
Boric oxide, B'l0 3
Boron carbide. B4 C
3.36
2.36 0.12
3.6 0.4
4.30 0.41
4.7 0.3
1.2 0.8
2.04 0.11
1.29 0.06
1.98 0.08
3.53 0.30
Fluorothene, c..zF 3 CI
Heavy water, D 2 0
6.66 0.8
2.76 0.11
1.01 0.05
1.89 0.10
0.99 0.10
Oil, CH,
1.31
1.07
3.49
0.97
0.81
0.05
0.06
0.35
0.10
0.05
Hevimet (90 wt.% W,

6 wt.% Ni, 4 wt.% Cu)
Lithium fluoride, LiF
Paraffin, C 30 H 62
Perfluoroheptane, C 7 F 16
3.22 0.18
2.43 0.34
2.84 0.11
80.5 5.2
26.3 0.8
tFrom E. P. Blizard (Ed.), Reactor Handbook, 2nd ed., Vol. III,

Part B, Shielding. p. 83, Interscience Publishers, a division of Jaho
Wiley & Sons, Inc., New York, 1962.
There is really no reason to expect the removal cross section to remain

constant with sample thickness since the removal concept is the result of a
crude application of theoretical principles; however, the variation should not
be very great up to about five relaxation lengths.
Another point to be emphasized is that the removal cross section for a
material can be applied only when that material is used in conjunction with a
hydrogenous shield since hydrogen is required to moderate and absorb the
scattered neutrons, as occurred in the experiments for determining the
removal cross sections. Thus a minimum thickness of hydrogenous material
following the sample is required to validate the use of removal theory. When
the hydrogenous material is water, moments-method calculations have
shown that about 60 em is required, whereas experimental results indicate
that about 45 em is adequate. If the hydrogen thickness is less than that
provided by this amount of water, the neutrons interacting with the heavier
elements will not be adequately "removed", and the effective removal cross
section will be decreased.
It follows from the removal-cross-section concept that the removal cross
sections of elements in a series of slabs or mixed together should be additive;
i.e., the number of relaxation lengths becomes
RADIATION TRANSPORT
181
where the index i refers to the various elements. This additive property has
been generally accepted, even though some discrepancies have been noted,
particularly in regard to compounds.
Removal cross sections can be predicted theoretically. Phenomenologically, the removal process can be considered equivalent to the total reaction
rate minus the forward component of the scattering process. This suggests
that an estimate of the removal cross section could be obtained from the
transport cross section. As it turns out, ~R ~ ~tr for neutrons between 6
and 8 MeV; therefore
~R
= ~tr = ~t -
~s
cos ()
(4.8-28)
where cos () is the average cosine of the neutron scattering angle per collision
in the laboratory system.
Removal cross sections can also be estimated from
(4.8-29)
where ~t is the average total macroscopic cross section between 6 and
8 MeV, and from
~R = 0.21XO.5 8
P
(4.8-30)
where p is the density and A is the atomic weight. Figure 4.10 compares
graphs of measured values of ~Rlp and ~tlp at 8 MeV as a function of
atomic weight. A reasonably good fit to the curve for A > lOis obtained by
Eq.4.8-30.
(b) Removal Cross Sections for Hydrogen-Deficient Shields. The traditional removal cross section as discussed in the preceding section is limited in
application to a shield configuration that has a hydrogen density of at least
6 g/cm 2 in its outer layer. Recognizing the usefulness of a removal cross
section that could be applied for shields that contain less hydrogen, Dudziak
and Schmucker' performed a series of calculations to investigate the effect
on the removal cross section of varying the surface density of the hydrogen.
Using a simplified P3 approximation to the transport equation, they
182
Table 4.3-FAST.NEUTRON REMOVAL CROSS SECTIONS

AND MASS ATTENUATION COEFFICIENTS!
"ER/P
Atomic
Element
Aluminum
Antimony
Argon
Arsenic
Barium
Beryllium
Bismuth
Boron
Bromine
Cadmium
Calcium
Carbon
Cerium
Cesium
Chlorine
chromium
Cobalt
Copper
Dysprosium
Erbium
Europium
Fluorine
Gadolinium
Gallium
Gerrnanium
Gold
Hafnium
Helium
number
0.0292 0.0012
5.730
3.500
0.0293
0.0136
0.0244
0.0173
0.0129
4
83
5
35
48
9.013
9.747
3.330
3.120
8.648
0.0678
0.0103
0.0575
0.0168
0.0140
0.1248
0.1003
0.1914
0.0523
0.1213
20
6
58
55
17
1.540
1.670
6.900
1.873
0.0230
0.0502
0.0126
0.0130
0.0252
0.0354
0.0838
0.0870
0.0243
24
27
29
66
68
6.920
8.900
8.940
8.562
4.770
0.0208
0.0194
0.0186
0.0117
0.Q115
0.1436
0.1728
0.1667
0.1003
0.0550
63
9
64
31
32
5.166
0.0120
0.0361
0.0119
0.0180
0.0176
0.0621
7.868
5.903
5.460
4.930
22.420
7.865
Lanthanum
Lead
Lithium
Lutetium
Magnesium
Manganese
Mercury
Molybdenum
82
3
71
12
25
80
42
11.347
0.534
Iron
Krypton
0.0792
0.0907
2.699
6.691
53
77
26
36
57
Iodine
em-I
13
51
18
33
56
19.320
13.300
Iridium
cm'/g
(exf')'
cm 18
(calc.),
79
72
2
67
49
Holmium
Indium
"ERlp
"E R
P.
g1cm 3
7.280
6.150
1.741
7.420
13.546
10.200
0.0993
0.0450
0.0717 0.0043
0.010 0.0010
0.0540 0.0054
0.0407 0.0024
0.020 0.014
0.0194 0.0011
0.0409 0.0020
0.0938
0.1060
0.0963
0.0106
0.0112
0.1135
0.0116
0.0139
0.2045
0.1484
0.0133
0.0107
0.0198
0.0165
0.0127
0.0654
0.2408
0.1560
0.0104
0.0840
0.0112
0.0307
0.0203
0.Q105
0.0151
0.1176
0.0449
0.1009
0.0214 0.0009
0.0783
0.0535
0.1505
0.1424
0.1543
!From L. K. Zoller. Nucleonics. 22(8): 129 (1964)
0.0103 0.0009
0.094 0.007
RADIATION TRANSPORT
183
Tabl. 4.3-(Continucd)
'f,Rlp
'f,R/P
Element
Neodymium
Neon
Nickel
Niobium
Nitrogen
Osmium
Oxygen
Palladium
Phosphorus
Platinum
Potassium
Praseodymium
Radium
Rhenium
Rhodium
Rubidium
Ruthenium
Samarium
Scandium
Selenium
Silicon
Silver
Sodium
Strontium
Sulfur
Tantalum
Tellurium
Terbium
Thallium
Thorium
Thulium
Tin
Titanium
Tungsten
Uranium
Vanadium
Xenon
Ytterbium
Yttrium
Zinc
Zirconium
Atomic
p,
number
g1cm 3
(oale.).
cm
/g
'f,R,
em-I
0.0124
0.0340
0.0190
0.0153
0.0448
0.0861
0.2432
12.160
1.820
21.370
0.0108
0.0405
0.0144
0.0271
0.0107
19
59
88
75
45
6.475
6.500
5.000
20.530
12.440
0.0237
0.0125
0.0100
0.0109
0.0145
0.1533
0.0812
0.0498
0.2238
0.1810
37
44
62
21
34
1.532
12.060
7.750
3.020
4.800
0.0163
0.0147
0.0121
0.0224
0.0170
0.0249
0.1777
0.0941
0.0676
0.0818
14
47
11
38
16
2.420
10.503
0.971
2.540
2.070
0.0281
0.0142
0.0322
0.0160
0.0261
0.0681
0.1491
0.0313
0.0407
0.0540
73
52
65
81
90
16.600
6.240
0.1838
0.0837
11.860
11.300
0.0111
0.0134
0.0118
0.0104
0.0098
6.550
4.500
19.300
18.700
0.0114
0.0137
0.0218
0.0110
0.0097
60
10
28
41
7
6.960
76
8
46
15
78
22.480
69
50
22
74
92
23
54
70
39
30
40
8.900
8.400
5.960
3.800
7.140
6.440
0,0213
0.0131
0.0113
0.0158
0.0183
0.0156
0.1693
0.1288
(oxf')'
em IS
0.0190 0.0010
0.031 0.002
0.1747
0.0493
0,2279
0.1238
0.1111
0.0898
0,0981
0.2120
0.1816
0.1267
0.0599
0.1306
0.1001
0.0082 0.0018
0.0091 0.0010
184

10 c--,-r-,-r,..,.,-rr--..--'-..-TTTTr,......-,.......,...,-r,..,.,..,.,
:r Total cross section at 8 MeV

.6.:r RI Removal cross section from
t
measurements on compounds
.L R , Removal cross section from

measurements on elements
0.01
0.OO5!-,-..,2!c-.J-"'-:!5,..u-'-':,~0--:2!,0,..... .....-;!50"u.J.+.,00!:,,-....,,200"""..J...."'5,.;OOb?"~,000
A (ATOMIC WEIGHT)
Fig.4.10-Removal cross sections per unit mass for fission neutrons as a function of
atomic weight. (From Chapman and Storrs. 3 6)
calculated the transport of neutrons from a Po-Be source through lead

followed by varying amounts of polyethylene in spherical geometry. They
showed that removal cross sections for large thicknesses of polyethylene
approached the asymptotic value of 0.116 0.01 cm- 1 (3.53 0.30 b)
reported from the Lid Tank Shielding Facility measurements with a fission
source. They also found that by extrapolating the cross-section curve back to
a zero thickness of polyethylene they obtained a value that was very close to
the value of 0.128 cm- 1 for lead alone reported by Price and Dunn" on the
basis of experiments with Po-Be neutrons.
Dudziak' 2 later performed calculations with better cross sections, more
neutron-energy groups, and an S S P3 approximation and reached essentially
the same conclusions. The results of the calculations by Dudziak and
Schmucker' 0 are shown in Table 4.4. The macroscopic removal cross
sections given in the table were fit to within 1.4% by a least-squares
procedure to an analytical expression given by
RADIATION TRANSPORT
~R(t) =
0.1106 (1 ~ 0.9836 e-O.lo9t)
185
(4.8-31)
where t is the polyethylene thickness in centimeters.

In calculations similar to those of Dudziak and Schmucker, but for a
fission source, Shure, O'Brien, and Rothberg43 determined removal cross
sections for both iron and lead followed by polyethylene thicknesses up to
50 em. They used two cutoff energies, E e , 302 keV and 0.625 eV, in their
dose calculations. The resulting microscopic removal cross sections, shown in
Table 4.5, indicate that, at least for small thicknesses of polyethylene, the
removal cross sections are relatively insensitive to the cutoff energy chosen.
Shure et al. also found that the asymptotic values for iron and lead are in
very good agreement with the experimental values of 1.98 0.08 band
3.53 0.30 b, respectively, obtained from Lid Tank measurements. 3 7 When
no polyethylene was present, the value for lead of 0.74 b (0.0243 em-I)
obtained for the fission source is nearly twice that for the Po-Be source,
which could be expected since the Po-Be spectrum is harder. The fact that
the asymptotic removal cross sections are in agreement for the two different
sources shows that the inelastic scattering in lead apparently degrades the
high-energy part of the spectra sufficiently to produce spectra that
equilibrate about equally after traversing approximately 10 em of polyethylene.
Table 4.4-EFFECTIVE MACROSCOPIC REMOVAL CROSS
SECTIONS FOR LEAD FOLLOWED BY VARIOUS
THICKNESSES OF POLYETHYLENE (Po-Be SOURCE)t
Polyethylene
thickness,
em
Lead
removal cross
-1
sectIon, em
7
9
15
0.0328
0.0473
0.0597
0.0701
0.0901
21
25
31
35
41
0.0999
0.1036
0.1069
0.1081
0.1093
tFrom Dudziak and Schmucker. 40
186
Table 4.5-EFFECTIVE MICROSCOPIC REMOVAL CROSS SECTIONS

FOR IRON AND LEAD FOLLOWED BY VARIOUS
POLYETHYLENE THICKNESSES (FISSION SOURCE)t
Removal cross section, b
Polyethylene
thickness,
em
0
1
2
5
10
15
20
25
30
40
50
Lead
Iron
Ee
=302 keY
0.76
0.94
1.08
1.37
1.66
1.84
1.91
1.95
1.97
1.99
1.99
Ee
=0.625 eV
0.61
0.78
0.91
1.24
1.60
1.79
1.89
1.95
1.97
1.99
1.99
Ee
=302 keY
Ee
0.88
1.17
1.41
1.92
2.47
2.83
3.06
3.20
3.29
3.37
3.41
=0.625 eV
0.74
1.00
1.22
1.74
2.36
2.77
3.03
3.19
3.28
3.37
3.41
tFrom K. Shure, J. A. O'Brien, and D. M. Rothberg,NuclearScienceandEngineering, 35: 373 (1969).
Shure et al. also investigated the use of these removal cross sections for
hydrogen-deficient shields with the technique that is normally used when
removal croSS sections are applied to large thicknesses of hydrogenous
shielding following the laminations of nonhydrogenous material. The normal
expression used to estimate the neutron dose transmitted through t em of
hydrogenous material and several slabs of nonhydrogenous materials is
Dlam (x,t) = Dealc(O,t) exp (-
f NjOiti)
(4.8-32 )
where Deale = calculated dose for the hydrogenous material alone

Ni = density
a i = removal cross section
ti =thickness of the ith nonhydrogenous material
Equation 4.8-32 was applied to three different shield configurations using
lead and iron followed by various thicknesses of polyethylene. Values of
Deale were obtained, and values of Ui from Table 4.5 were used. A
comparison of the resulting dose rates with dose rates obtained with a
separate, complete transport calculation showed that the frequently used
design technique exemplified by Eq. 4.8-32 does provide good estimates
RADIATION TRANSPORT
187
(within 10%) of the dose rates when the removal cross sections for
hydrogen-deficient shields are used.
(c) Albert-Welton Kernel. The experimentally determined removal
cross section provides a simple method for determining the attenuation
through nonhydrogenous portions of shield material if there is a hydrogenous portion of the shield. Albert and Welton" developed a semiempirical
theory of neutron attenuation which provides a simple method for
calculating neutron attenuation through the complete shield; this theory is
based on the removal-cross-section concept. Basic to the Albert-Welton
model is the assumption that any collision with hydrogen has the effect of
an absorption. This, in effect, neglects the buildup of scattered neutrons that
have undergone only small-angle scatterings by hydrogen. Inelastic scatterings with heavier nuclei are also regarded as absorptions because of the
characteristically large energy loss. Other collisions are mainly small-angle
elastic scatterings within the forward peak of the angular distribution, which
amount to virtually no collisions. Attenuation through the materials in the
shield is described in terms of removal cross sections. For hydrogen the
removal cross section is taken to be its energy-dependent total cross section,
and for the heavier nuclides it is taken to be an empirical energy-independent
removal cross section, such as the removal cross sections described in the
preceding paragraphs. Thus the Albert-Welton model provides a theoretical
basis for the removal-crass-section concept.
The Albert-Welton formulation for fission neutrons from a plane

monodirectional source which penetrate through a mixture of water and
heavy materials is given by
(4.8-33)
where (r) = number flux density at a distance r from the source
~ R j = macroscopic removal cross section of ith element (other than
hydrogen)
fj = volume fraction of ith nonhydrogenous material
S(E) dE = fraction of fission neutrons at E in interval dE for a total
source of 1 fission em-2 sec- 1
~ H = total macroscopic cross section for hydrogen
The proportionality constant included in the original Albert-Welton
derivation has been removed from Eq. 4.8-33 to avoid the implication that
the actual number flux density can be computed from this relation.
188
Integration of Eq. 4.8-33 yields the original Albert-welton kernel for

the hydrogenous portion of the shield, which is included in the first set of
braces in the following relation:
(r) a: {(jW r )0.2 9 exp[ ~O. 928(jw r)0'5 8 ]}
Xexp
where
[- fwr~Ro - (1- fw)rI;., f i ~R;l
(4.8-34 )
f w = volume fraction of water
l: R 0 = removal cross section of oxygen

~ R i = removal cross section of nonhydrogenous materials other than
the oxygen in the water
r = distance (in cm) from source
Although the derivation was for a plane source, Eq. 4.8-34 holds for a
point source when multiplied by the geometric factor 1/41Tr2 , and the
integral of S(E) is normalized to 1 fission/sec. The relation is also valid when
slabs of heavy material are laminated with the water. A minimum of about
50 or 60 cm of water is required between the dose point and the last of the
heavy materials (whether as slabs or in a mixture) to comply with the
limitations of the removal-cross-section concept.
On the basis of more recent experimental results, Casper44 evaluated
new constants for the Albert-Welton kernel. The result for a point
fission-spectrum source is
(4.8-35 )
where D(r) is the neutron dose rate r centimeters from the source [in
(rads hr" )/(neutrons sec" )]. When shield materials are inserted between the
water and the fission source, Eq. 4.8-35 is multiplied by
to obtain the dose rate at the shield surface.

4.9 COMBINATION REMOVAL-DIFFUSION METHODS
The removal-cross-section concept described in Sec. 4.8.2 provides a
method for calculating the dose rate due to high.energy neutrons that
RADIATION TRANSPORT
189
penetrate a hydrogenous shield; however, the technique cannot be used to

predict the dose rate due to neutrons that have been moderated to
epithermal and thermal energies or to predict the thermal-neutron-flux
density, which is used to obtain the capture gamma-ray source distribution
within the shield. t
The energy and spatial distributions of the moderated neutrons
throughout a shield have sometimes been calculated by using the elementary
theories of neutron diffusion and moderation (see Sec. 4.6). However, these
methods of reactor physics are normally used to predict the average behavior
of neutrons involved in reactor criticality problems, and, in the typical
shielding problem, the neutron of significance is the unusual fission neutron
that is born with an energy much greater than the average and contributes
very little to reactor criticality. This unusual neutron penetrates into regions
deep within the shield.
The inadequacy of both the removal concept and the elementary
methods of reactor-core physics in calculations of the whole shielding
problem has resulted in neutron transport's being regarded as a two-step
process: a step in which a high-energy neutron penetrates to a position deep
within the shield, where it suffers a collision that degrades its energy
significantly, and a step in which the resulting low-energy neutron enters a
diffusion process. Characteristically, the distance traveled by the neutron
during the diffusion process is very much shorter than the distance it
traveled as a fast neutron, and, once it has entered this second phase, the
methods of elementary reactor physics conceivably could apply. Such
reasoning prompted the first-flight correction to the age in Fermi age
theory.2 This correction was necessary because a neutron cannot enter a
process described as continuous slowing down (as required by Fermi age
theory) until it has had at least one collision.
The development of high-speed computers and the resulting extensive
use of multigroup diffusion theory for reactor criticality problems made the
development of a technique that utilized diffusion theory even more
attractive. In one of the first attempts to develop such a technique,
Haffner 4s in 1958 used diffusion theory to calculate thermal-neutron-flux
densities within a reactor shield and then normalized the results at each
space point according to the fast-neutron dose rate obtained with the
Albert-Welton kernel (see Sec. 4.8.2). Anderson and Shure 46 used a similar
technique when they applied a known pure-water kernel to normalize
tCalculations of capture gamma-ray doses are discussed in Chap. 6.
190
diffusion (actually P, multigroup) results for a metal-water mixture. In

general, they obtained good results for laminated iron-water shields.
(Shure 3 ,47 later showed that a straightforward P 3 calculation without the
use of a kernel also gave good results.) The main assumption in the
Anderson-Shure technique is that the multigroup procedure correctly
calculates the ratio between the flux densities in water and those in a
metal-water mixture.
After several attempts had been made to develop a technique by
correcting diffusion-theory results, a different approach to the problem
evolved: a first-flight correction was made before the diffusion-theory
calculation was performed. In the early calculations this was done by
computing the singly scattered neutron-flux density from the uncollided-flux
density and then using it as a source for the diffusion-theory calculation. A
difficulty inherent in this procedure, especially for hydrogenous media, is
that the penetrating component does not consist of uncollided neutrons
alone but rather is composed largely of neutrons that have had one or more
collisions and have suffered only small angular deflections. When these
neutrons were accounted for, the first successful two-step model for
neutron-penetration calculations became available. The method, in which the
fast-neutron removal concept and age-diffusion theory are combined, is
commonly referred to as the Spinney method, after its chief developer. The
remainder of this section is devoted to a description of the original version of
this removal-age-diffusion method and subsequent variations of it.
4.9.1 The Spinney Method
The Spinney method as first described by Avery, Bendall, Butler, and
Spinney 48 is characterized by the following basic physical assumptions:
1. The penetrating component of the source neutrons consists of the
high-energy neutrons that suffer only small energy loss through small-angle
elastic collisions and the uncollided neutrons.
2. Neutrons that suffer large energy loss through either wide-angle elastic
or inelastic scattering are regarded as being removed from the fast beam.
3. The removed neutrons are degraded in energy in accordance with age
theory and do not travel significantly from the point of removal.
4. The removed neutrons have a spectral and spatial distribution closely
described by the conventional age-diffusion theory near the source.
5. Neutrons removed after they have penetrated deep into a homogeneous medium develop an equilibrium spectrum and are attenuated at the
same rate as the penetrating component.
RADIATION TRANSPORT
191
6. The equilibrium spectrum of the degraded neutrons is disturbed near

the boundaries between dissimilar media. [Note: Assumptions 4, 5, and 6
arc not required in recent versions of the Spinney method.]
The neutron-flux density (removal flux) that corresponds to the

penetrating component of the source neutrons is given by the kernel
5
e-~Rr
<l>(r) = 2
47Tr
(4.9-1)
where 50 is the source strength of high-energy neutrons of energy E, ~R is

the removal cross section evaluated at the source energy E (determined
experimentally or approximated by the transport cross section (see
Sec. 4.8.2), and r is the distance traveled by the neutron to its first collision.
The removed neutrons are regarded as a local source of degraded
neu trons, the behavior of which can be adequately described by diffusion
theory. The intensity of this source is given by
5 ~ -~Rr
5(r)=(r)~R= 4Re2
7Tr
(4.9-2)
These neutrons (i.e., the removed neutrons) are then introduced into the
highest energy group of an appropriate set of multigroup diffusion equations
to calculate the distribution of the low-energy neutron-flux density. The
equations comprising the multigroup set are given by
2
2
~a!
5(r)
V <l>dr) - k l <l>dr) - <1>, (r) + =0
Dj
Dj
'1 2 <l>j(r) -
kl
~aj
<l>j(r) - Dj <l>j(r) +
Dj_l kr_! <l>j_, (r)

Dj
(i = 1)
(i> 1)
(4.9-3)
where <l>j = group flux density for the ith group

~aj = group-averaged macroscopic absorption cross section
Dj = group-averaged diffusion coefficient
ki! = slowing-down length for the ith group
The slowing-down length is calculated according to age theory and for
the ith group is given by
1,.)2
(k
rEi
JEi_l 3~(E)
dE
~s(E) ~tr(E) E
(4.9-4)
192
where ~(E) is the average change in lethargy per collision for neutrons of
energy E, ~s(E) is the macroscopic scattering cross section for neutrons of
energy E, and ~tT(E) is the macroscopic transport cross section for neutrons
of energy E.
In the original formulation of the Spinney method, five energy groups
were taken for the multigroup diffusion calculation. The bottom group,
which was a thermal group, had an upper energy of 2.81 kT
(k = 8.61 X 10-5 eV/oK), and the highest group (i = 1) had an upper energy
of 2 MeV. All removed source neutrons were assumed to be placed directly
into the highest group. Solution of the group-diffusion equations, of course,
required that boundary conditions be specified at the inner and outer
surfaces of the shield. A zero reentrant condition was imposed at the outer
boundary; this was stated in terms of the extrapolated boundary condition,
which requires that the flux densities vanish at a distance 3.13Dj beyond the
physical boundary. The boundary conditions at the inner surface of the
shield were established by requiring that the flux densities and current
densities be equal to those determined from reactor-core calculations.
The original formulation was used with some success in predicting the
distribution of low-energy neutrons in concrete shields for existing graphitemoderated reactors, but it was not suited for general application. Some of its
inadequacies were that (1) all the removed neutrons were placed in one
group, which neglected any additional diffusion-type transport that could
have been accomplished at energies greater than 2 MeV; (2) not enough
groups were used to adequately represent the continuous slowing-down
process; and (3) the transfer of neutrons from one energy group to the next
lower group did not describe the large energy losses experienced by neutrons
that had suffered an inelastic scattering or a collision with hydrogen.
4.9.2 Variations of the Spinney Method
Many modifications and variations of the Spinney method have been
developed, such as the RASH E, MAC, NRN, SABINE, and ATTOW codes.t
In the RASH E formulation of Bendall" 9 and Butler; 0 the modifications include an increase in the number of groups to 16 and a broader energy range
(0 to 10 MeV). Also, the ll1ultigroup equations have been modified to include
tRASH E is the latest member of the RASH family of codes utilizing the Spinney method,
RASH E is included in a FORTRAN code package known as COMPRASH and can be obtained from
till' Radiation Shielding Information Center, Oak Ridge National Laboratory, Oak Ridge, Tenn. Code
packages for MAC, NRN, SABINE, and ArrOW arc also available from this center.
RADIAnON TRANSPORT
193
a direct source of removed neutrons into the nine highest energy gtoups. The
equations so modified ate as follows:
(i = 1)
V <l>j(r) -
kf
Lai
<l>j(r) - J.): <l>j(r)

I
Dj_1
D
kJ_,
j
1/Ij(r)
<l>j_1 (r) + - - = 0
Dj
(i = 2. 3, ... , 9)
kaj
V 2 '"
'j(r) - k j2 <l>j(r) ~ J.):
<l>j(r)
I
kI
Di-z
<l>r(r) - -r(r) +
Dr
kl_ 1
Dj
D 15 kj 5
<l>j_, (r) = 0
<1>, 5 (r) = 0
(i = 10, 11, ... , 15)
(i=16)
(4.9-5)
Dr
where Tcorresponds to i = 16 and designates the thermal-group flux density.

The source term for the jth group resulting from removed neutrons is
designated as 1/Ij(r) and is determined in the following manner. The fission
spectrum is divided into 18 energy bands of 1-MeV width. Neutrons removed
from the jth energy band are given by
(j= 1, 2, ... , 18)
(4.9-6)
where So is a magnitude factor determined by the power level of the reactor

and F(E) is the normalized fission spectrum.
The neutrons from each of the removal bands in the energy range 0 to
8 MeV (j = 18, 17, 16, ... , 11) are introduced into the energy group whose
upper energy limit corresponds to the mid-energy of the band. Neutrons
from all the bands above 8 MeV (j = 10, 4, 8, ..., 1) have a mean energy of
about 10 MeV and are all introduced inro the highest energy group
(group 1), which has an upper energy of 10.5 MeV. This transfer scheme,
along with the removal-band and energy-group structures for RASH E, is
given in Table 4.6.
194

Table 4.6-REMOVAL-BAND AND ENERGY-GROUP STRUCTURES
USED IN RASH Et
Removal bands
Band
No.
Band-to
Diffusion groups
Energy limits, Me V
Group
No.
Energy limits, MeV
Upper
group
transfer
scheme
Upper
Lower
I
2
3
4
5
18
17
16
IS
14
17
16
IS
14
13
I
2
3
4
5
1.05
7.5
6.5
5.5
4.5
X la'
X 10
X 10
X 10
X 10
7.5
6.5
5.5
4.5
3.5
X 10
X 10
X 10
X 10
X 10
I ....
2 ....
3 ....
4 ....
5 ....
6
7
8
9
]0
13
12
11
10
9
12
11
10
9
8
6
8
9
10
3.5
2.5
1.5
5.0
5.0
X 10
X 10
X 10
X 10- 1
X 10-'
2.5
1.5
5.0
5.0
5.0
X 10
X 10
X 10- 1
X 10-'
X 10-3
6 .... 1
7 .... 1
8 .... 1
9 .... 1
] 0 .... I
11
12
13
14
IS
8
7
6
5
4
7
6
5
4
3
11
12
13
14
IS
5.0
5.0
5.0
5.0
5.5
X 10- 3
X 10-'
X 10-s
X 10-
X 10- 7
5.0 X 10-'
5.0 X 10-s
5.0 X 10-
5.5 X 10- 7
7.0X 10-8
11 .... 2
12 .... 3
13 .... 4
14 .... 5
15 .... 6
16
17
18
3
2
I
2
I
16
Lower
Thermal
I
1
1
1
1
16 .... 7
17 .... 8
18 .... 9
tFrom Bendall,49
In the MAC formulation of Peterson 5 I and Canali, Ilsemann, and

Preusch,52 the number of energy groups for the group-diffusion calculation
is increased to 31 over an energy range from 0 to 10 MeV. Again the fission
spectrum is divided into 18 removal bands of I-MeV width. The flux density
for the source of the removed neutrons (usually called removal flux)
corresponding to the gth removal band, which is introduced into the ith
energy group, is given by
o _
fEg+I
gi - SO JE g
PrE) e-~ R (E )r
41fr 2
dE
(4.9-7)
The removed neutrons are introduced into the five highest energy groups
only. The transfer scheme, along with the removal-band and energy-group
structures, is given in Table 4.7.
RADIATION TRANSPORT
195
Table 4.7-REMOVALBAND AND ENERGYGROUP STRUCTURES USED IN MACt

Removal bands
Band
Nn.
Diffusion groups
Energy limits, MeV
Upper
Lower
I
2
3
4
5
18
17
16
15
14
17
16
15
14
6
7
8
9
10
11
12
Energy limits, MeV
Group
No.
Lower
Upper
Band-togroup
transfer
scheme
13
I
2
3
4
5
6.065
3.679
2.231
1.353
8.208
X 10
X 10
X 10
X 10
X ] 0'1
1.000
6.065
3.679
2.231
1.353
X 10"
X 10
X 10
X 10
X 10
I -> I
2 -> I
3->1
4->1
5->]
13
12
11
10
9
12
11
10
9
8
6
7
8
9
]0
3.876
1.830
6.733
2.600
2.000
X 10"
X 10'1
X 10"
X 10"
X 10"
8.208
3.876
1.830
6.733
2.600
X
X
X
X
X
10'1
10"
10'1
10"
10"
6->1
8->1
9->1
10 -> I
14
15
8
7
6
5
4
7
6
5
4
3
II
12
13
14
15
9.118
3.355
1.234
4.540
3.I 99
X 10'3
X 10'3
X 10'3
X 10"
X 10"
2.000 X 10"
9.118XIO'3
3.355 X 10'3
1.234 X 10'3
4.540 X 10"
11 -> I
12 -> I
13 -> 2
14 -> 2
15 -> 2
16
17
18
3
2
1
2
1
0
16
17
18
19
20
2.255
1.120
6.147
3.374
1.515
X 10"
X 10"
X 10"
X 10"
X 10"
3.199
2.255
1.120
6.147
3.374
X 10"
X 10"
X 10"
X 10"
X 10"
16 -> 3
17 -> 4
18 -> 5
21
22
23
24
25
1.016
4.565
1.375
9.214
6.716
X 10"
X 10"
X 10"
X 10"
X 10"
1.515
1.016
4.565
1.375
9.214
X
X
X
X
X
10"
10"
10"
10"
10"
26
27
28
29
30
31
4.140
2.775
1.860
1.247
7.595
0
X 10"
X 10"
X 10"
X 10"
X 10'8
6.716
4.140
2.775
1.860
1.247
7.595
X
X
X
X
X
X
10"
10"
10"
10"
10"
10'8
13
7'""*1
t From Peterson. 5 I
The MAC formulation differs from the original Spinney method in two
major respects: (1) the removal flux density is added directly to the
group,diffusion flux density after the diffusion calculation has been
performed, and the combined flux density is then used to calculate source
196
neutrons for the lower energy groups; and (2) the general treatment of the
down scatter transfer of neutrons allows for a more accurate representation
of inelastic scattering and collisions with hydrogen.
The highest energy group (i = 1) in the diffusion part'of the calculation is
not actually treated as a diffusion group. The collision density,
~ k [~! "'~dr)]
I
which is based on the removal flux densities (corresponding to the energy

bands 1 to 12), provides neutrons by downscattering from the first group
into the kth group, k = 2, 3, .. .,K. The kinds of possible interactions, as
described by their reGpective group-to-group removal cross section, ~ 1 k, will
determine the extent of the downscatter. A diffusion calculation is then
performed on the second group, the neutrons removed from group 1 being
used as the source. Solution of the group 2 diffusion equation
(4.9-8)
"'2
yields the group-diffusion flux density

(r). The group 2 removal flux
densities are then added to the diffusion flux density so that the downscatter
source of neutrons from group 2 into the lower energy groups can be
calculated. The downscatter source into group k (k = 3, 4, .. . ,K) is given by
~2k [",>(r) + g=!3

E
"'gOz (r)]
The calculation proceeds in a similar fashion from one group to the next
lower group and so on. In general, for i> 2 the group-diffusion equations are
given by
K
Di \72 "'i(r) -
k=i+l
~ik "'i(r) - ~ai "'i(r)

!
E ~'i
;=1-1 J
[""(r) +
J
E
",0.]
g gJ
= 0
(i= 3, 4, ...)
(4.9-9)
and the downscatter source term from the ith group into the kth group is
~ik ["'i(r) + "'j;)

In the NRN formulation of Hjarne S 3 and Hjarne and Leimdorfer, S4 the
energy structure for the removal bands and energy groups differs signifi-
RADIATION TRANSPORT
197
candy from that used in the RASH and MAC formulations. The group
structure for the group-diffusion calculation consists of 24 groups over an
energy range 0 to 1'8 MeV, and the fission spectrum is divided into 30 bands
of varying widths. The removalband and energy-group structures are given in
Table 4.8.
Table 4.8-REMOVAL-BAND AND ENERGY-GROUP STRUCTURES USED IN NRNt
Removal bands
Energy limits, MeV
Band
No.
Diffusion group
Upper
Energy limits, MeV
Group
Lower
No.
Upper
Lower
1
2
3
4
5
1.8
1.43
1.136
9.021
7.166
X 10'
X 10'
X 10'
X 10
X 10
1.43
1.136
9.021
7.166
5.692
X 10'
X 10'
X 10
X 10
X 10
1
2
3
4
5
1.8
1.35
1.0
7.8
5.9
X 10'
X 10'
X 10'
X 10
X 10
1.35
1.0
7.8
5.9
4.4
X 10'
X 10'
X 10
X 10
X 10
6
7
8
9
10
5.692
4.521
3.591
2.853
2.267
X 10
X 10
X 10
X 10
X 10
4.521
3.591
2.853
2.267
1.800
X 10
X 10
X 10
X 10
X 10
6
7
8
9
10
4.4
3.4
2.6
2.0
1.5
X 10
X 10
X 10
X 10
X 10
3.4
2.6
2.0
1.5
1.2
X 10
X 10
X 10
X 10
X 10
11
12
13
14
15
1.800
1.430
1.136
9.021
7.166
X 10
X 10
X 10
X 10"
X 10"
1.430
1.136
9.021
7.166
5.692
X 10
X 10
X 10"
X 10"
X 10"
11
12
13
14
. 15
1.2
9.0
7.0
5.1
3.8
X 10
X 10'1
X 10"
X 10"
X 10"
9.0 X 10'1
7.0 X 10"
5.1 X 10"
3.8 X 10"
3.0 X 10"
16
17
18
19
20
5.692
4.521
3.591
2.853
2.267
X 10"
X 10"
X 10-'
X 10"
X 10'1
4.521
3.591
2.853
2.267
1.800
X 10"
X 10"
X 10"
X 10"
X 10"
16
17
18
19
20
3.0
1.0
3.10
1.10
1.10
X 10'1
X 10"
X 10"
X 10-2
X 10'3
21
22
23
24
25
1.800
1.430
1.136
9.021
7.166
X 10"
X 10"
X 10"
X 10"
X 10"
1.430
1.136
9.021
7.166
5.692
X 10"
X 10"
X 10"
X 10-2
X 10"
21
22
23
24
1.10 X 10"
1.10 X 10's
1.10 X 10'
26
27
28
29
30
5.692
4.521
3.591
2.853
2.267
X 10"
X 10"
X 10"
X 10"
X 10"
4.521
3.591
2.853
2.267
1.80
X 10"
X 10"
X 10"
X 10"
X 10"
t From Hjarne. S 3
1.0
3.10
1.10
1.10
1.10
X 10"
X 10-2
X 10"
X 10'3
X 10"
1.10 X 10's
1.10 X 10'
1.05 X 10'7
Thermal
198
The NRN method aUows for the transfer of removed neutrons from each
removal band to many diffusion groups. The source for the ith diffusion
group arising from aU removal coUisions is
1:g ~o'0
gl g
where 4>3 is the removal flux density in the gth energy band and ~3j is the
energy-averaged removal cross section for the transfer of neu trans from the
gth removal band into the ith energy group.
The calculation also allows transfer from each diffusion group to aU
lower energy diffusion groups. The group-diffusion equation for the ith
group is given by
K
Dj V2 4>j(r) -
1: ~jk
k=i+l
4>;(r) - ~a; 4>j(r)

!
1: ~'j<l>'(r)+~Oj<l>=o
j=j.! J
J
g
g
(4.9-10)
where the various diffusion-theory parameters have conventional definitions.

The SABINE code, described by Ponti, Preusch, and Schubart S5 and by
Nicks, Perline, and Ponti,S 6 represents later European technology than
RASH and NRN. SABINE uses 19 removal groups and 26 diffusion groups.
The code is also one dimensional and can solve problems in slab, cylindrical,
and spherical geometries. Particular attention has been paid to the coupling
of the removal flux with the diffusion equations. The assumed model in
SABINE makes use of the same transfer matrix for band-to-group and
group-to-group transfers, i.e.,
where the removal band g and diffusion group i correspond to the same
i. Within-group transfer of removal neutrons to the diffusion
energy and j
group of the same energy range is also included.
Gamma-ray transport is accomplished by using seven energy groups and
empirical region-dependent buildup factors based on transport calculations.
Both primary sources (fission gamma rays) in the reactor core and secondary
sources (capture and inelastic-scattering gamma rays) generated within the
shield can be included.
The ATTOW code is a recent British development described by Collier
and CurrisS? and by Avery and Curtis. 58 It is a two-dimensional (finite
*'
RADIATION TRANSPORT
199
cylinder or 20 rectangular) diffusion code that can accept removal sources

prepared by a built-in subroutine. The removal-source subroutine calculates
sources at points determined by the routine and fits two-dimensional
polynomials, of order chosen by the user, to the results. These data are put
on tape in a form that can be used by ATTOW. The spatial integration over
the reactor core is performed by a Gaussian scheme using a stored table of
zeros and weigh ts.
Removal cross sections 1:g ; are input to the program, allowing the user
control of the removal assumptions. The full group-to-group transfer matrix
1:j; is also assumed for the diffusion treatment.
The ATTOW code has been used extensively in the United Kingdom for
solving problems associated with fast breeder reactors. It was found that the
results were somewhat sensitive to the energy-group structure chosen.s 8 A
23-group structure was judged best. This sensitivity is most pronounced
when materials such as graphite are present and is attributed to the
continuous slowing-down (age theory) assumption, although even a fullscatter matrix cannot preserve the energy-angle correlation in diffusion
theory. These difficulties should not be apparent in hydrogenous materials,
which generally have short diffusion lengths.
4.9.3 Differences in Current Methods
A comparison of the preceding formulations shows that, with respect to
the removal-band and energy-group schemes, RASH E and MAC are similar
in concept and identical in many respects. The NRN, SABINE, or ATTOW
approach is more general and should provide the most accurate model if the
required removal and transfer cross sections are known.
With regard to removal cross sections, RASH E and MAC use the cross
sections suggested by the original Spinney formulation, which has the
general form
(4.9-11)
where ~ R = removal cross section
~ t = total macroscopic cross section
I;el = elastic-scattering cross section
f = fraction of elastic collisions that can be regarded as glancing

If f is taken to be the average cosine of scattering in the laboratory system,
lio, the removal cross section becomes the transport croSs section originally
used by Spinney. In general, the parameter f cannot be determined
200
intrinsically, and thus a value must be assumed or determined empirically.

This has been done for a large variety of typical shield configurations, and
the removal cross sections thus determined are used with a high degree of
confidence.
The NRN removal cross sections are obtained by experimentally
determining the angles of scatter above which elastic collisions can be
considered as removals. The removal cross section is given by
(4.9-12)
where a(O) is the differential elastic-scattering cross section per unit solid
angle about the scattering angle 0 in the center-of-mass system and 0 R is the
scattering angle above which the collision is considered to be a removal. The
value of ORis determined by comparing predicted neutron reaction rates
with experimental values. Aalto S 9 obtained a best value of cos OR =0.45 for
hydrogen, and cos 0R = 0.60 was obtained for other nuclides. With these
values of 0 R, a full set of removal cross sections can be derived.
The NRN removal cross sections may not have any advantage over the
Spinney cross sections since each scheme involves only a single adjustable
parameter, OR andf, respectively.
The MAC scheme for transferring removed neutrons into energy groups
differs significantly from that used by either RASH E or NRN. The
procedure in MAC is to add the removal flux density to the newly calculated
group-diffusion flux density to establish the group-to-group down scatter
source. In contrast, RASH E and NRN introduce the removed neutrons into
given groups as source neutrons to that group, a more natural procedure for
the group-diffusion calculation. The RASH E has a very restricted transfer
scheme wherein the removed neutrons from a given removal band are
introduced into a prescribed energy group and into no other. The NRN
provides for a much more general scheme, using a removal matrix to describe
the transfer of removed neutrons from a given removal band into any of the
lower energy grou ps.
of the methods mentioned, the slowing-down model embodied in
NRN, and especially in SABINE, gives the most accurate description of the
slowing-down process. It involves a general-group to any lower energy-group
transfer matrix using detailed elastic and inelastic scattering cross sections
for all nuclides. A similar scheme is employed by MAC; however, some

inaccuracy is allowed in the description of the nonhydrogen elastic
scattering.
RADIATION TRANSPORT
201
The RASH E uses a group-to-gtoup transfer cross section based on the

continuous slowing-down (age) model, which allows transfer to the next
lower energy group only. This could lead to serious inaccuracies, particularly
with respect to inelastic scatterings and collisions with hydrogen.
In conclusion, we should note that the Spinney method, used much
more in Europe than in the United States, provides rcliable results. As with
any empirical method, it should be applied with some cau tion. Before being
used extensively for design, removal-diffusion results should be compared
with rigorous calculational methods or experiments for similar configurations. The simplicity and speed of computation give the Spinney method a
significant advantage over more-sophisticated merliods in design work.
REFERENCES
1. I. Kulcer, Developments in Transport Theory, E. Inanu and P. F. Zweifel (Eds.),
p. 243, Academic Press, London, 1967.
2. A. M. Weinberg and E. P. Wigner, The Physical Theory of Neutron Chain Reactors,
University of Chicago Press, Chicago, 1958.
3. K. Shure, P-3 Multigroup Calculations of Neutron Attenuation, Nucl. Sci. Eng., 19:
310 (1964).
4. W. D. Lanning, Application of the Spherical Harmonics Technique to Problems in
Gamma Transpoet, Nucl. Sci. Eng., 15' 259-267 (1963).
5. G. C. Wick, Dber ebene Diffusionpcobleme, Physick, 121, 702 (1943).
6. S. Chandrasekhar, Radiative Transfer, Clarendon Press, Oxford, 1950.
7. B. G. Carlson, Solution of the Transport Equation by the 5 n Method, USAEC Report
LA-1891, Los Alamos Scientific Laboratory, 1955.
8. S. Preiser, G. Rabinowitz, and E. deDufour, A Program for the Numerical Integration
of Boltzmann Transport Equation, NIOBE, Report ARL-TR-60-314, Nuclear Development Corp. of America, Aeronautical Research Laboratories, 1960.
9. B. Carlson, C. Lee, and J. Worlton, The DSN and TDC Neutron Transport Codes,
USAEC Report LAMS-2346, Los Alamos Scientific Laboratory, Feb. 12, 1960; C. E.
Lee, The Discrete 5 n Approximation to Transport Theory, USAEC Report LA~2595,
Los Alamos Scientific Laboratory, Mar. 9, 1962.
10. K. D. Lathrop, DTF-IV, A FORTRAN-IV Program for Solving rhe Multigroup
Transport Equation with Anisotropic Scattering, USAEC Report LA-3373, Los
Alamos Scientific Laboratory, July 15, 1965.
11. F. R. Mynatt, A User's Manual for DOT, A Two~Dimensional Discrete Ordinates
Transport Code with Anisotropic Scattering, informal notes included in RSIC
Computer Code Collection CCC-89, Oak Ridge National Laboratory; see also F. R.
Mynatt, F. J. Muckenthaler, and P. N. Stevens, Development of Two-Dimensional
Discrete Ordinate Transport Theory for Radiation Shielding, USAEC Report
CTC-INF-952 (AD-692168), Union Carbide Corporation, August 1969.
202
12. H. ]. Kapp. Synthetic Method Solution of the Transport Equation, Nucl. Sci. Eng.,
17: 65 (1963).
13. E. M. Gelbard and L. A. Hageman, The Synthetic Method as Applied to the S"
Equation, Nucl. Sci. Eng., 37: 288 (1969).
14. L. V. Spencer and U. Fano, Calculation of Spatial Distributions by Polynomial
Expansion,]. Res. Nat. Bur. Stand., 46: 446 (1951).
15. U. Fano, L. V. Spencer, and M. J. Berger, The Penetration and Diffusion of X-Rays,
Encyclopedia of Physics, S. Fliigge (Ed.), p. 724, Springer-Verlag, New York, 1959.
16. H. Goldstein and J. E. Wilkins, Jr., Calculations of the Penetrations of Gamma Rays,
Final Report, USAEC Report NYO-3075 (NDA-15C-41), Nuclear Development
Associates, Inc., 1954.
Company. Inc., Reading, Mass., 1959.
18. L. V. Spencer, in Radiation Shielding, Vol. I, W. R. Kimel (Ed.), Office of Civil
Defense, Technical Report TR-40, 1966.
19.]. Certaine, A Solution of the Neutron Transport Equation. Introduction and Part I,
USAEC Report NYO-3081 (NDA-15C-43), Nuclear Development Associates, Inc.,
July 25, 1954.
20. V. A. Amharzumian, Diffusion of Light by Planetary Atmosphere, Astron. Zh., 19:
30 (1942).
21. R. Bellman, R. Kalaba, and G. M. Wing, Invariant Imbedding and Mathematical
Physics. I. Particle Processes,]. Math. Phys., 1(4): 280 (1960).
22. G. M. Wing, An Introduction to Transport Theory, John Wiley & Sons, Inc., New
York,1962.
23. D. R. Mathews, K. F. Hansen, and E. A. Mason, Deep Penetration of Radiation by the
Method of Invariant Imbedding, Nucl. Sci. Eng., 27: 263 (1967).
24.J. O. Mingle, Applications of the Invariant Imbedding Method to Monoenergetic
Neutron Transport Theory in Slab Geometry, Nucl. Sci. Eng., 28: 177 (1967).
25. A. Shimizu and H. Mizuta, Application of Invariant Imbedding to the Reflection and
Transmission Problems of Gamma Rays (II), ]. Nucl. Sci. Technol. (Tokyo), 3: 10
(1966); also Nucl. Eng. Des., 32: 184-194 (1968).
26. J. J. Taylor, Application of Gamma Ray Build-Up Data to Shield Design, USAEC
Report WAPD-RM-217, Westinghouse Electric Corp., Atomic Power Division,
June 28, 1954.
27. M. A. Capo, Polynomial Approximation of Gamma Ray Buildup Factors for a Point
Isotropic Source, USAEC Report APEX-510, General Electric Co., Atomic Products
Division, August 1959.
28. M. J. Berger, in Proceedings of Shielding Symposium Held at the U. S. Naval
Radiological Defense Laboratory, October 17-19, 1956, U. S. Naval Radiological
Defense Laboratory Reviews and Lectures No. 29.
29. D. K. Trubey, A Survey of Empirical Functions Used to Fit Gamma-Ray Buildup
Factors, USAEC Report ORNL-RSIC-10, Oak Ridge National Laboratory, February
1966.
30. J. H. Hubbell, A Power-Series Buildup Factor Formulation Application to Rectangular and Off-Axis Disk Source Problems,]. Res. Nat. Bur. Stand., C, 67: 291 (1963).
RADIATION TRANSPORT
203
31. J. H. Hubbell, R. L. Bach, and R. J. Herbold, Radiation Field from a Circular Disk
Source,]. Res. Nat. Bur. Stand., C, 65: 249 (1961).
32. D. K. Trubey, Oak Ridge National Laboratory, unpublished calculations.
33. J. H. Hubbell, R. L. Bach, and J. C. Lamkin, Radiation Field from a Rectangular
34. R. D. Albert and T. A. Welton, A Simplified Theory of Neutron Attenuation and Its
Application to Reactor Shield Design, USAEC Report WAPD-15, Westinghouse
Electric Corp., Atomic Power Division, Nov. 30,1950.
35. E. Solomito and J. R. Stockton, Modifications of the Point Kernel Code QAD-P5A:
Conversion to the IBM-360 Computer and Incorporation of Additional Geometry
Routines, USAEC Report ORNL-4181, Oak Ridge National Laboratory, 1968.
36. G. T. Chapman and C. L. Storrs, Effective Neutron Removal Cross Sections for
Shielding, USAEC Report ORNL-1843 (AECD-3978), Oak Ridge National Laboratory, 1955.
Interscience Publishers, a division of John Wiley & Sons, Inc., New York, 1962.
38. L. K. Zoller, Fast-Neutron-Removal Cross Sections, Nucleonics, 22(8): 128 (1964).
39. D. K. Trubey and G. T. Chapman, Effective Neutron Removal Cross Sections for
Carbon and Oxygen in Continuous Mediums, USAEC Report ORNL2197, Oak Ridge
National Laboratory, Sept. 19,1958.
40. D. J. Dudziak and J. E. Schmucker, Hydrogenous-Material-Dependent Removal Cross
Section of Lead for Fast Neutron Biological Dose, USAEC Report WAPDTM-662,
Westinghouse Electric Corp., June 1968.
41. W. J. Price and W. L. Dunn, Dose Attenuation of Polonium-Beryllium Neutrons by
Iron, Aluminum, and Lead, Air Force Institute of Technology, Tech. Report 57-7,
1967; also UsAEC Report ORNL-2497 (VoJ.1), pp.1L-12L, Oak Ridge National
Laboratory.
42. D. J. Dudziak, Fast-Neutron Biological Dose Attenuation by Lead and Polyethylene
Shields, Nucl. Appl., 6: 63 (1969).
43. K. Shure, J. A. O'Brien, and D. M. Rothberg, Neutron Dose Rate Attenuation by Iron
and Lead, Nucl. Sci. Eng., 35: 371 (1969).
44. A. W. Casper, Modified Fast Neutron Attenuation Functions, USAEC Report
XDC-60-2-76, General Electric Co., Atomic Products Division, Feb. 3, 1960.
45. J. W. Haffner, Neutron Energy Spectrum Calculations in Reactor Shields, Preprint
V84, Nuclear Engineering and Science Conference, April 6-9, 1959, Cleveland, Ohio,
Engineers Joint Council, New York.
46. D. C. Anderson and K. Shure, Thermal Neutron Flux Distributions in MetalHydrogenous Shields, Nucl. Sci. Eng., 8: 260 (1960).
47. K. Shure, Few-Group and Multigroup Calculations of Neutron Penetration, Nucl. Sci.
Eng., 27: 468 (1967).
48. A. F. Avery, D. E. Bendall, J. Butler, and K. T. Spinney, Methods of Calculation for
Use in the Design of Shields for Power Reactors, British Report AERE-R-3216,
February 1960.
49. D. E. Bendall, RASH D-A Mercury Programme for Neutron Shielding Calculations,
British Report AEEW-M-261, August 1962.
204
50. J. Butler, The Status of Theoretical Methods for Reactor Shield Design, British
Report AEEW-R-361, March 1964.
51. E. G. Peterson, MAC-A Bulk Shielding Code, USAEC Report HW-73381, Hanford
Atomic Products Operation, April 1962.
52. U. Canali, H. Ilsemann, and H. Preusch, MAC-RAD, A Reactor Shielding Code,
Report EUR-2152.e, European Atomic Energy Community, 1964.
53. L. Hjaroe (Ed.), A User's Manual for the NRN Shield Design Method, Report AE-145,
AB Atomenergi, June 1964.
54. L. Hjarne and M. Leimdorfer, A Method for Predicting the Penetration and Slowing
Down of Neutrons in Reactor Shields, Nucl. Sci. Eng., 24: 165 (1966).
55. C. Ponti, H. Preusch, and H. Schubart, SABINE: A One-Dimensional Bulk Shielding
Program, Italian Report EUR-3636.e, European Atomic Energy Community, 1967.
56. R. Nicks, G. Pecline, and C. Ponti, SABINE: A One-Dimensional Shielding Programme
Which Employs Experimental Removal Cross Sections, in Proceedings of the
Conference on the Physics Problems of Reactor Shielding, September 1967, British
Report AERE-R-5773, Vol. 1, pp. 5-40, 1968.
57. W. D. Collier and G. C. Curtis, ATTOW, A Two-Dimensional Shielding Program,
British Report TRG-1466(R), 1967.
58. A. F. Avery and C. C. Curtis, Neutron Attenuation Studies in Fast Reactor Shielding,
in Proceedings of the Conference on the Physics Problems of Reactor Shielding,
September 1967, British Report AERE-R-5773, Vol. 2, pp. 515-553, 1968.
59. E. Aalto, The Fine Adjustment of the Neutron Penetration in the NRN Method, Nucl.
Sci. Eng., 22: 443 (1965).
EXERCISES
4.1
Explain why the expression
(r,n,E) dn dE (nn) dS
in Eq. 4.2-1 gives the net (outward minus inward) leakage through the differential
surface dS.
4.2 Transform the particle-flux-density form of the Boltzmann equation with energy as
an independent variable (Eq.4.2-8) into its equivalent form with the energy-flux
density, 1= EcfJ, as the dependent variable and the wavelength, A, as an independent
variable. [Hint: I(A) = I(E).J
4.3 Starting with the Boltzmann equation (Eq.4.2-8), show that for a one-speed
one-dimensional slab geometry with azimuthal symmetry the equation reduces to
I!
f.2. f'-I ~,(n,n ,) (x,n).dl!.d,p, d,p
a(x,l!)
1 (2.
ax + ~t (x,l!) = S(x,l!) + 21T Jo
0
in which Il = cos (J, where {J is the polar angle of
verifying Eq. 4.3-1.
n with respect to the x-axis, thus
RADIATION TRANSPORT
205
4.4 Show that for the onedimensional spherical geometry the Boltzmann equation can
be reduced to
II
o<l>(r,Il,E) (1 -II') o<l>(r,Il,E) + ~ <1>(

E)
,
+
,
"'t
r,ll,
vr
uIJ
(~~,
, dE' dII'
+ 8(r,Il,E) + f.1l JO
",,(r,E -+E,llo) <1>'
(r,E,jl)
4.5
thus proving Eq. 4.4-3.

Using the form of the Boltzmann equation obtained as a result in problem 4.2, show
that for a one-dimensional slab geometry photon transport is described by
in which IJ. = n-j (i is the unit vector along the x-axis), Lt(A) is the total attenuation
coefficient, and Ls(A',StO) is the differential macroscopic Compton cross section
for scattering from wavelength A' through an angle whose cosine is 0'
This
proves Eq. 4.5-2.
n.
Monte Carlo ~lethods

for Radiation Transport
W. E. SELPH and
c. W. GARRETT
Monte Carlo techniques have been so important in shield analysis in the past
decade that a separate chapter on this subject is warranted. For a more
formal and detailed development of general Monte Carlo techniques, the
reader is referred to one of the texts devoted to the subject. J -3 We will be
concerned primarily with the step-by-step development of radiationtransport applications of the method based on a knowledge of microscopic
cross sections and elementary probability theory. A brief discussion on use
of the method to solve integral equations is given as an aid to understanding
the equivalence of this method with other means of solving the Boltzmann
transport equation.
We assume that the reader has some grasp of the elementary concepts of
probability, namely, that the probability of one of several possible events
occurring will be approximately equal to the ratio of the number of
times the desired event occurs to the total number of events observed in an
unbiased manner. As the number of observations increases, this ratio should
more closely approximate the true probability. Beyond this simple concept,
more specific defmitions will be developed as they are required.
Much of the information available on the physics of individual nuclear
interactions is obtained experimentally by observing the fate of large
numbers of particles. In like manner, Monte Carlo may be considered as a
means of repeatedly applying interaction probability data to individual
particles selected randomly until a sufficient number of particles have been
observed to allow conclusions to be drawn concerning the macroscopic
multicollision behavior of the total population of particles within a material
region. Dependence on such a mathematical experiment for the macroscopic
probabilities (shield leakage, particle absorption density, etc.) is in many
ways analogous to the laboratory determination of microscopic interaction
cross sections. To be sure, the laboratory experiment may be based on the
observation of a much larger sample than can be processed economically in

207
208
the mathematical game, but the principles are the same. Research on the
problem of increasing the sample size or employing other methods to reduce
the uncertainty associated with Monte Carlo data has led to the development
of special techniques combining analytical estimates with random particle
tracking.
It is possible to show how the various steps involved in a Monte Carlo
solution correspond to terms of the Boltzmann transport equation. Although
Monte Carlo may be considered a means of solving the Boltzmann equation,
it is more properly a modeling of the principles from which the
Boltzmann equation was developed; that is, the method may be developed
and properly applied to radiation-transport problems without reference to a
differential transport equation. The general course followed in this chapter is
to develop individually, and in a logical sequence, the steps required to solve
a radiation-transport problem by the Monte Carlo method.
Before getting into the detailed operations involved in generating a
particle track, we will first discuss the general organization of the tasks.
Approaches to accomplishing these tasks then will be developed through the
chapter. Since our objective is to explain the logical processes involved in
applying the Monte Carlo method, litrle space is devoted to explaining how
these operations are programmed for the computer. Programming is a strong
function of the nature of the computer, its auxiliary equipment, and the
language used. Further, a large number of general-purpose computer codes
using the Monte Carlo method for shielding applications are readily available,
and the probability is rather high that developing a new code for a particular
application will not be required.
Simply stated, the Monte Carlo approach requires that we construct case
histories of the travel of individual particles through the geometry and then
analyze these histories to derive relevant data, such as flux density and dose
rate. One particle history includes the birth of a particle at its source, its
random walk through the transporting medium as it undergoes various
scattering interactions, and its death, which terminates the history. A death
Can occur when the particle becomes absorbed, leaves the geometric region
of interest, or loses significance owing to other factors (e.g., low energy).
If we assume for the moment that the source-shield geometry has been
mathematically modeled, the major steps involved in generating a particletrack history are shown in Fig. 5.1.
The loop 2 through 4 is continued until the particle parameters fall
outside some predetermined limit of values, such as geometrical bounds,
minimum energy, or minimum statistical weight (a concept to be developed
MONTE CARLO METHODS FOR RADIATION TRANSPORT
209
later). This process only generates particle histories. Separate provision must
be made for estimating desirable output quantities, such as flux density,
current density, or interaction densities. The first three operations in Fig. 5.1
involve the selection of parameters at random from a probability distribution
of all possible values of these parameters. Thus a sufficiently large number of
~
Choose
source
parameters
111
f--
Select
path length
(2)
--
Select
collision
parameters
(3)
f-
Compute
particle
parameters
after collision
(4)
Fig. S.l-History generator.
selections of a given parameter would be distributed in the same manner as

the corresponding probabilities. The steps in making a random selection
from such probability distributions are based on the use of numbers
randoml y distributed between the limits of 0 and 1.
5.1
SAMPLING FROM PROBABILITY DISTRIBUTION FUNCTIONS

As suggested earlier, all physical processes, including the emission of
radiations from sources and their subsequent transport through material, arc
probabilistic; that is, one cannot predict, with certainty, exactly what will
occur for every individual particle in the process. Nevertheless, such
stochastic, or random, processes can be effectively characterized and
predictions can be made by describing the average behavior of many
elements or by estimating with a known degree of confidence (but never
with certainty) the behavior of one element. Mathematically this characterization is accomplished through the use of various probability functions.
Processes dependent on one or many independent variables may be so
handled.
An event is a physical occurrence, for example, a coin toss resulting in a
head or a photon assuming a specific energy Eo. An event space is the set of
all possible (mutually exclusive) events within the process under consideration. Event spaces, sometimes called sample spaces, may be finite and
discrete (the coin-toss problem where it contains only two points, heads and
tails), continuous (the visible electromagnetic spectrum), and/or infinite in
extent.
210
Clearly, events such as heads and tails require numerical representation

to be handled effectively from a mathematical standpoint. Thus events are
mapped onto real numbers by a function called a random variable. For
example, a coin-ross random variable might be a function that assigns 0 to
tails and 1 to heads. To be precise, one must define the random-variable
functions used. For the subject matter of this text, however, the event and
the value assigned it by the random variable (e.g., a photon produced at
energy E is assigned the value E) will be quite clear, and we will not
explicitly define the mapping function.
Inherent in the Monte Carlo procedure is the concept of the probability
density jimction (PDF) (also called frequency functions and probability
distributions). This concept is an extension of differential distribution
functions introduced in Sec. 2.2.1. The probability density function, f(x),
descrihes the relative frequency of occurrence of its random variable, x. Its
domain (all possible values of x) constitutes the event space.
Let P(x<xo<x+t>x) be the probability that a random variable Xo lies
within the interval (x,x + t>x) of its event space. The PDF f(x)t is related to
this probability by the equality
P(x<xo<x+t>x) = f(x) t>x
(5.1-1 )
as t>x becomes vanishingly small. Thus the PDF is central to the prediction
problem, and knowledge of the PDF enables one to obtain the probability of
occurrence of a specific event. Equation 5.1-1 shows that the PDF f(x) gives
the probability of occurrence per unit interval of the random variable x.
Over larger intervals PDF's are used with differentials. Thus
P(a<x<b) = f~ f(x) dx
(5.1-2)
Since PDF's describe relative frequencies of occurrences of events within an

event space, two properties of PDF's emerge; namely,
o <;f(x)
(5.1-3)
Cf(x) dx = 1
(5.1-4)
and
t As with differential distributions, the concept is easily extended to two or more variables and
may he changed to phase-space notation by substituting p for x. where P represents (r,n,E).
21 I
Equation 5.1-4 states that the probability of finding a chosen event

somewhere within its event space must be unity, and Eq. 5.1-3 accrues from
the fact that relative frequencies (i.e., probability values) must be positive.
The cumulative distribution function (CDF) gives the probability that
the random variable x has a value less than or equal to some fixed value. It is
given by
P(x~o) = F(xo) =
I.:" f(x) dx
(5.1-5)
Two restrictions placed on F(x) that follow from Eqs. 5.1-3 and 5.1-4 are
lim F(x) = 1
(5.1-6)
x~-
and
lim F(x) = 0
x--+
(5.1-7)
_00
The probability P(a<x<b) of x lying within an interval (a,b) is thus given

by
P(a<X<b) =
J:
f(x) dx = F(b) - F(a)
(5.1-8)
Clearly, if the variable x can only take on values within some finite range,
then one may substitute the lower bound of that range for _00 and the upper
bound for +00 in the preceding expressions.
Various distribution functions may be encountered which ate not
probability density functions. For example, a variable may be limited to the
region 2';;; x .;;; 4 and be distributed proportional to x 2 in that region. The
function x 2 violates both of the preceding restrictions on probability values
but may, with proper normalization, be converted to a PDF. In this case the
PDF f(x) is given by
/(x)
f4
x2
2 X
(5.1-9)
X
and
j{o x 2 dx
F (x 0) = --=f-4---:-2-d2 x
X
(5.1-10)
212
Or, stated mote generally, where q(x) is any function of x that is positive for
all values of x,
q(x)
[(x) = r:q(x) dx
(5.1-11)
and
F (x 0)
JX. q(x) dx
=::.=-~~
-'-'--'c--or: q(x}dx
(5.1-12)
Where the event space is discrete, the random variable can only take on
selected values, and the PDF [(x;) is defined by
(i = 1,2, ...)
(5.1-13)
where
I: [(Xi) = 1
i
The CDF becomes

n
P(Xi<Xn} = F(x n ) = I: [(Xi)
(5.1-14)
;=1
The mean value of a distribution is given by

n
X = Ilx =
I: X;[(Xi}
(5.1-15)
;=1
in the discrete case, or by
x= Ilx = J: x [(x) dx
(5.1-16)
in the continuous case. Since the average of a large sample of values of x

tends to this point. 11 is sometimes called the expected value of x.
The expected value of g(x) is then
n
i = I: g(Xi) f(Xi}
i=1
(5.1-17)
MONTE CARLO METHODS FOR RADIAnON TRANSPORT
213
and in the continuous case
g; r: g(x)f(x) dx
(5.1-18)
Note that
_
,,2 = 1: x7 f(Xi)
(5.1-19)
i=l
The variance of a distribution function is a measure of how closely the

distribution is grouped about the mean. The variance, denoted a2 , is defined
as
n
ai ; 1:
(Xi - 11)2 f(Xi)
(5.1-20)
;=1
in the discrete case or as
ai ;
f.: (x _11)2 /(x) dx
(5.1-21)
in the continuous case. A little manipulation will show that
(5.1-22)
in either of the preceding expressions. The standard deviation is often used
instead of the variance as a measure of the dispersion about the mean. The
standard deviation, ax, is given, logically enough, by
(5.1-23)
In analyzing and solving problems involving random processes, one must
randomly select events from distributions for which the PDF's are known,
exactly or approximately. Thus the problem is first to choose a value of f(x)
[or its equivalent, F(x) J by some random process. Then, having obtained
/(x), find x, which, of course, defines the desired event. Comprehending this
inverse process is often difficult for the beginner, who is accustomed to
calculating values of f(x), given x. It is common practice to use random
numbers equally distributed between 0 and 1 to obtain random samples of a
variable that will be distributed in the same manner as f(x). If a random
214
number n (0 < n < 1) is chosen, the value of the associated random variable
x may be obtained by solving the inverse of the equation
n=l'(x)
(5.1-24)
for x, where F(x) is a cumulative distribution function. We emphasize that

the random number n is a value chosen from the range of the function l'(x),
whereas the desired random variable x is a value within the domain of F(x).
--j------------1
I
I
--
I
I
f(xi
I
I
I
I
I
Xl
Fig. S.2-Rejection technique.
It is also possible to select randomly from a continuous probability

distribution by a powerful method known as a rejection technique. Consider
the function [(x) as plotted in Fig. 5.2. The objective is to generate a
sequence of values of x such that [(x) is approximated by the relative
frequency of occurrence of values of x within the collection formed by the
sequence.
We first choose a value K that exceeds all values of [(x) within rhe
region. Then, using appropriate techniques, we obtain a random number n,
which is used to select a value Xi (XI < Xi < xz) by the relation
Xi
Xl
+ n (xz -
XI )
(5.1-25)
215
The value of f(Xj) is evaluated at the selected value of Xj. Taking another
random number n', we next select a number N from a random distribution
between 0 and K by setting
N;n'K
(5.1-26)
If N > f(xj), the value of Xj is rejecred. The process is repeated by generating

new values of Xj and N until N ,,; f(Xi)' When this occurs the value of Xj is
accepted and used as the random sample. The probability of a particular Xi
being accepted is given by f(xj)/K and the probability of rejection is [K ~
f(xiJl /K. Thus, after a large number of determinations for various values of
x, the distribution in x of the values accepted would approximate the
function f(x).
The efficiency of this technique for selecting from f(x) is given by the
ratio of the area under the curve to the total area, or
E;
I:; f(x) dx
K (X2 ~ x.J
(5.1-27)
For highest efficiency, the value K should be the smallest possible value that
always exeeedsf(x) in the region.
Within a computer program, random numbers may be taken from tables
called into the machine memory or they may be generated by a subroutine
as needed. Random-number-generator routines are available as standard
software with most general-purpose computers. Numbers obtained in either
way arc more properly termed pseudorandom because they are systematically obtained.
Despite their availability and short length, the codes that generate
pseudorandom numbers are not trivial. In fact, a precise test to verify the
randomness of a generator has yet to be defined. Various criteria can be
applied which test for necessary conditions; the test for sufficiency awaits
further research. It is sound practice when selecting a random-numbergenerator code to examine the tests to which it has been put. In fact, some
generators that had been used for years were found to be not so random
after all when sophisticated tests were applied. Appendix I is an introduction
to random-number generators, including descriptions of some in general use
and a discussion of tests for randomness.
216
5.2 THE EVALUATION OF INTEGRALS

A theorem of statistics called the law of large numbers provides the basis
for evaluating integrals with the Monte Carlo method. To illustrate this
theorem, we
will consider the integral
z = J:
Z(x) f(x) dx
(5.2-1)
where f(x) is a probability distribution. The term Z is called the expected

value of the function Z(x). The law of large numbers states that this integral
may be approximated by
,
1
ZN = -
I; Z(Xi)
N]
(5.2-2)
and that ZN will approach Z as a limit as N approaches 00. The Xl' x2,' . .,xN
are randomly selected from the PDF f(x) (the rejection technique could be
used), and Zw is called an estimate of z.
Biasing may be defined as any' means of distorting the sampling
technique to advantage and may be introduced into the evaluation of the
preceding integral by means of a biasing function, h(x)
Z=
r- Z(x) f(x) h(x) dx
h(x)
(5.2-3)
This equation is then evaluated by sampling from a PDF given by the

product If(x) h(x)] , and each sample is adjusted by ljh(x). A reasonable
choice of the function h(x) is one that will encourage choices of X most
likely to give values of Z lying at or near the expected value; that is, the
variance of the distribution If(x) h(x) 1 should be less than that of f(x). A
goal is that the product If(x) h(x) 1 approach a constant such that there is
zero variance and
-
Z(x)
Z = h(x) for any choice of X
(5.2-4)
The integral Boltzmann transport equation may be expressed in a form

similar to the preceding equation for Z. To evaluate the density cf>(P) of
217
particles emerging from collision in a unit of phase space dP, (e.g., dP = dE

dn dr), we can write the Boltzmann equation
(P) =
J K(P' -+ P) (P') dP' + S(P)
(5.2-5)
The term K(P'-+P) dP is the transfer kernel giving the probability of a

particle emerging from a collision in increment dp about P given that one
particle emerged from a collision at P'; SIP) is the density of the source
generating particles in dp directly. In generating random-particle tracks
through a medium, we are essentially generating a (P') by straightforward
sampling. Performing a statistical estimation of the contribution to dP from
each collision point is a means of evaluating the kernel K(P'-+P). Evaluation
of the estimates from each collision point then is equivalent to performing
the integral over volume, energy, and angle. The transfer kernel may be
decomposed into the form
K(P'-+P) = T(r'-+rl',n') C(E'-+E,n'-+nlr)
(5.2-6)
That is, the transfer kernel K is expressed as a product of T, a transport

kernel, and C, a collision kernel. The vertical line in the arguments of the
tWo functions indicates quantities that are held constant; in T, r' --+ r for a
given ', n'; in C, both E' -+ and n' -+ n for a given r. The function T
would contain a delta function that vanishes everywhere except where r lies
along the ray n'; otherwise it is essentially the probability of interaction per
unit length along the ray.
Functions that are introduced into this equation to provide biasing of
parameters are often called importance functions because they provide
sampling in those areas most important to the answer. For example, an
importance function depending on position alone may be introduced into
the function T by defining
T* = T( '-+ IE'
r
n') 1(r')
1(r)
(5.2-7)
The 1(r) function often assumes an exponential form in penetration

problems. This function can then be combined in whole or in part with the
exponential in the transport kernel. A Monte Carlo application of position
biasing is discussed in the section on selection of path lengths.
218
Other importance functions of energy, angle, or position may be

introduced into the equation in a comparable manner. Goertzel and Kalos'
developed the Monte Carlo application to radiation transport in phase-space
notation.
A more systematic and generally successful approach to importance

sampling is to use the value junction, a solution to the equation adjoint to
the Boltzmann equation. The value function has been shown by Coveyou,
Cain, and VastS to be a very good, sometimes optimum, importance
function for biasing the Monte Carlo procedure. Cain 6 has shown that the
value function can be readily obtained by 5 n methods and applied to the
Monte Carlo problem.
5.3 SOURCE PARAMETERS

With this background in selecting from distributions, we are now ready
to consider the first block in the history-generator routine: choose source
parameters. The source parameters will be assumed to include the energy,
spatial point of origin, and direction of motion of the source particle, as well
as biasing parameters.
These parameters may be independent, or they may be interrelated in
various ways. For example, where the energy distribution is dependent on
the direction of motion, the order of selection would call first for selection
of the initial angle and then for selection from the energy distribution
applicable to that angle. When biasing parameters are used, they are tied to
one or more of the other parameters.
5.3.1
Selection from an Energy Distribution
The most convenient form for expressing an energy distribution depends

on the nature of the distribution. Some of the forms of distribution that
may be available for input to the problem are discussed in Chap. 2. Where
possible, a cumulative energy distribution should be input at frequent energy
points such that interpolation may be used with reasonable accuracy. Most
distributions can be input in this form without an undue amount of data
handling. The source-energy selection process proceeds as shown in Fig. 5.3.
For some types of problems, energy groups may be defined by using
boundary values: E" ... , E.,+ 1 for n groups. The particles may then be
classified both by their energies and by the 11 groups to which they belong.
This procedure adds one parameter which must be specified at the source
219
but which does not necessarily alter the procedure for selecting the source
energy.
5.3.2 Selection of Spatial Point of the Source Particle
The problem of locating and tracking particles from point to point in
material regions is primarily one of geometry. Physics becomes involved only
Generate a
random number
Evaluate E from
Set
stored values of
FIE)
E vs. FIE)
Fig. 5.3-Selection of initial energy.
in the determination of track lengths or in the fate of the particle at the

interaction point.
The means of expressing geometric parameters depends on the choice
of the coordinate system; for example, infinite-medium calculations lend
themselves to spherical geometry, but many reactor shielding problems are
best handled in cylindrical or rectangular geometry. Some examples of the
selection of source spatial position are discussed in the following paragraphs.
For sources uniformly distributed on a disk, the probability density
function describing the distribution of source points as a function of radius
is 2trr/trRj, where R 1 is the radius of the disk, and the cumulative
probability distribution is
P(r)
1:'
,
dr
f ' J(r)
=,
r
R,
(5.3-1)
Here a random number n used to select a value P(r) can be used to select
a random value of r directly. Since
2
n =P(r) =rRj
it follows that
(5.3-2)
220
At a given, the source positions are equally distributed in azimuthal angle </J;
thus
</J
n=P(</J)=21t
(5.3-3)
= 21tn
(5.3-4)
and
If </J is to be measured only to 1t, then

= 1t(2n - 1)
(5.3-5)
+1t
n=-21t
(5.3-6)
and
This may be extended to the case of a uniform cylindrical volume

distribution by next selecting a Z coordinate equally distributed between the
limits on Z.
For a nonuniform spherical-shell distribution, consider a spherical shell
bounded by radius limits ,[ and'2 in which the volume density of source
particles is given by
p = kr
(5.3-7)
The distribution function in r is given by

~
41tr2 (kr)
J(r)- k(4
4)
_
1f
'2 -
(5.3-8)
'1
The cumulative distribution is given by
J:
f(r) dr
P( r) - :..,-;-'----:-c,-----,-
dr
,
J,"f(r)
(5.3-9)
And the randomly selected radius is given by
r= lrj +n (r~ - rj]*
(5.3-10)
221
The corresponding
and '" coordinates would be selected for uniform
distribution over the shell at radius r (Sec. 5.3.3).
A variety of other distributions may be expressed analytically so that
coordinates may be selected from a continuous distribution. In practical
application it may not be possible to express the distribution analytically. In

such cases the distribution may be numerically integrated to provide a
Fig. 5.4--Integral for isotropic distribution.
cumulative distribution in each dimension. Where the dimensional probabilities cannot be separated, i.e., where 1'(r,O,"') cannot be expressed in terms of
1'(r) 1'(0) 1'("'), it may be necessary to specify 1'(0) and 1'("') or 1'(0,,,,) for
each interval in r. Two approximations arc then possible in making a random
selection: either a value of r is classified as being within a given .6.r increment
and the 1'(0) and P(",) values for that increment are selected or the 1'(0) and
P(",) values are interpolated between the two adjacent values of r for which
they are input.
5.3.3 Selection of Initial Direction of Source Particle
We turn next to the problem of assigning a direction to a source particle.
Consider first sources that are emitting isotropically. Picking a uniformly
distributed direction is tantamount to selecting unit vectors terminating
uniformly on the surface of a unit sphere. The cumulative probability

function is therefore given by the integral over the spherical surface area
shown schematically in Fig. 5.4.
222
The distribution in
e is given by
t: 21T sin e' de'

=P( e) =--'f''''""-----,-----,--,
21T sin
e' de'
1
-(I-cose)
(5.3-11)
and
e=cos- I (1-2n)
The angle

(5.3-12)
will be uniformly distributed between 0 and 1T, or
(5.3-13)
= 1T(2n - 1)

Note that the direction cosines are functions of these two angles and can
be computed once the angles are known.
For a surface source emitting isotropically, the direction would be
selected randomly from a 21T half-space, or
n
t:
=....:..JI
J:12
e' de'
21T sin e' de'
21T sin
= 1 - cos
(5.3-14)
and
e=
COS-I
(1 - n)
(5.3-15)
or, since (1 - n) is distributed in the same manner as n,
= cos- 1
(5.3-16)
In like manner, for a surface source emitting a cosine distribution, the PDF is
given by sin e' cos e' and the angles e would be selected from
cos 2
e= n
(5.3-17)
or
(5.3-18)
Any arbitrary distribution in direction may be put in terms of tabulated
values of p(e) and P(<I. When various source distributions are to be
223
considered incident on the same shield configuration, it may be advantageous to run individual unit source problems. The results obtained from
the unit source problems can then be combined with various weightings to
stimulate a variety of source distributions.
5.3.4 Source-Biasing Parameters
In certain calculations it may be desirable to prejudice the selection of
one or more source parameters to favor those most
likely
to contribute to
---e8
Fig. 5.S-Source direction biasing.
the quantity of interest, such as shield leakage or detector response. This can
be done by sclecting a larger number of the important source particles and
assigning to each particle a number, called its weight, to adjust for the bias
that was introduced. Figure 5.5 illustrates this concept.
Consider an isotropic source located at A and a detector at B. Clearly,
source particles leaving A in the direction of B will contribute more to the
flux density at B than those leaving in opposite directions. (We assume, of
course, that scattering is permitted in the transporting medium.) Suppose, in
following 10,000 source particles, we estimate that 2000 would lie in.::ln in
the unbiased isotropic case. However, because of their greater importance,
we desire to force 6000 of the 10,000 to lie in .::In. The weight assigned to
them is Y3 to remove this bias. Further, the weight of the remaining 4000
outside of .::In must be increased to 2 since each history in the biased case
represents two particles in the unbiased case. Table 5.1 summarizes these
facts. As each particle history is generated, its contribution, >j, to the total
flux density at B can be determined. This total flux density is given by
10,000
4> = K
1; Wj4>j
(5.3-19)
;=1
where Wj is the weight (Y3 or 2) of the particle contributing 4>j and K is a

source-normalization constant. (For a unit source strength, K = 1/10,000.)
224
When biasing is used, the summation of particle weights equals that of

the unbiased case where the weights are identically 1. However, because of
the altered statistical sample, a more representative answer is usually
obtained for a fixed total number of source particles.
Table S.l-WEIGHTS FOR BIASING ILLUSTRATION
Unbiased case
Histories in ~n
Histories outside
Total No. of histories
.an
Biased case
No. of
histories
Particle
weight
No. of
histories
2,000
8,000
10,000
1
1
6,000
4,000
10,000
Particle
weight
Y3
2
To amplify the example, suppose the 180 range of 0 is divided into 18

intervals of 10 each and the cumulative probability emission of a source
particle in thejth increment is given by Ll.Pj(O),j= 1, 2, ... ,18:
18
1:: Ll.Pj(0)
= 1 = PI 8 (0)
j=1
(5.3-20)
This distribution may be adjusted in any arbitrary manner by assigning

probability values Ll.Pj(O) and weights Wj to each angular group such that
(5.3-21)
provided one docs not violate the rule of probability that Ll.Pj(O)
that
< 1 and
18
1::
j=1
~]y'(0) = 1
}
(5.3-22)
As a second example, consider the case of a plane isotropic source where
it is desired to bias the histories toward directions near the normal to the
plane. The unbiased cumulative angular distribution at 0 = 10 (measured
from the normal) as computed by Eq. 5.3-14 is 0.015. This value is to be
increased by a factor of 5 such that
P~ (0) = 0.075
(5.3-23)
MONTE CARLO METHODS FOR RADIATION'TRANSPORT
225
(
Particles generated within this interval will have a weight 0.2, and particles
generated in all other categories will have probability increments
(5.3-24)
and will have weights given by
(5.3-25)
Thus the total particle weight generated by selecting N particles will be

0.2(0.075)N + 1.065(1 - 0.075)N = N
(5.3-26)
Situations where biasing might be used include:

1. Selecting more source points near the periphery of a reactor.
2. Selecting more particles with an initial direction toward the shield or
detector.
3. Selecting more particles with a high energy or with an energy
corresponding to a low total cross section.
Probability and weight adjustments can be developed for biasing the
starting position or energy for a particular situation. It is important to
remember that an adjustment in one category (e.g., angle interval or energy
group) affects not only that group but the entire distribution.
In summary, if the number of particles within a particular category is
increased by a factor k, then their weight is decreased by a factor 1/ k. The
weight of particles in other categories is increased such that the summation
of particle weights is equivalent to the number of particles generated.
5.4 PATH LENGTH

Thus far the geometry has been defined and a set of source-particle
parameters has been selected. Next we determine the particle path length
from the source to the point of interaction. The path length along with the
parameters of initial direction defines the point at which an interaction
occurs.
226
Assume initially a single infinite region through whieh particles are being
traeked. If the total maeroscopic cross section of the region is kt(E), then
the number of particles of energy Ii penetrating to depth L out of No trials is
and the probability of penetration to depth L is

P(L) ;.number penetrating to L; e-~t(E) L
number of trials
The probability of particle interaction at some path length

by
(5.4-1)
<L is thus given

(5.4-2)
If path lengths L to first collision are to be sampled at random, the random

number, n, is set equal to P/(L), or
(5.4-3)
from which
1
1
L; kt(li) In (1 - n) ; kt(E) In n
(5.4-4)
since 1 - n is distributed in the same manner as n.

Suppose now that a single region is bounded such that the boundary is
encountered at path length L; A. If the randomly selected path length
exceeds the value A, then the particle will be assumed to have escaped the
system. Thus values of L may be selected as shown above for an infinite
medium and tested to determine whether they exceed A. If they do the
particle has escaped and the history is terminated; if they do not, a collision
is assumed to have occurred at the selected point, and collision parameters
are calculated. The value of A will be a function of the initial direction of
the particle and the geometry of the region boundary, and it should be
calculated after the source direction has been selected so that it will be
available for the test against L.
Consider next a multiregion geometry with each region homogeneous. A
ray extending in the direction of particle motion encounters region
MONTE CARLO METHODS FOR RADIAnON TRANSPORT
227
thicknesses in the various regions given by AI, A z , ... , Ax with associated

total cross sections given by ~tl' ~t2' , LtX'
The effect of successive probabilities is such that the escape probability
from the entire system is eq ual to the product of the individual
YES
Ib-(I:tlAl+I:r2A2)1 ETC.
> I: t3 A 3
ETC.
Fig.
5.6~Selection
of path length.
zone-penetration probabilities. Thus the fate of the particle along the track
could be determined in the following way:
Select a path length at random assuming an infinite medium with cross
section ~ t, .
If L, exceeds AI, move the particle to the point where its track
intersects the boundary between region 1 and region 2.
Select a path length at random, in the original direction, assuming an
infinite medium with cross section ~t2'
If L 2 exceeds A 2 , translate the particle to the region 2 - region 3
boundary. This procedure is repeated until one of the Li < Ai or until the
particle escapes.
The total escape probability will be given by
(5.4-5)
Instead of selecting path lengths in each of the materials successively, as
described above, it would perhaps be simpler to decide the fate of the
particle by selecting only one random number. For this purpose we will
define b = 1: kti Ai, where Ai is the thickness of material i lying along the
i
particle path between the source point and the first collision. The value of b
is selected at random from
b = In n
(5.4-6)
The process of finding the path length corresponding to the selected value of
b is symbolized in Fig. 5.6. This search is continued until L is defined. Flow
228
charts for calculating boundary-intercept distances for several types of

geometry are given by Cashwell and Everett. 3
Biasing may be introduced into the path-length selection in a mannet
similar to that in source-particle selection; i.e., the true distribution is
distotted in some arbitrary fashion, and the particle weight is adjusted to
account for this distortion. The mechanics of ttack-Iength biasing depend on
the motive for biasing. Where it is desired to improve the statistics on
collisions internal to a region, particles may be prevented from escaping by
picking collision points from a cumulative distribution which becomes 1.0 at
the boundary; the weight of the colliding partide is then reduced by
subtracting the fraction that would ordinarily escape the region. Where it is
desired to improve the statistical data on particles penetrating a thick shield,
particle track lengths may be stretched in a direction favoring penetration
and reduced to the opposite direction with a commensurate adjustment in
particle weight. In some cases the approximation is made to replace
transport in an adjacent region by a surface effect or albedo, and one forces
collisions to occur only at the boundary.
To illustrate the first type of biasing in its simplest form, we return to
the case of a single limited region. If the initial weight of a source particle is
Wo and the path length to escape is A" then a particle of weight We = Wo
e -A I ~ t is allowed to escape, and a particle of weight
(5.4-7)
is forced to have a collision prior to escape. The cumulative probability of _
collisionP(L) is given by
t.e., n::. PtL)!?I",) ~ L "'-4, p"....... c4- ...
I
P(L) = n P(A,) = n(l -
e-~.A,)
~
~.4-8)
C'4.' ,.I
1-0
where P(L) is given by P(L) = 1 - e-~tL. Solving for L,
(5.4-9)
In a multiregion problem, this procedure could obviously be repeated for
paths beyond the first, forcing the particle to remain within the region of
interest until the history is terminated by minimum-weight or minimumenergy criteria.
229
Biasing of track length to affect deeper penetration can be done by

sampling from a linear rather than an exponential distribution with track
length. For example, for a thick-slab shield with the incident radiation
normal to the shield, the first-collision depth could be sampled from
L = nA I, where L is the depth to first collision, n is a random number, and
A 1 is the total shield thickness. For the linear distribution the probability of
collision between some arbitrary path length L, and L, + dL is given by
dLlA" whereas the probability in the true distribution is 1:, e-k,L I dL. A
weighting function must be assigned to the particles such that the product of
the number colliding times their weight is equivalent in the two systems, or
(5.4-10)
From this equality the weighting function
W
o
I,
_"
,
W= A 1: e""L
(5.4-11)
is given, where W is the weight assigned to particles at depth L, and Wo is

the initial particle weight in the unbiased system. Although the particle
weight is less, a larger fraction of source particles penetrates to great depths,
which increases the accuracy of the averaging there.
Similarly, any other arbitrary distribution of collision densities may be

used with the appropriate weighting schemes. One of these, the exponential
trausform, is designed to maintain approximately the same population of
particles at any point within the shield. Since the particle population and
flux density decrease approximately exponentially with travel through the
shield, these quantities are weighted by a function that increases exponentially with shield penetration. In this scheme a solution is sought for the
quantity
(r,w,E) = N(r,w,E) .,g(r,w,E)
where N =
r=
w ::
E=
unbiased flux density

spatial position
direction of particle motion
particle energy
(5.4-12)
230
In a source-free region this amounrs to biasing of track lengths only, but in a

distributed source the same importance function is applied both to source
starting positions and to track length.
We will consider only the implications to initial track-length selection
and particle weight in this section.
For an exponential adjustment we may define a pseudo cross section ~;
given by
~; = ~t (E) - g(r,w,E)
(5.4-13)
and select from a distribution of first collisions given by

(5.4-14)
The weight applicable to a particle suffering a first collision at depth Lean

be obtained by equating the products of first-collision density times the
weight in the biased and unbiased systems
(5.4-15)
from which
W(L) =
Wo
~t
e-g(r,w,E)L
~t -- g(r,w,E)
(5.4-16)
Thus, for posItIve values of g, the following observations may be made

concerning the collisions occurring at a large value of L:
1. A larger number of particles are available for interaction.
2. The number of interactions occurring per unit flux density is smaller.
3. The weight of interacting particles is smaller unless ~t 0( 1.
One form of the transform, first proposed by Kahn,7 is given by
g(r,w,E) =
wC
(5.4-17)
where E is a unit vector in the direction for which penetration is to be

studied, W is the direction of particle motion, and C is a constant governing
the magnitude of the transforming factor. In this form the magnitude of the
transform is proportional to the cosine of the angle between E and W [i.e.,
};(r,w,E) = C when the particle motion

g(r,w,E) = -C in the opposite direction] .
Eriksson" has proposed a form
is
in
the
231
direction e and
g(r,w,E) = C [(x - XO)2 + (y - YO)2 + (z - ZO)2]%
(5.4-18)
which would give spherically inward or outward biasing (depending on the

selection of sign) about the point (xo,yo,zo). Perkins and Burre1l 9 have
pro posed the form
~t = ~t
[1 - M( w e)]
(5.4-19)
and have made efficiency comparisons between this and the form
~7 = ~t
C (w e)
(5.4-20)
No clear preference was indicated for either of these two methods from their
study.
One of the most persistent difficulties in applying the exponential
transform is the possibility that the effective cross section ~* may become
negative. Some users S - 1 0 have prevented this by restricting the choice of gsa
that ~t is positive for the minimum value of ~t encountered in the
problem. However, such a restriction in the choice ofg may result in a severe
limitation on the efficiency of application of the transform.
Leimdorfer l l proposes two alternatives. The first is to set
~* =~'
>0
(5.4-21)
whenever ~ * becomes less than zero, where the value ~' is arbitrarily chosen
by the user. The second method is simply to accept the negative value and
select from the resulting distribution. This method is generally restricted to
problems that have finite external boundaries at distance 1"1 in the direction
of transport. Negative values arc then allowed only for penetration <;;1",;
beyond the boundary they arc replaced by an arbitrary positive value as
above. Thus, along a particular track crossing the boundary, two weighting
functions might be applied, one for each of the segments lying on either side
of the boundary.
232
Extending these concepts to a track along which l: * varies with the track
length and negative values are permissible, the track-length distribution is
given by
(5.4-22)
where l:t(S) is the total cross section as a function of track length, or, if
collision is to be forced between some limits 0 < L < T, then
,
I'(L)
ll:i(L)1 exp[- It l:i(s) dS]

IHIl:i(L)I exp H~ l:t(S) dsHJi
(5.4-23)
As in previous cases, the weighting function is given by

W=Fo(L)W~
(5.4-24)
F(L)
where the subscript zero refers to the unbiased values,

Clark' 2 has prepared an excellent review of the exponential transform in
which he discusses the efficiency of the method as a function of the form
and magnitude of g(r,w,E). For high-energy (7 MeV) gamma-ray sources
uniformly distributed through a 100-cm slab of ordinary concrete and for
the transform given by
(5.4-25 )
g=W'G
he found that the estimated relative errort reached a minimum ncar GIl: = 1.
Where the sources were distributed with slab thickness, 1':, in accord with the
function e- kz and Leimdorfer's first scheme for treating negative values of l:
was used, the minimum relative error was obtained at values of GIl: lying on
either side of 1.0 (e.g., 0.6 and 1.6) over a wide range of values of the
parameter k (0.5 " k " ~). As slab thickness is increased, the minimum error
again occurs near GIl: = 1.0, although this minimum is a larger value owing
to the increased penetration. When negative cross sections were dealt with
and Leimdorfer's second scheme was used, increased efficiency was noted,
and the minimum relative error in all cases was GIl: = 1.
tEstimated relative error"" deviation of the mean/estimate of the mean
=:
alp.,
233
With various particle-weighting techniques, it occasionally happens that a

particle arrives deep within the shield with a weight far in excess of the
average weight of particles at that depth. The presence of these high-weight
particles adversely affects the convergence of the value of flux density per
unit source particle for the region. This can be prevented by usc of the
technique of splitting, which decreases the particle weighr and increases the
number of particles by the same factor. Splitting may be instituted at the
point of collision or where previously defined boundaries are crossed. A
parallel problem is that of partieles scattering back into a region where their
weight is so far below the average weight that further tracking would have
little effeet on the average flux density. This is adjusted by a game of
Russian roulette. Whether or not the particle survives is determined by the
ratio 'of particle weight to some arbitrary weight typical of the region. As
discussed in the following text, a random number is used to determine the
future of such particles.
One may define certain minimum and maximum permissible particleweight values for a given region. When a particle appears with W < Wmin, a
random number is obtained and eompared with W/Wmin. If tl < W/Wmin,
the particle is saved and its weight is increased by the ratio Wmin /W so that
it has weight Wmin. If tl > W/Wmin, the particle perishes; i.e., its history is
terminated. Any arbitrary value W between Wmin and Wmax (e.g., an
average of the two limits) could be set as the weight of particles surviving
Russian roulette provided the corresponding ratio (W/W) is permitted to
survive. When W> Wmin a further test is made to determine whether
W> Wmax ' If it is, the particle is split into a number of particles given by
the integer value of the ratio W/W max , denoted as [W/W max J. Each particle
resulting from the split is assigned a lower weight, Wo , given by
l
W
W0 = ,-;-;,,-;-;---,
I W/W max J
(5.4-26)
and then each is tracked separately to its death.

The boundaries at which splitting and Russian roulette occur may be
arbitrarily defined without regard to region boundaries. The values of Wm ax
and Wmin assigned to regions should be based on the expected behavior of
the flux density. Where strong biasing is used in a particular direction, the
flow of particles in the opposite direction may be adequately controlled by
the Qiased track lengths so that Russian roulette is not necessary. Splitting
may still be required, however, to cover cases where the weight-adjustment
234
quantity (kt/kiJ e-g L becomes very large owing to approximately equal

values of kt and g.
An alternative to the splitting and Russian roulette techniques is a
technique that also performs the biasing function. In this method boundaries
arc defined at which particles crossing in one direction arc automatically
split into an arbitrary multiple, M, and those passing in the opposite
direction suffer a fractional mortality equivalent to 1 - (11M).

For example, such boundaries could be defined at surfaces where the
flux density would be expected to decrease by a factor of 2, and the
particles crossing the boundary toward the reduced.fluxdensity region
would be doubled. In this manner the particle population would remain
approximately at the same level through the shield, and the average particle
weight would reflect the degree of attenuation. The efficiency of this
method would depend in part on how well the flux.density behavior is
predicted at the outset.
5.5 COLLISION PARAMETERS

When it has been determined on the basis of the total material cross
section that a collision has occurred, it is then necessary to determine which
of the possible nuclear species was involved and which of the possible
interactions of that species took place. For this determination both the total
and the individual interaction cross sections must be available to the
computer for each nuclide over the energy range of interest.
In a material containing several clements of atomic density Pi (nucleil
cm 3 ), and total cross section ati(cm 2 /nucleus), the total material macro
scopic cross section is given by
(5.51)
and the probability of a particular species' being involved is given by the
fraction of the total cross section represented by that species PiOt;!kt. The
nucleus involved in a collision can be obtained by selecting a random number
r and comparing it to a cumulative distribution,
k
I:
P(k); i~l
PiOti
kt
(5.5.2)
and solving for k, where k can be any integer between 1 and 11, inclusive.
235
In a like manner, the type of interaction can be determined by

comparing anothcr random number to
I
P( I) ~
1; a.'
-,j-"1'-....':;..}_ ~ r
(J ti
(5.5-3)
wherc the ajj arc the various components of the different interactions
making up the total cross section of clement i.
For neutrons the ajj would ordinarily be the cross sections for elastic
scatter, absorption, fission, (n,n'), (n,2n), and (n,3n). Other identifiable
reaction types can ordinarily be classed into one of these categories; e.g.,
(n,p), (n,,,,) , (n,d), etc., arc absorptions, and (n,n',,,,) could be classed as
(n,n') since the resulting charged particles are of no interest except as they
may affect the energy of the neutron given off.
Secondary neutrons and gamma rays are associated with most of these
interactions. Ordinarily the secondary neutrons would be included by
tracking them in turn after the history of the incident neutron is terminated.
The secondary gamma-ray source data would be stored for later processing
by a gamma-ray transport program.
For gamma rays only three cross-section components, photoelectric
effect, compton scattering, and pair production, would be involved. Two of
these, photoelectric effect and pair production, would terminate the history,
although pair production would result in secondary gamma rays from
positron annihilation, which should be included as a new isotropic source of
a.5II-MeV gamma ray.
At first glance the choice of collision parameters would appear to be one
of the simpler tasks in tracking a particle. This is deceptive, however, because
the task of amassing the reg uired cross-section data in the proper format can
dwarf that of coding the problem. A distinguishing feature of the better
Monte Carlo codes is the degree to which the cross-section preparation task
has been automated (or the availability of an auxiliary code for that
purpose). In addition to data at individual energy points, provision must be
made (through interpolation or analytic fit) to cover all possible energies of
interest. Schemes often used in this task include linear or logarithmic
interpolations or polynomial fits to the data within certain energy bounds.
An additional complicating factor is the use of biasing in the selection of
the interaction type. For example, incident neutrons may be assumed to
scatter at every interaction point, and the weight of each after scattering will
be reduced by the fraction "i:.J1:." the ratio of the scattering to the total
236
cross section. In this manner the average scattering kernel will be the same as
though the proper number of neutrons were allowed to scatter with no
weight degradation, but the computing efficiency will generally be greater
because histories will not be terminated by capture events.
In practice the entire transport problem can be run with only scattering
interactions being allowed to occur (making the proper weight reductions)
and with the collision parameters at each interaction being stored for future
processing. The collision parameters are then analyzed to determine
absorption density, to make flux -density estimates (as discussed later in this
chapter), or to calculate energy-absorption density.
5.6 PARTICLE PARAMETERS AFTER COLLISION
The next task is the determination of the parameters of the particle(s)
that survive an interaction. These parameters include the type, number,
energy, and direction of the surviving incident particle and of any
secondaries created in the interaction. Depending on the purpose of the
calculation, one may also wish to determine the energy deposited in the
system. The determination of these parameters involves the application of
differential cross-section data in even greater detail in the selection of the
interaction type. The details of the method will necessarily depend on the
format used in specifying cross sections.
Rather than catalog the large number of possible nuclear interactions
involving neutrons and gamma rays and the optional methods that may be
employed in treating these in a Monte Carlo code, we will give examples
illustrating the principles involved.
Reference should be made to Chap. 3 for a discussion of the types of
interactions and of some of the expressions for particle energy following
collision.
5.6.1
Neutron Elastic Scattering
We recall that the angular distribution of elastically scattered neutrons is

approximately isotropic in the center-of-mass system for the lighter elements
and becomes more complicated for the heavier elements in which forward
(and occasionally backward) scattering is usually favored. The anisotropic
behavior generally intensifies with increasing energy of the incident neutron.
Most Monte Carlo programs use scattering angles measured in the center-ofmass system for selecting the scattering angle and computing energy after
237
scattering. The angle of the scattered neutron is then converted to the

laboratory system for tracking.
The scattering angle J/; in the center-of-mass system, measured from the
direction of travel of the neutron, will be selected randomly from an
isotropic distribution by the formula
cos J/;
=Jl = 2n -
(5.6-1)
where n is a random number. This is identical to choosing the polar angle of

emission from an isotropically emitting source.
Where the isotropic scattering approximation cannot be used, the
differential angular scattering cross sections, a( J/;,E) , must be used to
determine the probability distribution of scattering angle. A customary
procedure is to tabulate values of the cumulative scattering probability
P( J/; E)
It a( J/;,E) dJ/;
I~
a( J/;,E) dJ/;
(5.6-2)
at discrete values of J/; and E and use interpolation in both dimensions. There
are disadvantages to this method, however, from the standpoint of
interpolation inaccuracies and machine-storage requirements. As an alternative, the differential cross section a( J/;,E) may be fit by a function such as
a( J/;,E) = AE + BE cos J/; + CE cos 2 J/;
(5.6-3)
and the selection may be performed using the differential probabilities in

conjunction with a rejection technique. t Values of the coefficients AE, BE,
and CE may be input for various values of energy where the function
provides an adequate fit. Where the function cannot provide an adequate fit,
the differential form may still be used by selecting values at fixed values of J/;
and E and using interpolation in conjunction with a rejection technique. This
latter method would be inferior to using the cumulative distribution where
computing time is concerned, although input requirements for the differential form are simpler.
Once the scattering angle J/; has been obtained in the center-of-mass
system, the energy loss can be computed from the energy and momentum
tWc should note that O"(I/I,E) is often given by Legendre polynomial coefficients. providing an
alternative to Eq. 5.6-3.
238
Select 1J;
based on O(I/;,E)
Calculate E'
(Eq.5.6-41
Calculate ()
IEq.5.6-5)
Select (/J
<p=n 211
Fig. 5.7 -Elastic scattering.
balance equations and the laboratory scattering angle can be calculated. The
equations involved in these two operations are given in Sec. 3.2.2. If, as
before, we assume that the mass ratio of the target nucleus to the neutron is
A,
E' = 1 + A[ + 2A j cos IjJ

E
(A j +1)'
(5.6-4)
1+AjcosljJ
cos Ii = -:-:--c--:--;--c--'::-,---'---;--,,;
(1 +A[ + 2A j cos 1jJ)~
(5.6-5)
and
where E' = energy after scatter

E = energy before scatter
Aj = mass of a nucleus of type i (neutron mass = 1)
Ii = scattering angle in the laboratory system
IjJ = scattering angle in the center-of-mass coordinates
With E' and Ii known, the only parameter needed to resume tracking the
neutron is the azimuthal angle q" which is, in all cases (except where
polarization is considered), selected from a distribution equally probable in
the region 0 .;; q, .;; 211, or q, = n211.
In summary, the basic operations in the neutron clastic-scattering routine
are illustrated by Fig. 5.7. The directions of the neutron may then be
translated into direction cosines taken relative to the geometry axes based on
the known direction of the incident neutron and the deflection (Ii,q,) from
that direction.
5.6.2 Neutron Inelastic Scattering
An added parameter is included for neutron inelastic scattering, i.e., the
amount of energy deposited in the target nucleus. The loss of energy by this
means prevents the energy of the neutron after collision from being a unique
239
function of the scattering angle. The equations derived from energy and
momentum conservation t corresponding to the clastic-scattering equations
(Eqs. 5.6-4 and 5.6-5) then become
E' = (Ai-':1)2
1
[ ] + Ai ( 1 -- ~ ) + 2A; cos iJ;
If
(5.6-6)
(5.6-7)
where E is the excitation energy of the target nucleus. These equations

reduce to the elastic-scattering relations when E::;: O. For E > 0 and for Aj
large, they become approximately E' = E -- E and cos = cos iJ;. As with
clastic scattering, the distribution in I/J must be an input quantity.
A particular nucleus may have various excited states, each with its own
excitation energy, E. The neutron energy required for excitation of a
particular level is E> I(A + l)/A I E. When the energy of the neutron is
sufficiently high that anyone of several levels may become excited, the
interaction probabilities may be expressed in a total inelastic-scattering cross
section, a"n'(E), with an associated probability distribution prE) among the
levels, or in terms of separate cross sections onn'(E,) associated with the
levels. In light nuclei the levels may be dealt with individually. This may
also be possible for some levels in heavy nuelei, but in many cases (notably
at higher excitation energies) a large number of closely spaced levels may
appear which require an alternate treatment. One procedure is to split the
possible values of incident and deflected neutron energy into bands and use
experimentally determined probabilities for group-to-group transfer. These
are expressed as cumulative probabilities over the possible E' (energy after
scatter) groups for a given incident group. Then E' is selected by search
and interpolation as discussed previously. An alternative approach that can
be used in some cases is to assume a distribution in the continuum region of
E'
Ji( E')
E'
=-
1'2
~,
e -E IT
tEquations 5.6-4 to 5.6~7 are developed in detail by Cashwell and Everctt. 3
(5.6-8)
240
where T is a constant (or a function of E). This is a result of the statistical

model of the nucleus developed by Blatt and Weisskopf ' 3 and frequently
used in neutron-crass-section estimates. This function could be integrated to
obtain 1'(E'), but it would be more appropriate to select E' from the [(E')
distribution function using a rejection technique.
More than one secondary energy law may be used for particles of a given
incident energy. For example, the probability of exciting anyone of several
widely spaced levels may be 0.3, and the corresponding probability of falling
into the continuous range may be 0.7. By generating a random number, the
code could choose the technique to be used in calculating E'.
5.6.3 Compton Scatteting
The treatment of Compton scattering collisions is similar to that for
neutron elastic scattering in that the scattering angle is ordinarily picked
from stored distributions and the photon energy is computed as a function
of incident energy and scattering angle. Center-of-mass coordinates are no
simpler in this case; thus laboratory coordinates are used.
The photon energy after collision (discussed in Sec. 3.2.1) is given by
E' =
J
-:--=-=-1 + E( 1 - cos e)
(5.6-9)
where both E and E' are expressed in terms of the electron rest-mass energy
(mc 2 = 0.511 MeV). Extensive graphs and tables of the angular distribution
,
are available '4 S for input to the program. Some investigators have elected
to select scattering angles exactly by a rejection technique, using the
Klein-Nishina formula (see Chap. 3) as the probability density function.
Methods of making such an exact selection are given by Kahn I 6 and by
Cashwell and Everett. 3
5.6.4 Particle Absorptions

There are no particle parameters to be calculated following an absorption
interaction. This event simply terminates the particle history. However, a
record of the particle parameters at the time of collision may be desired for
collision analysis.
As with neutrons, it may be desirable to increase the average number of
photon scatterings per particle by forcing a scattering at each interaction
241
point and making the appropriate weight adjustment to account for the
portion of photons that, in reality, are absorbed.
In an absorption event secondary particles of the same type radiation as
the incident particle may be tracked from the absorption point in the same
calculation. J f the secondary particle is of another type (e.g., gamma rays
from neutron capture), the capture parameters will be used in formulating a
source distribution for a separate calculation of the transport of the
secondary particles.
Fission of heavy nuclei is a special case of absorption in which secondary
neutrons and gamma rays will be generated with a known yield, vo, per
interaction and a known energy distribution. With secondaty neutrons,
account must be taken of the weight, W, of the neutron initiating the fission.
For example, one neutron could be genetated with a weight Vo w.
Alternately, a number of neutrons could be selected from a distribution of
integral values whose average is Vo and each of these assigned a weight W. In
both cases the neutron energy would be determined by selection from a
cumulative energy distribution, and the direction would be chosen from an
isotropic distribution.
5.6.5 Calculation of Emergent-Direction Cosines
Usually particle directions are defined by recording the direction cosines
of the particle-track vector. As discussed previously, scattering angles at a
collision point are characterized by calculating the polar angle IJ (determined
by the differential scattering cross section) and the azimuthal angle r/J (picked
at random between 0 and 27T) of the scattered track relative to the incident
track. If (a,/3;y) are the direction cosines of the incident particle, the
direction cosines of the scattered particle (a',fJ';y') are given by the set of
equations:
-y' = -y cos IJ - (1 -- -y2)~ sin IJ cos r/J
(5.6-10)
242
except in
the case where 1 - '"'1 2 approaches zero, in which case the
degenerate form
C;'
~ sin
0 cos
</J
{3' ~ sin sin </J

"'I' ~ "'I cos </J
(5.6-11)
is used.
5.7 PARTICLE SCORING

The discussions to this point have concerned the generation of source
particles, tracking to the first collision, and the determination of the new
particle direction and energy following collision. The process of picking

track lengths and analyzing collisions may be continued until the particle
history is terminated. The termination may as previously mentioned, be the
result of exceeding predetermined bounds placed upon energy, spatial
position, and/or particle weight.
We next consider the output of the Monte Carlo calculation. Specifically,
answers arc required for the following questions: What is the desired form of
the output data? How do particle histories contribute to obtaining these
answers?
Consider the case of a point source and an infinite-slab shield of
thickness 1'. Typical of the results that might be desired are the following:
1. Flux density as a function of position, direction, and/or energy inside
the shielding material.
2. The penetrating dose or flux density.
3. The energy and angular distribution of the penetrating particles.
4. The distribution of penetrating particles relative to the number of
collisions encountered before penetrating.
5. The distribution in time of arriving particles.
1
There are obviously many other possibilities.

The flux density per unit source particle incident per unit area inside the
infinite slab can be obtained by the sum
<l>
~.!
n
""..f1iL
LJ; cos OJ
(5.7-1)
243
of particles crossing a plane sutface parallel to the sides of the slab, where 0
is the angle between the particle direction and the normal to the plane. The
result is normalized to a unit source by dividing by the number of case
histories, n.
The flux density at a particular location can be estimated by describing a
volume region about that location and calculating the weight times the track
length per unit volume of particles passing through the region. The flux
,
5'
R
Fig. S.8-Statistical estimation.
density at a point can be obtained by a method known as statistical

estimation. This technique is used to estimate the scattered-radiation
contribution to point detectors from scattering events within the shield and
is often used with the forccd-scattering technique discussed in Sec. 5.4.

Consider, for example, a track within the shield in Fig. 5.8. The flux density
at point D is estimated by determining the particle weight with which the
particle would arrive at D (within a unit spherical detector) if it were forced
to scatter in that direction from each point of interaction, 5. If the angle
between the incident-particle direction and the line of sight (5 to D) is Os
and the particle is incident on 5 with energy E, then the weight Wj of the
particle forced to scatter to D is
Wj
as(E)
Wo at(E) f(Os,E)
(5.7-2)
where Wo = weight of the incident particle (at 5)

as(E) = total scattering cross section of the element involved 111 the
collision
a t(1i) = total cross section of that element
f(Os,E) = probability density per steradian of scattering through angle Os
Note that f f(Os,E) dn = 1 and that the product as f(O,E) is equivalent to the
differen tial angular scattering cross section a( 0,E).
244
The contribution <t>j of the scatteted patticle to the flux density at D is

given by
(5.7-3)
whete ~t(E') is the total macroscopic ctoss section of the scattered particle.
The flux density at point D is estimated by summing the conttibutions ftom
all scatteting interactions:
1
<t> = - 1:: <t>j

n ;
(5.7-4)
This estimate is independent of the particle-ttacking procedure and does

not terminate the history. Ordinarily the only calculation common to both
the estimator and the tracking routine is the detetmination of which nuclear
species in the material was involved in the collision. In the estimating
process, every collision is assumed to be a scatter in the direction of the
detector, but, fot ttacking purposes, completely independent determinations
are made of interaction type and particle parameters after collision. A better
justification for this technique is given in Sec. 5.2.
The problems of convergence using statistical estimation of this type
have been examined by Kalos.' 7 He shows that an infinite variance and a
preferential tendency to converge from below (as the number of histories is
increased) result. Nevertheless the process does con~erge, and it has been a
very useful technique in many situations.
Where surface detectors are of primary interest, a different form of
estimator may be used. In this form flight-path length and direction are
selected normally just as in statistical estimation, but estimates are made by
extending flight paths in a straight line until a surface intersection is
obtained. In one form, known as the last-flight estimate, this estimator
consists in stretching every flight path that will provide an intersection. In
another form, known as the first-flight estimate, the ftrst-flight path length is
adjusted, where possible, such that all collisions lie on the surface. Of course,
appropriate weight adjustments must be made.
To illustrate the last.flight estimator, let us consider a point ~ource in an
infinite medium with a spherical surface detector at radius R. Collision
points in Fig. 5.9 are assumed at points A to D, and estimates are made by
stretching each track until an intersection with the sphere is obtained. Note
245
that intersection is always possible for tracks inside the sphere. Projection of
tracks from collision points outside the sphere may not intersect, but, when
they do, there will be a double score since the particle penetrates the sphere
~
2
-III B
I
0--.....'
I
R
I d ~~,
'-I
C
A
---I
/
/
/
51
~.
I
/
-.J
Fig. 5.9-Last-flight estimator.
twice. The estimate of flux density, <l>i(R) , at the sphere for one particle
flight will be given by
(5.7-5)
where Wi = particle weight on the ith flight
di = total distance along the selected flight direction to the point of
intersection
13i = angle between the normal to the surface and the track at the
point of crossing
~i = total cross section of the medium for particles on the ith flight
Thus the weight of the particle is adjusted by the attrition probability of the
total flight to the surface. For the path originating at point D (the 5th
flight), the estimator will be
41TR 2 <l>s (R) = Ws
(e-~,d, +e-~,d~)
leas 13,1
(5.7-6)
to account for the double encounter with the surface. Here again the
tracking continues independently of the estimates, but particle parameters
after collision are used as a basis for making the estimates.
The first-flight estimator is similar to the last-flight estimator except that
adjustments are made in the first flight so that, where possible, each collision
point lies on the surface of the sphere. In Fig. 5.10, the simulated tracks
246
REAcroR SHIELDING FOR NUCLEAR ENGINEERS
involved in the estimate for each collision in the history shown in (a) are
shown in (b) through (d). In (b) the first flight is adjusted to make the first
collision point lie on the surface. In (c) the Hrst flight is extended so that the
randomly determined end point of the second flight is shifted (parallel to the
----Ibl
lal
"'a'"
'"
"'" '" -
Ie)
ldl
Fig. 5.10-First-flight estimator. (a) True particle track. (b) Adjustment for first flight.
(c) Adjustment for second flight. (d) Adjustment for third flight.
Hrst flight path) until it coincides with the surface. In (d) the Hrst-f1ight
length is reduced so that the third collision lies on the sphere. These tracks
have been confined to a plane for ease of illustration, but adjustment of
collision points lying outside the plane must still be in a direction parallel to
the Hrst flight.
247
Numerically, the flux value of the particle scored in each case is

'"
-1: o df
41TR 2 cI>i(R) = W i "'0 e
~ilcos
where
ad
(5.7-7)
df = first-flight length after adjustment

ex = angle between the first-fligh.t path and the normal at the point of
intersection
~o = total cross section at the source particle energy
~i =total cross section on the ith flight
This expression is similar to the last-flight estimator in that the path-length

weighting factor is e-1:d. Primary differences are that the angle of
intersection with the sphere is determined by the first flight rather than the
last flight and that the weight is adjusted by ~o/~i to account for the fact
that interaction occurs at energy during the ith flight rather than the source
energy. In Fig. 5.10(d) the interaction point is outside the sphere; thus the
first flight must be adjusted backward to provide intersection. The estimate
cannot be made for points external to the sphere if projection in line with
the fIrst flight does not intersect the sphere or if the adjustment requires a
negative first flight. The estimator on the fIrst leg [Fig. 5.10(b)] will be
numerically equivalent whether the first- or last-flight estimators are used.
Trub ey I8 performed a series of calculations for point sources of fission
neutrons and 14-MeV neutrons in water to compare the results of these two
estimators. He concluded that either of the two forms may be applied with
equal ease, but the first-flight estimator appeared to give better results,
particularly in the low-energy (0.01 to 4 MeV) fission-neutron range.
5.8 STATISTICAL VARIANCE
The number of histories that must be followed to yield acc~ptable results
in a given problem depends to a large extent on the nature of the answer
desired. For example, a calculation could yield a value of total dose per
source particle accurate to 1% and at the same time have a very large
uncertainty associated with the current of neutrons in the range
3 MeV < E < 4 MeV and 100 < < 30 0.
A theorem of mathematical statistics that may be invoked as a means of
estimating statistical uncertainty is the central limit theorem. Consider an
experiment that can result in K different ways, designated as WI' .WK with
248
associated probabilities PI . .. PK, where Z; Pi = 1. In this case the average

j==1
(or mean) result may be defined as p. = Z; PKWK and the dispersion (variance
about the expected value) as
(5.8-1)
Suppose now that in a Monte CarIo penetration calculation N source
particles are started and that the total penetrating weight (summed over N
trials) of particles in a certain category is M. The central limit theorem states
that
(5.8-2)
This relates the probability that our answer will fall within some variation
about the true average to the number of trials and the desired maximum
value of
erf(x)
2 IX
=y1i
0
e-U du
(5.8-3)
(5.8-4)
The value of PN would be given in terms of the third moment of the W

distribution and for large N would not be a significant contribution. Thus,
for a given level of probability (given value of f), the size of is directly
proportional to ,,;o;;TN., Thus doubling the number of histories decreases the
value of only by a faaor of 2-~ This is a very general result; i.e., Eq. 5.8-2
is true provided only that p. and ae exist. Spanier and Gelbard 1 discuss this
and other aspects of statistical variance in some detail.
In cases involving direct analog calculations with no particle weighting, it
may be prohibitively expensive to increase N sufficiently to bring within
acceptable limits. Then an alternate means of decreasing would be to
decrease the dispersion, ae , by introducing biasing and weighting schemes
that increase the number of samples for a particular output category per
source particle, or give contributions that on the average lie closer to the
average value, or that do both.
249
To illustrate how 0e might be estimated in a particular problem, let us

suppose N source particles are generated and their contributions to
penetrating particles of a certain category are Xl ,X 2 , ,XN. The average of
this sample is defined as
1
U=-1:
N,
(5.8-5 )
Xi
and the sample variance is given by

1
02 = -
1:
N,
1
N
=-
(Xj - U)2
1:,
(5.8-6)
It should be noted that U is not the true expected value; it is a computed

average for the sample, which should approach the expected value for
large N if the problem has been formulated properly. Thus Eq. 5.8-6 is only
an estimate of the variance.
The variance 0; about the expected value a is related to the variance 0 2
about the average U of a series of observations by
(5.8-7)
Thus an estimate of
0; is given by
(5.8-8)
Xl
In some Monte Carlo programs, the Xj and

are tabulated for use in
calculating the variance of the answers obtained in one or more categories. In
others qualitative judgments of the output are made on the basis of
convergence and stability of the average as N increases.
Some investigators have shown apparent reductions in the estimated
variance by grouping the data, although this is not a !teal improvement in the
certainty of the answer. Suppose, for example, a penetration calculation is to
250
be made with 4000 particle histories. The data may be subdivided into 40
groups of 100 histoties each and averages, lJ.i, computed. The variance of
these IJ.j about an overall average, Ji, would ordinarily be much smaller than
the variance of the individual data.
The question freq uently arises concerning the effect on variance of some
biasing scheme. This question is not easily answered. As noted in Sec. 5.2,
the optimum importance function gives zero variance, but we do not know it
is optimum unless we have prior knowledge of the result we seek to
calculate. Obviously, the importance function must reduce the variance of
the result below that obtained from an unbiased calculation or it is useless.
Furthermore, there are cases where the apparent (computed) variance is
reduced, but the answer is clearly erroneous. Consequently the task of
proving or demonstrating the adequacy of the importance function rests
heavily on the shoulders of its advocate. Numerous schemes are used, e.g.,
comparisons with nonstochastic calculations, with measurement, and with
unbiased, straight analog Monte Carlo.
Systematic sampling is a technique used to obtain more representative
answers when only a few histories are to be run-such as may be done in
the check-out of a code. In this system the cumulative probability distribution of an occurrence (e.g., source angle, track length, or scattering
angle) may be divided into increments, and, instead of the random selection
of variable values, selections are made by systematically rotating through the
increments. Instead of
F(X)
=r
!(X) dx
=n
(5.8-9)
one may use

F(X) = i - (1/2)
N
(5.8-10)
where i = 1, 2, ..., Nand N is the number of probability increments defined

between 0 and 1.0. To relieve any bias that may be introduced by sampling
from the midpoint of the increment, one may select from
F(X) = i - n
N
(5.8-11)
where i = 1,2,. . .,N and n is a random number. In this system the increments
are rotated but the position within an increment is selected at random.
251
5.9 DEMONSTRATION MONTE CARLO PROGRAM

The previous secrions of rhis chapter introduce a number of concepts
used in a Monte Carlo calculation. A simple and obvious way to interrelate
various segments of the problem is to formulate a Monte Carlo program.
Following standard practice in devising a computer pr"h"'am, we will define
the problem and list the major divisions of the program; in Appendix J we
give a logic flow chart and a FORTRAN listing of the program. Appendix J
also lists the input-data requirements and presents a test problem, including
input and output, which an interested reader can use to verify that his copy
of this program is running properly. Although this program does not include
all the techniques discussed in earlier sections, its use should provide a
first-hand appreciation of Monte Carlo methods. We have compiled and run
this program on a small (8K core of 16-bit words) computer with only the
card reader and printer and its FORTRAN compiler.
The program listing contains a number of comments that describe what
each set of statements accomplishes. Although the computer will ignore
these comments (and they need not be included in working copies), they will
be useful to the inexperienced programmer as explanatory notes to relate to
the text.
PROBLEM DEFINITION
Neutron source:
Shield:
Assumptions:
Point, monoenergetic, cosine current located on

the face of the shield.
Homogeneous semi-infinite slab of thickness t
composed of a light element of atomic weight
A.
(1) Only capture and isotropic (center of mass)
scattering are allowed.
(2) No biasing.
(3) No energy or geometry ,cutoff.
(4) Cross-section handling will provide ~t(E)
and ~s on request.
(5) A random-number-generating routine will
provide random numbers on request.
(6) Appropriate values of constants, such as
source energy, shield thickness, maximum
number of histories, flux-density-to-dose-rate
conversion factors, and energy-dependent
macroscopic cross sections, are input data.
252
Output requirements:
Geometry:
Symbols used:
Leakage energy and angle-dependent current,

fluence, and dose for both shield surface and
neutrons absorbed as a function of energy and
z position within the shield. Neutrons leaving
the shield at z = t are called transmitted
neutrons, and those leaving at z = 0 are called
reflected neutrons.
Fig. 5.11.
n
random number
incident polar angle
q,
incident azimuthal angle
k
cutoff angle below which 0 is assumed to be zero
source energy
energy after scattering
H max
L
N(E,O)
F(E,O)
D(E,O)
Na(E,O)
0I.,(3,'Y
Z
K(E)
cosine of scattering angle in centerof-mass system

laboratory scattering angle
azimuth of scattered dire orion
history number
number of histories to be run
path length
number of leakage neutrons
flux density of leakage neutrons
dose transmitted by leakage neutrons
number of absorbed neutrons
direction cosines of particle
coordinate of particle in direction
normal to shield (z = 0 is the
shield surface on which particles
are incident)
Flux-density-to-dose-rate conversion
factors
No attempt is made to account for particle positions in the x-y plane. The
only information desired from the leakage particles is their energy and the
angle at which they penetrate either shield surface. Calculation of the x,y
coordinates could easily be added to the diagram, if desired, since the
direction cosines and track length are known for each particle flight.
MONTE CARLO MISTHODS FOR RADIATION TRANSPORT
253
The cross-section handling subroutine, XSEC, and the random-numbergenerating routine, RAND, which are included, are intended only for
illustrative purposes and to allow the reader to duplicate results obtained in
the text. The subroutine XSlEC reads a set of cross-section data the first time
-z
Incident particle
~:""-r-;--x
Fig.Soll-Geometry for demonstration Monte Carlo program.
it is called. Then a linear interpolation is performed to determine the total

and scattering cross sections as the energy transferred to XSEC through its
argument list. After the Hrst call of XSEC, no data are read, and only the
interpolation is performed. The subroutine RAND is actually not a
random-number generator but a table of 23 numbers and a stepping
procedure to alter the sequence of using them. For a realistic computation,
XSEC could be used provided the cross-section variation is approximately
linear with energy; however, RAND should be replaced with an actual
random-number generator.
Major divisions of the program logic are numbered below in the same
way as the flow chart (shaded connector symbols) in Appendix J and may be
categorized as follows:
1. Source-particle generation
A. Select direction angles IJ and <1>.
B. Calculate direction cosines 0<,13, and 'Y.
C. Step history counter.
D. Test for history completion.
E. If all histories are complete, branch to print: area.
254
2. Path-length determination
A. Calculate path length L.
B. Calculate z coordinate.
c. Test for penetration through either slab surface.
D. If penetration occurs, store results and terminate history.
3. Collision-parameter calculations
A. select scatter or absorption.
B. If absorption, store absorption parameters and terminate history.
C. If scattering, calculate scattering angle IJ..
D. Calculate neutron energy 8' after collision.
E. Calculate direction angles IJ sand rf>s of scattered neutron.
4. New-particIe-direction calculation
A. Determine new direction cosines ex', /3', and -y' using coordinate
rotation equations (see second chart in Appendix J).
B. Return to (2) for next flight calculations.
5. Analysis of leakage radiation
A. Calculate dose rate per bin.
B. Print out flux density and dose rate.
C. Print out absorptions per increment.
A listing of the FORTRAN coding of this problem is shown in
Appendix J. Designation of input variables has been changed in some cases
from those previously given for ease of coding. A table has been incIuded to
define input variables, limits, and input format requirements. Finally, the
test problem is defined with specimen input and output.
5.10 PROGRAMMING SUGGESTIONS

A few additional remarks are in order concerning usc of the Monte Carlo
method for solving shielding problems.
Monte Carlo codes tend to become quite large and complex as the
various options on geometry, biasing, and analysis are incorporated. The
programs are more manageable if they are broken into subroutines which arc
called into the computer fast memory only when needed (unless the fast
memory is large enough to contain the whole program and working data).
They have also been coded as completely separate routines linked by data
tapes. For example, one code generates source-particle parameters and stores
these on tape. Another code reads the source parameters and generates
MONTE CARLO MEtHODS FOR RADIATION TRANSPORT
255
collision histories. A third code analyzes thc tape output of the collisionhistory processor and calculates the desired output data, such as flux
density, dose, and current. Each of those programs must have compatible
features and use overlapping input data for geometry description, cross
sections, etc.
Checking or debugging Monte Carlo codes is a tedious and difficult task

since the nondeterministic nature of the calculation may mask errors large
enough to influence the answer but not large enough to make the answer
appear unreasonable. Debugging is the step that separates the men from the
boys. It has been found useful to debug each subroutine separately and in
various aggregates prior to assembling the entire program. Simple problems
can be devised to check individual calculations by allowing only one
interaction or by inputting an extreme in geometry or energy. Additional
output features can be inserted into the program for use during debugging
and complete particle histories can be printed out. If the sequence of path
lengths, particle weights, and parameters after collision is available, errors are
easier to detect.
The energy cutoff specified in a fast-neutron calculation to a large extent

will determine the average computing time required per history and thus
should not be set lower than required by the accuracy of the problem. In
problems designed to calculate fast-neutron dose, an energy cutoff near 0.1
MeV is ordinarily used.
A reasonably economical calculation can be made of thermal-neutron
transport and diffusion from a source of thermal neutrons, but the
epithermal region between fast and thermal is much more difficult to treat.
The application of Monte Carlo to reactor criticality studies has been quite
limited. The Monte Carlo method cannot compete with analytical and
semiempirical methods because of the large number of collisions required to
thermalize fast neutrons. It has found limited application,19, z 0 however,
because of the ability to treat complex three-dimensional geometries. Monte
Carlo has been successfully used to determine the thermal slowing-down
density for the input to a thermal-neutron diffusion or transport code. The
choice of a low-energy cutoff point is less important for gamma rays because
of the rapid rise in the photoelectric cross section with decreasing energy.
There is usually little or no incentive to track photons with energies below
0.01 MeV.
Monte Carlo techniques may be designed to reproduce a physical model
in as much detail as is necessary and so provide a powerful tool for solving
256
problems with very few compromises with the physics. The Monte Carlo
method can incorporate any geometry. To use Monte Carlo successfully,
however, one generally must invest a considerable amount in analysis,
programming, and computer machine time. The importance of machine time
is often overemphasized, and analysis and programming are underemphasized. The user should keep in mind that a well-developed theory exists
which specifies, in principle, a near optimum procedure for solving a given
problem. This procedure consists in obtaining the best possible approximation to the value function for the problem and then using this function to
obtain parameters for importance-sampling techniques or to guide development of new sampling techniques.
As an aid to the programmer, the concept of a Monte Carlo programming
system was developed. For example, the 05R system 21 and its updated
version 06R can, in principle, be used to solve any neutron-transport
problem. The framework is there (cross-section handling, geometry-solving
routines, random-walk procedures, etc.), but the programmer must incorporate the special features he desires by adding subroutines to the framework.
More recently a highly versatile and easy to use multipurpose neutronand gamma-ray-transport code, the MORSE code,22 has been developed at
the Oak Ridge National Laboratory. Some of its features include the ability
to treat the transport of either neutrons or gamma rays or to simultaneously
treat the transport of neutrons and secondary gamma rays; the incorporation
of multigroup cross sections; an option for solving either the forward or
adjoint problem; modular input-output; cross section, analysis, and geometry modules; debugging routines; time dependence for both shielding and
criticality problems, and albedo option at any material boundary; one-, two-,
and three-dimensional geometry packages; and several types of optional
importance sampling.
Traditionally, Monte Carlo codes for solving neutron- and gamma-raytransport problems have frequently been separate codes because of the
physics of the interaction processes and the corresponding cross-section
information required. However, when multigroup cross sections are used, the
energy group-to-group transfers contain the cross sections for all processes.
Also, for anisotropic scattering each group-to-group transfer has an associated angular distribution that is a weighted average over the various cross
sections involved in the energy-transfer process. Thus these multigroup cross
sections have the same format for both neutrons and gamma rays. In
addition, the generation of secondary gamma rays may be considered as just
257
another group-to-group transfer. Therefore, when multi!,'toup cross sections

are used, the logic of the random-walk process (th(, process of being
transported from one collision to another) is identical for both neutrons and
gamma rays.
In general, the Monte Carlo method is not the best method for
one-dimensional problems since discrete-ordinates codes are likely to be
much faster than Monte Carlo codes. For two-dimensional problems, Monte
Carlo and discrete-ordinates methods are somewhat comparable, but for
three-dimensional or two-dImensional time-dependent problems, there is no
competitor to Monte Carlo for accurate solutions to transport problems.
REFERENCES
1. J. Spanier and E. M. Gelbard, Monte Carlo Principles and Neutron Transport
Problems, Addison-Wesley Publishing Company, Inc., Reading, Mass., 1969.
2. H. G. Tucker, An Introduction to Probability and Mathematical Statistics, Academic
Press, Inc., New York, 1962.
3. E. Cashwell and C. Everett, A Practical Manual on the Monte Carlo Method for
Random Walk Problems, Pergamon Press, Inc., New York, 1959.
4. G. Goertzel and M. H. Kalas, Monte Carlo Methods in Transport Problems, in Progress
in Nuclear Energy, Physics and Mathematics, Series I, Vol. 2, p. 35, Pergamon Press,
Inc., New York, 1958.
5. R. R. Coveyall, V. R. Cain, and K. J. Yost, Adjoint and Importance in Monte Carlo
Application, Nucl. Sci. Eng., 27: 219(1967).
6. V. R. Cain, Application of SN Adjoint Flux Calculations to Monte Carlo Biasing,
Trans. Amer. Nucl. Soc., 10: 399{1967).
7. H. Kahn, Random Sampling (Monte Carlo) Techniques in Neutron Attenuation
Problems, [and I1,Nucleonics, 6(5): 27-37 and 6(6): 60-65(1950).
8. B. Eriksson, On the Use of Importance Sampling in Particle Transport Problems,
Swedish Report AE-190, June 1965.

9. J. Perkins and M. Burrell, Efficiency of Some Biased Sampling Schemes in Monte
Carlo Calculations, in Nuclear Engineering and Science Conference, April 6-9, 1959,
Cleveland, Ohio, Engineers Joint Council, New York.
10. M. Kalos, Methods in Monte Carlo Solution of the Radiation Transport Equation,
Report UNC-5014 (NDL-TR-36). United Nuclear Corporation, May 31, 1962.

11. M. Leimdorfer, A Monte Carlo Method for the Analysis of Gamma Radiation
Transport from Distributed SOUirces in Laminated Shields, Nukleonik, 6(2): 58-65
(March 1964).
12. F. H. Clark, The Exponential Transform as an Importance-Sampling Device, USAEC
Report ORNL-RSIC-14, Oak Ridge National Laboratory, January 1966.
258
13. J. M. Blatt and V. F. Weisskopf, Theoretical Nuclear Physics, J aho Wiley & Sons, Inc.,
New York, 1952.
14. A. Nelms, Graphs of the Comptan Energy-Angle Relationship and the :K1ein-Nishina
Formula from 10 keY to 500 MeV, Report NBS-542, National Bureau of Standards,
August 1953.
15. R. Latter and H. Kahn, Gamma-Ray Absorption Coefficients, Report R-170 (RAND),
Rand Corporation, Sept. 19, 1949.
16. H. Kahn, Applications of Monte Carlo, USAEC Report AECU-3259 (RM-1237-AEC),
Rand Corporation, April 1956.
17. M. Kalas, On the Estimation of Flux at a Point by Monte Carlo, Nucl. Sci. Eng., 16:
III (1963).
18. D. K. Trubey and M. B. Emmett, A Comparison of First- and Last-Flight Expectation
Values Used in the 05R Monte Carlo Calculation of Neutron Distributions in Water,
USAEC Report ORNL-RSIC-3, Oak Ridge National Laboratory, May 1965.
19. Proceedings of Conference on the Application of Computing Methods to Reactor
Problems, May 17-19, 1965, Argonne National Laboratory, USAEC Report ANL7050, August 1965.
20. H. Rief and H. Kschwendt, Reactor Analysis by Monte Carlo, Nucl. Sci. Eng., 30:
295-418 (1967).
21. D. C. Irving, R. M. Freestone, Jr., and F. B. K. Kaffi, OSR, A General-Purpose Monte
Carlo Neutron Transport Code, USAEC Report ORNL-3622, Oak Ridge National
Laboratory, February 1965.
22. E. A. Straker, P. N. Stevens, D. C. Irving, and V. R. Cain, MORSE Code: A
Multigroup Neutron and Gamma-Ray Monte Carlo Transport Code, USAEC Report
ORNL-4585, Oak Ridge National Laborarory, September 1970.
EXERCISES
5.1 Define the event (sample) space and specify a typical event for the following: (a) One
die randomly tossed once. (b) Two dice randomly tossed together. (c) One die
randomly tossed twice in sequence. (d) The fission-neutron spectrum. (e) A photon
undergoing a Compton scatter.
5.2 For 5.1(a), 5.1(b), and 5.1(c), describe and plot: (a) The probability density function
(PDF). (b) The cumulative distribution function (CDF). (c) Were you forced to use a
random-variable function? (Hint: Suppose the six sides of a die were labeled
a,b,c,d,e,f rather than with spots).
5.3 Figures 2.1 and 2.4 in Chap. 2 present spectral data. (a) Is either function a PDF?
(b) Can the other functions be converted to a PDF? How?
5.4 Using I-MeV intervals, compute and plot the CDF for your answer to 5.3(a).
5.5 A PDF is known to have the form f(x) = C sin x over the integral (0 < x < IT), where
C is a constant. (a) Comput: C. Does the PDF satisfy Eqs.5.1-3 and 5.1-4?
(b) Compute the CDF. Does it satisfy Eqs. 5.1-6 and 5.1-7? (c) On one set of axes,
plot the PDF and CDF. (d) Rather than generate random numbers, select an evenly
distributed set of numbers in the interval (0.1) (e.g., at 0.1 intervals). Using the CDP
MONTE CARLO METI-IODS FOR RADIATION TRANSPORT
259
plotted in 5.5(c), obtain the values of x for each selected number. Use n = F(x)
(Eq.5.1-24). (e) Is the distribution of events obtained in 5.5(d) consistent with the
PDF? (f) Would it have been so if the numbers had been selected randomly? (g) Note:
5.5(d), 5.5(e), and 5.5(f) illustrate the principle of biasing as well as random sampling.
Why?
5.6 The following experiment illustrates the rejection technique. As with any random
process, no two experimenters should obtain exactly the same answer. One side of a
die is labeled A, two sides B, and the remaining three sides C. The accompanying
graph is a proper distribution function for one toss of such a die, as the student
should verify.
6-------------- . . .
5-
41---------4
N
31-
21---<
1O'------I-----:!~---~----'
ABC
Graph for Exercise 5.6(a). Distribution function for one toss of a die.
<a) Is this a graph of a PDF? Why? (b) Use a normal die to randomly select with equal
probability a letter from A, B, or C. For example, let 1 or 2 = A, 3 or 4 = B, and 5
or 6 = C. From the graph find N associated with the selected letter. Then roll the
normal die again to obtain N', a number from 1 to 6. If N' ~ N, accept the letter
choice. If N' > N, reject it. Repeat the process until a sequence of accepted letters
has been established. Tabulate the distribution within the sequence. If both you
and your die were honest, the ratio of A's to B's to C's should approach 2:4:6.
Shield Attenuation
Calculations
W. E. SELPH and C. W. GARREIT
Now we can turn to some of the practical considerations for applying data
and empirical formulas to attenuation problems. Two previously mentioned
observations are worth repeating. There are several levels of attenuation
calculations, ranging from elegant and precise to improvised and approximate. Since no one method satisfies all applications, the reader should
become familiar with a number of techniques so that he can select the one
most appropriate for a particular application. The second observation is that
for most of these techniques a good bit of the calculation has already been
done; a particular application usually reduces to assembling the appropriate
data matrix from the existing literature and integrating over the particular
shield geometry and composition of interest.
6.1 ANALYSIS OF THE SOURCE
Although this topic was discussed in an earlier chapter, some additional
remarks are pertinent. There are, in general, two approaches to the analysis
of the radiation source depending on the methods to be used in analyzing
the shield. The direct approach is to treat the source and shield as an entity
and to obtain as a single calculation the distributions of radiations leaving
the shield which originated within the source, thus solving the sourcedescription and shield-penetration problems simultaneously. The direct
approach is used for a "one-shot" calculation where the system can be
precisely defined and only one set of answers is required. An example would
be a final, detailed analysis of a reactor shield "as built."
The alternate approach is to split the source description and shield
penetration into two distinct phases. The first phase involves calculating the
radiation leakage from the sCurce region to obtain an energy and angular
distribution incident ro the shield. This distribution can then be applied to
261
262
shield-attenuation calculations. This method provides a useful means for

comparing several designs based on parametric data and is more economical
than the direct method where a number of designs are to be analyzed or
where the effect of small variations in shield design is to be studied.
The task of analyzing the radiation emitted by a reactor source is often
as complex as the shield-attenuation calculation because of a plurality of
sources. The source density must be determined for fission neutrons and
gamma rays, fission-product gamma rays, secondary gamma rays from
nonfission neutron interactions, and, in some instances, secondary neutrons
from photoneutron interactions. This analysis requires a detailed knowledge
of the density of fission events, the material compositions, and the neutron
spectrum as a function of position within the reactor.
For example, the local source strength of gamma rays (except for the
fission-product source) in an energy group AEk will be given by
F' Gk + 1:j
1:j IPNa
J J J}
Y"k
J}
where F' = local fission density

Gk = yield of gamma rays per fission within t..Ek
IP j =local flux density of neutrons in' energy group t..E j
Nj =atomic density of element j
0ij = local microscopic cross section for radiative interaction of
neutrons of energy i with element j
Yjjk = yield of gamma rays per interaction
To this sum the fission-product sources should be added.
The method used in evaluating fission-product sources will vary with
their relative importance and with the operating history of the reactor. Their
intensity will increase with time of operation of the reactor at a given power
level until some maximum or "saturation" level is attained. It is common
practice to combine the saturation fission-product activity with the fission
gamma-ray source and consider only this maximum condition. A source
spectrum for this condition is given in Chap. 2.
The fission density and low-energy neutron spectrum, as a function of
position within the reactor core and reflector, are usually derived in an
iterative process using transport and diffusion theory. These data, obtained
in the process of reactor design, are ordinarily available to the shield
designer. If only the fission density is available, the low-energy neutron-fluxdensity spectrum may be calculated using a combination of transport and
diffusion calculations, or an Sn (or Pn ) calculation. In general, the methods
SHIELD ATfENUATION CALCULATIONS
263
used in shield analysis are not adequate for the task of determining the
distribution of fissions within the reactor core or for determining criticality.
The reverse is also true, most reactor analysis codes do not do an adequate
job of predicting the radiation leakage from the reactor. Although the theory
is analogous in reactor and shielding codes, the emphasis is different and the
accuracy with which a particular reaction type is analyzed varies significantly. Two methods that do provide adequate results in both reactor and
shield analysis are the Monte Carlo and the discrete ordinate Sno
Once the position, energy, and intensity of all sources within the reactor
have been evaluated, reactor leakage can be analyzed. The form used in
expressing the energy and spatial distribution of the sources will vary with
the intended method of performing the leakage calculation. The methods
used in performing the leakage calculation are the same as those used for the
analysis of shield penetration as discussed below.
For a direct Monte Carlo solution, a separate source generator routine is
used to generate particles with the appropriate distribution in energy, angle,
and position. Biasing can be used to generate more of the particles expected
to be important to shield penetration, such as those born at high energy,
traveling in a radial direction, and/or originating near the periphery. In a
discrete-ordinate Sn calculation, the spatial variation of source strength can
be expressed analytically. If existing data sets, either experimental or
analytical, are used, then it is sometimes desirable to express the source as an
equivalent simple source, such as a disk, an infinite plane, or a series of point
sources.
6.2 DIRECT SOLUTIONS

One means of obtaining shield-attenuation data is the direct application
of Monte Carlo, moments method, Sn' or other similar methods that directly
use the material cross sections. Chapters 4 and 5 are devoted to the general
theory of the methods involved in direct solutions. Applications of these
methods invariably involve the use of rather sophisticated computer
programs and an expenditure that some problems may not justify.
Consequently direct solutions are not widely used in shield design and
evaluation studies. They are mentioned here only to call attention to the fact
that they may be applied directly to a shield-attenuation problem.
To some extent a judgment based on analogy with similar designs or with
the available basic data will assist in the selection of a method by
determining which penetrating components are likely to be most important.
264
For example, where low-energy neutrons are not expected to be significant

in heating or in the penetrating-radiation component, a Monte Carlo
approach may provide the most accurate analysis. If gamma rays from
thermal-neutron capture are expected to be the primary factor, then a
moments method, Sn' or a combination of transport and diffusion methods
may be sele cted.
6.3 APPLICATION OF PARAMETRIC DATA

A variety of data is available from calculations and experiments that have
been performed to determine the penetration of radiation from basic, typical
sources of neutrons and gamma rays through materials of general interest.
These data maybe scaled to approximate the solution to a variety of
attenuation problems.
The concept of the unit source, introduced in Chap. 2, is integral to the
understanding and application of parametric data. The term unit source
generally refers to a source of unit strength, i.e., 1 particle/sec, 1 watt cm-' ,
1 MeV sec- t cm-2 , etc. All the radiation evolved from such a source need not
necessarily come from the same point, be traveling in the same direction, or
be of the same energy. Distributions in each of the independent variables
(position, angle of emission, and energy) commonly chosen for unit-source
characterization are listed below. Various combinations of the listed spatial,
angular, and energy distributions have been used.
The most frequently used spatial distributions are point, plane, disk, and
line; angular distributions are usually isotropic, monodirectional, cosine, and
(cosine)n; energy distributions are usually monoenergetic, uniformly distributed, and fission spectra.
For a given shield material, the number of dimensions involved in the
parametric data is determined by the calculational method used to obtain
the data, the shield geometry, and the type of detector assumed. For
example, in the geometries represented by a point source in a spherical shield
or an infinite-plane source and shield, the data can be applied to any point at
a given radius or thickness. On the other hand, for a point source and slab
geometry, the position on the exit face also becomes a parameter. Source
energy and angle and exit energy and angle may also be parameters. Thus
parametric data available for application to shield analysis vary from
one-dimensional sets of dose vs. thickness to nine-dimensional sets involving
all the parameters mentioned (two source aJl.gles, source energy, material
thickness, two exit coordinates, two exit angles, and exit energy). In
SHIELD ATTENUATION CALCULATIONS
265
specialized studies one might include additional parameters, such as the

number of collisions suffered by the particle before ex.iting the shield or the
time of flight from source to detector.
In addition to direct applications of parametric-data matrices, interpolation (linear, logarithmic, or other) is often required between particular values
of one or more of the parameters to improve accuracy. Graphical
interpolation is one of the most widely practiced methods because together
with simplicity it provides a more direct view of the behavior of the data
with respect to a particular parameter. Computer programs have been
written for complex interpolations between two or more variables where
nonautomated methods would become quite tedious. In addition, some of
the data sets have been fit with analytical functions. With these functions
analytical representations of the parameters can be used for calculating
transmission or reflection probabilities at any desired point.
Occasionally the shape of an energy or angular distribution does not
change with distance within some range. Such a distribution is said to be in
equilibrium for that range. It is always desirable to observe such equilibrium
conditions because of the resulting simplifications in representing the data.
The existence of equilibrium with distance is an indication that the distance
variable c,\n be separated in the distribution function.
The three types of data most likely to be available for application to a
shielding problem include data resulting from moments-method calculations,
Monte Carlo analyses, and experiments. These three categories are reviewed
in the following sections. Other types of analytical data, particularly from Sn
calculations, may be available and just as applicable to a particular shielding
problem. In such cases the methods of application are analogous to those
discussed in the following sections but have to be adjusted to the number
and form of the variables considered.
6.3.1 Moments-Method Differential Energy Spectra
(a) Neutrons. The moments method has been applied to neutron
penetration through a variety of materials. Data from these calculations have
been summarized by Goldstein 1 and more recently by Krumbein and
O'ReUly! Studies have included penetration of point isotropic fission
sources through water, hydrogen, lithium, lithium hydride, lithium hydroxide, beryllium, carbon, and various combinations of carbon and hydrogen. In
addition, data were obtained for penetration from monoenergetic sources
through water and hydrogen.
266
r-,.' - lem
I~
1
"-l.
"
10
r-......
I"
1\
.......
-""'::'0
2 \
r-......
30
r-......
............ t-..
1\
r--...
i'....
:"
r-...... Ir-......
r-... l"- I"
r-...
'"
...........
........
5 1\
90
...........
10-6
120
7
'\
,......,.,
6
10
12
E. MeV
Fig.6.1-Differential number spectrum of neutrons in water from a point isotropic

fission source emitting 1 neutron sec-1. (From Goldstein.. 1 )
A sample of the data from these calculations is given in Fig. 6.1, which is
a graph of the differential neutron spectrum for various penetration depths
in water resulting from a point isotropic fission source of 1 neutron sec-I.
The slope of the curves in Fig. 6.1 decreases with increasing penetration,
indicating a hardening of the spectrum (increasing fraction of higher energy
particles). This trend is followed in all materials containing significant
concentrations of hydrogen and is a consequence of the dependence of the
hydrogen cross section on neutron energy. As the neutron energy increases,
the hydrogen cross section decreases, and higher energy neutrons then suffer
fewer collisions in crossing a given thickness of hydrogenous material than
lower energy neutrons. The energy loss in hydrogen is so great that neutrons
will usually be thermalized very near their point of initial interaction. For
1000
,>
r-r--....---,----,--.---r-------,
100
"
l,
267
E ;;: 651 eV
10
E ;;: 6.04 keY
~.
10-1
E = 0.22 MeV
E = 0.49 MeV
E'" 1.1 MeV
o 10-2
X
:::>
...-'
2:
E = 2.44 MeV
10- 3
a:
I-
fi:
= 5.4 MeV
10-4
10-6
E=
'-.L_--'-_ _'-_..L_--l._ _L-
10 20
40
60
60
100
12.1 MeV
---J
120
PENETRATION (r), 9 cm- 2
Fig. 6.2-Differential neutron-flux density (x 4nr2 ) vs. penetration (r) in carbon due to
unit point isotropic fIssion-neutron source. (From Goldstein. )
this reason the initial flight is quite important to the distribution deep within
the shield. A larger fraction of the higher energy will be transmitted and will
thus contribute to a reduction in slope (or hardening) of the energy
distribution.
Figure 6.2 shows similar graphs for carbon except that the parameters
have been switched so that the flux density, as a function of depth, is
plotted for various energies. Here the slopes are essentially equal beyond
30 g cm-2 , which indicates that an equilibrium spectrum is established.
Moments-method neutron-penetration data for several materials are given in
Appendix K in tabular form.
In all these calculations the point source is assumed to be surrounded by
an infinite medium. Thus particle spectra calculated at a given depth include
radiation that has been scattered back to that depth from more distant
points. For a shield whose outer surface corresponds to that depth, the flux
density at the surface is overpredicted by the infinite-medium data
(particularly at lower energy where a large percentage of the flux density is
268
due to backscattering) ; thus corrections are necessary for applications of

infinite-medium data to shield layers. A severe limitation of the momentsmethod data is that angular-distribution information on the flux density is
not given.
This type of data can be applied to surface sources by breaking the
surface into small increments and treating each increment as an equivalent
point source. In like manner leakage from a volume-distributed source is
usually represented by volume-distributed point sources or by points on the
surface of the volume source. As an application of these data, consider the
following problem:
Find the flux density of neutrons lying above 2 MeV at a penetration
depth of 60 cm in water from a small (assumed point) fission source
generating 1 watt of power. The solution is obtained by multiplying the
60-cm differential spectrum curve (Fig. 6.1) by the source strength (converted to neutrons sec'l) and then integrating over energies above 2 MeV:
S (source strength, neutrons sec") = (3.1
= 7.56
_ 7.56 X 10
41f(60)2
'<'60 -
10
10' 0 fissions sec-' watt'l)
(2.47 neutrons fission'l)
...
fE
2
mox
10 10 neutrons sec-I
4
1fT No
(E) d
T,
(6.3-1)
where the integral is carried out over the 60-cm curve given in Fig. 6.1.
(b) Gamma Rays. The gamma-ray data generated by infinite-medium
moments-method calculations have found wide application in shielding
calculations. These data are quite extensive both in the number of materials
and the variety of source types that have been considered. Extensive
discussions of these calculations, their results, and their application are
given by Goldstein.'3 The gamma-ray penetration data are more extensive and also more precise than the neutron data. Goldstein has estimated the overall accuracy of the data at 20%. Since backscattering for
gamma rays is much less than for neutrons, interface effects are less
important for gamma rays, and infinite-medium data can be applied to finite
layers to a good approximation.
Examples of the differential spectra are shown in Appendix L for the
penetration of gamma rays of various energies through water and lead. One
significant difference between these data and the neutron data discussed in
SHIELD AITENUATION CALCULATIONS
269
the preceding section is that the photon data at a given range have been
multiplied by 41fr 2 el'or, where 1'0 is the linear absorption coefficient of the
material at the source energy. Since the exponential factor accounts only for
the attenuation of the unscattered component, the flux density at a given
or
energy falls off less rapidly than e-/l . Thus these data, which have been
2
multiplied by 41fr el'o r, increase with increased penetration.
These data have also been integrated over a flux-density-to-dose-rate
response function and expressed as a ratio of total dose rate to the dose rate
due to unscattered photons alone. These ratios, called dose buildup factors,
are a more compact representation of the data that are useful for
applications where biological dose is of primary importance and the
spectrum is of only secondary interest. Buildup factors were discussed in
Sec. 4.8.1 and values are found in Appendixes E and F. Applications of dose
buildup factors are discussed in Sec. 6.4.1.
The data for water indicate that an equilibrium spectrum is established at
very shallow penetrations and does not vary greatly, even at penetration
depths approaching 20 mean free paths. The spectra for lead exhibit a wider
variation in distribution with penetration depth. Differences in the shape of
the 1'0 vs. E curves for water and lead are instructive in understanding the
relative spectral shifts (see Chap. 3).
Differential spectra for gamma rays arc applied to shield attenuation
calculations in much the same manner as the neutron data. An added benefit
in the case of gamma rays is the broader range of materials for which data
are available and the apparent smooth variation of the data with atomic
number, which allows interpolation between elements for which data are
available. A variety of computer programs have been written to perform this
interpolation, also to interpolate between source energies and penetration depths. One method of performing this interpolation is discussed in
Sec. 6.3.4. These programs also integrate over a complex source geometry
and estimate the combined effect of a mixture of elements in a shield.
Integrations over a complex source are performed numerically by breaking
the source into individual point sources representing volume increments and
by breaking the source spectrum into individual monoenergetic sources
representing energy increments. The programs are limited in that they
consider only the materials lying in the straight-line path from a source point
to the detector and also in that they must apply infinite-medium data and
cannot account for the effects of interfaces along the path. The limitations
are not severe in that interface effects are small (-10%) and may be
compensated with appropriate corrections.
270
6.3.2 Monte Carlo

The Monte Carlo method has an advantage over the moments method in
that Monte Carlo techniques can accept a wider range of source types and
can predict the angular distribution of the neutrons and gamma rays at the
exterior surface or at any arbitrary internal point in the material. This
feature, however, spreads the existing calculations through a wider range of

possible parameter variations and thus decreases the probability of finding an
existing set of data to fit a particular need.
with few exceptions, Monte Carlo has been used more for obtaining
precise answers to individual problems than for the generation of parametric
data. This is probably the most economical approach since, as mentioned
previously, it would be difficulr and expensive ro obtain data for all
foreseeable paramerer combinations. Data have been generated, however, for
rhe penetration of unit sources through finite slabs and through infinite
media of some materials.
Monte Carlo neutron data have probably found wider application than
gamma-ray data. This is due partially to the fact that the momentsmethod gamma-ray data have proven to be widely adaptable and have
compared favorably with experiments and other analyses, thus obviating
the need for Monte Carlo photon data.
With Monte Carlo techniques very complex penetration and reflection
probabilities can be obtained. For example, consider a monoenergetic
monodirectional beam incident on a laterally infinite slab at polar angle IIi
and azimuthal angle <l>j (Fig. 6.3). Penetration probability values,
P(II i,t,Ek ,j;R,I3,1I ,<1>'), and reflection probability values, 01(11 i,Eko<l>j;
R' ,{3' ,0' ,rj/ ,E'), are needed. In general practice, however, data of this nature
are more expensive to obtain and more complex than most applications
warrant. Furthermore, in any Monte Carlo srudy, as the number of variables
increases, the accuracy of the differential data degenerates for a given total
number of histories. Consequently the reflected energy E' can be
eliminated by tabulating only the exit dose rate or the distance R' can be
eliminated by integrating over the surface, or, indeed, all the exit parameters
can be eliminated and only the total exit dose rate calculated.
If the calculational model described in the preceding text is being used to
simulate a collimated narrow beam of radiation incident on a shield, the
variation of intensity with the position of exit is sometimes an important
parameter to the design of the shield. The same model can be used for
broad-beam incidence with the principle of reciprocity. The reciprocal
relation of adjacent entrance and exit points indicates that the integral of the
271
Incident
particle
Normal
(energy j)
Reflected
particle
(energyE')
Transmitted particle
(energy )
Fig. 6.3-Particle history parameters.
current density from a I-cm z beam is equivalent to the leakage per square
centimeter from a broad beam of the same intensity per square centimeter as
the collimated beam. Thus the integral
over exit
surface
where the transmiSSIOn probabilities in the integral are for a collimated

source of 1 particle/sec, yields the probability of leakage per square
centimeter due to a broad-beam source of 1 particle sec- I cm-z .
The parameter E can be eliminated by integrating the leakage radiation
to obtain a total particle current density, energy current density, or dose
current. All these quantities are assumed to be dependent on the exit angle.
There is little or no advantage to performing a parametric study by Monte
Carlo unless the angular dependence is retained. The term dose current is
used for the quantity that is given by the product of the differential angular
current density (particles cm- z sec- I steradian-I) and the flux-density-todose-rate conversion factor. This quantity is a convenient way of expressing
the radiation intensity in terms of a dose response while maintaining the
angular dependence. Dose current, D(IJ',q,'), is given by
272
D(O',</>') = f J(O',</>',E) R(E) dE
(6.3-2)
where J(O',</>',E) is the current density per steradian about the direction
(0',</>') and R(E) is the flux-density-to-dose-rate conversion factor. An
application of the data reported in this form is illustrated by the fact that
the dose rate at a point outside the shield from a differential area, dA, of the
surface is
D = 12 J(O',</>',E) R(E) dE dA = D(O;,</>') dA
(6.3-3)
where the detector is at distance r in the direction (0' ,</>'). This quantity is
integrated over the surface of the shield to obtain a total dose rate at the
external point. The variation of the direction of the detector with position
on the shield surface usually necessitates a numerical integration. The
directional energy-current density or particle-current density can be applied
in the same manner as discussed for the dose current. However, the total
energy or particle-flux density at a detector is not generally as important as
the biological dose rate.
Sample results of a typical Monte Carlo study are shown in Fig. 6.4 from
a report by Allen, Futterer, and Wright.' These investigators calculate the
total neutron-energy distribution (integrated over angle) and a total angular
distribution (integrated over energy) but not doubly differential data, i.e.,
energy distribution within a particular angular interval. The data result from
a calculation for 2-MeV neutrons incident on a 60-cm-thick laterally infinite
slab of water at an angle of 45 from the slab normal. In this analysis the
calculation is performed for a thick slab, and the neutrons are tagged on
their first and subsequent penetrations through a given thickness plane. In
this manner both initial and multiple penetrations are tallied, and flux
density vs. thickness and flux density vs. depth are calculated simultaneously. The initial penetrations correspond to the flux density vs. thickness
as though the slab had nothing beyond the plane; thus there is zero
probability of scattering back into the material once the neutrons penetrate.
of course, the multiple-penetration calculations do contain the effects of
backscattering. Figure 6.4 shows the curves of total flux density vs. thickness
and flux density vs. depth resulting from these calculations. The data for
flux density vs. thickness remain 30 to 50% below the data for flux density
vs. depth over a wide range of material thickness, illustrating the effect at
each depth into the slab of backscattering from deeper regions.
273
C
X
::J
-'
u.
10-5 L-_ _--'-
10
L-_ _--'-
"'--L_-l.-"'-L_"'-_ _- '
50
60
Fig. 6.4-Energy-dependent and total flux density VB. depth and total flux density VB.
thickness. Slab, 60 em water; incident energy, 2 MeV; angle of incidence, 0. Flux density
vs. depth:
0,
2 MeV; V, 10 eV to 2 MeV. Flux density vs. thickness:
0,
10 eV to
2 MeV. Hollow symbols, run 1; solid symbols, run 2. (From Allen, Futterer, and
Wright.' )
274
6.3.3 Measured Data

Experimental data can be applied directly to some shielding problems by
properly scaling source intensities and by adjusting for geometry differences.
Experiments are essential for the verification of an analysis or design and for
obtaining cross-section information. In the early days of nuclear technology,
empirical data were widely used either directly in design work or as the basis
of semiempirical analysis methods (see, for example, Sec. 6.4.2).
The source energies available naturally limit the parameter range of
experimental results. Other parameters of the experiment, such as incident
angle, exit angle, and energy at the detector, can be controlled or measured
fairly accurately. Reactor-shield experiments are generally performed with
either reactor sources or the cleaner source of fission neutrons and gamma
rays from a fission-source plate. Other neutron sources include various
pseudo-fission spectra (obtained in accelerators from boil-off in heavyelement bombardment) and a few monoenergetic sources obtained from
low-energy charged-particle interactions. Notable among the more useful
particle interactions are the d-t interaction, obtained by accelerating
deuterons into a tritium target, and the d-d interaction, obtained by
accelerating deuterons into a deuterium target. The Q values (kinetic energy
release) of these interactions are 17.58 and 3.26 MeV, respectively, and the
neutrons evolved are approximately 14 and 2.5 MeV, respectively. Relatively
low intensity neutron sources formed by combining 0'- or -y-emitting isotopes
with light elements, such as deuterium, beryllium, lithium, boron, and
fluorine, have been used for shield experimentation on a limited basis.
Neutrons from these interactions are generally of little interest for shield
measurements because they comprise mixed energies that do not approximate any spectrum to which the shield will be subjected.
Gamma-ray sources available for shielding experiments include fission
gamma rays, fission-product gamma rays, and individual gamma-ray-emitting
isotopes. Difficulties are encountered with fission gamma-ray sources
because of the interference of secondary gamma rays from neutron
interactions. The spectrum from fission products gives a lower average energy
but is useful in testing containers for spent fuel elements and in
weapons-fallout shielding analyses. Individual isotopes can provide useful
monoenergetic photon sources for obtaining penetration data in the range
below 1.5 MeV; however, most of the sources emitting gamma rays of energy
above 1 MeV emit undesirable background energies. Three of the isotopes
frequently llsed for shield measurements are 60Co, which emits one
l.17-MeV and one 1.33-MeV gamma ray per disintegration; 137CS, whose
275
daughter, 137 Ba, emits one 0.662-MeV gamma ray per disintegration; and
144Ce, which emits 17 gamma tays of energy 0.134 MeV per 100
disintegrations accompanied by 2 at 0.1 MeV, 2 at 0.08 MeV, and 17 at
0.04 MeV.
Some of the experimental data quoted here were originally reported
before the adoption of the terms kerma and absorbed dose (as defined in
Chap. 2). We have converted the obsolete unit rep to the .kerma unit on the
basis that 1 rep = 100 ergs/g. Examples of data obtained with fission sources,
either in fission plates or reactors, are shown in Figs. 6.5 through 6.7.
Figure 6.5 shows the neutron dose rate vs. distance through water as
measured along the center line of the Lid T~nk Shielding Facility.s This
curve has been widely used as a check on the validity of methods for
calculating neutron penetration through hydrogenous media. Figure 6.6
shows typical results from an experiment using the Aerospace Shield Test
Reactor 6 housed in an aboveground tank. The angular distribution of
neutrons exiting slabs of borated polyethylene was measured as a function of
slab thickness. A narrowly collimated detector was rotated on a circle of
4.9-m radius centered at a point on the exit face. The reactor neutrons
incident on the shield were precollimated into a lS.2-cm beam. Figure 6.7
shows the variation in thermal-neutron-flux density as a function of range in
air from the Tower Shielding Reactor. 7 The influence of the ground is seen
as an increase in the thermal-neutron-flux density as the detector height was
decreased from 46 to 1.0 m.
Figure 6.8 is an exampleS of time-of-flight neutron measurements in
which a boil-off spectrum from a lead target in a linear accelerator was used.
Spectra are shown for the source and for the neutrons penetrating slabs of
zirconium hydride of various thicknesses. With this method a very fine
angular resolution of the penetrating neutrons can be obtained. The spectra
shown are for penetrations normal to the slab along the source-shielddetector axis.
Data from experiments to determine the transport of 60Co gamma rays
through air from an isotropic source are shown in Fig. 6.9, taken from
Burton! A collimated detector was used for measuring the relative
gamma-ray spectra for a variety of detector polar angles about the
source-detector axis. The data are therefore on a per-unit-solid-angle basis,
which means that the total counts per angular increment were obtained by
dividing by the solid angle intercepted by the increment. Figure 6.10
shows the geometry and results obtained from measurements by Serduke
and Smith 1 of 6Co gamma-ray penetration through iron. Here the angular
276

10'
,--..,--.,-----,:r--,----,--..,----, 10'
...,= 104
10"
.le
.le
,~
,~
l-
I-
10- 1
: 103
'""
"
"
'"
'"
II:
II:
II:
'"0
II:
II:
I:::l
l:::l
,
t;;
,
=
,
l-
10-2
102
10' ' -_ _-L_ _- '_ _----''"'-_ _--'o

20
40
60
80
10-3
'40
' -_ _-L_ _--'
100
120
DISTANCE FROM SOURCE PLATE, em
Fig. 6.S-Past-neutron kenna rates measured in H2 0 vs. distance along the center line of
the ORNL Lid Tank Shielding Facility. (From Casper. s )
distribution is also given but this time as a cumulative distribution. This

represents the integral of flux density over solid angle
as the value of IJ' ranges from 1r{2 to IJ.
277
10- 3 , _ , _ , _ ,_ _, _ , _ , _ ,_ _, _ , . _ , _ - ,
10-4
1
~
E 10-5
w
a:
w
g
Z
~ 10-6
I-
15.2
:J
22.8
30.5
10-8 L-_-'-_-J'-_.J...._--'-_ _.L--_-'-_-J_ _.J...._-'-_ _.L--_-'

-5
o
5
10
15
20
25
30
35
40
45
50
9, deg
Fig. 6.6-Angular distribution of fast-neutron dose rate from particular thicknesses of

3% borated polyethylene slabs. (From Western.")
6.3.4 Fitted-Parameter Data

The data from a parametric shield analysis or experiment may comprise a
bewildering collection of numbers that can be laborious to apply in design
situations. It is, therefore, desirable to find some means of reducing the bulk
of information without losing detail or accuracy. One solution is to fit
analytical functions to the data. An advantage offered by this approach, in
addition to the reduction of data, is that it may provide a means of
analytically integrating over one or more parameters. As an example,
consider the variation of exit current density with slab penetration thickness.
278

107 ~------~-----~------~---~
5 x 105 L
......l
---.JL
200
'00
...L
--l
300
R, m
Fig.6.7-CR 2 <l>th as a function of slant range (R) and detector height (D H ). Reactor
height, approximately 53 m; C = 41T (0.093). (From Muckenthaler, Holland, and
Maerker. 7 )
This often fits an exponential function of the form

t = J0 e
J()
-at+b
, I
-7
-[
partie es ern t' sec
(6.3-4)
or in rare cases by a simpler exponent (b = 0). The variation with exit angle
often approximates a function of the form
J(I,O)
=J(I) k[a + b (cos O)C]
particles sec-' steradian-'
(6.3-5)
where the constant k normalizes to unity the integral of the function over
the 211' solid angle at the exit face, that is,
279
10'
ZrH,.94 Thickness
10"
468
10
12
NEUTRON ENERGY. MeV
Fig. 6.8-Time-of-flight measurement of neutron spectra penetrating zirconium hydride.

[From W. E. Selph, Transactions of the American Nuclear Society, 7(1): 42 Gune
1964)1
k f~/2 [a + b (cos 0)'] 211' sin 0 dO = 1
(6.3-6)
Sometimes the angular distribution has been observed to be a function of

penetration. When this occurs, the constants in the angular-distribution
equation must be evaluated at each depth. At other times the distribution
280
12.2-m separation distance

0
10 collimator (50 half-angle)
103
0
Source
:;0
Detector
12.2 m
,.'
f-
in
f-
10'
20'
10' ' -_ _- '
0.2
--l.
0.4
-'-
0.6
.!-
0.8
' -_ _--l.
1.0
1.2
-'
1.4
ENERGY, MeV
Fig.6.9-Measured gamma-ray spectra and angular distribution from a 60Co source

penetrating through 12.2 m of air. (From Burton. 9 )
has reached an equilibrium after penetrating a depth of one relaxation length

or less, which suffices for most requirements.
The energy spectrum of the penetrating radiation is likely to require a
more complex fitting function than the other distributions and consequently
is usually left in numerical or graphical form.
The problem of fitting a function to the energy spectrum is avoided if
the calculation requires only the total particle-current density, energy-flux
density, or dose rate of the exit radiation. Use of these units may also
involve an angular distribution. For example, an angular distribution of the
//
Detector
/Ii
/e
-t---~
281
- ---
Slab
(a
..
>-'
~
~~/
3.0
='
:l'-
~ '~2.5
2.4C./ /"
~7
iJj~
o~~o
z_
~
5:!2.0
~Z
-,0
LL
0 1.5
oiE
/0
='0:
;J
w 1.0
/ / . /
>1=
-'"
"
_0
~.".--.--- _e-
..,-
0----
'
------",
oD~=-:..L
85
15
9.9~
0--
.,/
0/
I~/j/
~~
o 0.5
~
,---
~@
I-
7.4C~
// /
u.1-
12.4~cm_oo--'
L-_ _...L_ _----l_ _L-_...L--'

60
45
30
20
10 5
DETECTION ANGLE lei. deg
(b)
Fig. 6.10-A family of integral curves for five mean free paths of iron with 60 CO as the
irradiation source. Integral values of multiply-scattered photon number-flux density are
plotted as a function of detection angle. (a) Geometry. (b) Measurements. (From Serduke
and Smith.' 0)
282
dose rate can be obtained from that of the flux density or that of the current
density. The dose rate due to flux density at angle 0 is given by
(6.3-7)
where (O,E) is the flux density at energy E per unit solid angle about the
polar angle 0 (azimuthal symmetry is assumed) and R(E) is the flux-densityto-dose-rate conversion factor.
In addition, the quantity called dose current (Sec. 6.3.2) is defined as
Dc(O) =
f:
max f(O,E) R(E) dE
mm
(6.3-8)
where f(O,E) is the current density per steradian per MeV exiting at polar
angle O.
A set of data giving dose rate vs. depth and dose rate vs. angle is much
more compact than triple differential data in depth, energy, and angle,
particularly if one angular distribution can be used at most depths.
The variations of dose rate with thickness and angle can be combined
into a single equation. In rare cases a function can be found that fits dose
rate vs. angle, thickness, and energy of the incident particle.
In an attempt to fit all the parameters adequately, the empirical
equations can become so complex and so difficult to apply that the
advantage over raw tabular or graphical data is lost for hand calculations.
Such complex equations may still have an advantage for computer
applications in that the computer memory required for storing constants and
evaluating the equations generally is less than that required for storing the
original data and interpolating between data points.
An example of such a fitted function is one developed by Peterson to
represent the moments-method differential gamma-ray-energy spectra for use
in the C-17 point-kernel computer program. 1 1 The probability, f, of gamma
rays' of source energy Eo being degraded to penetration energy E in
penetrating a thickness JJ., t relaxation lengths was fitted by a quadratic in
the variables JJ., t and the effective atomic number, Zen, of the shield
material. The form of the equation is
283
where a set of coefficients, A j , is given for each combination of source and

final energy. For particulat initial and final energies, this equation provides a
continuous probability distribution over penetration thickness and material
composition. The effective atomic number, Zeff, is found by computing the
quantity
1:1'02
i
' ,
/I=--
1: z
.
(6.3-10)
summing over the constituents of the material. This value can be used to
select a Zeff from a plot of I'j vs. Z. For most cases the Zeff so determined
will be very close to the average atomic number of the material.
Although this form of curve fit results in a rather extensive library of
coefficients, it is more efficient in application than attempting to store all
the moments-method data. It can be compared to the use of the equation of
a surface as opposed to a numerical listing of points on the surface.
Equations can also be derived to fit parameters appearing in equations
for dose rate vs. material thickness. Discussions in Sec.4.8.1 and the
following section on the buildup-factor approach to the moments-method
gamma-ray data present examples of such equations.
6.4 SIMPLIFIED SOLUTIONS

It is natural that attempts should be made to simplify the solution to
problems as complex as those associated with radiation transport. It is also
inevitable that most simplifications result in some sacrifice in the number of
parameters or in the precision associated with the solution.
Simplifications of radiation-transport analyses can generally be classified
as those resulting from (1) analysis of detailed analytical or experimental
data which reveals a simpler means of obtaining some gross quantity, such as
dose rate or total flux density, or (2) attempts to simplify the mathematical
models of radiation-transport physics and thereby obtain (in an easier
manner) data which are a little less precise but which are possibly as detailed
as the more complex solutions. The use of gamma-ray-dose buildup factors
and removal theory fall into the first category; the other methods are
broadly classed in the second category.
284
6.4.1 Applications of Gamma-Ray Buildup Factors

The definitions and theory of exposure and energy-deposition buildup
factors were presented in Sec. 4.8.1. In this section we describe some typical
applications.
For a thin shield and narrow beam (Fig. 6.11), the buildup factor is
unity, B = 1, because scattering events deflect the gamma rays out of the
_ _ 1r
Sorce
I
I
ILI
.
~
. l~l
DoT'
I
I
...JII
Fig. 6.11-Narrow-beam geometry.
beam and thus away from the detector. Where t is small and the detector size
can be neglected, the gamma-ray-flux density at the detector is given by
(6.4-1)
where s is the source strength, M(E) is the linear attenuation coefficient
(em-I), and t is the shield thickness (em).
For the geometry of Fig. 6.12, as the thickness of the shield is increased
so that multiple scatters become more prevalent or as the width of the beam
_.Source
Detector
Fig. 6.12-Wide-beam geometry.
285
is increased so that single scatters may reach the detector. the buildup factor
increases to values greater than 1. The value of B continues to increase with
increasing beam width or shield thickness. A saturation point is reached
when the size of the beam forms a 2'lT solid-angle intercept at the detector.
The gamma-ray-flux density at the detector in this case is
 = _ $-
4'lTR2
B(E Bt) e-IJ.(Eo)t

0 ....
(6.4-2)
Note that the value of B is dependent on the angular distribution of the

source incident on the shield. the shield thickness. and the energy of the
gamma rays. We should also recognize that this buildup factor applies only
to the photon-flux density. As mentioned previously. the concept of buildup
can be applied to any detector response. but. for a given set of conditions.
the dose buildup factor. flux-density buildup factor, or energy-flux-density
buildup factor are not necessarily numerically equal. To be generally useful
therefore. buildup factors must be identified as to source-shield geometry.
source energy, shield thickness. shield material, and detector response. An
example is the dose buildup factor for an infinite-plane collimated source of
2-MeV gamma rays shielded by an infinite-plane slab oflead four mean free
paths thick.
As an example of the applicarion of buildup factors, consider the
penetration of gamma rays from an infinite-plane source (Fig. 6.13). The
source is assumed to emit monoenergetic gamma rays isotropically with a
source strength of So photons cm- 2 sec-! .
The dose rate at the detector is given by
r-
-J1R
D z = So F(E))o :'lTR 2 B(E,IJ.R) 2'lTrdr
(6.4-3)
Source
Tr
Detector
J.. -1--------==",.-z
Fig. 6.13-Buildup from infinite-plane source.
286
where r is the radius along the source plane to the source element, z is the
distance of the detector from the source plane, and F(E) is the flux-densityto-dose-rate conversion function. By changing the variable from r to Rand
replacing B(E,JJR) with the Berger formula, B(E,J,lR) = 1 + C(E) J,lReD(E)p.R
(see Sec. 4.8.1), the expression becomes
D
= So F(E)
2
f-ze_p.R (1 + Cp.e
R Dp.R) dR
R
(6.4-4)
Introducing the exponential integral

(6.4-5)
the expression becomes
D = 5 0 1'(E)
z
2
[E-1I1Z
( )+ ~
_(I-D)P.Z]
I_De
(6.4-6)
Graphs of the function En(x) for n = 1, 2, and 3 are available in

Appendix G.
6.4.2 Applications of Neutron-Removal-Theory Kernels
In the discussion of neutron-removal-theory kernels (Sec. 4.8.2), the
Albert-welton kernel' 2 was discussed as one method of handling neutron

attenuation in hydrogenous shields. The Albert-welton kernel is especially
useful in the following two applications:
1. This kernel can be used to correct measured or calculated data when
small changes are made in the heavy clements of a shield. For example,
suppose that a lead layer and a water layer surround a point source. If the
lead layer is increased and the water thickness remains the same, the new
dose rate is given by
(6.4-7)
where ~ R = removal cross section for lead
t = change in lead thickness = r, + t
T2 == new distance from source
D(r2) = new dose rate
r, = original distance
D, (r,) = original dose rate
287
Notice that the assumption is made that the water thickness (and its
effect) remains unchanged. Consequently the dose rates are evaluated at
different positions.
2. The kernel can be used to correct results obtained for one hydrogenous medium so that they apply for another hydrogenous medium. The
assumption is that the hydrogen effect remains constant for a given
hydrogen length, with the effects of other elements accounted on the basis
of removal cross sections. Thus the hydrogen attenuation kernel in one
medium is set equal to the hydrogen attenuation in the other, which gives
4'/fd D(r,) e~R, '2 = 4'/frf D, (r, ) e~R, "
(6.4-8)
with the constraint, to ensure the equivalence of the hydrogen effect, of

(6.4-9)
where P, = hydrogen density in medium for which V(r,) is unknown
PI = hydrogen density in reference medium for which V, (r,) is
known
~R 2
=removal
cross section for all elements except hydrogen

medium being analyzed
In
the
~ R, = removal cross section for all elements except hydrogen in the

reference medium
Combining the preceding equations yields
(6.4-10)
A word of caution is appropriate here. The preceding equations represent a
simple model of rather complex phenomena, and rather large errors are
possible.
Although the differences are not great, values of the removal cross
section measured in homogeneous-hydrogenous mixtures are slightly lower
than those measured in slabs preceding hydrogenous materials. The value
would be expected to vary slightly with the ratio of atomic densities of
hydrogen to the material being measured. Table 6.1 shows the extent of
these variations in some particular measurements given in the shielding
volume of the Reactor Handbook. 13
288

Table 6.1-COMPARISONS OF ~R AND ~DR
Measured in homogeneous mixture
~R measured in slab
Element
preceding Hz 0
~DR
Carbon
0.81 0.05
0.72 0.05
0.7 - 0.8
Lithium
Oxygen
1.01 0.05
0.99 0.01
0.9
0.92 0.05
Mixture
C12H22011 - H 2 0
CH and CH 2 (momentsmethod calculation)
LiH
Oil and water
In both neutron and gamma-ray calculations in all types of media,

estimates of the approximate change in detector response with material
thickness can occasionally be made with an effective relaxation length, AeH,
for the material. The effective relaxation length is defined as the material
thickness required to reduce the response by a factor of e- 1 , or
R _ -'/Adf
-e
Ro
(6.4-11)
neglecting any differences in geometric attenuation. The value of AeH is

determined from an examination of basic experimental or analytical data.
The selection of a value for Aeff is fraught with hazards because it may be a
function of many variables, such as distribution in energy, angle, and
position of the source and the total thickness of the shielding material. Also,
slight differences may be observed in the response as a function of thickness
or in the response as a function of depth due to spectral shift with increasing
penetration. Provided that data can be obtained under analogous circumstances, use of an effective relaxation length can provide a quick and
inexpensive estimate.
6.4.3 Other Point-Kernel Applications
The point-kernel method has been discussed in preceding sections, but
not all applications of the method have been covered. It should be
emphasized that this method is not a means of obtaining basic penetration
data but is rather a widely used means of applying these data. This method
can be used in combination with appropriate point-source, infinite-medium
data to determine either dose rate or spectrum at a point due to a distributed
source. It can also be used in combination with total-cross-section data to
obtain the uncollided component at a point due to a distributed source.
289
Fig. 6.14-Cylindrical point-kernel model.
Other applications are distinguished by the nature of the source being

analyzed. In the last-collision method (to be discussed at the end of this
section), the source is proportional to the local scattering density. In the
analysis of secondary radiation, the source is proportional to the local
density of radiative interactions.
(a) Numerical Techniques. The major steps in performing a numerical
integration over the source volume are: (1) divide the source into increments
that may be considered as point sources, (2) determine the source strength
to be associated with each increment, (3) determine the attenuation along
each source-detector ray, and (4) sum the contributions from individual
sources to obtain a total detector response.
These tasks are considered in order in the following discussion. As with
all forms of numerical integration, the number and size of the increments
selected and the means of averaging within an increment are critical to the
accuracy that can be obtained. Consider the three detector positions near the
cylindrical source shown in Fig. 6.14. Assuming uniform distribution of the
source strength throughout the cylinder, the nearest region will account for
most of the response at detector D 1 Consequently nearby increments
should be smaller in size than those at a greater distance so that the averaging
in this region is at least as precise as that for more distant points. The
increment size selected should be such that the distance from any point in
the increment to the detector would be a very weak function of position in
the increment. In addition, the material attenuation across the increment
should be small. Under these conditions the midpoint of the volume element
can be used as a representative average of all source points within the
increment. The maximum acceptable increment size is therefore influenced
both by the distance to the detector and by the relaxation length of the
source region. For detectors at great distances from the source, the
increment size is determined primarily by material relaxation length. For
290
nearby or internal detectors, particular caution should be exercised in laying

out local increments lest a source point be so nearly coincident with the
detector that its effect overshadows all others. t Other factors that will
influence increment layout include source-density variations and shield
inhomogeneities. The estimated relative importance of a given source region
to total detector response should determine the precision in layout of the
region.
The source strength to be associated with each of the assumed central
point sources is given by the integral of source density over the increment or,
if the source density is fairly uniform, by the product of source density
times volume associated with the increment:
n;
"v
p dv ; 7J ~v .
(6.4-12)
The attenuation along each source-detector ray is generally determined

from a set of basic penetration information. The material thickness along the
ray (g em -2) is the criterion used in normalizing to the basic data. The effect
of a material mixture or laminate along the ray is assumed to be the same as
that of each of the component materials taken individually and in sequence
with no perturbation due to interface effects.
Any of the various types of attenuation data that have been discussed in
Sees. 4.8 and 6.3 may be used in this manner. The choice depends on
whether spectrum or dose response is required at the detector and also on
the availability of data for the elements involved. Depending on the type of
attenuation data used, the summation of the contributions from the
individual source points may be as simple as adding dose-rate values or as
complicated as adding contributions to individual energy and angular
increments.
(b) Attenuation Kernels from Monte Carlo Calculations. Results from
Monte Carlo calculations (or from other transport methods) for dose-rate
transmission through slab shields with an incident beam of neutrons can be
quite useful when expressed in terms of attenuation kernels, i.e., as plots of
transmission factors or dose-rate attenuation as a function of slab thickness.
With such graphs it is a simple matter to estimate the fraction of the dose
rate that is transmitted for each incident-energy group of neutrons, the total
dose rate being the sum of the dose rates from all energy groups.
t 1/41Tr Z gets very large as r gets very small.
SHIELD ATIENUATION CALCULATIONS
291
Clark, Betz, and Brown [4 performed Monte Carlo calculations for

monoenergetic beams of neutrons normally incident on slabs of ordinary
concrete and also on a semi-infinite medium (half-space) of concrete. The
neutron energies were 0.7, 1.3, 2, 3, 4, 6, 8, 10, 12, and 14 MeV. The
density of the concrete was assumed to be 2.43 g cm- 3 , and its composition,
other than its water content, was representative of that given for ordinary
concrete 01 in Report ANL-5800 (Ref. 15). The resulting curves for dose
attenuation are shown in Appendix M (Figs. M-1 through M-10). In addition
to being useful directly, these results can be helpful in adjusting the large
body of infinite-medium neutron-attenuation data for concrete (for example, Trubey and Emmett' 6) to finite layers. After one or two relaxation
lengths, the penetrating characteristics of neutrons in an infinite medium of
concrete should differ very little from those of neutrons in a semi-infinite
medium. Therefore data such as is shown in Figs. M-1 to M-10, which in all
other respects appear to be appropriate for application to a particular
situation, can be adapted to a finite system by correcting the data in
proportion to the ratio of the curve for the semi-infinite medium (dashed
curve) to the curve for the slab configuration (solid curve) at the proper
penetration distance and energy.
Other useful results were obtained by Allen and Futterer,l 7 who
determined the attenuation of the multicollision dose rate in the materials
listed in Table M-l due to monoenergetic neutron beams incident at various
angles. The neutron energies used were 5, 3, 2, 1, and 0.5 MeV, and the
results are plotted in Figs. M-11 through M-15. In order to use these curves,
we must know the multicollision dose rate at the inner surface of a slab of
one of these materials due to neutrons incident in a broad beam at an angle
(or angle band) and energy (or energy band) close to the angle and energy
for which the attenuation data are given. The attenuation factor appropriate
to the material, thickness, energy, and angle is read from the curve, and the
incident dose rate multiplied by that factor should approximate the dose
rate that has penetrated the slab.
(c) Formulas Derived from Kernel Integrations. Transformations of
source geometries with kernel integrations are elucidated by Blizard,
Foderaro, Goussev, and Kovalev.' s They have derived equations to relate
dose rate (or other nondirectional radiation quantities) from one type of
source geometry to that of another. For example, the dose rate from a finite
disk source may be related to either that from a point source or that from an
infinite-plane source. A cylindrical source may be shown equivalent to a line
source buried within the cylinder. Additionally, the attenuation of various
292
sources in passing through slab shields can often be expressed in terms of

srandard integral equations whose values are well known as a function of the
parameters involved. The formulas derived in this manner are useful in the
evaluation of rhe recurring regular geometry problems that may require only
a preliminary estimate of shield effectiveness. The exponential integral
funcrions (Eq. 6.4-5) are involved in the formulas and are given for reference
in Appendix G. The function
F(O,p.t) = foB e- Mt sec 0' dO'
(6.4-13)
is known as Sievert's integral and is occasionally involved. Numerical values

(detailed plots) are given by Blizard et al. 13 1 B We will summarize some
useful cases using the following symbols for source intensities:
Volume sources (Sv)' particles cm- 3 sec-I
Surface sources (Sa)' particles cm-' sec-I
Line sources (51), particles cm-, sec-I
Point sources (So), particles sec-I
Shield penetration depth is denoted by (p.t) as though only one material
were involved. Where there is more than one shield material, the symbol
should be interpreted as :E p.jt j .
Basic point-source kernels are
(6.4-14)
for the direct-beam (uncollided) component of an unshielded source,
(6.4-15)
for the direct-beam component with an intervening absorbing shield, and
e-M t
cI> = So B(p.t) 47TR'
(6.4-16)
for the total (direct and scattered) component with an intervening absorbing
and scattering shield where B(p.t) is the buildup factor for shield thickness
(p.t) at the source energy.
293
An element dl in an unshielded line source (Fig. 6.15) gives
d<l> =
S/ dl
41T(a sec 0)2
(6.4-17)
but, since
dl = a sec 2 0 dO
(6.4-18)
S/
<1>=-(0, +( 2 )
41Ta
(6.4-19)
T--I
I
I
L
I
I
I
Fig. 6.15-Line source.
L __
If a shield is placed between the line source and the detector, the uncollided
flux density is given by
S dl e-lJ.t sec 0
d<l> = -4
7/ 1T-(:-a-s-e-c-;;07')2'---
= S/
(6.4-20)
-J.1t sec ()
1Ta
dO
(6.4-21 )
(6.4-22)
S/
=-4 [F(0"ll t )+F(02"ut)]

1Ta
(6.4-23)
In like manner the integral over a disk source (Fig. 6.16) can be
evaluated in terms of the exponential functions, En(x).
294
Fig. 6.16-Disk source.
At point P on the axis of the disk, the uncollided flux-density

contribution of the source increment in a differential ring at radius r from
the center is given by
d<l> =
Sa(2ltr dr) -Ilt sec 8'

4ltp 2 e
(6.4-24)
Substituting p2 = r 2 + a2 , p dp = r dr, and sec O' = pia
Sa dp -Ilt pia
d<l> = - - e
2 p
(6.4-25 )
or, if p.tp/a '" t,
Sa
 = 2
jJt
~t
sec 8
e- t
-dt
t
(6.4-26)
(6.4-27)
Similar although more complex expressions can be derived for volumedistributed sources, such as a cylinder or an infinite slab.' 8
(d) Special Applications. A first estimate of the scattered radiation can
be obtained by performing a kernel integration over the density of first
collisions within a medium. Consider Fig. 6.17 for a point source and slab
shield.
295
SOL.._.J....;.
+-__+
__
..::L_~D
Fig. 6.17-Single-scattering geometry.
The uncollided flux density at dV is
(6.4-28)
where ~t is the total cross section.
The density of first collisions within dv is
(6.4-29)
and the flux density at D due to single scatters within dv is
(6.4-30)
where Ls(lI) is the value of the angular-distribution function (macroscopic
differential-scattering cross section) for scattering at angle II,
~s(lI) = N us(lJ). Integrating Eq.6.4-30 over the entire scattering volume
296
yields the singly scattered flux density at point D. This method gives a
reasonable estimate of the total scattered flux density in cases where the
scattering medium is thin with respect to the relaxation length of the
radiation, i.e., where there is a small probability of multiple scattering within
the medium. In such cases, especially for problems of air scattering, it is
common practice to neglect the attenuation within the scattering medium of
both the unscattered and the singly scattered radiation, which approximately
corrects for the buildup of the multiply scattered component.
The last-collision method 19 can be used in estimating the angular
distribution of the radiation penetrating a shield. Prior knowledge of the
approximate flux density vs. shield thickness must be available from other
sources, however. The approach is similar to that for the single-scattering
model except that a total flux density (rather than an uncollided flux
density) is used in determining the density of scatters. The total-fluxdensity-vs.-penetration data are taken from any of the basic sources
discussed in Sec. 6.3. In this method we assume that the total flux density
within the slab is collimated in the direction of the incident beam and is of
the same energy as the source. For a broad-beam flux density <1>0 normally
incident on a slab (Fig. 6.18), where the flux density vs. penetration depth
---
</>0-
Fig. 6.18-Lastcollision geometry for a slab.
can be expressed in terms of an exponential function utilizing the removal

cross section, exp [-~ R (Eo) t] , the local scattering density is given by
(6.4-31)
If we assume that the energy after scatter, E, is given by an elastic scatter of
particles at the initial energy and that the cross section for scattering into a
unit solid angle about the direction of the detector is ~s(Eo,II), then the flux
density at the detector due to scattering in dv is given by
297
(6.4-32)
The term ~s(Eo ,II) is the differential scattering cross section. Equation 6.4-32 can be integrated over that portion of the shield lying within
solid-angle increments as seen by the detector. A differential angular
distribution at the detector can be obtained by expressing dcf> on a per-unit
solid-angle basis. This is done by dividing by the differential solid angle
(21T sin 0 dO) and integrating along r.
The last-collision method can be applied to any geometry using
numerical techniques, and in certain simple geometries the integration can be
carried out analytically. In the slab-geometry broad-beam normal-incidence
case, the flux density per steradian incident at angle 0 per unit incident flux
density at energy Eo is
~s(Eo,O)
1'(0'0) =
~t(E)
exp [-~R(Eo) tmJ

- ~R(Eo) cos 0
x(l-exp{-secO[~t(E)-~R(Eo)cosO]tm}) (6.4-33)
where t m is the slab thickness. Evaluation of this equation at representative

angles provides an estimate of the differential angular distribution at the
detector. Where a spectrum of radiation is incident, an evaluation can be
made at each angle for each incident-energy group. The energy spectrum at
the detector is estimated by logging the contributions from the particular
incident-energy-scattering-angle combinations into appropriate energy bins
at the detector. This approach ignores the spectral shift likely to be
encountered as the radiation penetrates the shield. However, energy loss due
to the last collision is included.
Consider another geometry: a detector on the axis of a cylindrical shield
as shown in Fig. 6.19. Broad-beam incidence is assumed with all particles
incident at angle ~ with respect to a plane through the axis of the cylinder.
In this case the scattering angle is 0 and the reference angle at the detector is
lx. The integral equation for scattered flux density per steradian per incident
neutron at energy Eo and direction ~ is
298
INCIDENT
NEUTRONS,
"
Fig. 6.19-Last collision for cylindrical geometry.
+ ~t(E) t z J}d4> dT
where t, = [(T~ - TZ sin z 4~
t z = (T - TI) csc '"
4>max =
(6.4-34)
- T cos 4>] csc 13
~+ cos-' (TI /T Z )
In stating Eqs. 6.4-33 and 6.4-34, the flux density vs. penetration depth
was given in terms of an exponential function. The method can also be
applied with a basic set of spectrum-vs.-depth data, although the integrations
could no longer be carried out analytically but would require a numerical
solution. French l9 demonstrates that the simplified approach (ignoring
spectral shift) gives results in reasonable agreement with experiment and
Monte Carlo analyses for LiH shields of about 10- and 20-cm thickness.
6.4.4 Methods for Estimating Low-Energy Neutron-Flux Density
Estimating the thermal-neutron-flux density as a function of position
inside the shield presents a dilemma. The diffusion methods for analyzing
SHIELD ATIENUATION CALCULATIONS
299
neutrons are generally inadequate for predicting the transport of higher

energy components, which are the source of thermal neutrons deep within
the shield. Likewise, some of the transport methods are extremely inefficient
for obtaining thermal-flux density because of the many decades of energy
degradation that must be considered. Various combinations of transport and
diffusion methods have been used to resolve this dilemma. The assumption is
made that ar some energy around 1 ke V the neutron angular distribution
changes from forward-peaked, which is best treated by transport theory, to
nearly isotropic, which is characteristic of diffusion. One of the earliest
corrections to diffusion theory was the first-flight correction to the age in
Fermi age theory.2 0 The correction is applied because the neutron cannot
enter a process described by continuous slowing down (as Fermi age theory
requires) until it has collided at least the first time.
A calculational model formulated by Blizard 2 I equates the slowingdown source to the rate of removal from the fast beam. The resulting
equation for the eq uilibrium thermal-neutron-flux density (ignoring diffusion at thermal) is
(6.4-35)
where <l>th (z) =thermal-neutron-flux density

](z) = fast-neutron current density
X =relaxation length of ](z)
:I; a ;;;: thermaI-neutron-absorption cross section
T = average age of neutrons from removal to thermal
Obvious limitations of this model are the need for empirically determined
values of ](z), X, and T. These values are dependent on penetration depth,
material, and geometry; thus values from one measurement do not
necessarily apply to another set of conditions.
The development of more sophisticated multigroup diffusion models
made it possible to obtain reasonably good estimates of the thermal-to-fastneutron ratio in a material, even though the penetration of the fast
component could not be treated in this way. Haffner 22 used such an
approach and normalized the results to the fast-neutron distribution given by
the Albert-Welton kernel. In this approach the corrected thermal-neutron
flux is given by
300
f Kd E ) (E,R) dE
 t h
=D (R) f'---K---'-,('---E'---)~'---(E-,R--')-dE-:-
(6.4-36)
where K 1 (E) = response of detector used to measure thermal-neutron-flux

density
K, (E) = response of a fast-neutron dosimeter
D(R) = Albert-welton kernel dose rate at position R
(E,R) = flux density computed by the multigroup diffusion code
Good results were obtained with this method, but it is limited to cases where
a reliable D(R) kernel is available.
A refinement of the preceding method is to perform a multigroup
diffusion calculation using as a source rhe neutrons recoiling from first
collisions within the shield. Although this method has been used with some
success by Schreiber and Kodras' 3 in predicting thermal-neutron distributions in beryllium oxide, difficulties are encountered in hydrogenous media
owing to neutrons which have suffered one or more small-angle collisions but
which have retained their directional characrer. Spinney" used a removal
kernel rather than an uncollided kernel for the diffusion source to solve this
difficulty. The removal concept was applied to individual source energies in
order to approximate a transport kernel for individual energy groups which
accounted for both the uncollided and the small-angle scattered components.
Spinney first used age theory to predict slowing down of a removed neutron
and later used the concept in multigroup diffusion calculations.' 5 This
combination of removal and diffusion theory has come to be known as the
Spinney method, the theory of which is discussed in Sec. 4.9.
Trubey' 6 proposed a direct coupling of a multigroup diffusion code to
the results of a transport code such as RENUPAK (moments method),
NIOBE (numerical integration of the Boltzmann equation), or Monte Carlo
codes. In a trial calculation of the transfusion (transport diffusion) method,
Trubey chose the RENUPAK transport code for fast-neutron calculations
and the MODRIC diffusion code for the analysis of neutrons below 1 keY.
MODRIC is a one-<limensional multigroup diffusion code written for the
IBM 7090 computer. With RENUPAK, an energy spectrum of neutrons was
calculated at various spatial points between 0 and 160 em. These values were
then interpolated to provide spectra at 226 spatial points. Two separate
determinations were made with lower energy limits of 1 and 10 keV to test
sensitivity. The spectral values were written on magnetic tape and used as a
source distribution for the low-energy portion of the transfusion calculation.
301
Variation of the cutoff energy from 1 to 10 keV affected the thermal-flux

density by only 1 to 5%. Figure 6.20 compares the thermal-neutron-flux
density calculated by Trubey using the transfusion method with measurements in water at 'the Lid Tank Shielding Facility. The experimental data
have been converted to an equivalent point source, and all data have been
normalized to unit-source strength. The transfusion method provides a much
closer fit to the measurements than does the application of diffusion theory
alone.
6.5 APPLICATION OF KERNEL TECHNIQUE TO CALCULATIONS

OF SECONDARY GAMMA-RAY DOSEt
Often a large fraction of the radiation dose rate behind reactor and
shelter shields is the gamma-ray dose rate due to neutron capture and, in
some instances, to inelastic scattering within the shield. If the spatial
distribution of the neutron-flux density is known, the gamma-ray dose rate
can be calculated for a large number of configurations by integrating the
dose kernel over the source volume. With the kernel technique as
exemplified by Eq. 4.8-4 and the slab geometry shown in Fig. 6.21, the dose
rate, qt,a,b), on the shield surface due to a distributed monoenergetic
isotropic gamma-ray source, S(x), bounded by planes at a and b is given by
qt,a,b) = G(E)
b
a
S(x) dx
fooo
0
e-IlR
2 21rp dp
1rR
B,(IJ.R) 4
(6.5-1)
where x = one-dimensional spatial coordinate measured from the reference

plane
t = shield thickness
p = radial distance to source point measured from the detector axis
R = distance from the source point to the detector
IJ. = total macroscopic cross section for gamma rays of source energy E
G(E) = flux density to dose rate in tissue conversion factor, which
for conversion to rads/hr is 5.767 x 10-5 [lJ.a(E)/p]E, where Il a /p
is the mass energy absorption coefficient for tissue
B,(IJ.R) = dose buildup factor for gamma rays of energy E
tThis section by P. A. Stevens and D. K. Trubey is adapted from Methods for Calculating Neutron
and Gamma Ray Attenuation, in Weapons Radiation Shielding Handbook, Chap. 3, USAEC Report
DASA-18923, Oak Ridge National Laboratory, 1968.
302

10-3 " , - - - - , - , - - - , - , - - - - , - , - - , - , - . , - , . . . . . , . . . - , - , - - - - - , , - ,
10....
e
s
c
10-6
Measured data transformed

to point source
10- 13
10-1. L---l.._.L-l._.L.-.J._..L....J_..L_L--L_L---L_.L-.J._..L....J
o
20
40
60
80
100
120
140
160
DISTANCE FROM SOURCE, em
Fig. 6.20-Comparison of measured and calculated thermal-neutron-flux densities as a

function of distance from the source. (From Trubey.2 6 )
303
I---~_-z--~
DETECTOR
a
b
Fig. 6.21-Geometry for integration over exponential source distribution.
Since R 2 = p2 + z2 ,
(6.5-2)
The gamma-ray source term usually can be represented quite well either
by fitting with several terms or by piecewise fitting of the thermal-neutronflux density distribution t (or of the fast-flux density distribution if inelastic
scattering is being considered) with a function of the form
S(x) = Sae
J.x
(6.5-3)
tThe production of secondary gamma rays by the capture of nonthermal neutrons is usually
insignificant for thermal reactors. It can be important in fast reactors and in thermal reactors where
absorbing layers deplete the thermal flux.
304
where Sa is the gamma-ray source at x = 0, which gives S(x) in the interval

(a,b), and k is the reciprocal of the effective neutron relaxation length.
Usually Sa can be calculated by
(6.5-4)
where y is the number of photons of energy E released per neutron capture
(or per inelastic scattering), (a) is the neutron-flux density (usually
thermal-flux density for capture and fast-flux density for inelastic scattering), and L is the macroscopic neutron cross section for thermal-neutron
capture (or for inelastic scattering).
When exponential or polynomial forms of the buildup factor are used
(see Sec. 4.8.1) together with the source description given by Eq. 6.5-3, then
Eq.6.5-2 can be integrated analytically, and very useful results can be
obtained. In the following paragraphs examples of such integrations are given
for two cases of interest: a slab shield of finite thickness t and a semi-infinite
shield (t =00), the latter corresponding to a real problem in which the shield
is very thick.
6.5.1 Calculation for Slab shield
Trubey' 7 calculated the secondary gamma-ray dose rate for a slab shield
using the Berger form of the buildup factor
(6.5-5 )
used in Eq. 6.5-2. The equation then becomes
G(E)
f(t,a,b) = -2- Sa
e- x dx
[f-
J.l(t-x)
f-
-IlR
~
d(ilR)
J1
~(t_x)
C(E) ell (D_I)R d(il R )]
(6.5-6)
where the uncollided dose rate, ro(t,a,b), is represented by the first term
and the scattered.dose rate, rs(t,a,b), is given by the second term.
Letting il(t - x) = y and integrating the first term of Eq. 6.5-6 by parts,
we find the dose rate from the uncollided gamma-ray dose rate is given by
305
ro(l,a,b) =
G(E) S ea
Qilt
-y
(eQY Edy)
11 (t-bi
20111
IIl(t-a~ + f.~(t-a) eay~ dy)

~(t-b)
G(E) Sae-Qllt{
eQIl(t-a) Edll(1 - a)) - Ed(1 - Dl)Il(1 - a)]
2Dlll
(6.5-7)
+E 1 [(1- Dl)Il(t - b)J - eQIl(t-b) E 1 [1l(1 - b))}
where 01 = kill and E 1 is an exponential function of the first order and is

defined by
Ed x ) =
oo
e-Y
-dy
x
(6.5-8)
Appendix G contains graphs and other details of the exponential functions.

If Ol = 1 or 0 (case of uniform source distribution) or if b = I,
indeterminate forms result which may be resolved by I'Hospital's rule, by
series expansions, or by integrating Eq.6.5-6 for k = Il, k = 0, and b = I,
respectively. These cases are as follows:
For b < I and
ro(l,a,b) =
01
= 0
G(E) Sa {
21l
Il(b-a) E.IIl(I-b))
+ III Edll(l-a)J -Ill Edll(l-b)J}
For b < I and
For b
01
(6.5-9)
= 1
=I and 01 ". 0 or 1
ro(l,a,l) =
G(E) Sa ( -Qila .
2
IlOi
E, [1l(I-a))
- e-
Olilt
{E 1 [1l(1-0l)(I-a)] + In 11 -
Oil})
(6.5-11)
306
For b
= t and a = 0
ro(t,a,t)
G(E) Sa {
1l(1 a)}
1 + p(t-a) Edp(t-a)] - e- 2p
(6.5-12)
G(E) Sa { Ila
Ill}
2p
e- E 1 [p(t-a)] - e- In 'YP(t-a)
(6.5-13)
For b = t and a = 1
ro(t,a,t) =
where 'Y = 0.577215665..., Euler's constant.

For the special case of b = t and a = 0, Eq. 6.5-11 can be represented by
(6.5-14)
where
1/Jo(pt,a) = -..L (E 1 (pt) - e2a
CXIlI
{E 1 [(1 - a)pt] + In 11 -
al})
(6.5-15)
Equation 6.5-15 is plotted in Fig. N.1 as a function of the number of

mean-free-paths pt with a as a parameter. [Equations 6.5-9 through 6.5-15
can be used to calculate the total gamma-ray dose (uncollided + scattered)
when the Taylor form of the buildup factor is used (see Sec. 4.8.1).]
Since z = t - x, the scattered dose rate behind a slab shield can be
determined by expressing the second term of Eq. 6.5-6 (i.e., the Berger term)
as
_ G(E) C(E) Sae-CXlllj t_a
rs(t,a,b) 2
t-b
(6.5-16)
Integrating Eq. 6.5-16 gives
G(E) erE) S e -CXIl I

a
[_(l_D_CX)Il(/_b)
-(l-D-CX)Il(I-a)]
s(t,a, -2(1-D)(1-D-a)p e
-e
b-
(6.5-17)
307
0.25
0.20
0.15
0.10
0.05
0
0
Sn
-0.05
Pb
-0.10
-0.15
0.2
0.5
20
'0
PHOTON ENERGY, MeV
Fig. 6.22- The coefficient D for the Berger form of the gamma-ray-dose buildup factor.
(Based 00 data from Chiltoo. 2 8)
Equations 6.5-16 and 6.5-17 reveal that, unless D is less than 1, negative dose
rates are obtained. However, D is always significantly less than 1, as is shown
in Fig. 6.22 from the work ofChiiton. 28
When", + D = 1, Eq. 6.5-17 gives an indeterminate form, which, when
resolved, becomes
rs(t,a,b) =
G(E) C(E) Sa e-cxjJ.t (b - a)

2(1 _ D)
(6.5-18)
For the special case when b = t and a = 0, Eq. 6.5-18 can be expressed as
G(E) C(E) Sa eDjJ.t
rs(t,O,t) =
11(1 _ D)
'iJ, (Il t ,"')
(6.5-19)
where
(6.5-20)
and "" = '" + D. The function given by Eq. 6.5-20 is shown in Fig. N.2.
308
6.5.2 Calculation for Semi-Infinite shield

Solutions of Eq. 6.5-2 for a semi-infinite shield, i.e., for b = 00, give useful
results that are generally applicable for the special case in which a = 0,
particularly if one is interested in a gamma-ray heating rate within a shield.
Using the polynomial form of the buildup factor
3
Br(/.IR) =
1:
n=O
(6.5-21)
An (/.IR)n
Claiborne 29 determined solutions to Eq. 6.5-2 for this case which were all in
the form
(6.5-22)
The dose rate from the uncollided-flux density is represented by the first
term and is given by
(6.5-23)
and the sum of the next three terms represents the scattered contribution.
The terms are
(6.5-24)
(6.5-25)
1]
Ilt
(/.It)2
/.It
e+
+
- [ 2(I-a) (l_a)2 (l_a)3
6
=-",-'-----,
(6.5-26)
When a = 0, an indeterminate form occurs in Eq. 6.5-23, and, when

a = 1, indeterminate forms occur in Eqs.6.5-23 through 6.5-26. The
following equations result when the indeterminate forms are evaluated:
309
For a = 0
(6.5-27)
Fora= 1
(6.5-28)
(6.5-29)
(6.5-30)
(6.5-31)
The functions given by Eqs. 6.5-28 through 6.5-31 for the semi-infinite
medium are plotted in Figs. N.3 and NA for various values of a. These
figures can be compared with Figs. N.1 and N.2, which are the corresponding
functions evaluated for a slab shield.
These solutions contain the contribution from the gamma-ray sources
between the detector position at t and infinity since integration of Eq. 6.5-2
from x = 0 to x = "" produces two integrals: one gives the contribution from
the interval 0 <;; x <;; t and the other gives the contribution from the interval
t <;; x <;; "". In the usual shield, however, the contribution from the second
interval at deep penetrations is small, and the gamma-ray dose rate outside a
shield of thickness Ilt will be only slightly less than that calculated for a
distance Ilt within a semi-infinite shield.
If Eq. 6.5-22 is used for gamma-ray heating calculations within a shield,
the coefficients An must correspond to the polynomial fit of the
energy-absorption buildup factor, and the conversion factor for expressing
the heating rate (in watts g-' ) becomes
a
G(E) = 1.6 x 10. ,3 Il E

p
(6.5-32)
where Ilal p is the mass energy absorption coefficient of the material in which
heat is generated.
310
REFERENCES
Company, Inc., Reading, Mass., 1959 (out of print); now available from Johnson
Reprint Corporation, New York.
2. A. D. Krumbein and B. D. O'Reilly, Neutron Attenuation in Hydrogenous Media by
the Method afMaments, in Engineering Compendium on Radiation Shielding, Vol. 1,
Sec. 3.3.1, R. G. Jaeger (Ed.), Springer-Verlag, New York, 1968.
3. H. Goldstein and J. E. Wilkins, Jr., calculations of the Penetrations of Gamma Rays,
USAEC Report NYO-3075 (NDA-15C-41), Nuclear Development Associates, Inc.,
June 1954.
4. F. J. Allen, A. Futterer, and W. Wright, Neutron Reflection and Flux Versus Depth
for Water, Report BRL-1204, Ballistic Research Laboratories, June 1963.
5. A. W. Casper, Modified Fast Neutron Attenuation Functions, USAEC Report
XDC-60-2-76, General Electric Company, Feb. 3, 1960.
6. G. T. Western, Energy and Angular Distribution Experiment, Angular Distribution of
Reactor Radiation from Slabs and of Emergent Secondary Gamma Rays, Report
NARF-62-16T, Volume I, General Dynamics, Dec. 31, 1962.
7. F. J. Muckenthaler, L. B. Holland, and R. E. Maerker, In-Air Measurements in the
Viciniry of the Tower Shielding Reactor II, USAEC Report ORNL-3288, Oak Ridge
National Laboratory, April 1963.
8. W. E. Selph, Experimental Determination of Neutron Attenuation in Zirconium
Hydride, Trans. Amer. Nucl. Soc., 7(1): 42 Uune 1964).
9. B. S. Burton, Jr., Air Scattering of Co 60 Gamma Radiation, Report FZM-I016,
General Dynamics Corporation, October 1957.
10. J. T. Serduke and C. V. Smith, Penetration of Plane Normal Gamma Rays Through
Slabs of Aluminum and Steel, Report USNRDL-TR~730, Radiological Defense
Laboratory, Feb. 17,1964.
11. D. M. Peterson, Shield Penetration Programs C-17 and L-63, Report NARF61-39T,
General Dynamics Corporation, Dec. 29, 1961.
12. R. D. Albert and T. A. Welton, A Simplified Theory of Neutron Attenuation and Its
Application to Reactor Shield Design, USAEC Report WAPD-15, Westinghouse
Electric Corporation, Nov. IS, 1950.
14. F. H. Clark, N. A. Betz, and J. Brown, Monte Carlo Calculations of the Penetration of
Normally Incident Neutron Beams Through Concrete, USAEC Report ORNL-3926,
Oak Ridge National Laboratory, January 1967.
15. Argonne National Laboratory, Reactor Physics Constants, USAEC Report ANL-5800
(2nd ed.), July 1963.
16. D. K. Trubey and M. B. Emmett, Some Calculations of the Fast-Neutron Distribution
in Ordinary Concrete from Point and Plane Isotropic Fission Sources, USAEC Report
ORNL-RSIC-4, Oak Ridge National Laboratory, June 1965.
17. F. J. Allen and A. T. Futterer, Neutron Transmission Data, Nucleonics, 21(8): 120
(1963).
311
18. E. P. Blizard, A. Foderaro, N. G. Goussev, and E. E. Kovalev, Extended Radiation

Sources (Point Kernel Integrations), in Engineering Compendium on Radiation
Shielding, Vol. I, Chap. 6, R. G. Jaeger (Ed.), Springer-Verlag, New York, 1968.
19. R. L. French, A Last Collision Approach to Calculating the Angular Distribution of
Fast Neutrons Penetrating a Shield, Nucl. Sci. Eng., 23: 28-33 (1965).
20. A. M. Weinberg and E. P. Wigner. The Physical Theory of Neutron Chain Reactions,
The University of Chicago Press, Chicago, 1958.
21. E. P. Blizard, Nuclear Radiation Shielding, in Nuclear Engineering Handbook,
Chap.7-3, pp.7-89, H. Etheringtnn (Ed.), McGraw-Hill Book Company, Inc., New
York,1958.
22. J. W. Haffner, Use of Program G-2 for Shielding Calculations, USAEC Report
DC-59-7-170, General Electric Company, July 20, 1959.
23. P. W. Schreiber and F. D. Kodras, Measured and Calculated Radiation Levels Within
and Behind Beryllium Oxide, USAEC Report XDC-61-1-149, General Electric
Company, Dec. 15,1960.
24. K. T. Spinney, Neutron Attenuation in Concrete. British Report AERE-T/R-2507,
1957.
25. A. T. Avery, D. E. Bendall, J. Butler, and K. T. Spinney, Methods of Calculations for
Use in the Design of Shields for Power Reactors, British Report AERE-R-3216, 1960.
26. D. K. Trubey, Calculation of Fission-Source Thermal-Neutron Distribution in Water
by the Transfusion Method, USAEC Report ORNL-3487, Oak Ridge National
Laboratory, August 1964.
27. D. K. Trubey, A Survey of Empirical Functions Used to Fit Gamma-Ray Buildup
Factors, USAEC Report ORNL-RSIC-10, Oak Ridge National Laboratory, February
1966.
28. A. B. Chilton, Two-Parameter Formula for Point Source Buildup Factors, Nucleonics,
23(8): 119 (1965).
29. H. C. Claiborne, Analytical Solutions for Heat Generation Distributions in Regular
Geometries, in Engineering Compendium on Radiation Shielding, Vol. 1, Sec. 7.3,
J. G. Jaeger (Ed.), Springer-Verlag, New York, 1968.
EXERCISES
These exercises can be done with data given in this and previous chapters and the
appendixes.
6.1 Two isotropic fission point sources, A and B, are contained in a tank of water, as
shown by the plan view in the accompanying sketch. Sources A and B both emit 10. 0
neutrons/sec. (a) Estimate the spectra at points 1 and 2 at the edge of the tank. (b)
Estimate the total energy-flux density (MeV em-2 sec-) at points 1 and 2. (c) Make a
rough estimate of the thermal-flux density at points 1 and 2.
6.2 An ordinary concrete slab is to be used to shield a 15.25-cm-diameter beam tube from
a fission reactor. The fast-neutron-flux density at the face of the core is known to be
isotropic with a value of 10 12 neutrons em--2 sec-1 at the core face. Estimate the
thickness x required to reduce the neutron dose rate at point A to 1 mrad/hr. Neglect
scattering in the walls of the beam tube.
312
~42.5--1--42.5--l
Geometry of Exercise 6.1.
Core
, ------,52 c m ' - - - - - - _ +
1..
Beam-tube geometry, Exercise 6.2
6.3 In the beam-tube slab shield of Problem 6.2, estimate the thermal-flux density
emerging at point A.
6.4 A thin 60eo disk source 10 em in radius with a strength of 5000 Ci is shielded in a
large pool of water. Using the Berger form of the buildup factor: (a) Compute the
dose rate (DR) at a point 75 em from the disk along its axis. (b) Compute the buildup
factor for this situation. Note: At 1.25 MeV, (1l,/p)H.O=0.064 em'/g and
Ila /p = 0.03 em'/g.
6.5 It was decided to add 5 em of lead to the inside face of the concrete slab covering the
beam tube of Problem 6.2 to lower the gamma dose rate. How does this addition
affect the fast-neutron dose rate calculated in Problem 2? Use removal theory.
6.6 Approximately what was the prompt fission gamma-ray dose-rate reduction that
resulted from adding the 5 em of lead in Problem 6.5?
Albedos, Ducts,
and Voids
W. E.SELPH
We turn now to a number of subjects we might euphemistically call special

topics. They are special only in the sense that they are necessities. In
contrast to previous chapters that deal fundamentally with one topic,
transport in dense media, this chapter and Chap. 8 deal with a number of
topics equally important to shield design. We might call them engineering
considerations, perhaps another euphemism. At any rate some of the
following considerations will arise in any reactor-shield design.
L1 this chapter we are concerned exclusively with an interface between
two media, usually air and a shield material. The geometrical shape of the
interface may be an infinite plane, a wall of a room, a cylindrical tube with
several bends, or a spherical cavity.
Because the most successful methods for analysis of these in terfaces use
an albedo concept, the first half of the chapter is devoted to the definition
and description of albedos. The most important application of an albedo for
a reactor-shield designer is the analysis of ducts that penetrate the shield.
Accordingly, duct analysis occupies most of the last half of the chapter. A
brief final section covers a closely related topic, voids.
On the assumption that these topics are considered to be advanced in
shielding and on our opinion that they are still in a state of development, we
present this material in review form with a discussion of approaches taken by
several laboratories. Experimental results play an important role here and are
included in greater number than in earlier subjects for comparison with
analysis and to illustrate trends.
7.1 INTRODUCfION TO ALBEDOS
Backscattering of radiation incident to a dense medium has been treated
with some success as a reflection phenomenon. Borrowing the optical
concept of albedo, we represent the scattering from interfaces, such as ducts
313
314
and walls, strictly as a surface effect. So long as the surface dimensions are
large compared with the relaxation length of radiation in the medium, the
approximation is valid and useful.
When applied to nuclear radiation, the definition of albedo is much
broader than the traditional one used in optics. Nuclear-radiation albedos
include radiation that is scattered at depths of the order of a relaxation
length rather than just from the surface of the medium since particles
scattering from these depths contribute significantly to the total radiation
emerging from the surface. The tacit assumption made in most applications
of albedo theory is that particles emerge from the medium at the same
points on the surface at which they were incident. Some nuclear-radiation
albedos refer to mixed radiations; Le., emergent particles are sometimes of a
type different from the incident particles. Such albedos more properly carry
the modifier effective, the emergent radiation being identified as a particular
type of secondary radiation. For example, effective capture gamma-ray
albedos are albedos specifying the gamma-ray dose emerging from a medium
which is caused by incident neutrons being captured within the medium.
This type of albedo is especially important since for some duct configurations the capture gamma-ray doses can exceed the scattered-neutron doses.
Theoretically, calculations of nuclear-radiation albedos should be
straightforward since a large body of information is available on interaction
probabilities, the angular distribution of scattered radiation, and the
emergent energy vs. scattering angle for a variety of incident energies and
materials. However, even though the single-event probabilities are well
known, the solution of the macroscopic multicollision albedo problem
becomes complex. Consequently the value of the single-event probabilities
lies primarily in their usefulness in predicting trends. For example, if the
ratio of the scattering cross section to the absorption cross section is high, as
it is for neutrons diffusing in concrete, the resulting albedo will tend to be
high. If, however, the scattering is predominantly in the forward direction, as
it is for high-energy gamma rays, the albedo will be low. The albedo
properties of concrete have been thoroughly investigated since this material
is used in virtually every type of fixed shield. Other materials have been
studied, such as water, iron, lead, borated polyethylene, aluminum, and
various soils, but the albedo data for each are generally more limited.
Investigations have consisted primarily of calculations with the Monte Carlo
machine programs. In addition, a number of measurements have been made
to establish the validity of the calculated data. In nearly all cases the
calculated results have been fitted to empirical expressions.
ALBEDOS, DUCTS, AND VOIDS
315
7.2 DEFINITIONS
Traditionally, albedo refers to the ratio of the radiation current reflected
from a surface to the current incident on that surface. Consider a
monodirectional source of radiation of energy Eo incident on a surface at
polar angle 0 0 (see Fig. 7.1). The reflected current of energy E per unit
energy per unit solid angle at polar angle 0 and azimuthal angle I/J is given by
(7.2-1)
where i(E o ,( 0 ) is the incident current and a(E o ,0o,E,O ,I/J) is the albedo.
RECEIVER
Fig. 7.1 ~Geometry for calculating neutron and gamma-ray reflection from a surface.
When applied to nuclear radiation, albedo is not always expressed as the

ratio of the reflected current per incident current but instead may be given
as the ratio of reflected current per unit incident flux density, of reflected
dose per unit incident current, of reflected dose per incident dose, etc.
Neutron and gamma-ray albedos are available in several forms. The form
used in Eq. 7.2-1 is doubly differential; i.e., it is differential with respect to
both the reflected energy, E, and the reflected direction (as determined by 0
and I/J). A singly differential albedo is integrated over either energy or
direction. Dose albedos obtained experimentally as a function of exit
directions are also examples of singly differential albedos since dose is an
integral quantity with respect to energy. The term total albedo always
implies that integration has been performed over both energy and direction.
316
In this text differential albedos are denoted by the symbol a, as in Eq. 7.2-1,
and total albedos by the symbol A.
Three different types of differential and total albedos have been used by
various investigators in reporting their results on material reflectivity. The
first type, denoted here by the subscript 1, is an albedo that represents an
incident flux density of particles and an emergent current. The second and
third types, denoted by subscripts 2 and 3, are albedos for which the
incident and emergent particles are considered to be the same, current for
the type 2 albedo and flux density for the type 3 albedo. Flux densities and
currents are related functionally by the cosine of the entrance or exit angles.
For example, if <P(E,O,q,l is the reflected differential flux density per unit
energy per steradian, then the reflected current per unit energy per steradian,
J(E,O,q,) , is equal to <P(E,O,q,) cos O.
If the albedo being considered represents some weighting of the
particle-flux density, such as dose or energy-flux density, then the
subscripted letter D or E will precede the numerical subscript. In the
following albedo definitions, the term dose or dose rate is used in the generic
sense. The albedo definitions are not affected by the various types or names
of doses that are used. The choice of the dose definition and the flux-to-dose
conversion ratio govern whether the quantity is exposure, absorbed dose,
kerma, or dose equivalent. A functional notation is used in conjunction with
the albedo symbols to designate the independent variables for the particular
albedo being considered.
7.2.1 Differential-Dose Albedos

The three types of differential albedos for which the particle-flux density
has been weighted by a dose response function are defined as follows:
The form aD 1 (Eo,O 0,0,q,) represents differential current out (in dose
units) per incident flux density (in dose units). If the dose due to particles of
energy Eo incident at angle 0 0 is Do, then the particle current (in dose units)
per steradian reflected in the direction O,q, is given by Do aD l ' The
reflected-particle current in dose units (or dose current) has no physical
meaning but is merely a computational convenience. It is expressed
mathematically by
D(O,q,l =
f K(E)J(E,O,q,) dE
(7.2-2)
where J(E,O ,q, l = q,(E,O,q,) cos 0 and K(E) is the flux-to-dose conversion
ALBEDOS, DUCTS. AND VOIDS
317
factor for particles of energy E. The current measuring plane in these

definitions is the material-interface plane.
The form OlD 2 (Eo,Oo,O,<I is used for differential current out (in dose
units) per incident current (in dose units). If the dose due to particles of
energy Eo incident at angle 0 0 is Do (flux density in dose units), then the
particle current (in dose units) per steradian reflected in the direction 0,<1> is
given by Do cos (00)OID2, where Do cos 0 0 is the incident-particle current in
dose units. This type differs from the traditional current albedo only in that
the current is weighted by a dose response function.
The form OlD' (Eo,Oo,O,<I is used for differential flux density out (in
dose units) per incident flux density (in dose units). If the dose due to
particles of energy Eo incident at angle 0 0 is Do, then the dose per steradian
due to particles reflected in the direction 0,<1> is Do OlD ,. If the incidentparticle current per unit surface area is J(E o ), then Do = K(E o ) J(E o ) sec 0 o.
If the reflected-particle current per unit surface area is J(E,O,<I, then the
reflected differential dose is
D(O,.
These three types of albedos are related by
OlD 1
=OlD 2
cos 0 0
=OlD'
cos 0
(7.2-4)
7.2.2 Total-Dose Albedos

Total-dose albedos are obtained by integrating differential-dose albedos
over the solid angle represented by the exit hemisphere. Thus the three types
of total-dose albedos corresponding to the differential albedos described in
the preceding section are defined by
(7.2-5)
(7.2-6)
318
fOlD 1 (EO ,flo ,IJ ,q,) sec IJ dn

(7.2-7)
where dn = sin IJ dIJ dq, and the limits of integration are from IJ = 0 to rr/2
and from q, = 0 to 2rr.
Differential data must be available when data for AD 3 are being
compared with the other two types of total albedos, whereas AD I and AD2
are directly related; i.e., AD 1 = cos IJ oAD2 .
7.2.3 Other Albedos
Particle flux density or current- and energy-flux density or current
albedos, which refer either to particle or energy flow, have also been used. In
keeping with the previous nomenclature, these are
A I or OIl = particle current out per unit particle flux density in
A 2 or a2 = particle current out per unit particle current in
A 3 or 0I3 = particle flux density out per unit particle flux density in
For the case in which energy flow is considered, ,hese particle-flow
quantities are weighted by the energy and AE 1 or OlE 1 = energy current out
per unit energy-flux density in, etc.
All the parameters involved in these albedo definitions are the same as in
the dose albedo definitions except that neither the incident nor the reflected
flux density (or current) is converted to dose units.
An effective albedo may be defined for the emergence of secondary
gamma rays from a material due to incident neutrons. Examples are capture
gamma-ray albedo, inelastic gamma-ray albedo, or activation gamma-ray
albedo.
7.3
NEUTRON ALBEDOS
The fundamental mechanisms that result in neutron backscattering are

elastic and inelastic scattering, the two processes being distinguished by the
condition of the target nucleus following its collision with the neutron.
319
Owing to the relatively weak variation of the dose response function

with energy, neutron-dose albedos are more nearly proportional to the
particle-flux-density albedo than to the total-energy albedo. Calculations of
neutron albedos have been successfully performed and experimentally
verified in a number of cases. Investigations fall into three major categories,
distinguished by the energies of the reflected neutrons: fast neutrons,
intermediate-energy neutrons, and thermal neutrons. Studies of thermalneutron albedos have been further categorized as relating to those resulting
from incident thermal neutrons and those resulting from neutrons incident
at energies higher than thermal energy. Since neutrons incident at thermal
energy scatter in a more orderly process than do higher energy neutrons, this
category has yielded to direct analytical approach more readily than the
other categories.
7.3.1
Fast-Neutron Albedos
Major contributions to the data on fast-neutron albedos have resulted

from studies made at Oak Ridge National Laboratory (ORNL) by Maerker
and Muckenthaler 1 and at the U. S. Army Ballistic Research Laboratories
(BRL) by Allen, Futterer, and Wright. 2 Both groups performed detailed
Monte Carlo calculations to determine the reflection of fast neutrons as a
function of incident energy and angle. The results of Maerker and
Muckenthaler are more detailed than those of Allen et al. in that the
reflection data are differential with respect to both the reflected direction
(see Fig.7.1) and to the reflected energy. The data of Allen et al. are
differential with respect to the direction only. The ORNL results are limited
to concrete; the BRL studies include several additional materials.
The calculations by Maerker and Muckenthaler were performed as part
of an extensive analytical and experimental program that covered a wide
range of neutron energies (see Sees. 7.3.2 and 7.3.3) and included an
investigation of secondary gamma-ray albedos (see Sec. 7.5). In the
measurements a 23-cm-thick concrete slab that was reinforced with steel bars
at a depth of 3.8 em from either side was used. For the fast-neutron
calculations the steel was not considered, and the concrete composition was
assumed to be a typical concrete of the composition shown in Table 7.1.
Calculations were performed for six incident-energy bands covering the
energy range between 0.2 and 8 MeV. In a given problem neutrons incident
to the concrete were sampled uniformly from each incident-energy band,
and a statistical estimation technique was used to obtain estimates of the
320

Table 7.1-COMPOSITION OF CONCRETE USED
IN MONTE CARLO CALCULATIONS
Composition,
10 21 atoms em- 3
Maerker and
Muckenthaler 1
Allen et al. 2
Element
Hydrogen
Oxygen
Silicon
Calcium
Density, g em- 3
9.43
47.6
11.85
7.8
2.35
13.75
45.87
20.15
2.26
current emerging from the surface at various angles from a normal to the
surface.
Emergent angles were determined by the intersection points of a grid
formed by nine space-fixed polar angles and six azimuthal angles. The
results, obtained for distinct values of 0 0 ,0, and if!, were grouped into energy
bands !l.E o and !l.E. There were 10 reflected-energy bands, which, like the
incident-energy bands, covered the range between 0.2 and 8 MeV. (Albedos
that include neutrons reflected at energies less than 0.2 MeV were
determined separately and are discussed in Sees. 7.3.2 and 7.3.3.)
The differential albedo CXD, (Eo,O 0 ,E,O ,if!) calculated by Maerker and
Muckenthaler is in units of reflected current (in single-collision dose units)
per MeV per steradian per incident current (in single-collision dose units) of
a "gun-barrel" beam source. The average statistical uncertainty associated
with the Maerker-Muckenthaler data is about 10% for the doubly
differential albedos and about 3% for singly differential albedos.
Results from these calculations are shown in Figs. 7.2 through 7.5.
Figure 7.2 shows the variation of the total albedo (integrated over both the
reflected energy and the reflected angle) as a function of the incident angle
and incident-energy band. Figures 7.3 and 7.4 show the dependence of the
differential-dose albedo on the reflection angles 0 and if!; Fig. 7.4 clearly
illustrates that an assumption of no dependence on the reflected azimuthal
angle if! can lead to considerable error in the differential albedo for some
conditions. The dependence on the azimuthal angle is strongest for
high-energy neutrons at grazing angles of incidence and emergence; it
becomes very weak for low-energy neutrons or for values of if! greater than
45. This trend is consistent with what would be expected since the first
321
0.7
/1E o . MeV
,.4
00.2 - 0.75
0.8 r-'O'O.75 -1.5
~
:..-~ /
~ ~ ---
61.5 - 3
A6
03-4
0.5 f - . 4 - 6 - f-.
06-8
:;:-0
oi'
Z
'<:(
0.3
r:::---'
.....-
;::::---
.....-
//
x~
~/
r"
.,.-
~~/
./
--
0.4
~~
----::::::: ~
C ....-v
0.2
o.
o1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
cos 80
Fig. 7.2-Total single-collision dose albedo as a function of cos (J 0 and till o for fast
neutrons (>0.2 MeV) reflected from concrete. [From R. E. Maerker and F. J. Muckenthaler, Nuclear Science and Engineering, 22: 458 (1965 ).1
scatterings of high-energy neutrons are in the forward direction and since

neutrons that have scattered more than once tend to have "forgotten" their
initial direction and thus emerge from the material in a random manner.
Figure 7.5 shows how the ratio of the total-dose albedo for singly scattered
neutrons to the total-dose albedo for singly plus multiply scattered neutrons
increases with increasing values of the polar angle of incidence.
Maerker and Muckenthaler developed an expression to fit their angular
differential-dose-albedo data which reproduces their Monte Carlo results
within 10%. The expression involves two terms: the first accounts only for
singly scattered neutrons and the second includes all multiply scattered
neutrons; it is assumed that the configuration is effectively a semi-infinite
body of concrete. The expression is
(J(D2(dEo ,{)o,{),rJ =
cos ()
cos
() +K (AE)
1 .l.l
COS
()
1:
Gm(dEo)Pm(cos{)s)+
cos ()
m=O
I:
k=O
Bk(dE o ) Pk(COS ()s)
(7.3-1)
322
where cos Os = sin 0 0 sin 0 cos

polynomials of orders m and k;
q, - cos 0 0 cos 0; Pm and Pk are Legendre

I
K 2 (flE o ,Oo,O)
= 1::
j=O
(7.3-2)
(cosO)'.1:: aij(flEo)cosJO o
)=0
and the constants G m , Bk, K, , and aij are given in Appendix 0, Table 0.1.
I
.
~
'--.
'~
~~
'~
-""
1'-.
"'-..,,
" ~~ r-.
~
~.
"S ......
'~~
~'"
'~~
- - - - - = 15
4 f - - - - - - ~ = 45
- - - =7So
3 f----- ____ q, = 1050
2 f - - . _ - q,=13So
r--.
~'"
"\
~\\
,\
"~~\
- - - =165
1
o1.0
0.8
l\
0.4
0.6
0.2
cos 8
"
o
Fig. 7.3-Differential single-collision dose albedo per steradian as a function of cos 8 and
f/J for 1.5- to 3-MeV neutrons incident on concrete at 8 0 = 60. [From R. E. Maerker and
F. J. Muckenthaler, Nuclear Science and Engineering, 22: 458 (1965 ).1
To provide experimental verification of these data, Maerker and

Muckenthaler used their differential albedos to predict reflected fast-neutron
doses from a collimated beam of reactor neutrons incident at various angles
to a 1.8-m-square, 23-cm-thick concrete slab and compared calculations with
corresponding measurements at the ORNL Tower Shielding Facility (TSF).
The values of the incident angles 0 0 covered in the calculations and the
el'periment were 0, 45, 60, and 75. The calculations were weighted by
an incident spectrum that had been previously measured at the TSF, and the
incident dose rate used was the dose rate determined by integrating the
ALBEDOS, Duers, AND VOIDS
323
measured dose rates over the effective cross-sectional area of the incident
beam.
Agreement was considered good in that the root-mean-square deviation
between the predicted and measured values was 3.1 % and the largest single
deviation was 9%. Maerker and Muckenthaler subsequently incorporated
these results in a duct calculation for comparison with a series of
J
18 -
I
I
I
----- 15
____ .45
- - - q,'=' 750
16
- - - - .'105
4 _ _ _ .'135
_ _ _ .'165
,,
".,----
''\
V'
\,
!./,,,,
"'" .:----
>~-
\,
------ ---- - -'-
'-
-- r---.
,
-c::::
~ ,~~
~~
,,
,
~\
1.0
0.8
0.6
0.4
0.2
cos 9
Fig. 7 A-Differential single-collision dose albedo per steradian as a function of cos 8 and
for 6- to 8-MeV neutrons incident on concrete at (Jo = 85. [From R. E. Maerker and
F. J. Muckenth.ler, Nuclear Science and Engineering, 22: 458 (1965 ).1
measurements in a three-legged duct. Their results are described later in this

chapter.
The Monte Carlo calculations performed by Allen et al. 3 were used to
determine the fraction of neutrons from monoenergetic sources that was
transmitted through and reflected from infinite slabs of various materials,
including concrete. The source energies were 0.10, 0.25, 0.50, 1.0, 2.0, 3.0,
5.0, and 14 MeV, and the angles of incidence were 0, 30,45, and 70
from the normal to the slab. The low-energy cutoff was 10 eV. Transmitted
and reflected neutrons were accumulated in energy and angle intervals, the
reflected data yielding the multicollision-dose albedos OlD 3. (Eo,ll 0,11) and
324

0.7
t1E o . MeV
//
00.2 - 0.75
,///'
0.6 1---'17 0.75 - 1.5

A
1.5-3
./
0.5 1-03-4
x4- 6
o6 -
Q
:;:
l:l
/0.2-0.75~
0.4
V ~
V
~
'.
~ 0.3
o. 2
6 MeV
o. 1
0
1.0
0.9
0.8
0.7
0.6
0.5
0.'
0.3
0.2
0.1
Fig.7.S-Ratio of total singlecollision dose albedo for fast neutrons (>0.2 MeV) singly
scattered from concrete to total albedo for singly plus multiply scattered neutrons as a
function of cos 8 0 and Eo. {From R. E. Maerker and F. J. Muckenthaler, Nuclear Science
and Engineering, 22: 458 (1965).1
AD3(Eo ,Oo)' Energy spectra of the reflected flux were also determined for
each incident energy-angle combination, but the albedo was not calculated
as a function of the energy of reflected particles.
In addition to concrete, the materials included in the calculations were
water, iron, borated polyethylene (8% B.C by weight), and three Nevada
Test Site (NTS) soils that differed only in moisture content. The densities
and elemental content of the materials are given in Appendix 0, Table 0.2.
In all cases the slabs were assumed to be sufficiently thick to yield albedo
data approximating those for a semi-infinite geometry.
Some results from the calculations of Allen et aU are shown in Figs. 7.6
through 7.9. Figures 7.6 through 7.8 are plots of the total-dose albedos for
the various materials as a function of the hydrogen content for incident
source energies of 0.1, 2.0, and 14.0 MeV. Figure 7.9 gives the angular
distributions of reflected neutrons from a l.O-MeV source.
French and Wells' analyzed the differential data of Allen et al. 3 and
obtained a fit that is a function of the incident and reflected polar angles
only. The lesser detail of the BRL results led to the conclusion that the
325

1.4
I
1.2
Eo'" 0.1 MeV
~ Iron
o 8 0 '" 0
9 0 =30
4 80
&
1.0
r~
r
I
0.6
M
'"
o.6
.".....;
o.4
.-
45
8 0 '" 70
Soil (saturated)
[~"_._~
-Soil (dry)
~~
Wal]
POIYethy:.ne
~.--.
o. 2
10
20
30
1~_
40
50
60
70
HYDROGEN, at. %
Fig. 7.6-Total multicollision dose albedo for O.lMeV neutrons incident on various
materials. (From Allen, Futterer, and Wright. 3 )
dependence of the reflected azimuthal angle rp was weak and for the most
part irregular; thus the dose reflection data were averaged over rp. The
dependence on the reflected polar angle was found to fit a cos 0 function,
and the dependence on the incident angle was approximated by cos% 0 0 ,
which yielded an expression of the form t
(7.3-3)
where k (Eo) is a normalizing constant that includes the effect of incident
energy and reflecting material. Values of k(E o ) are shown in Appendix 0,
Table 0.3, for concrete, the three NTS soils, and iron for eight monoentThe original data of Allen et al. 3 were converted by French and Wells to a type 1 albedo (see
Sec. 7.2).
326
REAcrOR SHIELDING FOR NUCLEAR ENGINEERS

1.2
rim"
I
1.0
- Concrete
~ Soil (saturated)
~
"
0.8
"
~ 0.6
..
o.4
Soil (50% saturatjl

-Soilldrv)
~~
..............
-"-
EO
= 2.0 MeV
WaterI
Polyethylene _
..
06 0 = 0
0.2 t-- .8 =300
0
680 =45
.. 8 0 = 70
I
10
20
30
40
50
60
70
HYDROGEN, at. %
Fig. 7.7-Total multicollision dose albedo for 2.0-Me V neutrons incident on various
materials. (From Allen. Futterer. and Wright. 3 )
ergetic sources and a fission source. Equation 7.3-3 is assumed to be valid for
all materials of low to moderate hydrogen content (''ERlkt < 0.5). (The
water data of Allen et al. show a less pronounced dependence on () 0 and are
not represented by Eq. 7.3-3.)
French and Wells found that, except for incident energies near
cross-section peaks of the elements in the material, the total-dose-albedo
data of Allen et al. 3 could be correlated by a linear function of the ratio of
the macroscopic hydrogen cross section of the material to its macroscopic
total cross section. There is also an excellent correlation when the total
albedo is averaged over the fission-neutron spectrum, as is shown in Fig. 7.10
for normally incident and normally reflected fission neutrons. This should be
a useful correlation in extrapolating to other materials for which calculations
have not been performed.
SongS used the Monte Carlo data of Allen et al. 3 to obtain values of an
energy-dependent parameter that would give the best fit to a semiempirical
formula he had derived for the fast-neutron differential-dose albedo for
327
o.
J-,mn
0.5
- Concrete
- Soil (saturated)
0.4
Soil (50% saturat~)
..,
~
O. 3
-...;
o. 2
IrSoilldo-VI
Water-
~~
I--!..
..........
~~
~
E Q = 14.0 MeV
o. 1
o 60 ::: 0
r---.- (Jo:=
POIV.thvl~ll
30
6 00 ,=45
6. 9
10
70
20
30
40
50
60
70
HYDROGEN. at. %
Fig. 7.8-Total multkollision dose albedo for 14.0-MeV neutrons incident on various
materials. (From Allen, Futterer, and Wright. 3 )
concrete. The formula, analogous to that used by Chilton and Huddleston'

for gamma rays (given in Sec. 7.4), is
F(E o ) cos IJ
cos IJ o + cos IJ
(7.3-4)
where F(E o ) is the energy-dependent parameter. Song obtained values of this

parameter from a least-squares analysis of the Monte Carlo data for concrete
which gave the best fit to the equation. The values were then empirically
correlated as a function of energy by
F(E o ); Eo exp (0.9719 - 2.895A + 0.3417E o )
(7.3-5)
Another investigation of fast-neutron albedos was performed by Henry,

Mooney, and Prevost, 7 who studied the reflection of fast neutrons normally
328

0.240'~---~----~---""T""-----r------'
Slab: 24-i'n.thick concrete
Eo = 1.0 MeV
o 80 = 0
0
0
eo ;30 ---+-------t-----f---AI.I
0.200
!:J.O o =45 D
~8o=70
O.l60I------l-----+-----t--~=_9Jq=----__l
O.1201------l-------jh-L~~~---_+----__I
O.080'I----+--/-~___,4-----__+---__+---__l
Note: single curve drawn through
points for (] 0
0 and 30
0.04011---+-..te,,,(...----+---__+-----I--------1
_ _---l
o~
0.2
----L
,.L--____::_~--~
0.4
0.6
0.8
1.0
cos 8
Fig. 7.9-Angular distributions of fast neutrons reflected from concrete. (From Allen,
Futterer, and Wright. 3 )
incident on various thicknesses of steel and 6% borated polyethylene. Their

work included both Monte Carlo calculations and experiments utilizing a
well-collimated reactor beam. In the experiments total-dose albedos (AD3)
were evaluated from data obtained by traversing the beam area with a
dosimeter in a plane adjacent to the slab and repeating the traverse in the
same plane with the slab removed. Figure 7.11 shows experimental and
calculated results.
The data of Henry et al. are of particu lar interest because they show the
dependence of the albedo on material thickness. In addition, F~..z..12 shows
the reduction of the steel albedo caused by facing the steel sla with various
thicknesses of polyethylene. In Fig. 7.11 it appears that the albedo for steel
is approaching a value of nearly 0.6, which is lower than the value of 0.84
329
0.6
0.5
0.4
f~
Concrete
..... Dry soil
Soil 150% saturatedt-
~O.3
o. 2
~N
0.435 - 0.430
turated)
~H/~t " "
o. 1
I'---- 1'--.
Water
I'--- ............
0.1
0.2
0.3
0.4
0.6
'i:.HIE t
0.6
0.7
0.8
1.0
0.9
Fig. 7.10-Dependence of fast-neutron total multicollision dose albedo on hydrogen

content of reflecting material (fIssion neutrons, Eo > 0.2 MeV). (From French and
Wells.' )
0.6
0.5
0.4
v-
1---.
/1
3
rl
o. 2
o.
Measured
Calculated
'~
10
15
20
25
30
THICKNESS, em
Fig. 7.II-Total single-collision dose albedos for ftssion neutrons normally incident on
steel slabs: comparison with experiment. (From Henry, Mooney, and Prevost.')
330
obtained when the data of Allen et al. are put in the AD 3 form. The finite
detector used in the experiments to traverse the interface approximates the
slab detector assumed in the calculations of Allen 'et al. except for the
low-energy cutoff, which was 0.2 MeV in the experiment and in the
calculations performed by Henry et al. but was 10 eV in the calculations by
0.5
,{ Polyethylene
0.4
Incident--f
beam
0.3
1\
CI
'"
0.2
'I
1T
10.3cm
steel '
l!I
polyethylene
~"--
I-
o. 1
"15.2cm
k.e
o Measured
Calculated
0 2 4
POLYETHYLENE THICKNESS, em
Fig. 7.12-Comparison with experiment of total single-collision dose albedos for fission
neutrons normally incident on a laminated slab of steel and 6% borated polyethylene.
(From Henry, Mooney. and Prevost.?)
Allen et al. When the contribution below 0.2 MeV is subtracted from the
data of Allen et al. for representative energy groups, good agreement with
the results of Henry et al. is obtained.
The Monte Carlo results of Maerker and Muckenthaler l are the most
complete available for neutron scattering from concrete, especially for
grazing angles of incidence and emergence of high-.,nergy neutrons where the
dependence on azimuthal angle was found to be important. For water, soil,
iron, and polyethylene. the Monte Carlo data of Allen et aL 3 can be used.
The best accuracy is obtained when the Monte Carlo data are used
directly in calculations; if the requirements on accuracy are not too
stringent, the empirical expressions for the data can be applied, especially
when an azimuthal dependence is not expected to be great. The analytical
fits will generally provide less information on the distribution of the
331
scattered radiation, but they will allow the reflected dose to be expressed in
compact form, which will greatly reduce the bulk of data needed for
calculations. A less precise but expedient calculation can be made by using
the fission-spectrum total albedo shown in Fig. 7.10 with an assumed
cos'% eo cos e angular dependence.
7.3.2 Intermediate-Neutron Albedos
Coleman, Maerker, Muckenthaler, and Stevens calculated albedos for
neutrons whose incident and reflected energies both are in the intermediateenergy range for steel-reinforced concrete. Using a Monte Carlo technique
similar to the one used for the fast-neutron albedos, 1 they determined the
distribution in energy and angle of neutrons reflected from the concrete for
5 incident directions and 10 incident-energy groups in the energy range
0.5 e V to 200 ke V. The reflected distributions are given in terms of a doubly
differential albedo for each of 54 different emergent directions for each
group lying between and including the incident group and the lowest group
(0.5 to 1.8 eV).
Reinforced concrete was used in these calculations because they were a
part of the ORNL albedo program mentioned earlier. In the experiment a
23-cm-thick concrete slab that had steel reinforcing bars at a depth of 3.8 em
from either side was used; therefore all the calculations except the
fast-neutron calculations were performed for a mock steel configuration in
which each depth interval containing steel reinforcing rods was taken as a
homogenized region of concrete and steel that was 2.5 em thick. The result
was a five-region slab that had ordinary concrete and reinforced-eoncrete
layers. The effect of the iron on the intermediate-energy-neutron albedos
was found to be negligible.
A complete tabulation of the doubly differential albedos calculated by
Coleman et al. is too lengthy to be included here. The results, however, were
numerically integrated over all exit energies to produce singly differential
albedos and over all exit energies and angles to produce total albedos. The
integrated results were fitted by the following expressions to within 15% for
differential albedos and to within 3% for total albedos:
x {I + (1 -11)(1 - 1I0)[a(2 cos' 4> - 1)
+ b cos 4> + C cos 3 4>]}
(7.3-6)
332
and
(7.3-7)
where 11 ; cos Ii, 110 ; cos Ii 0' and the constants are given in Appendix 0,
Table 0.4, for each incident-energy group. When the expression for the
differential albedo was determined, all exit neutrons were assumed to have a
liE energy distribution within each energy group. The expression for the
total albedo was obtained by integrating Eq. 7.3-6 over all exit angles.
7.3.3 Thermal-Neutron Albedos
As mentioned previously, two types of thermal-neutron albedos can be
considered: the purely thermal albedo, for which both the incident and the
reflected neutrons are at thermal energy, and the albedo for emergent
thermal neutrons that result from the moderation of neutrons that are
incident at energies higher than thermal. These two categories are treated
separately.
(a) Neutrons Incident at Thermal Energy. Various approximations to
the purely thermal albedo have been derived analytically, with isotropic
scattering and capture being the only interactions allowed. In some
approximations only the total albedo is derived, and it is expressed as a
function of the incident angle, assuming that the reflected neutrons will
emerge with isotropic or cosine distributions. In other approximations
differential albedos that are functions of both the incident angle and the exit
angle are obtained.
In analytical treatments of purely thermal scattering, the exit current is
independent of the azimuthal angle by virtue of the assumption of isotropic
scattering. Monte Carlo calculations made with this assumption have shown
reasonable agreement with the other forms of analysis; however, Monte
Carlo calculations in which anisotropic scattering was assumed for hydrogen
contained in the material have shown that the albedo exhibits an azimuthal
dependence, although to a lesser extent than was shown for fast neutrons.
Results from Monte Carlo calculations in which both types of scattering
assumptions were used follow the discussion of various analytical
approaches.
Probably the first investigation of thermal-neutron reflection was
performed by Fermi! who showed that for large values of N (N ; J:.t/J:. a , the
ratio of the total cross section to the absorption cross section) the total
333
albedo for thermal neutrons incident on an infinitely thick isotropically

scattering medium bounded by a plane is given approximately by
..IN -1
"IN +y'3 cos 11 0
(7.3-8)
where A,(l1 o ) is the number of thermal neutrons reflected per incident

thermal neutron and 11 0 is the angle of incidence.
A rigorous calculation was carried out later by Halpern, Lueneburg, and
Clark,! 0 who obtained the formula
A, = 1
(7.3-9)
where k = 2.91, 2.31, and 2.48 for normal, isotropic, and cosine angular
distributions of incidence, respectively.
Glasstone and Edlund! I derived a formula by use of diffusion theory
which is given by
1- 2KD
A, = 1 + 2KD
(7.3-10)
where K is the reciprocal of the thermal-neutron diffusion length and D is

the diffusion coefficient.
Chandrasekhar!' treated radiation backscattering by a method that can,
in general, be applied to any type of radiation that scatters isotropically with
a relatively constant cross section, a condition that is approximately met by
thermal-neutron scattering. The resulting backscattering angular distribution
can be written in the form
",,(l1 o ,I1,p) =-
2 cos
cos 11
11 H(p, cos 11 0 ) H(p, cos 11)
+ cos 0
11
(7.3-11)
where p is the probability of nonabsorption in a single interaction

(p =T, siT, t, the ratio of the scattering cross section to the total cross section)
and Chandrasekhar's H function is tabulated!' for various values of p and
cos 11 o' The total albedo as a function of 11 0 and p is obtained by integrating
Eq. 7.3-11 over the exit angle 11. This integration yields
A, (l1 o ,p)
=1 -y'l=p H(p, cos 11 0 )
(7.3-12)
334
Spencer, Diaz, and Moses' 3 have applied Chandrasekhar's albedo to neutron

duct penetration and compared it with the Monte Carlo data.
Mockel' 4 studied alternate approaches for determining the total
thermal-neutron albedo for strongly absorbing media as a function of slab
thickness. His objective was to develop a method that would be readily
adaptable to computer usage and still provide good results with less
0.02
/I~
0.0
II
a
a
~5/I:t
...
'" 0.1
Semiempirical formula
Numerical integration
Variational method (exponential

trial function)
...
0.5
Approximation of Chandrasekhar's functions
Variational method (constant trial function)
1.0
1.5
2.0
THICKNESS, mean free paths
Fig. 7.13-Total albedos obtained by various methods for thermal neutrons isotropically
incident on a strongly absorbing medium ("f.s/Lt:::: 0.1). [From A. Mackel, Nuclear
Science and Engineering, 22: 346 (1965).]
computing time than is required for integration of the Boltzmann equation.

A variational method with a constant trial function was found to give poor
results except for thin slabs. In a search for better methods, three solutions
were tried which were based on an approximation to the moments of the
Chandrasekhar function, a variational solution with an exponential trial
function, and a diffusion-like semiempirical formula. Figures 7./3 and 7.14
compare the results from the three methods for isotropic incidence and two
values of T-s/T- t . Also shown are results obtained with a constant trial
function and with numerical integration of the Boltzmann equation.
The semiempirical formula provides a fit within 2% for an isotropically
incident current (plane isotropic source) and within 1% for a current either
normally incident or incident with a cosine distribution (isotropically
incident flux). It has the form
335
1 - e-'ox
A,
=1 _
[1 - f(p)
(7.3-13)
Ae-'~x 1 + f(p) + g(p)
where p = '2:.s!'I;t, as before, x is slab thickness in mean free paths, and values
of the coefficients and functions are given in Appendix 0, Table 0.5, for
three conditions of neutron incidence.
0.15
/'
0.10
""
0.05
~:
Es/1;t
!IT
0.5
Semiempirical formula
(:,
Numerical integration
Variational method (exponential

trial function)
...
Approximation of Chandrasekhar's functions
Variational method (constant trial function)
0.5
1.5
1.0
2.0
THICKNESS, mean free paths
Fig. 7.14-Total albedos obtained by various methods for thermal neutrons isotropically
incident on a strongly absorbing medium CT.sl'i:. t = 0.5). [From A. Mackel, Nuclear
Science and Engineering, 22: 346 (1965).1
Pomraning l5 proposed a variational solution with an exponential trial

function which yields for the total albedo the expression
2
= (1 + v) In( 1 _ v') [In( 1 + v) - v]
(7.3-14)
for normally incident thermal neutrons and the expression
A,
v2 In( 1 _ v2) [ n( 1 + v) - v J
(7.3-15)
336
for isotropically incident thermal neutrons. Here v is a positive quantity

satisfying the transcendental equation
2vc =In (11-v

+ v)
(7.3-16)
and c is the mean number of secondaries per collision.

Values obtained with these relatively simple formulas compare very
favorably with the results of exact solutions by Chandrasekhar,' 2 the
difference being generally less than 2%.
Equations 7.3-14 and 7.3-15 are derived from the more general
expression
2f
A 2 =In(1-V 2 ) [In(l +v) -v]
(7.3-17)
where
(7.3-18)
in which Jl is the cosine of the incident angle and B(Jl) is the general
expression for the angular distribution of the incident flux. The success of
Eqs.7.3-14 and 7.3-15 for normal and isotropic incidence suggests' that
Eq. 7.3-17 could be applied to other angular distributions with equal success.
Wells' 6 determined total and differential thermal-neutron albedos for
portland concrete by analyzing Monte Carlo results obtained in a calculation
that he had originally performed to establish the distribution of capture
gamma-ray sources in concrete and air due to thermal neutrons incident on
the concrete.' 7 Expressions that he derived to fit the Monte Carlo results are
(7.3-19)
and
at = 0.21 cos* 0 0 cos 0
(7.3-20)
which, when converted to type 2 albedos, are given by
A 2 = 0.66 cos-\> 0 0
(7.3-21)
337
and
=0.21 cos-1300 cosO
0<2
(7.3-22)
A type 3 total albedo is obtained by integrating the differential albedo over

the exit hemisphere as follows:
A 3 (portland) =
0<1
sec 0 dn
12 f2.
0 0.21
= f0
2l'
COS'30 0
sm 0 dO d</>
= 1.32 cos% 0 0
(7.3-23)
Expression 7.3-23 is to be compared with an expression obtained by

Wells for a particular concrete used in structures built at the ORNL Tower
Shielding Facility (TSF):
(7.3-24)
A 3 (TSF) = 1.3 cos 0 0
In both the preceding calculations the thermal neutrons were assumed to

scatter isotropically in the center-of-mass sysrem with no energy loss, and
histories were rerminated afrer a fixed number of interactions. Statistical
estimates were made for each interaction point. The c:oncrete compositions
used in these calculations are given in Table 7.2.
For their thermal-neutron albedos, Maerker and Muckenthaler' 8 performed single-velocity Monte Carlo calculations for a 23-cm-thick steelreinforced concrete slab. Two different scattering laws were used for the
Table 7.2-ASSUMED CONCRETE COMPOSITIONS USED

IN MONTE CARLO CALCULATION BY WELLS 16
Element
Hydrogen
Oxygen
Carbon
Ma.gnesium
Aluminum
Silicon
Calcium
Iron
Portland concrete,
1021 atoms cm- 3
TSF concrete,
10 2 t atoms em- 3
2.868
43.260
6.507
15.6
39.6
5.42
0.40
1.32
10.00
7.40
0.31
9.899
8.736
338
water contained in the concrete: (1) isotropic scattering in the laboratory

system and (2) anisotropic scattering with an approximation based on
experiments by Greenspan and Baksys 19 at 0.0358 eV. Scattering from all
other constituents in the concrete was assumed to be isotropic in the
laboratory system.
)
Measurement
2. 4 ~
2.
Monte Carlo. H2 0 anisotropic scatter

___ Monte Carlo. H 0 isotropic scatter
2
19 0 =0
... ~
... ...
'~
...
1. 6
N
"
1. 2
~~
o.8
"':l
r,
"~
o.4
0
1.0
0.9
0.8
0.7
0.6
0.5
00,8
0.4
0.3
...........
0.2
1'-.. . .o
0.1
Fig. 7.1S-Differential albedos for thermal neutrons normally incident on steelreinforced concrete. A comparison of single-velocity Monte Carlo calculations and ORNL
TSF experiments. [From R. E. Maerker and F. J. Muckenthaler, Nuclear Science and
Enilineering, 26: 345 (1966).]
Typical results from the calculations are shown in Figs. 7.15 and 7.16,
both of which illustrate that the best fit to the experimental data is obtained
when the scattering in water is assumed to be anisotropic. Figure 7.16 also
shows that even for thermal neutrons there is some dependence of the
albedo on the azimuthal angle.
From an analysis of their data, Maerker and Muckenthaler fitted their
results for the differential albedo as follows:
0.08751'
"'2 = I' + 1'0
(1 + 1.281')(1 + 1.621'0 - 0.421'~)
x [1 + (1 -1'0)(1 -1')(-0.10 + 0.43 cos
</>
+ 0.20 cos 2 </] (7.3-25)
339
where Il = cos 0 and Ilo = cos 0 0 , A much simpler fit was obtained for the
total thermal-neutron albedo (integrated over all exit anl~les):
A2
=0.86 -
=75
M~asurement
12
-
I--
0.19 cos 0 0
(7.3-26)
1 I I
Monte Carlo, H 2 0 anisotropic: scatter

Monte Carlo. H 2 0 isotropic scatter
10
r-
,8
N
"
---
f.Q.
o
o
20
40
f--
0,278
cos (}
0,167
0
0
r--.....
r--
cos (}
60
80
100
120
.- C'.
. ~s (} - 0.056
140
160
180
., dog
0
Fig. 7.16-Differential albedos of thermal neutrons incident at 75 on steel-reinforc.ed

concrete. A comparison of single-velocity Monte Carlo calculations and ORNL TSF
experiments. [From R. E. Maerker and F. J. Muckenthaler, Nuclear Science and
Engineering, 26: 346 (1966).J
Results obtained for concrete with the various equations for thermalneutron albedos are compared in Table 7.3. Comparisons are made between
the total-current albedo for various conditions of incidence. For all cases
except 0 0 = 75, there would appear to be 10% agreement between all the
values. The formula due to Wells should be limited to angles of 0 0 < approximately 60 because of its tendency to overpredict at grazing incidence.
(b) Neutrons Incident at Nonthermal Energies. The only estimates
available for albedos of reflected thermal neutrons resulting from incident
neutrons of higher energy are those from the Monte Carlo calculations of
Coleman et az.s for 0.5-eV to 200-keV neutrons incident on the same
steel-reinforced concrete described previously (see ,especially Sec. 7.3.2).
340
Expressions that reproduce the Monte Carlo values to within 15% for the
differential albedos and to within 10% for the total albedos are given in
Appendix 0, Table 0.6. The reflected angular distributions (differential
albedos) for the six highest energy groups have a shape that is independent
of lID (the cosine of the incident polar angle) and identical to the shape
Table 7.3-COMPARISON OF VARIOUS VALUES OF THE TOTAL ALBEDO FOR

THERMAL-NEUTRON REFLECTION FROM PORTLAND CONCRETEt
Values of A 2 obtained with the formulas of
Maerker
and
Source
Isotropic
distribution
Cosine
distribution
Monodirec-
Mucken
thaler l8
0.78
Glasstone
Wells l6
0.792
Porn
raning 15
0.8
0.734
Halpern
et al. 1O
0.719
and
Edlnnd"
CbandraFermi9
0.74
sekhar l2
0.79
0.698
tional
source
8 0 = 0
0 0 = 45
0 0 = 75
0.67
0.72
0.855
0.66
0.74
0.996
0.698
0.645
0.745
0.78
0.82
0.691
0.730
0.82
t};s/~t = 0.9849 except for the Maerker-Muckenthaler results, which are based on the
five-region concrete slab discussed in the text.
derived by Fermi for the emergent angular distribution from a plane surface
in the simplified case of thermal neutrons diffusing in a noncapturing and
isotropically scattering semi-infinite medium.
Some contribution is also made to the emergent thermal-neutron current
from the moderation of incident neutrons with energies greater than
200 keV. Coleman et al. accounted for these higher energy neutrons by
extrapolating the results of Table 0.6 to obtain expressions for energies up
to 9.57 MeV (see Appendix 0, Table 0.7). For the spectrum of neutrons
from the ORNL Tower Shielding Reactor II, which was the source used in
verifying tbe calculations, it was estimated that a consistent error of 20% in
the extrapolated results would lead to an error of only about 8% in the
predicted values of the differential thermal-neutron albedos.
341
7.4 GAMMA-RAY ALBEDOS

The primary interaction that contributes to the backscattering or albedo
of gamma rays is Compton scattering from electrons. In this interaction the
photon rebounds with an energy that is directly dependent on the scattering
angle and the incident energy and is given by the Kldn-Nishina formula
(Sec. 3.2.1). The higher the incident energy, the more strongly will the
forward direction be favored. Thus the gamma-ray albedo has an inverse
relation with the incident-photon energy. Owing to the strongly peaked
forward scattering, the gamma-ray albedo also has a strong dependence on
0
the azimuthal angle t/J; i.e., the scattering angle Os (see Fig. 7.1) at t/J: 0 is
0
smaller than the scattering angle at t/J: 180 , and hence the albedo decreases
with increasing t/J. The magnitude of the difference increases with increasing
values of the incident polar angle 00 ; thus at grazing incidence the difference
0
0
lies between forward scattering (t/J: 0 ) and backscattering (t/J: 180 ).
Another interaction that contributes to the reflected energy for incident
gamma rays of high energy is pair production (Sec. 3.2.1). The positron
created in this reaction is annihilated by combining with an atomic electron
and releasing energy in the form of two new gamma rays. This reaction is
possible only if the energy of the incoming gamma ray is greater than
1.022 MeV, and it is predominant only at energies above about 5 MeV. Each
of the photons created has an energy of 0.511 MeV, which is greater than
the maximum energy possible for gamma rays scattered backward
(Os: 1800 ) by Compton scattering.
Leimdorfer' 0 investigated the relative contributions by positron annihilation and Compton scattering to the total gamma-ray albedo for concrete
in a calculation that considered single scattering only. The results for
normally incident gamma rays are shown in Fig. 7.17, in which the fraction
of the albedo due to annihilation is plotted as a function of the incident
gamma-ray energy. Also plotted is the same fraction from a calculation by
Wells' 1 in which multiple scattering was considered.
As with neutrons, most of the studies of gamma-ray albedos have been
carried out either by Monte Carlo analysis or by experiments and most of
them have been for concrete. Where both experimental and calculated data
are available, there is good agreement (within 10%); however, experiments
have been limited to low gamma-ray energies 2 MeV) that can be obtained
from isotope sources, and definitive Monte Carlo calculations have been
performed for only a few materials.
342
One of the earliest Monte Carlo calculations was performed by Berger

and Doggett,2 2 who obtained the total-dose albedo, AD 3' for monoenergetic
sources incident on iron, tin, lead, and water. From these calculations a
quantitative measure of the dependence of the albedo on the thickness of
1.0,.-----,---,-----,.----,------,
o MUltiple scattering
- - Single scattering
~ 0.8
....
:I:
Z
Z
"~O.6
:>
W
al
;i 0.4
u.
o
z
5" 0.2
a:
u.
oO.!--=----:2~-----!4,......---:!-6-----:8!----1:-!.O
INCIDENT GAMMA-RAY ENERGY. MeV
Fig. 7.17-Ratio of pair-production annihilation albedo to total albedo for gamma rays
normally incident on concrete. (From Leimdorfer20 and Wells.2 1 )
the scattering material was obtained. This dependence is illustrated for iron
and water in Table 7.4.
Berger and Raso H . , . carried out an extensive series of Monte Carlo
calculations to determine the total-energy albedo for monoenergetic gamma
rays incident on a variety of materials, and in some cases they obtained
energy and angular distributions. An interesting result of their work is the
analysis of the variation in the total albedo with the atomic number of the
scattering material. plots of the albedo for normally incident gamma rays are
shown in Fig. 7.18. These data may be fitted by smooth curves; however,
data at intermediate Z values would greatly increase confidence in
interpolation, particularly at the 2.0-MeV level.
343
Raso' 5 performed additional calculations for concrete for source

energies of 0.5, 1, 2, 4, 6, and 10 MeV and polar angles of incidence of
cos 0 0 = 0.10, 0.25, 0.50, 0.75, and 1.0. The emergent gamma rays were
divided into 8 polar and 12 azimuthal angular increments, and differentialdose albedos, cxD, (Eo ,0 0 ,0,4, were obtained for all exit-angle combinations.
Total albedos, AD,(Eo,Oo), were obtained for each incident energy and
Table 7 A-DEPENDENCE OF GAMMA-RAY TOTAL-DOSE
ALBEDO, AD', ON MATERIAL THICKNESSt
Fraction of reflected
photons reflected
within depth of
Material
H,O
Iron
0.5
1.0
2.0
Eo,
MeV
0.,
mean free
mean free
mean free
deg
paths
paths
paths
0.66
0.66
1.0
1.0
0
60
0
60
0.65
0.61
0.79
0.89
0.88
0.96
0.93
0.98
0.99
1.0
1.0
1.0
t Adapted from Berger and Doggett.2 2
angle. Values of the differential- and total-dose albedos for normal incidence
are shown in Figs. 7.19 and 7.20, respectively.
Figures 7.19 and 7.20 also show Monte Carlo results obtained by
Wells,' 1 who calculated differential- and total-dose albedos for gamma-ray
reflection from concrete by analysis of prior Monte Carlo data' 6 on
gamma-ray scattering in air and concrete. The results of Raso and Wells are
essentially in agreement on the total albedo for energies of 2 MeV and
below; there are, however, some differences in the differential albedos at
these energies. At energies above 2 MeV, there is a substantial disagreement
between Raso and Wells on the magnitude of the total albedo, as shown in
Fig. 7.20. The divergence of agreement with increasing energy could be
atttibuted to the different manner in which pair production was treated in
the two calculations.
As is apparent from Fig. 7.19, the incident energies used in the
calculations by Wells were 0.6, 1, 2,4, and 7 MeV. The incident polar angles
were 0, 30, 45, 60, and 75. The emergent angles varied with the individual
problems, and no regularly spaced grid was used; however, by extrapolation
344
2 x 10-
'-----r-----.,-----..,...-----,
o
2 -
Eo = 0.2 MeV
Ii Eo '" 0.5 MeV

Eo'" 1.0 MeV
Eo'" 2.0 MeV
/concrete (Zeffective
13.41
50
70
90
ATOMIC NUMBER (ZI
Fig.7.18-Total-energy albedo for normally incident monoenergetic gamma rays as a

function of the atomic number of the material. (From Berger and Rasa? 3.24)
ALBEDOS, overs, AND VOIDS
345
4 X 10- 2
-- ""
""-
6
4
..
r.....
::::::-r--.....
>-
""'-
"
"
),.
..........
EoMeV
1\
\ \
\
\0.6
~ ~\ \
"
1.0
, 2.0 -
'\
I--
R.so
4f-O 1.0 MeV

tJ. 2.0M.V
1\\4.0-
- . 4.0MeV
, 1.0--
-Wells
2
10-
20
40
60
ao
B.d..
Fig. 7.19-Differentialdose albedos for gamma rays normally incident on concrete.

(From Wells" and Rasa.")
and interpolation, smooth curve fits were obtained for aD 1 VS. 0 for
azimuthal angles of 0 and 1800 Complete data sets are given in Ref. 21.
Chilton and Huddleston 6 developed a serniempirical formula for the
differential-dose albedo for concrete of the form
_ C K(Os) X 1026 + C'
aD2(E o ,00,O,tj -- 1 + [(cos Oo)/(cos 0)]
(7.4-1)
346
where K(8 s } is the Klein-Nishina differential-energy scattering coefficient

for the scattering angle 8s, and C and c' are adjustable parameters that are
dependent on the initial energy. Values of K(8 s } can be obtained from
Eq.3.2-8 [K(8 s}=a(8)]. Values of C and C were obtained by Chilton,
1
Raso
-Wells
2
i
1
\
8
.
N
\
\
-- r-
'\ "-
r-
,
2
10
12
EO' MeV
Ftg. 7.20-Total-dose albedos for gamma rays normally incident on concrete. (From
Wells" and Raso.")
Davisson, and Beach' 7 from an analysis of earlier Monte Carlo calculations

performed by Davisson and Beach. 28 Parameters were obtained by normalization to the total calculated albedo rather than to the differential data. The
resulting values of C and C' are given in Appendix 0, Table 0.8, for water,
concrete, iron, and lead. Chilton 29 later revised Eq.7.4-1 by altering the
347
denominator and multiplying by a fraction with five adjustable constants.

The newer formula has been applied only to concrete but is accurate within
10% for 13 7Cs and oOCo. Chilton's formula is given in Table 0.9.
In most of the experimental studies of gamma-ray albedos, the sources
used were 6Co and 137 Cs. Examples are the data obtained for concrete by
0.010 I'-...
0.008
~
---
.~
~ Chilton-Huddleston
formula (Eq. 7.4-1)
-'Wells
(CalculationS)~
0.004
[, ,
\
1\:\\
'\~ ~\
Measurements
l:J. Barrett and Waldman
Haggmark 1ft .,.
0.00 2
aa
00
8. dog
Fig. 7.21-Comparison between calculated and measured differential-dose albedos for

gamma rays (1.25 MeV) normally incident on concrete. (From Wells.21 )
6Co
Clifford;30 Haggrnark, Jones, Scofield, and Gurney;:ll Barrett and Waldman;32 and Berger and Morris. 33 Comparisons of typical data from these
investigations with results from some of the calculations discussed previously
are shown in Figs. 7.21 through 7.23. For these figures all the values were
converted to type 1 albedos, and, where necessary, the analytical data were
interpolated to match the experimental energies.
348
Except for small angles of reflection, the experimental and calculated

values generally are within 10%. For the single case shown in Fig. 7.22, the
weighted values of Raso and the calculations by Wells are very similar.
Figure 7.23 points up the strong dependence on '" for obliquely incident
radiation.
0.020
11
Rasa (weighted average of 1.0- and O.5-MeV Monte Carlo datal
0.016
~ I--
l
-- -~
"0.._'
1--...- __
Chllton-Huddl,,'on
formula (Eq. 7.4-1)
~~ IS
0.012
~ to..."< dY o
Wells (interpolated
fr~~
0.6- and l-MeV datal
0
0
0
~ "
0.008
Berger and Morris
~~\o b
~\\
.----"'\~
",\0'\Oil
0.004
Measurements
Haggmark et al.
o CI"ff0j"
o
o
20
40
60
80
e, deg
Fig. 7.22-Comparison between calculated and measured differential-dose albedos for
1 37CS gamma rays (0.662 MeV) normally incident on concrete. (From Wells. 21 )
Haggrnark et al. 3 I found that their experimental data on differentialdose albedos for concrete, iron, and aluminum could all be represented by
the expression
(7.4-2)
where (J s is the effective scattering angle shown in Fig. 7.1; OlD 3 is the
differential-dose albedo described in Sec. 7.2 except that the exit direction is
349
expressed in terms of (J s, which is a function of (J and </J; and b is a constant

for a given (Jo, Eo, and material. Figure 7.24 shows a comparison of
Eq. 7.4-2 with the experimental data for iron, concrete, and aluminum
reduced by the appropriate b value given in Appendix 0, Table 0.10.
0.024 r - - - - - r . , . . - - - - T " - - - - r - - - - - , - - - - ,
e!fl'''' 1800} Haggmarketa/.
o tP '" 0
(measurements)
A tP ~ 0 } Barrett and Waldman

tP =1800 (measurements)
0.020
--Berger and Morris (Monte Carlo)

-Wells (Monte Carlo)
0.0161-------Jf----.-,----7"l-+--~-__t
~ 0.0121------f---~A'--~~.-,-----_++__t
",,/
O.OO8I.-~~rl---+---+----l-___1
0.0041-----+----"1-.::..:"""''''"..,c---+--j
0 ' - - - -......- - - -.......- - - - ' - - - -......- - '
00
00
e,deg
Fig. 7.23-Comparison between calculated and measured differential~dose albedos for

gamma rays (1.25 MeV) incident on concrete at 60. (From Wells. 21 )
60 CO
Equation 7.4-2 is a useful tool for extrapolating albedo data, but it is not
valid for grazing angles since the measurements do not include 8., < 30.
Leimdorfer used the Monte Carlo method for calculating the total
gamma-ray-energy-flux albedo ftom concrete slabs2 0 and also from spherically
concave concrete walls. 3 4 An interesting result of the latter studies is shown
350
in Fig. 7.25, which gives the energy-flux albedo, AE3' for 1-MeV gamma rays
from a concrete wall surrounding a spherical cavity as a function of the
radius of curvature of the wall. The point isotropic source is located in the
center of an evacuated cavity in an infinite concrete medium. The AE 3 is
defined as the ratio of the collided to the uncollided energy-flux density at
0.08-
0.06-
0.02-
8 s ' radians
Fig. 7.24-Gamma-ray differential-dose albedo for aluminum, iron, and concrete as a

function of the scattering angle Os' [From L. G. Haggmark, T. H. Jones, N. E. Scofield,
and W. J. Gurney, Nuclear Science and Engineering, 23: 145 (1965).1
the wall surface. The infinite-radius (flat-surface) albedo is calculated to be

only about 2.5% higher than the albedo for a 500-cm radius spherical cavity.
Other interesting conclusions drawn by Leimdorfer from these calculations
are that reflections beyond the third or fourth order do not contribute
appreciably to the energy-flux albedo in the spherical geometry and that at
least 90% of the reflections occur within the first 24 g cm- 2 thickness for
gamma-ray energies below 10 MeV.
7.5 SECONDARY-GAMMA-RAY ALBEDOS
Secondary gamma tays, i.e., gamma rays produced by the interactions of
neutrons with nuclei, can appreciably increase the number of gamma rays
351
reflected from a material surface. The most important neutron interaction

involved is radiative capture, described in Chap. 3, which results in capture
gamma rays. Since these gamma rays usually create the greatest problem for
shield designers, the few studies of secondary-gamma-ray albedos that have
been performed involve captures. Activation gamma rays and inelastic0.10
0.09
0.08
..
:>
0.07
0
0
0.06
-M
."'
-'
x 0.05
::>
-'
u.
).
0.04
'-'
a:
w
zw
0.03
0.02
0.01
o0!----:I~OO:----:c200~---:3:-!OO=----:400~--::-!500
RADIUS,em
Fig. 7.25-Gamma-ray-energy flux albedo as a function of the radius of the concrete

reflecting wall (I-MeV source). (From Leimdorfer. 34 )
scattering gamma rays could be of concern in special situations. Albedo

definitions given in Sec. 7.2 do not apply precisely to secondary-gamma-ray
albedos. To emphasize that the incident particles are neutrons and the
emergent particles are gamma rays, we have included the notation (noy) in the
subscripts associated with the albedo symbols. All albedos are based on an
incident-particle current or flux density and an emergent gamma-ray dose
current.
352
For capture gamma-ray albedos to be determined, the capture gammaray source density within a medium must be known. The density at a point
is given by
Sy(E') = 1,:
J Uci(E)
Yj(E') Nj 4>(E) dE
(7.5-1)
where Uci(E) = the capture cross section of element i for neutrons of energy
E
Nj = the atomic density of element i
4>{E) =the scalar neutron-flux density at energy E
Yj(E') = the yield per capture in material i of gamma rays of energy E'
Equation 7.5-1 displays the quantities that must be known as a function
of position within the absorbing material for a calculation of the gamma-ray
production in the medium It also illustrates that a number of parameters
must be duplicated if secondary-gamma-ray data obtained for one situation
is to be applicable to another situation. In many cases sufficient accuracy
can be obtained by considering only the thermal-neutron-flux-density profile
and the capture cross sections averaged over the thermal group.
Data on capture gamma-ray albedos are quite limited. Wells' performed
a series of Monte ('..arlo calculations from which he obtained capture
gamma-ray albedos for the portland and TSF concrete compositions given in
Table 7.2. Assuming that only thermal neutrons were incident, he arrived at
the following expressions for the differential-dose albedo, which is given in
rads hr- 1 steradian-' per unit thermal-neutron-flux density incident at angle
Ii 0:
(7.5-2)
for portland concrete and
(7.5-3 )
for TSF concrete. In calculating the capture gamma-ray dose rate along the
axis of a concrete-lined cylindrical hole, Wells found that the results
obtained with the TSF albedo were higher by a factor of 40 to 50% than
those obtained with the albedo for portland concrete, which demonstrates
that variations in local aggregates cause variations in the capture gamma-ray
albedo because of changes in both capture density and gamma-ray yield per
capture.
353
Maerker and Muckenthaler l8 also obtained capture gamma-ray dose

albedos from Monte Carlo calculations for thermal neutrons incident on
concrete. In their case the concrete was the same five-region slab containing
steel in two regions. The differential-albedo data were fitted to the
expression
(7.5-4 )
where (XD2 is given in rads hr- ' steradian-I per incident unit thermal-neutron
current and Jlo and Jl are the cosines of the incident and reflected polar
angles, respectively. The corresponding expression for the total albedo is
(7.5-5)
given in rads hr- I per incident unit thermal-neutron current. These equations
fit the Monte Carlo data to within 15%.
Figure 7.26 compares the differential albedos obtained with the
Maerker-Muckenthaler expression with those obtained with the Wells
expression for TSF concrete, the latter being converted to type 2 albedos.
The data differ by a factor of approximately 2, which may have resulted
from several causes. In the calculations by Wells, neutron histories were
terminated after 50 collisions, whereas 200 collisions were allowed in the
calculations by Maerker and Muckenthaler. According to additional studies
performed by Maerker and Muckenthaler l8 and later by Coleman et al., 8
this would account for no more than 20% of the difference. Other
differences may appear in the gamma-ray yields, the minimum gamma-ray
energy considered, and the concrete composition. In particular, the layered
concrete slab used by Maerker and Muckenthaler had more iron nearer the
surface than did the homogenized system used by Wells. The additional iron
at this position probably resulted in more iron-capture gamma rays emerging
from the surface.
Coleman et al. 8 calculated, also by Monte Carlo methods, the capture
gamma-ray albedos resulting from the slowing down and capture of neutrons
incident on concrete at intermediate energies (0.5 eV to 200 keY). As was
the case for the other calculations by Maerker and Muckenthaler, the
concrete was the same five-region slab containing two regions with iron;
however, at most the iron contributed only 20% of the capture gamma-ray
dose.
The empirical equations representing the results of Coleman et al. are
shown in Appendix 0, Table 0.11. The departure from unity of the last
354
term in parentheses in the expressions for the differential albedos reflects the
contributions from captures occurring at nonthermal energies. These
expressions reproduce the Monte Carlo values to within 15% for the
differential albedos and to within 10% for the total albedos.
(Xl0-'lr------r----r------r----r-----,
- - - Maerker and Muckenthaler
2.0
-- -
-Wells
~
a
o
.~.~~
o
'------j-----+----+7"'---t--7.---l
1.6,
1.2f------t----l7"'--------:7!":.----f---:;;...."---j
:2
.~
&
80
~
:c
j~~~~
....
0
__
~ 0.81-------j-r'-7"'--+-7""'--+:;
........'""----t....
:;;;;-o-''''----j
~..",
.,.,." .......
,.,..
-0.2
0.8
0.4
60
.-J...-.-- - 80
0.8
1.0
cos 8
Fig. 7.26--Capture gammaray differential albedos for thermal neutrons incident on

concrete. (From Wells I 6 and Maerker and Muckenthaler. 18 )
Using thermal-neutron fluxes measured in TSF concrete, French, Wells,

and Schaeffer" calculated the capture gamma source strength as a function
of depth and obtained the resulting gamma-ray leakage from the concrete
with a Monte Carlo code. The resulting angular dose rates were found to fit
D(IJ) = 6.14 x 10-8 cos IJ
(7.5-6)
where D(IJ) is the dose rate in rads hr- ' steradian-I per unit thermalneutron-flux density at the concrete surface.
355
To approximate the capture gamma dose rate from this concrete, we

integrate over the solid angle that the surface subtends at the detector
(7.5-7)
where d(lI) is given by Eq. 7.5-6 and
i(>
is the thermal-neutron-flux density.
7.6 APPLICATIONS OF ALBEDOS

Albedos are invaluable in estimating the streaming through ducts, as will
be seen in the remainder of this chapter. They are also useful in dealing with
DETECTOR
SOURCE
Fig. 7.27-Wall scattering.
other interfaces, such as scattering from the ground in an air-ground

geometry or scattering from the inside surfaces within a shielded room. As
an illustration, consider the scattering from a wall, as shown in Fig. 7.27,
where the detector is shielded from the direct radiation but is exposed to a
wall that is irradiated by the source.
We assume a monoenergetic isotropic point source of N particles/sec.
The dose rate incident to wall-area increment ds is
D _N K(E o )
o-
411' rt
(7.6-1)
where K(E o ) is the flux-to-dose conversion factor. If we use type 1 dose

albedo data, then the dose reflected to the detector from the area increment
will be
(7.6-2)
356
The dose due to scattering from the entire surface will be
(7.6-3)
The integration over surface area in these equations would ordinarily be
carried out numerically. The dose at the wall surface has two components.
the uncollided dose arriving from the source and the scattered dose from the
wall:
D=D o [l +AD3(E o ,00)]
= Do
[1 + f02~ f: 12 CtD3 (Eo ,00,0,1j
sin 0 dO dlj>]
(7.6-4 )
The interface effect is then expressed by the integral.

The albedo of radiation from the floor, ceiling, and walls is often
significant to the total dose inside shielded compartments. For example, the
neutron dose at the center of an underground shielded compartment due to
neutrons penetrating the cover shield is almost a factor of 2 higher than
would be predicted on neutrons traveling directly from the shield to the
detector. Most of this factor can be predicted by performi ng an albedo
calculation over the inside walls.
7.7 DUerS
Passages through the bulk attenuation layers of a reactor shield are
ubiquitous. The working fluid must enter and leave the core. Controls and
instruments must have mechanical or electrical conduits leading outside.
Maintenance implies access to every region, and access implies door, plug, or
labyrinth. The shield designer is inevitably faced with the problem of
maintaining the integrity of the shield while providing workable means for
the necessary communicating pathways to the outside world. Each shield
penetration may be considered a duct, whether it is a 3-m length of i-em
tubing through the primary shield or a 1- by 2-m passageway with three
turns into the control foom.
Design and analysis of a straight duct is relatively uncomplicated. The
line-of-sight component is obviously the most important. Radiation entering
the duct from surrounding shield materials, either uncollided or scattered, is
ALBEDOS. DUCTS. AND VOIDS
357
another potentially significant component. Generally these components may

be treated with a ray-analysis technique (described in Sec. 7.9).
More-sophisticated approaches are needed for treating ducts with two or
more legs (one or more bends) where the scattered component predominates. Because of the importance of duct analysis in a shield design. a
number of techniques have been advanced which range from simpleminded
(with limited success) to quasi-mystical (adjust enough parameters and it will
fit) to thoroughgoing. detailed procedures.
As with other facets of shielding. notably those involving three spatial
dimensions. Monte Carlo seems the best method for accurate results.
However. a complete analog Monte Carlo calculation for a muItibend duct
geometry is usually an expensive undertaking. especially if parameter
variations are sought in an effort to optimize the design. Consequently
modified Monte Carlo programs have been developed which follow particle
paths only within the duct (not in the surrounding shield medium) and
employ albedo theory to estimate reflectance from the duct walls. In fact.
some simplified albedo methods that do not require Monte Carlo (or any
computer programs) have been devised with useful results.
We shall review the current literature on each of these methods to
provide some examples of the alternative approaches. The applicability and
validity of each method is usually demonstrated by comparison with
measurements.
7.8 LINB-OF-SIGHT COMPONENT

For a duct that penetrates straight through a thick shield and has
dimensions that are large compared with the mean free path of the
penetrating radiations. the line-oj-sight radiation (that which travels directly
from the source to the detector through the duct) usually is the most
significant portion of the total radiation reaching the detector. If one
assumes a plane source at the duct opening and a ratio of the duct length to
the diameter >1, all line-of-sight points are approximately equidistant
from the other end of the duct and may be treated collectively as an
equivalent point source. If the plane source is treated as having isotropic
current. the equivalent point intensity is the product SNo where No is the
particle emission rate per unit area and time and S is the area on the source
plane that is viewed through the duct from the point of interest. For
nonisotropic emitters each differential area element must be weighted by the
358
angular distribution when integration is performed over the source area. For
many geometries and source angular distributions that can be expressed
analytically, it is possible to derive formulas for the line-of-sight current
through an opening.
Several methods for computing the line-of-sight current in simple ducts
of various geometries have been developed. The source is assumed to be a
plane normal to the axis of the duct that covers the duct entrance. Two
types of source angular current distributions are considered: an isotropic
distribution and cosine distributions. Here isotropic distribution (as used in
the previous paragraphs) refers to radiation that is given off in all directions
in the forward hemisphere with equal probability but with no emission in
the backward hemisphere. Thus, if a surface emission is No cm- 2 sec-I, then
the number emitted into any unit solid angle is N o /2tr cm- 2 steradian- I
sec.- 1
For cosine distributions the radiation leakage through a unit area on the
surface of a self-absorbing volume-distributed source is most intense in the
direction normal to the surface, and the variation with the angle 8 from the
normal is approximated by a cosine function, in which case the number
emitted per unit solid angle is (No cos 8)ltr cm- 2 steradian- I sec-I.
Other functions have been used to describe sources that have more
strongly peaked angular distributions than that given by cos 8. In general,
most of these can be represented by cos" 8, where n may be as large as 20.
The cos" 8 function has been useful in representing a current (or a flux) at a
shield surface from an absorbing source region in which the activity per unit
volume increases with depth.
Two types of plane sources have been used in analytical expressions
derived for the transmission of uncollided particles through ducts. In one
type the source strength is defined by an emission rate, which we designate
No (cm- 2 sec-I). In the other type the source strength is defined by a flux
density that will be designated here as <1>0 (cm- 2 sec-I). The latter source is
the more natural one for use in shield design.
When calculating the flux density in a duct, one must take care to ensure
that the proper source angular distribution is used in a given situation since
there are two basic source descriptions. A particle emission rate is a current
and therefore is referenced to a unit of the source area, whereas a flux
density is always referenced to a unit of area normal to the particle
direction. Thus, to correct for the different orientation of the flux-source
reference area, one must multiply the flux density passing through that area
359
by cos IJ to project the reference area on the source plane; i.e., the relation
between the angular current J(IJ) and the angular flux (IJ) is
J(IJ)
=(IJ) cos IJ
(7.8-1)
The effect of the difference in the two source descriptions can be shown
by the following example. The uncollided flux density in a duct is to be
calculated for a given angular distribution g(cos IJ) cm- z steradian-I sec'
specified on the source plane. If g(cos IJ) is the normalized angular current
[the normalization condition on g( cos IJ) is f J g( cos IJ) 2IT d( cos IJ) = 1],
then the flux density at any point in the duct is
f.s
dS
r
No g(cos IJ)""2
(7.8-2)
where No is the total emission rate, dS is a differential area on the source

plane, S, and r is the distance between ds and the detector. If, on the other
hand, g(cos IJ) is the normalized angular flux density, then the flux density
in the duct is
i.s
cos IJ
<1>0 g( cos IJ) -z- dS

r
(7.8-3)
where <1>0 is the total or scalar flux density at the source plane.
A general relation between the emission rate and the flux density can be
derived for the very useful case of the cosn (} distribution by defining the
angular emission rate (angular current) as
J(cos IJ)
n+1
=z;;No
cosn IJ
(7.8-4)
That Eq. 7.8-4 IS properly normalized is shown by calculating the total

emission rate:
,1 n + 1
. - - No cosn IJ 2IT d(cos IJ) = No
.o
2IT
(7.8-5)
The corresponding expression for the total flux density, <1>0, at the source
plane is then
360
<1>0 =
J'
n+1
cos n e
n+1
- - No - - 2" d(cos e) = - - No
02"
cose
(7.8-6)
It is evident that Eq. 7.8-6 does not hold when n =0; in fact, the flux cannot
be defined at the source plane when the emission rate is isotropic.
For all the different duct geometries included in the following
discussion, the flux density, <1>, or leakage current, j, at the exit plane of the
duct is expressed in the same units as the source strength (particles cm- 2
sec-I). To obtain the total leakage flux density or current through the exit
plane of the duct, one must integrate or j over the exit area.
7.8.1 Rectangular Ducts
The work of Hubbell, Bach, and Lamkin'6 on radiation from uniform

rectangular sources with arbitrary angular distributions is directly applicable
to the calculation of the uncollided flux density along the axis of a
rectangular duct. Radiation crossing a plane S (cross-sectional area of duct
normal to axis) is assumed to be known and is designated by an arbitrary
function, g(cos e), which is the angular flux density and represents the
number of particles traveling in the direction e with respect to the duct axis
per unit solid angle, unit time, and unit area normal to direction (neutrons
cm- 2 sec-I steradian-I). The flux density at any point along the axis is
expressed by
(7.8-7)
where dO-see) is the differential solid angle subtended from an isotropic

detector by the surface ds (see Fig. 7.28) and is cos e dS/r 2
A solution to Eq. 7.8-7 was developed by Hubbell et al. in terms of
completely separable source and geometry functions and is expressed as
 =
'('\ 21 + 1
l.J
--2- g/PI(a,b)
(7.8-8)
1=0
where a is the ratio of the half height (H/2) to the length (Z) of the duct, b is
the ratio of the half width (W/2) to the length, and g/ and PI are coefficients
of Legendre polynomial expansions of the source and geometry functions,
respectively. With these definitions of a and b, Eq. 7.8-8 gives the flux at a
corner for the quarter plane limned in Fig. 7.28. The flux density at the
361
dS
I
I
I
I
I /
I .
/1
I
I
I
I
II
I
-r-----WI2
,DETECTOR
L ---
----,-"7
I
I
~PLANE
OF S
I
)
I / /
//1/
[.../
Fig. 7.28-Schematic drawing of source plane and detector position for a rectangular
duct.
center line for the full plane will be four times as great as that given by
Eq.7.8-8.
The coefficients are calculated by
gz
= f+1 g(cos 0) PI(cOS 0) d(cos 0)

-1
(7.8-9)
and
PI (a,b) = fs PI(cOS 0) dns(O)
(7.8-10)
where the values of PI( cos 0) are the Legendre polynomials.

Although the solution of Eq. 7.8-10 is in terms of a finite series, the
terms become progressively more complicated with increasing I, and only the
first four PI values were evaluated analytically. Hubbell et al. numerically
evaluated PI values for 0 '" I '" 13 over a grid of a and b values of
362
0.1.0;; b.o;; a.o;; 20. With these tabulated values and the g/values determined
for any arbitrary g(cos 0), the flux density along the axis or at the corner of
a duct can be calculated with Eq. 7.8-8.
For the special and useful case of the cos n 0 distribution,
g(cos 0) =
(n + 1) <1>0
cosn 0
2IT
(7.8-11)
where <1>0 is the source strength that is represented by a scalar flux density
obtained by integrating the angular flux density over all angles in the
direction of the duct entrance. Equation 7.8-11 is properly normalized since
n+1
-2- cosn 0 2IT d(cos 0) = 1
o
IT
(7.8-12)
For integral values of n, Eq. 7.8-11 can be expanded into a finite number
of Legendre polynomials, and, in principle, an analytical solution to
Eq. 7.8-8 can be obtained. However, the solution rapidly becomes unwieldy
for progressively higher values of n. The flux densities along the axis of a
rectangular duct for the isotropic source (n = 0) and for more forwardly
peaked sources (n = 1 and 2) from Maerker, Claiborne, and Clifford" are
given below. The flux density at a corner of a rectangle with dimensions
W/2 x H/2 is one-fourth that calculated by Eqs. 7.8-13 through 7.8-15.
For n = 0:
... _ 2<1>0
_,
... ---tan
IT
ab
'v!l + a 2
(7.8-13)
+ b2
For n = 1:
 _ 2<1>0 (
--:;r
~h +a 2
tan-'
~
+ J b 2 tan-' ~)
1 + a2
1+b
1+
(7.8-14)
For n = 2:
...
_'
... =3<1>0
- - [tan
IT
ab
_r==a~b==;;;;
VI + a2 + b2 + VI + a' + b'
As the distance between the source and detector becomes large with
respect to cross-section dimensions, the inverse tangent in Eq. 7.8-13
363
approaches ab and the terms in the parentheses in Eq. 7.8-14 and in the
brackets in Eq. 7.8-15 approach 2ab or WH/2Z 2 . In general, it can be shown
that for long thin ducts of rectangular cross section 3 s
 = (n + 1) WHo
211"Z2
(7.8-16)
Equation 7.8-16 represents a lower limit for a point source (emitting

only into the forward hemisphere) and also for a plane source (for a detector
far away from the duct entrance) since the flux scattered from the wall of
the duct is not included. In many practical cases the contribution to the dose
by the scattered flux will be of the same order or smaller than the
contribution by the uncollided flux. (The dose is calculated by summing the
product of the flux density and the appropriate dose response function for
each energy group of neutrons.)
7.8.2 Rectangular Slots
The geometry used in the derivations of the approximate equations for a
rectangular slot is shown in Fig. 7.29. The dimensions of the duct adjacent
to the source plane are W by H, and the distance from the source plane to
the exit end of the slot is Z. When H> W (say, H/W> 5), the exposed
source can be approximated by a line source, and the line-of-sight flux
density, <1>, and the leakage current, J, at any point P along the horizontal
axis on the exit end of the slot can be approximated by
(7.8-17)
(7.8-18)
<l>c=NoW (sinO, +sin02)

1I"Z
Jc
=~~;
[sin 20, + sin 20 2 + 2(0, +02)]
(7.8-19)
(7.8-20)
where No is source emission rate (or current) per unit area, 8 is in radians,
and the subscripts i and c refer to isotropic and cosine sources, respectively.
364
SOURCE
PLANE
Fig. 7.29-Rectangular slot geometry. Source plane and exit plane are normal to Z
direction.
If the slot dimensions are such that Ii 1 + li 2

become
.p.=NoW
J
2Z
.=NoW
JJ
"Z
..... ",
the preceding equations
(7.8-21)
(7.8-22)
(7.8-23)
Jc
=NoW
2Z
(7.8-24)
The rectangular slot can be considered a special case of the rectangular

duct for the condition when W ..... 0, and Eq. 7.8-19 can be derived by taking
the limit of Eq. 7.8-13 as W ..... O. In the limiting process an apparent anomaly
ALBEDOS, DUcrS, AND VOIDS
365
occurs; the answer is in terms of 4>0, the flux density at the source. However,
the flux densities on a line source for both isotropic and cosine distributions
are always infinite. Consequently, when the limit of Eq. 7.8-13 is taken as
w.... 0, 4>0 should be converted to an emission rate (2N o ).
7.8.3 Cylindrical Ducts
Consider a cylindrical duct of radius a normal and adjacent to a plane
source with the detector a distance Z from the source plane and on the duct
SOURCE
PLANE
Fig. 7.30-Geometry for cylindrical duct adjacent to plane source.
axis (Fig. 7.30). For an isotropic current or emiSSIon rate on the source
plane, the line-of-sight flux density at any point P is given by
and the current parallel to the axis of the cylinder is given by
1]
l;=NO [1 y'1 + (a/Z)'
(7.8-26)
For a cosine current source
4> = 2No
c
[1 - ./1
_;===I=;;;;;'ll]
+ (a/Z)2]
(7.8-27)
366
and
N
a'
Je = Z' [1 +o(a/Z)']
(7.8-28)
Equations 7.8-27 and 7.8-28 can be expressed in terms of an isotropic flux

density at the source plane by letting 2No = <Po.
For Z}> a, say, Z/a > 5, these equations approach
(7.8-29)
which is the same result for a point source of strength No (rra' ), and
(7.8-30)
If the condition of a long duct of small cross-sectional area is met,
Eqs. 7.8-29 and 7.8-30 will also hold for a straight duct of any crosssectional geometry if rra 2 is replaced by the cross-sectional area of the duct.
7.8.4 Cylindrical Annulus

The geometry for a long cylindrical annular duct is shown in Fig. 7.31.
The duct is bounded by two cylindrical surfaces having radii a[ and a,. For
the conditions Z a2 - at, a2 >- a2 - at, and an isotropic source No, the
average line-of-sight flux density,~, along the duct axis can be approximated
by
<P
=~
2rrZ'
[ ( 2a 2
2
- aJ2) cos-I
(aa,
l
- aJ .Ja2' - a[']
(7.8-31)
The average leakage current, J, for the isotropic source No can also be
obtained from Eq. 7.8-31 by setting M = 1. The equation given by Price,
Horton, and Spi nney 38 for the flux density at any radius r(a J ,,;; r";; a2)
within the annulus is
{1_(::)2 cos-[ C:)

+ cos-
C:) -(::) [1 -(::) T}
(7.8-32)
367
where M = 1 for an isotropic source, M = 2 for a cosine source, and rand Z

are the cylindrical coordinates of the point P at which the flux density is
given. Equation 7.8-32 is equivalent to
MNoS
4>( r,Z) = 211'Z2
(7.8-33)
where S is the area of the annulus on the source plane that is viewed from
the point P(r,Z) (Ref. 38).
Source plane
Fig. 7 .31-Geornetry for cylindrical annulus adjacent to plane source.
7.9 WALL-PENETRATION COMPONENT

In addition to the line-of-sight radiation, the flux reaching a detector
located near the exit end of a simple duct includes radiation that enters the
duct through its walls. t This contribution consists of two components:
tRadiation that enters the mouth of the duct in a direction other than the direction of the
detector but subsequently scatters in the duct wall toward the detector also contributes to the flux.
For simple ducts where the solid angle subtended by the walls at the detector is less than that
suhtended by the source, the contribution from the walls is usually small, but for ducts with bends it
becomes the dominant component (see Sec. 7.10).
368
radiation that travels directly from the source to the detector through
shielding material without an interaction (uncollided flux) and radiation that
scatters in the direction of the detector as a result of interactions with the
shield. Ray analysis is a useful method for calculating these components.
The assumption of the ray-analysis technique (actually a point-kernel
method) is that the radiation transmission is a function only of the path
SOURCE
PLANE
I~~~~~"-IT
a
'0 ----_..:::::::::J.i
p
Fig. 7.32-Geometry for ray-analysis calculation of radiation component arriving

through wall of duct.
lengths through each material or void encountered along a straight line

between the source point and the detector. This assumption is entirely
correct for predicting the uncollided flux density; for geometries in which
the scattered flux density becomes important, the collided flux densities can
be estimated to a first approximation by using buildup factors for gamma
rays and removal cross sections for neutrons. Although the accuracy of the
method is highly dependent on the particular configuration being considered, an approximation of limited accuracy for the scattered component can
be applied to simple ducts since the radiation reaching the detector consists
largely of the uncollided component.
To illustrate the method, we will consider a point detector P removed a
distance r from a point source at Q, where the straight-line path between the
two points passes through both shielding material and void (see Fig. 7.32).
369
The flux density at the detector is given by
(7.9-1)
where N p is the point-source strength and K is the material attenuation
kernel for all the materials located between the source and the detection
point.
If only rhe uncollided flux density at P is considered,
(7.9-2)
where l; is the mean free path of a particle and ti is the straight-line path
through the ith material. When the radiation being considered consists of
neutrons, 1/li is usually expressed in terms of the total macroscopic cross
section, ~t (cm-'), and ti is given in centimeters. When gamma rays are being
treated, 1/li is expressed as fJ.i.
Equation 7.9-2 can be modified to include an approximation of the
scattered gamma radiation by multiplying the right-hand side by a buildup
factor:
K = E(t, ,t2 ,
...)
exp (-
~ tiT.)
.
(7.9-3)
The equation can be modified similarly to include an approximation of

scattered neutrons by substituting the neutron removal cross section t ~ R
for ~t:
(7.9-4)
In this discussion we assume that the flux density at the point P is the
same as it would be if all the particles traveling from Q to P, including those
that are scattered, encountered the amount of attenuating material that
exists along the line from Q to P. If significantly less material is encountered
by some particles, the total flux density at P determined by Eq. 7.9-1 may
tRemoval cross sections are valid only when the shielding materials are followed by a thick layer
of water or, to a lesser extent, when mixtures containing hydrogenous materials are used (see Chap. 6).
370
be underestimated. Conversely, if significantly more material is encountered,

the flux density may be overestimated. Also, Eq. 7.9-1 neglects the effects of
material-void interfaces or interfaces between dissimilar materials
encountered along the ray and in the surrounding regions.
When the ray-analysis method is used for integration over a source plane
or source volume, any inaccuracy involved in the contribution of a source
region may be masked by the contributions from other points. There are, of
course, complex geometries and material combinations for which the simple
ray-analysis technique does not predict the radiation fluxes with a
satisfactory degree of accuracy. However, ray analysis is adequate in many
cases and is easily applied.
There are also cases in which the accuracy requirements and the
complexity of the geometry may prohibit the use of the approximations
given here except as first estimates. In such situations numerical integration
can be used to obtain a ray-analysis solution either by hand or by machine
calculations.
7.9.1 Application to Cylindrical Ducts
If the duct shown in Fig. 7.32 were a cylindrical duct of radius a
penetrating an infinite-slab shield adjacent to an infinite-plane isotropic
source, the uncollided flux density at P that would arrive through the duct
walls would be
<1>; = No
dr
e- I1 /(r)_
'0
r
(7.9-5)
where No :;;; surface emission rate or current

I(r) = path length in the shield that lies along r
=r
(1 __
-,=;pa~",,)
.Jr
Z2
2 _
11 :;;; linear attenuation coefficient
The corresponding flux density for a cosine source is given by

(7.9-6)
371
Equations 7.9-5 and 7.9-6 are not amenable to straightforward analytical

solutions, although values of the integrals can be found in tables of
mathematical functions. Certain approximations, based on a Taylor
expansion of I(r), have been given by Chase,' 9 but they are valid only for
large values of aiZ. Trubey'o evaluated the equations numerically for a
unit-surface-source intensity (for No = 1). The results are shown in
Appendix P, Tables P.1 and P.2, as a function of the ratio of the duct radius
to the length (aIZ) and the shield thickness (j.!Z), where j.!Z is measured in
relaxation lengths (or mean free paths determined from material cross
sections).
The data for the uncollided flux density given in Tables P.1 and P.2 can be
used to estimate the total radiation arriving through the duct wall provided
that the parameters al Z and j.!Z are measured in terms of an effective
relaxation length, A = 1/j.!, which accounts for the radiation scattered in the
direction of P by the shield. Relaxation lengths that include scattering can be
obtained from experimental data or from basic calculations of penetration
through the material. Relaxation lengths are necessarily a function of the
radiation spectrum as well as of the material surrounding the duct. Neutron
removal cross sections or gamma-ray buildup factors, both of which are
derived from basic penetration data, can be used in obtaining values of A. In
this use of the data, the approximation is made that the total attenuation is
exponential; i.e., the relaxation length does not vary with penetration depth.
Figures 7.33 and 7.34, which are based on data by Trubey,'o show rhe
variation with shield thickness and duct size of the total flux density,
including the line-of-sight component, reaching the center of the duct exit.
All parameters are given in terms of the shield-material relaxation length.
The flux density is given for an infinite-plane source of unit source strength
(e.g., 1 particle cm- 2 sec-I).
The relative importance of the line-of-sight radiation discussed in Sec. 7.8
and of the uncollided radiation that leaks through a duct wall is illustrated
by some additional calculations by Trubey in Appendix P, Table P.3. From
these calculations it is apparent that the ratio of the line-of-sight component
to the total uncollided flux density depends strongly on the ratio of the duct
radius to the length (aIZ) and on the shield thickness (j.!Z). These ratios
represent an upper limit since the buildup of the flux due to scattering has
not been considered. The ratios are applicable to either fast neutrons or
gamma rays provided the shield thickness is defined in terms of the number
of relaxation lengths. Data in Table P.3 show, as one would expect, that the
radiation that penetrates the side wall dominates over that from the source
372

2 x 10
lrP
IJZ = 0.5 Relaxation lengths
6
4
~Z=
10- 1
i...
6
4
10-2
6
4
I.,._L....l..:~;Z~.~5;::::::::;::::::r~Ll.,....-!-_-\-.j,.J~_+-_-\-l.JJ
1O- 3L
10.3
6 10-2
2
4 6 10'
2
4 6 10"
DUCT RADIUS, relaxation lengths
10'
Fig. 7.33-Effect of duct radius and shield thickness on total uncollided flux density
reaching exit of cylindrical duct (isotropic source).
in the mouth of the duct for small diameters penetrating thin shields. As the
duct diameter increases or the shield thickness increases, or hath, the

line-of-sight component increases in importance.
If the duct is filled with a lightly attenuating medium, formulas similar to
those in the preceding sections could he derived which include exponential
attenuation through the material filling the duct in the kernel integration.
This condition would tend to decrease further the importance of the source
at the mouth of the duct.
7.9.2 Application to Partially Penetrating Cylindrical Ducts
The radiation transmitted through the walls of a cylindrical duct that
penetrates only part of the shield and is not adjacent to the source can he
calculated hy the ray-analysis technique in much the same manner as for a
completely penetrating duct. Consider the duct shown in Fig. 7.35(a).
Equations 7.9-5 and 7.9-6 can he applied to this geometry merely hy
changing the lower limit of integration from ro to Z. As hefore, I(r) is that
portion of r lying inside the shield.
373
2 x 10
10
p.Z'" 0.5 Relaxation lengths
6
4
p.Z"'1
10. 1
0
~
"-
p.Z"'2
6
4
10- 2
6
4
p.Z"'5
2
1O-3'L."~~-+--!:..LL,,.-J,--+---!-..LL.,--J,--!--+..LL::---!:--!--+..LJ
10-3
10-2
6 10.1
100
101
DUCT RADIUS, relaxation lengths
Fig. 7.34-Effect of duct radius and shield thickness on total uncollided flux density
reaching exit of cylindrical duct (cosine source).
SOURCE
PLANE
SOURCE
PLANE
(a)
(b)
Fig. 7.35-Geometries for ray-analysis calculations of radiation component arriving

through walls of ducts that partially penetrate shield.
374
An alternative to numerical integration of these equations for each case

would be to determine the flux density penetrating an infinite slab of
thickness t (see Sec. 7.11.1), use this as a source-strength input to the due ted
shield of thickness Z - t, and evaluate using the data given in Figs. 7.33 and
7.34. In the absence of more definitive data, a more forwardly peaked
distribution, such as a cosine distribution, should be used at the artificial
interface if t is greater than one relaxation length. Even when the original
source is isotropic, the radiation will have taken on a more directional
character after penetrating one relaxation length.
Simpler equations may be given for the flux density through partially
penetrating ducts if the assumption is made that only the radiation
penerrating the end of the particle duct need be considered. 3 9 In this case
the flux density at P in Fig. 7.35(a) can be estimated by
(7.9-7)
where 'It = tan- 1 [a/(Z - t)l. Similarly, the flux density at P due to radiation
entering the mouth of a duct that penetrates only part of the shield but is
adjacent to the source [see Fig. 7.35(b)] is given by
(7.9-8)
where 'It' = tan-I (a/Z).
For the same shield and duct dimensions, the angle 'It' is smaller than 'It
and the peak flux density increase at P due to the presence of the duct is
greater in case b than in case a. For the approximation Ilta 2 <I:: 2Z2 , the
integrated emergent current can be approximated by lI'a 2 N o e-llt [a 2 /2(Z t)2) for either case a or b. Thus, although the peak current tends to be
greater in case b, the integrated current increase due to the void is
approximately the same in both cases.
7.9.3 Comparison with Experiment
The ray-analysis technique has been used successfully many times in
predicting measured values for both gamma-ray and neutron transmission in
solid un penetrated shields. In complex geometries involving ducts, the
technique should give better results for gamma rays than for neutrons since
neutrons on the average undergo more scatterings before they are absorbed
or escape. However, as is demonstrated by the following comparisons, the
375
technique yields good results for neutrons transmitted through straight ducts
>\butting a reactor.
Benenson and Fasano' I used the ray-analysis technique for analyzing
eocperiments at the Brookhaven National Laboratory Shielding Facility on
the transmission of fast neutrons through straight cylindrical ducts in a water
shield. Neutrons from a fission source plate in the shielding facility entered
the bottom of a water tank in which a duct had been vertically positioned,
and neutrons that reached the detector at the opposite end of the duct
arrived there either through the base of the duct or through its walls after
passing through the surrounding water. The ducts were 5.1 to 40 em in
diameter. The 32 S(I1,p)32 P reaction was used to measure the
fast-neutron-flux density, and an experimentally determined relaxation
length was used to analyze the penetration of neutrons above the 32 S
threshold from the fission source. Figure 7.36, which is typical of the results
of this work, shows that the ray-analysis method apparently yields good
results when applied to streaming of very penetrating radiation from a
diffuse source through simple ducts.
Piercey and Bendall 42 also used the ray-analysis technique for
calculating the flux density of fast neutrons transmitted through straight
cylindrical ducts. Their calculations corresponded to measurements made for
2.5-, 5-, and 9.S-cm-diameter aluminum ducts positioned in the water of the
LIDO Shielding Facility at Harwell, England. The duct lengths were up to
200 duct radii. As in the Brookhaven experiment, the fast flux density was
determined by the 32 S(I1,p)32 P reaction. The calculations predicted the
abs(!)lute sulfur reaction rate to within a factor of 2 over the experimental
values. The poorest agreement occurred in the region of the duct nearest the
reactor. Piercey and Bendall concluded that most of the discrepancy resulted
because of the inhomogeneity of the system and particularly because of the
aluminum layer between the duct face and the reactor. Figure 7.37 shows a
comparison of the calculated and experimental results as an example of this
work.
7.10 WALL-SCATTERED COMPONENT
We have seen in the preceding sections that for straight ducts the most
important components contributing to the flux density at the detector are
line-of-sight radiation and radiation that enters the duct through its wall at
the proper angle to reach the detector. For more complicated geometries
376

10 5......_-,-_ _- , , - - _ , -_ _.--_ _, -_ _......_ - ,
o Experimental
Li. Ray analysis
~
z
in
w
C
:>
-'
w
~
>
~
-'
w
a:
1021-_~_--;':__
_:!::-_:'::---~;__;:!:::--~
10
20
50
100
DISTANCE ALONG OUCT AXIS, em
200
500
Fig.7.36-Comparison of ray-analysis calculations with measurements of fast-neutronflux density along axis of 20-cm-diameter straight cylindrical duct in a water shield;
fission source at duct mouth. (From Benenson and Fasano.41 )
involving ducts with bends, the line-of-sight component disappears, and the
dominant component becomes radiation that is transmitted through the duct
by successive scatterings from the surrounding shield. Calculations of this
contribution require more involved techniques than those described
previously since on the average many scatterings are involved and several
boundaries are crossed.
7.10.1 Analog Monte Carlo Calculations
Two types of Monte Carlo methods can be applied to duct transmission
problems: analog Monte Carlo methods, described here, and albedo Monte

1O- 19
377
,------,.----------r--------,
0
10-20
'::::
,
E
,
E
10-21
."
~
w
ta:
>=
10-22
a:
a:
::>
~
-'
::>
"'
0
10-23
1()"24<-:--
10- 1
Experimental
Ray analysis
-'-;;-
-'-;
10
10 1
-'
102
DISTANCE FROM DUCT MOUTH, duct radii
Fig. 7.37 -Comparison of ray-analysis calculations with measurements of fast-neutronflux density along axis of 9.5-em-diameter straight cylindrical duct in a water shield; duct
mouth adjacent to reactor core. (From Piercey and Bendall. 42 )
Carlo methods, described in the next section. The term analog is used to
designate calculations in which the model solved is an analog of the
interactions occurring as the particles traverse the duct and confining walls as
opposed to the albedo calculations in which the particles traversing a duct
are followed by a random-walk process but wall interactions are represented
by a reflection coefficient.
378
The analog Monte Carlo method is not always practical because of the
machine computing time required; nevertheless, several machine programs
that have been developed arc applicable to this type of calculation.
Collins and McCl eary 43 used a Monte Carlo code (called LOS) to
calculate the transmission of neutrons from a Po-Be source (3.08 X 10 7
neutrons/sec) through 3D-em long, 7.6- and 1S.2-cm-diameter straight
10 2
Experiment
LOS calculation
E
w
I-
"
a:
w
'"
0
0
10'
I-
a:
::>
w
,:.
'"
u-
"
2
100!:---~
__-f:-_-:!::-__~__-:-_--!
10
15
20
25
30
DISTANCE FROM DUCT AXIS, em
Fig. 7.38----Comparison of LOS analog Monte Carlo calculations with measurements of

fast-neutron dose rates along the axes of 30-cm-tong, 7.6- and 15.2-cm..cliameter straight
cylindrical ducts in water; Po-Be neutron source. (From Collins and McCleary,43)
cylindrical ducts in water. The detector was positioned 7.6 em from the duct
mouth. The agreement they obtained between the measured and calculated
data (Fig. 7.38) shows that the LOS procedure predicts correctly for this
simple case.
The LOS code was then used by Marshall 44 for calculations
corresponding to measurements made of radiation transmitted through
one-bend cylindrical ducts in water. The ducts were reconstructed of
7.6-cm-diameter aluminum tubing and had one bend midway along their
379
lengths; the bend. angle vatied between 30 and 60. The sources were
14-MeV neutrons and 60Co gamma rays positioned on the duct axis 7.6 em
from the mouth, The detector at the opposite end of the duct was also
7.6 cm from the duct mouth.
10 3,
_ _, -_ _. ._ _, -_ _--,-_ _--,,-_---,
'e"
"0
w
I-
a:
8z
oa:
~ 10 2
w
Z
u.
8
6
o Experimental
L05 calculation
2 x 10~~50::-----::3~O----71O:---!:-O-:', O::---3~O::---;50'.:--"-.J
DiSTANCE FROM DUCT AXIS, em
Fig. 7.39-Comparison of LOS analog Monte Carlo calculations with measurements of

the fast~neutron dose rate transmitted through a bent (30) duct (7.6 em in diameter by
46 em long) in a water shield; 14MeV neutron source. (From Marshall.44 )
Figures 7.39 and 7.40 compare some typical experimental data and
calculated results for a duct bent 30 which penetrates a 46-cm water shield,
In general, Marshall found that for shield thicknesses that exceeded 46 cm
the calculated dose rates were in good agreement with the experimental data
in shape but were too low in magnitude.
380
lO J ,....--"'T""----.---'lO:>--.,..---,,..---...
""~E
w'
I-
"'w
~
o
>-
"'
'~"
10
"
6
o Experimental
.. L05 calculation
4
3 X 101':---~---:':---.J,------:':---'---~
-60
-40
-20
0
20
40
60
DISTANCE FROM DUCT AXIS, em
Fig. 7.40-Comparison of LOS analog Monte Carlo calculations with measurements of

the gamma-ray dose rate transmitted through a bent (30) duct (7.6 em in diameter by
46 em long) in a water shield; 60 Co gamma-ray source. (From Marshall. 44 )
7.10.2 Albedo Methods

Perhaps the simplest albedo application to a duct is the case where only
one reflection need be considered between the source and the detector. The
albedo is integrated over duct surfaces that have an unobstructed line of
sight to both the source and the detector. Consider the point source and
simple duct shown in Fig. 7.41. The dose at P due to firstorder reflection
from the wall is
D=
f
wall
surface
Np R (100) O'.Q '-4(-"# ~ ,1J-,</>L"o~.1J

7Trlr2
dS
(7.10-1)
381
=source strength
Eo = source energy
R (Eo) = flux-to-dose conversion factor
OlD 2 (Eo ,0 0 ,lJ ,<1 =differential-dose albedo (reflected current per incident
current) (see definitions in Sec. 7.2)
 = azimurh angle separating the incident and reflected rays
where
Np
and '" '2' 0 0 , and 0 are defined in Fig. 7.41. The reflected dose must, of
course, be added ro the line-of-sight dose to obtain the total.
s~r,
1
r~p
L
"I
Fig. 7.41-Geometry for albedo calculation of singly reflected radiation in a simple duct
(point source).
The Simon and Clifford45 equations for neutrons reJected through

ducts having cylindrical geometry (discussed later in this section) resulted
from an integration of Eq. 7.10-1 for an assumed functional form for 01. For
cases in which use of their expression for Oi is possible, the Simon-Clifford
equations can be used. Other expressions can be derived which would allow
the integration to be performed for simple geometries.
If either the nature of Oi or the geometry prohibits analytical integration,
Eq. 7.10-1 can be integrated numerically by dividing the duct wall into small
scattering areas and assuming that all points within an elemental scattering
area are the same distance from the source or detector and that they have
the same angles of incidence and reflection. A numerical evaluation
performed by LeDoux and Chilton 4 6 led to a set of equations for calculating
singly reflected gamma rays propagated through bent rectangular ducts. (The
LeDoux-Chilton technique is discussed later in this section.)
382
"
~P
Fig. 7.42-Geometry for albedo calculation of doubly reflected radiation in a simple

duct (broad-beam source).
For a broad-beam source incident on the duct at an angle such as that

shown in Fig. 7.42, no line-of-sight radiation reaches the detector and all
reflected radiation must be evaluated, including that which has been
reflected more than once. The singly reflected component at P is given by
Dp, =
Dj DiD 2 (EO
ov"
illuminated
, 0 ,0,1
cos 00 dS,
,2
(7.10-2)
area
where Dj is the incident dose and, is the distance from the surface area
element to the detector. The doubly reflected component is estimated by
DP2 = f
ov"
illuminated
area
fall
{[Dj DiD,(Eo,Oo,O,1>Ilcos 00 dS,
area
x DiD 2 (E"O,,02,1>2) cosO, dS 2 ] (rj,~)}-l (7.10-3)
Complexities of the problem will almost always require the use of a

computer program for numerical evaluation of these equations. Cain' 7 has
shown that for low-energy neutrons numerical evaluation of these integrals
becomes impractical because the number of reflection events that must be
383
considered to adequately calculate the penetrating radiation is so large. As an

alternative, a method was developed that traces particles through successive
reflection events by an albedo Monte Carlo technique (discussed later in this

section).
r-
w,
'1
Fig.7.43-Geomctry for determining corner-lip transmission effect in two-legged

rectangular duct. This geometry also illustrates the LeDoux-Chilton technique (discussed
later in this section).
In most albedo calculations of the transmISSIOn of radiation through

rectangular ducts with bends, the corner lip (the corner formed by the
intersection of two legs) is assumed to be opaque to radiation in the main
parts of the calculation, and a correction is made for its effect with a
separate approximate calculation. This correction is usually broken down
into two contributions: a transmission effect and an inscattering effect.
Corner-lip transmission is illustrated in Fig. 7.43 by the dashed line that
originates at the duct entrance and passes through the inside corner. The
radiation penetrating the lip scatters from the opposite wall and contributes
384
to the dose at the detector. [n accounting for this radiation. the albedo
integrations of Eqs. 7.10-2 and 7.10-3 can be performed with the incident
dose represented by a variable that depends on the path length in the
duct-wall material of a ray that is parallel to the initial direction of the
radiation and terminates at the scattering point. An approximation can be
made by assuming that all particles that penetrate the material with a path
length less than about a relaxation length are unattenuated and that all
particles that penetrate more than a relaxation length are completely
absorbed. (LeDoux and Chilton used the energy-absorption coefficient for
estimating this effective relaxation length for gamma rays.) On this basis, the
effect of the corner-lip penetration can be approximated by an increase in
the number of scattering areas beyond the primary scattering areas. The
areas designated as AI, A., As, A 6 All' and AI, in Fig. 7.43 show some of
these additional scattering surfaces.
Corner-lip inscattering is the process by which radiation is redirected
toward the detector by scattering in the corner. This component can be
evaluated with the last-collision technique of French,' 8 for which the
geometry is shown in Fig. 7.44. The flux density at a distance t I in the
corner lip is given by
(7.10-4)
where A(E o ) is the effective relaxation length for the incident radiation in
the material. The radiation per unit solid angle scattered toward P by the ith
element in the material is given by <P Kj(E,Oj Nj dv, where Kj(E,O) is the
differential angular scattering cross section of material i for radiation
incident at energy E and Nj is the atomic density of element j in the
duct-wall material. The flux density at P due to radiation of energy E
scattering in the lip is then
where E I is the energy after scattering and V is the scattering volume. This
technique is analogous to integrating a point kernel with the strength of the
source point being used as the differential scattering density at that point.
The scattering volume includes that portion of the corner for which t 1 + t,
is between one and two relaxation lengths.
385
(a) Simon-Clifford Technique for Cylindrical Ducts. The albedo

method of Simon and Clifford 4 s is one of the least complicated of the
methods available for calculating the transmission of thermal- and intermediate-energy neutrons through cylindrical ducts. One considers only the
neutrons entering the open end of the duct, and transmission is assumed to
fig. 7.44-Geometry for last-collision calculation of corner-lip inscattering effect in

two-legged rectangular duct.
be either by line of sight or by scattering from the duct walls. The wall
scattering is calculated with a single-energy spectrum-averaged material
albedo.
For a straight cylindrical duct, the albedo, A z , is defined as the fraction
of incident neutrons that is scattered from the wall (see Sec. 7.2). In this
form of the albedo, the reflected radiation has been integrated over all exit
directions, the reflected angular distribution being expressed in the general
form (il + 2'Y cos 8)/2.", where il is the fraction scattered isotropically, 'Y is
the fraction scattered with a cosine distribution, and il + 'Y = 1. The flux
density at distance Z along a duct of radius a is then given by
386
(7.10-6)
where <1>(Z) is in the same units as the surface source strength No.
The first term in this equation is the line-of-sight component given in
Sec. 7.8. The second term, which accounts for the wall-scattering component, is a strong function of the material albedo. Since A 2 is of the order
of 0.1 for fast neutrons and the second term is small for long thin ducts
(a ~ Z), Eq. 7.10-6 simplifies to that for the uncollided flux density.
However, for thermal neutrons, which have an albedo of about 0.8, the
second term becomes important, and the scattered flux can dominate.
'2a
Fig. 7.45--Geometry for calculating neutron transmission through cylindrical duct by

Simon~Clifford method.
Artigas and Hungerford'" improved the results of Simon and Clifford by

developing a more accurate version of Eq.7.10-6 in terms of integral
functions of the parameter A = a2 (Z2. The newer results are particularly
betrer for values of A greater than about 0.1.
Simon and Clifford' 5 also covered the case of a bent cylindrical duct
(see Fig. 7.45) by assuming that the albedo (in the area of the bend) of the
neutrons transmitted through the first leg of the duct is the source for
calculating transmission through the second leg. When both legs are of equal
diameter and their angular separation is 8, the total transmission at a
distance Z2 along the second leg is given by
<1>(Z2)
where
NoA;
4
(~) 2 (~) 2 (~ + 2{ sin

L,
Z2
SIn
A; is an empirically determined albedo parameter.
8)
(7.10-7)
387
If a series of bends is involved, the flux density at a distance Z from the

last bend may be expressed as
... [A~ (--"-V

(f3 + 2-y sin On-I)]
Ln-1J
8 Sin
(7.10-8)
where the duct segments are of lengths L I ' L 2 , . . , L n and are separated by
angles e1,8 2 , . 8 n- 1 .
Ducts that follow a smooth curve through the shield may be treated by
this method provided the radius of curvature is large with respect to the
radius of the duct. 3 8 The curved path may be divided into a series of
equivalent straight sections of length equal to the maximum chord that can
be drawn internal to the duct. The angular separation between successive
chords can then be used as rhe angle between the equivalent sections.
Horton S 0 treated helical ducts with an albedo approach similar to the
Simon-Clifford method. For a helical path divided into n equivalent sttaight
sections of radius a and length 1 deviating successively by constant angle 1/1,
the flux density is approximated by
'J
(7.10-9)
This can be compared with Eq.7.10-8, which would be the equivalent
Simon-Clifford expression provided all values of 0 equal 1/1 and all values of
L equal 1. The usefulness of this and other semiempirical albedo techniques
depends on the availability of suitable values of the albedos and angular
distributions of the reflected neutrons.
Horton and Hallidays 1 used the Simon-Clifford albedo method for
calculating the transmission of neutrons through straight cylindrical ducts.
The calculations corresponded to an experiment in which 5-, 10-, and
20-cm-diameter steel-lined ducts penetrated a water shield that was adjacent
to an extended plane source of thermal neutrons. A current albedo, f3, of
0.55 was used, and a cosine distribution was assumed. The comparison
shown in Fig. 7.46 demonstrates that the Simon-Clifford equation gives
good agreement with experiments in which the transmission of thermal
neutrons through cylindrical ducts is measured provided that reasonable
assumptions are made for the thermal-neutron albedos.
388
lO,-,r;--r-",---,--,--,-,,--,....-,-r-n
o
Simon-Clifford
(~" 0.551
... 3
I-~
"'l-Z 10-2
X
:JW
-' 0
U
~
o 20-cmdiameter duct
6.
15-cmdiameter duct
5-cm-diameter duct
_ _L-...L....J..:f:c_...J_ _..L.....J....J."*_ _.L..._...J.._.L...'"=:!
1O-4~_...J..
0.1
1.0
10
100
LENGTHTOD1AMETEA RATIO (Zla)
Fig. 7.46-Comparison of albedo calculations by Simon-Clifford technique with

thermal-neutron-flux densities measured in straight cylindrical steel*wall ducts through a
water shield; plane source of thermal neutrons. (From Horton and Halliday,5 I )
(b) LeDoux-Chiiton Technique for Rectangular Ducts. A method of

calculating gamma-ray transmission through two-legged rectangular ducts
was formulated by LeDoux and Chilton,'6 extended by Chapman, and the
same general technique was used by SongS for neutron transmission. This
group of investigators has applied a maximum of insight with a minimum of
input parameters to obtain reasonable agreement between calculation and
389
measurement. A review of this work is instructive because the role of each

component of the calculation is easy to follow, and the necessity fur adding
second- and third-order contributions becomes apparent. Equations were
derived for calculating radiation streaming down two-legged concrete ducts
based on (1) a single-reflection albedo for surfaces that are viewed by both
the source and the detector and (2) scattering by material in the corner lip of
the duct.
In the absence of sufficient differential albedo data at the time of this
work, a total albedo was used, and radiation was assumed to emerge
isotropically. To simplify the equations, they assumed that all points on a
rnajar scattering area (n umbered areas of Fig. 7.4 3) could be considered to
have the same angle of incidence from the source and the same reflection
angle to the detector. Formulas were then written for the detector response
due to first-order scattering from each of the surfaces. Later work 25 . 5 2.53
with more accurate differential albedo data showed that this method tends
to underpredict because multiple reflections are neglected. That the method
predicts measurements as well as it does is attributed to a compensating
overprediction caused by the assumption of isotropic scattering.
Chapman S4 extended the LeDoux-Chilton approach to include double
reflections and second-order effects, such as a wall backscatter followed by a
corner-lip inscatter. In piace of the total albedos with the assumption of
isotropy of the reflected radiation used by LeDoux and Chilton, Chapman
substituted the semiempirical formula for the differential-dose albedo that
was derived by Chilton and Huddleston 6 (described earlier in this chapter).
The two parameters in this formula were determined by a least-squares fit to
the Monte Carlo data of Rasa. 25
The number of the interaction combinations was enlarged (e.g., one
backscatter, one inscatter ,one backscatter plus one penetration, two
backscatters, and one backscatter plus one inscatter).

The singly reflected component was evaluated by integrating over each
of the numbered areas in Fig. 7.43. When penetration through the corner lip
was evaluated, the assumption was made that rays penetrating less than one
relaxation length of material are unattenuated and that those encountering
more than one relaxation length of material do not penetrate. Scattering in
the corner lip was calculated with the singly reflected formula that LeDoux
and Chilton derived on the basis of Klein-Nishina differential scattering
probabilities.
To perform the double-reflection calculations, Chapman divided the duct
into a larger mesh consisting of 12 scattering areas as shown in Fig. 7.47, and
combinations that would allow transport from the source to the detector
390
with two scatters were considered. Combinations whose fractional con-
tribution to the total dose was considered to be small were deleted.

Specifically, these were reflection from surface 1 to surfaces 10 and 12 and
reflection from surfaces 2 and 4 to surfaces 10, 11, and 12. The incidence of
singly reflected radiation on the corner lip of the duct was combined with
2,4
DUCT
ENTRANCE
10,12
DETECTOR
Fig.7.47--Scattering areas for albedo calculation of doubly reflected gamma rays in

two-legged rectangular duct.
the lip inscatter calculation to determine the contribution due to one

backscatter plus one inscatter. In all cases the energy degradation in a
reflection was assumed to be that of a single Compton scattering.
Comparisons with a number of measurements were made by
Chapman,S4 who used the extended LeDoux-Chilton technique for
calculating gamma-ray dose rates at the exit end of two-legged L-shaped
rectangular ducts, the source in each case being a gamma-ray-emitting
isotope at the mouth of the duct. Chapman also calculated the gamma-ray
dose rates along the axes of two- and three-legged rectangular ducts, and
comparisons are shown in Figs. 7.48 through 7.50 with the dose rates
measured at several points along the duct axes for three sources of different
energy.s s-s 7

10 0
L,=1.07m
391
L 2 "'1.22m
L 3 "" 1.68 m
If-
Experimental
o Albedo analysis
2f-
31--
4f-
51--
61--
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
DISTANCE FROM SOURCE ALONG DUCT AXIS, m
Fig. 7 AS-Comparison of albedo calculations with measurements of gamma-ray dose

rates along the axis of a O.3-m-square V-shaped concrete duct; 80-Ci 137 Cs gamma-ray
source. (Experimental data from Terrell et al. S 5; calculations from Chapman. 54 )
Figure 7.49 also includes results of an empirical formula developed by

Ingold and Huddleston S 8 from the experimental and analytical data of
LeDoux, Chilton, and Chapman. By data fitting, the formula was found to
be
D _
(H/W)0.907 W 2 . 864
Do - 3.44 LTs'4 L~667 Eg7IO
(7.10-10)
392
lOI=---r--,--,-,.--,---,--,--,.--=
L1 =2Am
Experimental
o Albedo analysis
I::J. Simplified formula
1O-5!:--~_--!:---'--!:--~:--~:---!:--7--~
o
2345678
Fig. 7.49-Comparison of albedo calculations with measurements of gamma-ray dose

rates along the axis of a 1.8-m-square L-shaped concrete duct; 3.67..ci 60 CO source. The
results designated simplified formula correspond to the Ingold-Huddleston equation
(Eq. 7.10-10). (Experimental data from Terrell et al. S 5; calculations from Chapman. s 4)
where D ~
Do ~
H~
W~
dose rate on axis of duct at point of interest

dose rate 0.30 m from the source
height of duct (m)
width of duct (m)
L [ ~ length of first leg measured from the source to the center of the
first corner (m)
L 2 ~ length of second leg measured from the center of the corner to
the detector (m)
Eo ~ gamma-ray source energy (MeV)
Use of this formula is limited to problems that meet the following

criteria: Point source at duct entrance, 0.662 MeV";;; Eo ,,;;; 6.0 MeV,
393
10"
L, '" 5.2m
L 2 ""5.8m
10-1
Experimental
Albedo analysis
10- 2
"Ci
10-3
_ _~_..L_~l-_L-_~_~_...I,...._~_~
2
3
4
5
6
7
8
9
'0
11
1O-6~_....L.._...J,
Fig. 7.50-Comparison of albedo calculations with measurements of gamma-ray dose

rates along the axis of a 1.8-m-.square L~shaped duct; 4.2-Ci 24Na source. (Experimental
data from Terrell et al. 5 7 ; calculations from Chapman. s 4)
0.3 .;; H';; 1.8 m,
0.3';; W';; 1.8 m,
0.6';; L , ';;11 m,
1';; H/W.;; 2,
LI/H';; 6, L 2 /H.;; 6, LI/W;;;' 2, L 2 /W;;;' 2. Although this formula is useful
for point sources, it should be applied with caution to a broad-beam source.

In calculating neutron transmission through rectangular ducts, SongS
used Chapman's modification of the LeDoux-Chilton method. The reflected
neutron spectrum was assumed to consist of two neutron energy groups, the
number and energy in each group being based on the Monte Carlo data of
Allen et aU In one analysis the duct was assumed to be 0.91 m square in
cross section, and the sources considered were 14.7- and 2.S-MeV neutrons.
The analysis included all single and double scattering events to determine the
394
energy and angular dependence of the neutron-dose albedo. A comparison of

the results with measurements made by Doty' 9 (Fig. 7.5 J) shows good
agreement. (The solid line represents the calculations for both the 2.5- and
the J4.7-MeV sources since there was very little difference between the two
sets of results.) Song derived a semiempirical formula for use in the
calculations in terms of a single-energy parameter for the differential
10-9
9
'"::>0
'"
"
'"
"
0
u.
...
'"
Albedo model calculation
'"
0
...
"...
a:
'"
0
0
0
"'"
Experiment. 14.7 MeV
Experiment. 2.5 MeV
10-8
100
200
300
400
500
DISTANCE ALONG SECOND lEG. em
Fig. 7.51-Comparison of albedo calculations with measurements of neutron dose rates

in the second leg of a O.91 m-square two-legged concrete duct; 2.5- and 14.7-MeV
neutron sources. In the experiment the source was 30 em outside the duct and
L 1 = 4.6 m; in the calculation the source was assumed to be at the duct entrance and
L, = 4.9 m. (From Song.')
M
neutron-dose albedo in a manner analogous to that used by Chilton and

Huddleston 6 in their gamma-ray calculations. Values for the energydependent parameter were obtained by a least-squares analysis that gave the
best fit to the data of Allen et ai. 2
395
(c) Albedo Monte Carlo Methods. Experiments and calculations have

shown that gamma rays produced by neutron absorptions in a concrete
shield surrounding a rectangular duct make a significant contribution to the
total dose in the duct, and analysis of this component requires a knowledge
of the distribution of low-energy neutrons throughout both the duct and the
shield. These distributions can be determined with accuracy by analog Monte
Carlo calculations, but, as pointed out in Sec. 7.10.1, the machine time
required to trace neutrons until they approach thermal equilibrium can
become excessive. To circumvent this, Cain developed a technique 47 that
uses a random-walk Monte Carlo approach in which particles are traced by
selecting random paths inside the duct but prohibits particle penetration into
the shielding material. Thus tracking through a large number of interactions
prior to emergence from the wall material is averted, and no time is wasted
in tracking particles that never emerge. Each time a particle encounters the
wall, it is assumed to reemerge with a reduced weight (given by the incident
weight times the albedo of the wall). Flux densities at various points along
the duct are then calculated either by making a statistical estimate for each
encounter to various point detectors or by determining the track length per
unit volume in a detector region. The most satisfactory detector was found
to be a thin region extending across the duct and perpendicular to the duct
axis.
The method was modified by Maerker and Cain 60 to include doubly
differential thermal-neutron current and fast-neutron-dose albedo data
(differential in reflected solid angle and energy) previously determined I .8.18
by Monte Carlo calculations for infinite slabs of concrete (described in
Sec. 7.3). In addition, a provision was included for calculating the
capture-gamma-ray dose rate arising from neutron capture in the duct walls.
In the modified code the incident neutrons are reflected with reduced
weight into various energy groups and directions, the probabilities for which
are predetermined from the doubly differential albedo data. The neutrons
are followed in the random-walk procedure until they either escape out the
front or the rear end of the duct or are killed by Russian roulette (Chap. 5)
when their weight falls below a predetermined value. Statistical estimations
of the fluxes to point detectors are made to obtain the results.
Good accuracy is obtained when the calculations are done separately for
three energy ranges: (1) Neutrons with energies from 200 ke V to 8 MeV are
divided into six energy groups and are sampled to obtain the fast flux
contribution to the dose rate. (2) The dose rate from neutrons of
intermediate energy (0.5 eV to 200 keV), including all neutrons that have
396
slowed down to intermediate and thermal energies from higher energy

groups, is determined with a 13-group calculation that involves sampling the
entire energy range from thermal to 8 MeV but scoring only those neutrons
with energies below 200 ke V. (3) The dose rate from source thermal
neutrons (subcadmium, <0.5 eV) is then determined in a separate calculation using a single-velocity model.
Monte Carlo albedo data of Maerker, Muckenthaler, Coleman, and
Stevens l 8 18 are used in each energy range as described earlier in the
chapter. To recapitulate, five sets of albedo data are used: one for each of
the three neutron-energy ranges given in the preceding paragraph plus two
for capture gamma rays (for incident thermal neutrons and for incident
intermediate neutrons). In this method no neutron is allowed to penetrate a
corner during the course of its random walk, although a statistical estimate
of the flux that includes a corner-lip penetration in the last flight can be
made.
Maerker and Muckenthaler18.61.62 carried out an extensive series of
experiments and calculations to test the accuracy of the Monte Carlo albedo
method as programmed in the AMC code. 6 Three configurations of a
0.91-m-square duct were used: a straight duct 13.7 m long; a two-legged duct
with a right-angle bend located 4.6 m down the first leg, the second leg being
9.8 m in length; and a three-legged duct with two right-angle bends, the two
bends being located 4.6 m down the first and second legs and the third leg
being 6.1 m in length. In the experiments a collimated neutron beam from
the Tower Shielding Reactor II entered each configuration through a 30- by
30-cm area located at the geometric center of the duct mouth. The neutrons
were incident at an angle 45 to one of the duct walls. This angle of
incidence was chosen to serve as a rigorous test of the calculation since the
fluxes and doses for detector locations greater than about a meter from the
duct mouth would then be due entirely to wall-scattered radiation. The measured absolute energy spectrum of the beam was used for the source spectrum in the calculations.
Figure 7.52 shows a comparison of the calculated and measured results
for fast neutrons. In general, the agreement averages better than 20%
through five orders of fast-neutron dose attenuation. For subcadmium or
thermal fluxes that result from an incident beam of sub cadmium neutrons
(Fig. 7.53), the agreement is within 20% in the second and third-legs and
within 5% for most of the straight ducts.
Calculations were also made of the thermal-neutron-flux density resulting from all source neutrons with energies above cadmium cutoff to test
397
the accuracy of the slowing-down model used in the code. The comparison
between calculations and experiment shown in Fig. 7.54 indicates that the
agreement is about the same as that obtained fot incident subcadmium
neuttOns. The statistical ettOts, number of wall backscatterings, and running
times fot the same number of source histories wete compatable in the two
5)(
1O-9!:-_-:,..-_~
__~__:':-_-:'~_--:'::::J
2
4
6
8
10
12
CENTER-LINE DISTANCE FROM MOUTH. m
Fig. 7.52-Comparison of AMC albedo calculations with measurements of fast-neutron

dose rates in one and twolegged O.91-msquare concrete ducts. (From Maerker and
Muckenthaler. 62 )
calculations. It should be noted that the calculations were normalized in

each case to the measured number of neuttOns incident on the wall of the
duct from the source beam.
Calculations were also made of the epicadmium or nonthermal spectra at
several locations along the center line of the three-legged duct, and these
398
,0
~Efm~
.
.=====
'0 0
5
g~
'D
I"'"
--1 46ml----
""'-
T=
"ll:
10~ 2
::l
-'
""
4=
Measurements
=====
-0
Calculated, straight duct
5=~
===0
I.Q.
Calculated, two-legged duct

Calculated, three-legged duct
1
10
12
'4
CENTER-LINE DISTANCE FROM MOUTH, m
Fig. 7.53-Comparison of AMC albedo calculations with measurements of thermalneutron-flux densities in one-, two-, and three~legged O.91-m-square concrete ducts due to
incident sub cadmium neutrons. (From Maerker and Muckenthaler. 62 )
spectra, together with the flux-to-multicollision dose factors of Snyder and

Neufeld: 3 were used to calculate the epicadmium multicollision dose rates
at the same locations. The results are compared in Table 7.5 with
measurements made with a dosimeter designed to give a response proportional to the Snyder-Neufeld multicollision dose curve over the range of
interest. The dosimeter was a spherical BF 3 counter placed inside a 25- or
3D-em-diameter polyethylene ball covered with cadmium.
The method of Cain and Maerker (AMC) may be extended to cover any
other duct arrangements by providing an appropriate geometry analysis
routine. Calculations in cylindrical geometry have been made by Selph,
Newell, and Bayer,64
7.10.3 Additional Experimental Investigations
A selection of experimental data has been presented in the preceding
discussions for the purpose of validating analytical or empirical methods of
duct analysis. Some additional experiments can be cited which help to
illustrate principles involved in duct penetration.
399

10 1
10
~.5+m
--!4.6m!- T
'" --0...
.......
'l
....... :--.....U
x
::>
'"
'"
::>
~"'-
"'::>
1---0
10- 51===6
Measurements
Calculated, straight duct
ri:
""'--'
"'-
Calculated, twolegged duct
~o
::;
~
""-
lO- 4
0 _0 _
10- 3
Calculated. three-legged duct
lO- 6
10- 7
10
CENTER-LINE DISTANCE FROM MOUTH,
12
14
Fig. 7.54-Comparison of AMC albedo calculations with measurements of thermalneutron-flux densities in one-, two-, and three-legged O.91-m-square concrete ducts due to
neutrons having energies above cadmium cutoff. (From Maerker and Muckenthaler.
62
Table 7.5-CALCULATED VS. MEASURED EPICADMIUM MULTICOLLISION

DOSE RATES IN A THREE-LEGGED DUCT
Distance
from mouth
Duct
to detector.
leg
2
2
2
5.8
7.0
8.8
10.0
12.2
2
3
Measuredt
25-cm ball
2.5
5.5
1.28
1.45
1.08
30-cm ball
1.35
2.85
6.6
6.4
X 10'8(1.2 X 10'8)
X 10'9(9.0 X 10'10) 6.0
X
10'6(2.1
10'6)
X 10'6(4.9 X 10'7)
X 10'7(1.2 X 10'7)
Calculated
X 10'6(1.1 X 10'6)
1.9
X 10'7(2.45 X 10'7) 4.5
X 10'7(6.1 X 10'8)
1.05
X 10'9(4.8 X 10'9)
9.2
X 10'10(5.1 X 10'10) 6.5
X 10'6
X 10'7
X 10'7
X 10'9
X 10'10
tThe dose rates in parentheses are those obtained when the center of detection was
assumed to be at the center of the polyethylene ball rather than at its leading edge.
400
Terrell, Jerri, and Lyday" performed experiments to investigate the

effect of geometry and source energy on the attenuation of gamma rays
through rectangular concrete ducts with two right-angle bends. They found
that for l.B-m-square concrete ducts it made little difference whether the
gamma-ray source was 60Co or '37 Cs or whether the right-angle bends
formed a U-shape (Fig. 7.55a) or a Z-shape (Fig. 7.55b). These data, plotted
in Fig. 7.55c, show a maximum deviation of a factor of 1.5 when properly
Fig. 7 .SSa-Geometry of U-shaped duct.
normalized. The effect of geometry and energy becomes more noticeable for
a 30-cm-square duct.
Chapman 6 ' calculated the contribution of a given scattering area on the
gamma-ray dose rate in an L-shaped 0.91-m-square concrete duct. The
scattering area chosen was the one labeled S, in Fig. 7.56a. Gamma rays
from a 60 Co source at the mouth of the duct were scattered from thi' area
to the detector at the exit end of the duct. The dose rate due to gamma rays
that were initially scattered from S, was calculated to be 0.379 mr/hr, which
compares favorably with a measured dose rate of 0.404 mr/hr.
During the experiment Chapman measured the gamma-ray spectrum at
the exit end of the duct both with and without the scattering area S,
shadowed from the source. The results are shown in Fig. 7.56b. Shielding of
401
s,
had a more pronounced effect on the higher energy radiation than on the
multiply reflected low-energy component. This effect would be expected
since the percentage reduction in the single-reflected area that results from
blocking s 1 is greater than the percentage reduction in possible combinations of multiple-reflection areas.
~~~iiJ:~~lf~:~l~'~~~~~J.l~1J~~(4~~1r~~!~>~
r--
2 4m
.
--l
3m
Fig. 7.55b-Geometry of Z-shaped duct.
In a subsequent experiment Chapman and Grant 66 investigated the

relative effectiveness of coplanar and noncoplanar three-legged rectangular
ducts. For the cases investigated (28-cm-square duct, 60Co source,
L, = 75 em, L 2 = 80 em, L, = 91 em), the attenuation provided by the
noncoplanar configuration was found to be greater by a factor of 2 than that
provided by the coplanar configurarion.
Clifford s 2 used differential albedos 2s to calculate dose rates corresponding to those measured in experiments designed to determine the importance
of multiple reflection on the transmission of " 7 Cs gamma rays through 20-,
402
10 3
137 Cs source:
'10
~
10'
'0
~ !J 1ls"I:!
I!l!. !Jl!
..'"
11 Z-shape
U-shape
1l
BOCa source:
o U-shape
Z-shape
I-
10'
"'00
w
.
>
;:::
-'
10
'"
10-2 L_---l,.-_...L_ _.J,-__!:-_--:':-_-:':,.-_:'.

o
2
4
6
B
10
12
14
Fig. 7.55c-Gamma-ray dose rates measured in l.B-m-square U- and Z-shaped concrete

ducts; 60 CO and 137 Cs gamma-ray sources. (From Terrell et al. S 5 )
s
I
"" "
'j
"", /
L,
SOURCE
L,
Fig. 7.56a-Schematic draw ing of a O.91-m-square L~shaped concrete

duct (L, ,L 2 = 2.3 m).S, is t he area on the long side of L 1 which can be
VIewed from D; It 15 0.91 by 1.15 m.

10'
,----,---r--...,--...,..--,---..---,......--,
j \
"/'
,/\ '\
\
, \.
"
\
i.'
- -
Unshadowed
- - - - Shadowed
- - - Difference
\'
\ \\
.f \"\,\\
,,'--
/,
C
,
403
"
h
'I
-,"
o
u
'
'\
\,\
""
P
II
\\
I/~
,
,
,
,,"
"''''
'I "
'II,
7 x 10- 2 L..._--'_ _.....L_ _....._ _.,.,.,_ _..L-_-:~__' - _ ~
100
200
300
400
GAMMA RAY ENERGY, keV
Fig. 7.56b~Gamma-ray energy spectrum at exit of O.91-m-square L-shaped duct with

and without scattering area S 1 (see Fig. 7 .56(a)] being shadowed from
(From Chapman"')
60
Co source.
404
30-, and 91-cm-square concrete ducts. In the experiments collimated gamma

rays were incident on the duct entrance at an angle to the duct axis such that
the area in which the first reflection could occur was near the entrance of
the duct (see Fig. 7.57a). The dose rate along the duct center lines was
measured first with the unmodified concrete (condition A). The second
measurement (condition B) was with a lead lining over all inside surfaces
except the first scatter area. The third measurement (condition C) was with a
lead lining over the entire inside surface. Figure 7.57b shows the results of
the measurements for all three conditions in the 20-cm duct together with a
single-reflection analysis based on the differential albedo for condition B,
and Fig. 7.58 shows the fraction of the total dose which is due to reflection
from surfaces not exposed to the primary radiation for all the ducts. It is
apparent from the results of all three duct sizes that the multiply reflected
component increases to become as much as 20 to 30% of the total dose as
the duct length is increased.
7.11 VOIDS
An irregularity that increases the radiation penetration through a shield
but does not extend to either surface is called a void. Under this definition
we consider single voids in a shield, such as an air bubble in a concrete pour,
or multiple small voids created by filling a container with lead shot. Interest
in an accidental void is likely to be centered about repair or determining the
patch size to restore shield integrity.
7.11.1
Single Voids
The technique most frequently applied to single voids is the ray-analysis

technique, which is described in Sec. 7.8 in connection with duct transmission problems. As is pointed out there, the assumption of the ray-analysis
technique is that the radiation transmission is a function only of the path
lengths through each material or void encountered along a straight line
between the source point and the point of interest. When applied to voids
the increase in the flux density at the shield surface is determined by the
difference in intensities calculated with and without the void.
Consider a disk-shaped void in an infinite-slab shield adjacent to an
infinite-plane gamma-ray source with the planes of the disk parallel to the
slab planes as shown in Fig. 7.59. Let <P(L) be the flux density of energy Eat
a point on the surface of the shield with no void and <P(L - t) be the flux
FIRST SCATTER AREA
405
.,..-DUCT WALLS. CONCRETE

/'
DR LEAD-LINED
Fig. 7.57a-Schematic drawing of square

concrete duct showing first scatter area.
1ooo,----,-----,-----,---,..---,---..,.--......,
600
400
200
AU lead-lined except first scatter area (B)
~
~
100
8
w
60
"
40
Calculated (8)
>
;::
-'
0:
20
10
6
42!::5--~5;';0;----:!7::-5----:,0:!:0;--~1:-!2:::5----:,!:o50:--~1~7:::5---::!2oo
DISTANCE ALONG DUCT AXIS, cm
Fig. 7.S7b-Measurements and single-reflection analysis of gamma-ray dose rates along

the axis of a 20-cm-square concrete duct with and without lead linings; 137 Cs source.
(From Clifford.' 2)
406

0.'
20-em duct
30-em duct
91-cm duct
0.3
....
'a:"
Ww
'"0a:
....
0",
Ow
~ ~O.2 f-
1=-'
'>-" ....'"
u ....
",0
-'
i=
-'
:::>
:;;
<It
WO
0.1 t-
tI
2
,I
10
DISTANCE FROM DUCT ENTRANCE, duct width
Fig. 7.58-Ratio of multiply scattered gamma-ray dose rate of total dose rate along the
axes of 20-,30-. and 91-cm-square concrete ducts; 137 Cs source. (From Clifford. s 2)
p'
Fig. 7.59-Geometry for calculating the effect of a disk-shaped void on radiation

transmission through a shield.
407
density with the void present. For the latter, the point of interest is the
intersection of the center line of the void with the shield surface. The
influence of the void is estimated by performing a point-kernel integration
over the source plane to obtain 4>(L) and 4>(L - I). To calculate the flux
density for the plane shield of thickness L, define b I = ~, Mjlj over L
normally through the shield. The contribution of a ring-shaped source with a
particle-emission rate of No at an angle (j from the detector would then be
(7.11-1)
Integrating over the plane source to an angle
(j
gives
e- f
-dl
(7.11-2)
where I = b I sec
(j.
Since by definition
EJlx) =
-f
'!- dl
(7.11-3)
then
(7.11-4)
integrating over the entire source plane gives
(7.11-5)
If the gamma-ray number buildup through hI is B I
then 4>(L) may be
estimated by
(7.11-6)
The buildup factor used here as a simple multiplier is only symbolic since
the buildup factor is really part of the kernel that is integrated over all space.
408
As is pointed out in Chap. 4, when the Taylor form of the buildup factor is
used, the form of these equations is unchanged, and the buildup effect is
simply included by modifying the arguments of the exponential functions.
Actually in these problems it is sufficiently accurate to use the equation as it
stands and to determine B, only on the basis of the slant penetration (Q to
P) that grazes the void. For B 2 , an average of B I and of the buildup factor
based on the minimum path length normal to the void is usually adequate.
Use of the arithmetic mean will always produce a conservative result since
this gives equal weight to the buildup factor from a region of lesser
importance.
To calculate the flux density for a shield in which a disk.shaped void

intercepts at an angle e from the point P, define b 2 = ~ IJ.jfj over (L - f).
Then for angles less than
(7.117)
and for angles greater than
NoB I
b
cI>2 = - 2 - E I ( I sec e)
(7.11.8)
Thus the total flux density is

NoB,
cI>(L - f) = cI>1 + cI>2 = - - E, (b, sec e)
2
N o B 2 [Ed b 2) - Ed b 2 sec e)]

+-2-
(7.119)
and the difference cI>(L - f) - cI>(L) is the increase in the flux density at
point P due to the presence of the void. Since P is centered over the void, it
will be the point of maximum increase.
For points offset from the center, for example, for point p' in Fig. 7.59,
Eq. 7.119 will not be valid. If the source is isotropic, the increase in the
contribution at p' by a source surface increment dS caused by the void is
given by
409
where B 3 and B 4 are the buildup factors along the path dS -+ p' with and
without the void, respectively. If the source is a cosine source, the increase is
given by
(7.11-11)
The integral of Eq. 7.11-10 or 7.11-11 over the appropriate source area will
give the total increase at p' due to the presence of the void. The appropriate
source area includes all points for which a ray p' passes through the void. In
most practical situations this integral must be evaluated numerically. The
value of the integral at representative points will serve to determine the
shape and thickness of the shield patch required to negate the effect of the
void.
An approximate form of ray analysis given by Tonks 7 can be applied to
certain problems involving voids or depressions in a very thick shield. With
this method we first calculate the flux density at a point on the surface of a
uniform infinite-slab shield whose thickness is equal to the ray of minimum
length through a void or depression in the shield. We then account for the
reduction in flux density due to the longer material path of nearby rays
terminating at the same point. The method is based on the fact that the
formula for the flux density through an infinite slab
(7.11-12)
can be approximated by
(7.11-13)
for jJ.t:>- 1, with an error that is approximately equal to l/jJ.t.
The application of Tonks' approximation is as follows: Consider a slab
shield with a void such as that shown in Fig. 7.60. Let the ray passing
through Q and P be the path of minimum material penetration, to, and let it
coincide with the z-axis. In many cases nearby rays may be expressed as
(7.11-14 )
410
p
SOURCE
PLANE
Fig. 7.60-Geometry for calculating by Tonks' approximation the effect of a void in a
shield.
where Ox and Oy are the angles between the minimum ray and a nearby ray
projected onto the x-y and y-z planes, respectively, and a and bare
determined by the specific geometry. Substituting Eq.7.11-14 into an
approximation analogous to Eq. 7.11-13 yields for the flux density <1>' at P
(7.11-15)
if both a and b are greater than zero. The ratio of the flux density estimated
by Eq. 7.11-15 to that estimated by Eq. 7.11-13 (for a slab of minimum
penetration) is
(7.11-16)
which, providedyav > 1, gives the increase in effectiveness at P of the shield
with a void over that of a simple slab whose thickness is equivalent to the
minimum ray.
The limit of Ox and Oy for which Tonks' approximation is valid is set by
Ox - Oy < C/, where C/ is given by C/ =..j2Tiit;;. As an example, for a spherical
void of radius R v whose center lies at a depth R from the outside surface of
a slab shield of thickness T, it can be shown that the length of a ray near the
minimum ray is given by
= t o + ( _T2
R2)
411
t
+ - fl2 +-2f12
Rv
(7.11-17)
2 Y
and the increased effectiveness at the minimum ray terminus is given by

(7.11-18)
Other expansions of t = to + [(flx,fl y ) may be derived which would lead

to equations of similar form to Eq. 7.11-15. Kouts 68 treated the problem of
p
SOURCE
PLANE
Fig. 7.61-Geometry for calculating by Kouts' technique the effect of a void in a shield.
voids on the basis of one-velocity transport theory. Instead of using the

boundary conditions of the differential transport equation to describe the
void, he accounted for it by the way in which the kernel of the integral
equation is defined. The effect of the void is then calculated as a
petturbation of the flux density in a homogeneous medium (no-void case).
The method is applicable to either gamma tays or neutrons.
Using this approximation to the transport equation, Kouts investigated
the effect of spherical voids in water and obtained good agreement with
experiment when he chose certain constants in the solution by fitting to the
experimental data at one point. The exptession derived for a spherical void
in the geometry shown in Fig. 7.61 has the form
412

2
(P)
where (P)
B V ('
=:2
Jo dl1l1 k exp (2RI1
~)
(7.11-19)
flux density at point P
B = constant in the approximation to the angular distribution of
the scattered flux used by Kouts, i.e.,
dn (np;t p(n ... n) =
Beas 0
(0 <0 ,.;;~)
=0 (~<0<1T)
~t ;;; total macroscopic cross section
p(n...n) = probability of scattering from direction
n'
to direction
n
V = parameter in the expression for flux distribution through the
shield in the absence of voids given by <1>0 = V(n) e- 3 / b
R = radius of void
L = shield thickness
b = number of relaxation lengths in the shield material
11 = cos 0
A best value of k = 4.5 was chosen by Kouts by fitting experimental data for
gamma rays at 2R/L = 1.
Results obtained with this equation for a water shield are compared in
Fig. 7.62 with experimental data for gamma rays.
7.11.2 Small Random Voids
For a large shield containing small randomly spaced voids (small
diameter relative to the relaxation length of the radiation, e.g., a gamma-ray
shield consisting entirely of iron shot), the effect of the voids on the
radiation transmission is obtained simply by increasing the relaxation length
used in the calculation of the shield attenuation by the ratio of the average
density of the shield to the density of the solid material used in the shield.
When the voids are not negligibly small, however, it can be shown by a
statistical argument that the penetrating radiation will be greater than would
be predicted by using an average density. Consider two normal rays through
the shield, one that is the average penetration t plus an increment S and one
413
that is shorter than average by the same increment. Their average penetration
is
Thus, although their average penetration distance is

penetration is greater than e-lll .
I, their average
2.2 ~--......---~--......---.,......--....,
-----
Cl 2.0
0
_ 0
>
0-
Theoretical (Kouts)
0-0
i=
18
'S"
------""
0-
v.1J
0>
",,,,
1.6
,/
I.
/.
;'
;'
;'
~
o
~ 1.4
a:
'"-'
U-
1.2
;';'
/'
Experimental
/,/'
1.0 ~--~---:L:---:':--:-~--~.
0.2
0.4
0.6
0.8
1.0
SPHERE DIAMETER, relaxation lengths
Fig. 7.62-Comparison of flux increase due to spherical void in a water shield calculated
by Kouts' technique with that determined by experiment. (From KoutS. 68 )
This problem has been treated statistically by Smith." 9 From his work
an effective thickness, Ie, can be derived:
I = 1(1 - v) _
e 1 + 0.55sv2
(7.11-20)
where I is the actual thickness, v is the fraction of the volume that is void,
and s is the average distance between voids measured in relaxation lengths.
414
An equivalent expression for an effective attenuation coefficient is
/1(1 - v)
/1e = 1 + 0.55 (s//1) v2
(7.11-21)
or, generalized to the case where the voids contain a material of attenuation
coefficient Il',
_
/1' + (1 - v)(/1- /1')
/1e - 1 + 0.55 (5//1) (/1 - /1') v'
(7.11-22)
REFERENCES
1. R. E. Maerker and F. J. Muckenthaler, Calculation and Measurement of the
Fast-Neutron Differential Dose Albedo for Concrete, Nucl. Sci. Eng., 22: 455 (1965).
2. F. J. Allen, A. Futterer, and W. Wright, Neutron Reflection and Flux Versus Depth
for Nevada Test Site Soil, Report BRL-1190, January 1963; ... for Concrete, Report
BRL-1l89, January 1963; ... for Iron, Report BRL-1l99, March 1963; Neutron
Reflection and Flux Versus Depth for Water, with an Appendix: Comparison with
Results of National Bureau of Standards, Report BRL-1204, June 1963; Neutron
Reflection and Flux Versus Depth for Aluminum, with an Appendix: Neutron Dose
Transmission Versus Thickness, Report BRL-1238, February 1964, Ballistic Research
Laboratories.
3. F. J. Allen, A. Futterer, and W. Wright, Dependence of Neutron Albedos upon
Hydrogen Content of a Shield, Report BRL-1224, Ballistic Research Laboratories,
October 1963.
4. R. L. French and M. B. Wells, An Angular Dependent Albedo for Fast-Neutron
Reflection Calculations, USAEC Report RRA-M31, Radiation Research Associates,
Inc., November 1963.
5. Y. T. Song, Fast Neutron Streaming Through TWO-Legged Concrete Ducts, Report
AD-457746 (TR-354), Naval Civil Engineering Laboratory, Feb. 2, 1965.
6. A. B. Chilton and C. M. Huddleston, A Semiempirical Formula for Differential Dose
Albedo for Gamma Rays on Concrete, USAEC
number NP-12455 (TR-228),
Naval Civil Engineering Laboratory, Nov. 16, 1962; also Nucl. Sci. Eng., 17: 419
(1963).
7. R. L. Henry, L. G. Mooney, and R. J. Prevost, Study of Radiation Penetration and
Reflection from Shield Materials, Report FZK-183, General Dynamics Corporation,
August 1964.
8. W. A. Coleman, R. E. Maerker, F. J. Muckenthaler, and P. N. Stevens, Calculation of
Doubly Differential Current Albedos for Epicadmium Neutrons Incident on Concrete
and Comparison of the Reflected Subcadmium Component with Experiment, Nucl.
Sci. Eng., 27: 411 (1967).
me
415
me
9. E. Fermi, On the Motion of Neutrons in Hydrogenous Substances, USAEC

number NP-2385 [translated from Ric. Sci., VII(2): 13 (1936)].
10. o. Halpern, R. Lueneburg, and O. Clark, On Multiple Scattering of Neutrons.
I. Theory of the Albedo of a Plane Boundary, Phys. Rev., 53: 173 (1938).
11. S. Glasstone and M. C. Edlund, The Elements of Nuclear Reactor Theory, D. Van
Nostrand Co., Inc., New York, 1952.
12. S. Chandrasekhar, Radiative Transfer, Clarendon Press, Oxford, England, 1950.
13. L. V. Spencer, J. A. Diaz, and E. Moses, Neutron Penetration in Cylindrical Ducts,
Report NBS-8542, National Bureau of Standards, Sept. I, 1964.
14. A. Mackel, Reflection and Transmission by a Strongly Absorbing Slab, Nucl. Sci.
Eng., 22: 339 (1965).
15. G. C. Pomraning, The Albedo Problem, Nucl. Sci. Eng., 21: 265 (1965).
16. M. B. Wells, Reflection of Thermal Neutrons and Neutron-Capture Gamma Rays from
Concrete, USAEC Report RRA-M44, Radiation Research Associates, Inc., June 1964.
17. M. B. Wells, Gamma Dose Rates Resulting from Neutron Capture in Air and Concrete,
USAEC Report NARF-59-31T, Convair, September 1959.
18. R. E. Maerker and F. J. Muckenthaler, Measurements and Single-Velocity Calculations
of Differential Angular Thermal-Neutron Albedos for Concrete, Nucl. Sci. Eng., 26:
339 (1966).
19. H. Greenspan and I. G. Baksys, Addenda to Newsletters Nos. 3 and 4, Argonne
National Laboratory Newsletter No.9, USAEC Report TlD-18481, March 1963.
20. M. Leimdorfer, The Backscattering of Gamma Radiation from Plane Concrete Walls,
Swedish Report AE-92, December 1962.
21. M. B. Wells, Differential Dose Albedos for Calculation of Gamma-Ray Reflection
from Concrete, USAEC Report RRA-T46, Radiation Research Associates, Inc.,
December 1964.
22. M. J. Berger and J. Doggett, Reflection and Transmission of Gamma Radiation by
Barriers: Semianalytic Monte Carlo Calculations, ]. Res. Nat. Bur. Stand., 56: 89
(1956).
23. M. J. Berger and D. J. Raso, Monte Carlo Calculations of Gamma-Ray Backscattering,
Radiat. Res., 12: 20 (1960).
24. M. J. Berger and D. J. Raso, Backscattering of Gamma Rays, Radiat. Res., 12(1):
20-37 (1960).
25. D. J. Raso, Monte earlo Calculations on the Reflection and Transmission of Scattered
Gamma Rays, Nucl. Sci. Eng., 17: 411 (1963).
26. M. B. Wells, Air and Concrete Scattering of Gamma Rays, USAEC Report
NARF-59-11T (MR-N-229), Convair, March 1959.
27. A. B. Chilton, C. M. Davisson, and L. A. Beach, Parameters for C-H Albedo Formula
for Gamma Rays Reflected from Water, Concrete, Iron, and Lead, Trans. Amer. Nucl.
Soc., 8: 656 (1965).
28. C. M. Davisson and L. A. Beach, A Monte Carlo Study of Backscattered Gamma
Radiation, Trans. Amer. Nucl. Soc., 5: 391 (1962).
29. A. B. Chilton, A Modified Formula for Differential Exposure Albedo for Gamma
Rays Reflected from Concrete, Nucl. Sci. Eng., 27: 481 (1967).
416
30. C. E. Clifford, Differential Dose Albedo Measurements for 0.66 MeV 1's Incident on
Concrete, Iron and Lead, Canadian Report DRCL-412, August 1963.
31. L. G. Haggmark, T. H. Jones, N. E. Scofield, and W. J. Gurney, Differential Dose-Rate
Measurements of Backscattered Gamma Rays from Concrete, Aluminum and Steel,
Nucl. Sci. Eng., 23: 138 (1965).
32. M. J. Barrett and J. Waldman, Experimental Gamma-Ray Backscattering by Various
Materials, Report TO-B-64-68, Technical Operations Research, July 1964.
33. M. J. Berger and E. E. Morris, Dose Albedo and Transmission Coefficients for
Cobalt-60 and Cesium-I3? Gamma Rays Incident on Concrete Slabs, Report
NBS-9071, National Bureau of Standards, July 5, 1966.
34. M. Leimdorfer, Backscattering of Gamma Radiation from Spherical Concrete Walls.
Swedish Report AE-93, January 1963.
35. R. L. French, M. B. Wells, and N. M. Schaeffer, Penetration of Neutron and Gamma
Radiation Through the Openings of Underground Structures, USAEC Report
RRA-T41, Radiation Research Associates, Inc., Nov. 30, 1964; also Neutron and
Gamma Radiation Penetration into Concrete-Shielded Underground Structures,
USAEC Report RRA-T51, Radiation Research Associates, Inc., May 1965.
36.J. H. Hubbell, R. R. Bach, and J. C. Lamkin, Radiation Field from a Rectangular
Source,]. Res. Nat. Bur. Stand., C 64: 121 (1960).
37. R. E. Maerker, H. C. Claiborne, and C. E. Clifford, Neutron Attenuation in
Rectangular Ducts, in Engineering Compendium on Radiation Shielding, R. G. Jaeger
(Ed.), Vol. 1, Springer-Verlag, New York, 1968.
38. B. T. Price, C. C. Horton, and K. T. Spinney, Effects of Ducts and Voids in Shields, in
Radiation Shielding, Sec. 4.12, p. 202, Pergamon Press, Inc., New York, 1957.
39. D. M. Chase, Ducts and Voids in Shields, in Reactor Handbook, Vol. III, Part B,
Shielding, E. P. Blizard (Ed.), Chap. 12, Interscience Publishers, a division of John
40. D. K. Trubey, A Calculation of Radiation Penetration of Cylindrical Duct Walls,
USAEC Report CF-63-2-64, Oak Ridge National Laboratory, Feb. 28, 1963.
41. R. E. Benenson and A. N. Fasano, The Transmission of Fission Neutrons Having
Energy Above the S32 (n,p)p 32 Threshold by Straight Cylindrical Ducts in Water,
Report WADC-TR-57-89, Wright Air Development Center, February 1957.
42. D. C. Piercey and D. E. Bendall, The Transmission of Fast Neutrons Along Air Filled
Ducts in Water, British Report AEEW-R-69, June 1962.
43. D. G. Collins and L. W. McCleary, A Systemization and Penetration Study for
Straight Cylindrical Ducts, Report NARF-63-3T, General Dynamics, May 1963.
44. J. D. Marshall, Analysis of Radiation Streaming Through Single-Bend Ducts, Report
AD-613480 (WL-TR-64-136), Air Force Weapons Laboratory, January 1965.
45. A. Simon and C. E. Clifford, The Attenuation of Neutrons by Air Ducts in Shields,
Nucl. Sci. Eng., 1: 156-166 (1956).
46. J. C. LeDoux and A. B. Chilton, Attenuation of Gamma Radiation Through
Two-Legged Rectangular Ducts and Shelter Entranceways-An Analytical Approach,
Report AD-264022 (TN-383), Naval Civil Engineering Laboratory, January 1961.
47. V. R. Cain, Calculations of Thermal-Neutron Flux Distributions in Concrete-Walled
Ducts Using an Albedo Model with Monte Carlo Techniques, USAEC Report
ORNL-3507, Oak Ridge National Laboratory. January 1964.
417
48. R. L. French, A First/Last Collision Model of the Air/Ground Interface Effects on

Fast-Neutron Distributions, Nucl. Sci. Eng., 19: 151 (1964).
49. R. Artigas and H. E. Hungerford, A Numerical Treatment of the Attenuation of
Neutrons by Air Ducrs in Shields, Nucl. Sci. Eng., 36: 295 (1969).
50. C. C. Horton, The Shielding of Helical Ducts, Nucl. Sci. Eng., 6: 525-529 (1959).
51. C. C. Horton and D. B. Halliday. The Attenuation of Thermal Neutrons in Cylindrical
Ducts Through a Water Shield, British Report AERE-SWP/P-28 , January 1956.
52. C. E. Clifford, Gamma Shielding Provided by Ducrs, Canadian Report DRCL-370,

May 1962.
53. w. C. Ingold. Some Applications of a Semiempirical Formula for Differential Dose
Albedo for Gamma Rays on Concrete, Report AD~291133 (TN-469), Naval Civil
Engineering Laboratory. November 1962.
54. J. M. Chapman. Computer Calculation of Dose Rates in Two-Legged Ducts Using the
Albedo Concept, USAEC fIle number NP-13652 (TR-264), Naval Civil Engineering
Laboratory, Oct. 24, 1963.
55. C. W. Terrell, A. J. Jerri, and R. O. Lyday, Jr., Radiation Streaming in Ducts and
Shelter Entranceways, Report ARF -11 58-A02-7 ,Armour Research Foundation, April
1962.
56. c. W. Terrell, A. J. Jerri, R. O. Lyday, and D. Sperber, Radiation Streaming in Shelter
Entranceways, Report ARF-1158-12, Armour Research Foundation, October 1960.
57. C. W. Terrell and A. J. Jerri, Radiation Streaming in Shelter Entranceways, Final
Report, Report ARF-1158-AOl-5, Armour Research Foundation, July 1961.
58. W. C. Ingold and C. M. Huddleston, An Empirical Formula for Calculating
Gamma-Ray Dose Attenuation in Concrete Ducts, Report AD-609049 (TR-349),
Naval Civil Engineering Laboratory, November 1964.
59. D. R. Doty, Dose Measurements for Neutron Streaming in Ducts, USAEC me number
NP-13805 (AD-436183; TR-282), Naval Civil Engineering Laboratory, March 1964.
60. R. E. Maerker and V. R. Cain, AMC: A Monte Carlo Code Utilizing the Albedo
Approach for Calculating Neutron and Capture Gamma-Ray Distributions in
Rectangular Concrete Ducts, USAEC Report ORNL-3964, Oak Ridge National
61. R. E. Maerker and F. J. Muckenthaler, Calculation of the Fast-Neutron Dose Rate
Along the Center Line of Both a Straight and a TWO-Legged Square Concrete Duct
Using the Albedo Concept and Comparison with Experiment, Trans. Amer. Nucl.
Soc., 9: 147 (1966).
62. R. E. Maerker and F. J. Muckenthaler, Monte Carlo Calculations Using the Albedo
Concept of Fast-Neutron Dose Rates Along the Center Lines of One- and Tv..u-Legged
Square Concrete Open Ducts and Comparison with Experiment, USAEC Report
ORNL-TM-1557, Oak Ridge National Laboratory, September 1966.
63. W. S. Snyder and J. Neufeld, Calculated Depth Dose Curves in Tissue for Broad
Beams of Fast Neutrons, USAEC Report ORNL-1872, Oak Ridge National
Laboratory, Apr. 13, 1955.
64. W. E. Selph, ]. Newell, and C. E. Bayer, Albedo Monte Carlo Analysis of Fast
Neutron Transport in a Missile Launch Tube, Trans. A mer. Nucl. Soc., 11: 1 (June
1968).
418
REACTOR SHlELDlNG FOR NUCLEAR ENG1NEERS
65.J. M. Chapman, Gamma-Ray Dose Rates and Energy Spectra in a 3-Foot Square Duct,
Report AD-450421 (TR-325), Naval Civil Engineering Laboratory, October 1964.
66. J. M. Chapman and J. S. Grant, Gamma Ray Attenuation in Coplanar and
Noncoplanar Three-Legged Concrete Ducts, Report AD-453022 (TN-658), Naval civil
Engineering Laboratory, November 1964.
67. L. Tonks, Enhancement of Leakage by Internal Voids in a Shield, USAEC Report
KAPL~107, Knolls Atomic Power Laboratory, January 1949.
68. H. J. Kouts, Theory of Flux Perturbations by Voids in Shields, USAEC Report
BNL-1357, Brookbaven National Laboratory, Sept. 15, 1952.
69. N. M. Smith, Jr., The Absorption and Scattering of Radiation in Random Aggregates
of Pebbles, USAEC Report CNL-21, Oak Ridge National Laboratory, March 1948.
Shield
Heating~ Air Transport~

Shield Materials~
and Shield Optimization
L. G. MOONEY and N. M. SCHAEFFER
Four additional design considerations-shield heating, air transport, shield

materials, and shield optimization-should be added to the special topics
needed to supplement the transport calculations in the design of a shield.
The grouping of the four topics is a matter of convenience and has no special
significance.
8.1
SHIELD HEATING
Heat transfer may be said to dictate the design of a reactor. Materials

selection is based on a compromise between thermal and nuclear properties.
Within the core the kinetic energy of the fission fragments is the principal
heat source. However, the design of the reactor structure, core vessel, and
thermal and biological shield is determined by the energy deposition of
radiation emerging from the core. Although significant heating rates may be
generated
by neutrons in the reflector, control rods, and inner shield regions,
most of the heating in the shield results from energy degradation and
absorption of gamma rays. Generally speaking, neutrons are considered in
shield heating only insofar as they produce gamma rays. Energy deposited by
charged particles is a consideration only in a limited number of materials.
Methods for calculating energy deposition are identical to those for
obtaining attenuation. Thus all the techniques and data described in Chaps. 4
through 6 are pertinent here. The principal difference between attenuation
and heating-rate analyses is the quantity of interest. For an attenuation
calculation, we are primarily interested in the quantity transmitted. In shield
heating we are interested in the quantity absorbed. This distinction makes
the shield-heating calculation less demanding. The first orders of attenuation
are obviously most important to heating rates; thus approximate methods
can be used to better advantage for shield heating than for radiation
transmission.
419
420
We should not oversell the acceptability of approximate results, however.

An underestimate of a biological dose rate by a factor of 2 is a nuisance:
working times are more limited or exclusion areas must be enlarged.
Consequences of a similar error in heating-rate estimates in the shield could
be severe; excessive temperature rises could wreck the reactor structure at
design power levels. A prudent design approach would therefore be to
estimate the heating rates two ways: one that is known to overpredict and
one that is known to underpredict.
It is evident that a calculation of heating rates is simply another response
function of the flux density. If one preserves his flux-density data from the
shield analysis, then the heating rate can be obtained simply by converting
the flux density to energy deposition.
8.1.1 Gamma-Ray Heatingt
The principal interactions of a gamma ray are pair production,
photoelectric absorption, and Compton scattering. In Chap. 3 the Compton
energy absorption cross section, uca , was defined on the basis of the average
energy loss per collision (Eq. 3.2-14). Since the other two processes involve
complete absorption, G ca is used with them to define a macroscopic cross
section (Eq. 3.2-16,1la =N[upe +Zuca +upp ]) that represents local energy
deposition. The photon-energy deposition gives rise to a heating rate
H(r)
=C f EE~ax
mm
E lJ.a(E,r) (E,r) dE
(8.1-1)
where H(r) =heat generation rate at spatial position r (watts cm- 3 )

E = gamma-ray energy (MeV)
(E,r) = photon-flux density at energy E (photons cm- 2 sec- 1 MeV-I)
lJ.a =energy-absorption coefficient or macroscopic cross section for
energy deposition (em-I)
C =conversion factor (1.60 X 10- 13 watt-sec MeV I )
We should emphasize that (E,r) includes the contributions from
secondary as well as from primary gamma rays.
Implicit in the use of the energy-absorption coefficient as defined here is
the assumption that some other possible second-order processes are
negligible and that the small amount of radiation that survives the
photoelectric effect and pair production through fluorescence, bremsstrahlung, or annihilation radiation is absorbed where it is produced. Such
tThis section follows a recent review by Claiborne. I
SHIELD HEATING, AIR TRANSPORT, ETC.
421
assumptions are justified for reactor design because of the low photon
energies involved in these secondary processes.
Calculation of the heating rate due ro gamma rays is then reduced to
obtaining the spectrum (E,r), the subject of Chaps. 4 through 6. We will
discuss buildup factors and mention some other methods in general use for
calculating heating rates.
(a) Buildup Factors. Because they are manageable with hand calculations as well as with machine application, buildup factors represent the most
popular technique, and they have been validated by comparison with
experiment. Perhaps the most valuable feature of the buildup-factor method
is the relative certainty with which the sign of the error in the calculated
heating is known. Even for extreme multilayered construction (assuming
that other sources of error are reasonably small), it is possible to bracket the
true answer by separate calculations using the largest and the smallest
individual buildup factors for the effective buildup factor of the composite
shield.
Recalling that buildup factors can be used as a type of point kernel, we
measure the path length along the line of sight from the source to the
receiver. The uncollided flux density at the receiver (the point of energy
deposition) is then multiplied by the buildup factor to account for the
scattered contribution. If the source is distributed (i.e., a reactor core), we
represent it as an array of volume elements, each represented by a point
source, and integrate the point kernel over the source volume for each source
energy. The integral over volume is then
(8.1-2)
where H(r) =energy deposition per unit volume (MeV sec-I cm- 3 )
Su(r') = isotropic source intensity in photons per unit time per unit
volume at energy Eo (sec-I cm- 3 )
V = source volume
Eo = energy of gamma ray emitted by source
Ila = energy-absorption coefficient of material at the point of energy
deposition
Ba(llx) = energy-deposition buildup factor
Il =total attenuation coefficient at energy Eo
422
REAerOR SHIELDING FOR NUCLEAR ENGINEERS
x = Ir - rl, the distance between the source and the point of

energy deposition
r = point of energy deposition
r' : ; : location of source in volume V
We note that Ba (lJx) is an energy-deposition (formerly called energyabsorption) buildup factor and not the energy-fluence buildup factor. The
latter would give the total energy flux, which would result in an overestimate
of the heating rate.
All we need to obtain heating rates for photons is a proper set of buildup
factors. For infinite homogeneous media we can use the energy-deposition
buildup factors given by Goldstein and Wilkins. 2 For finite media and, more
realistically, for the multilayered media to be found in any reactor shield, we
need other buildup factors or an additional assumption. The pitfalls
encountered in applying infinite-media data to finite geometries, described
briefly in the discussion of differential energy spectra in Sec. 6.3.1, apply
here as well.
The infinite-media data (either buildup factors or differential energy
spectra) contain the effect of the medium beyond the receiver point as well
as that between the source and the receiver. Therefore a receiver at the far
edge of a slab will intercept an intensity differing from the infinite-medium
case depending on the difference in backscattering from the adjacent
medium. If the adjoining medium is air, the infinite-medium buildup factor
produces an overestimate. If it is another dense medium, the infinite-medium
buildup factor will give more or less, depending on the relative albedo of the
two media. Since the albedo depends on the ratio of scattering to absorption
cross sections, a greater ratio in the adjoining medium would imply that the
infinite-medium buildup factor gives an underestimate of the actual intensity
at the receiver. A lower ratio would imply an overestimate. The saving grace
is that the magnitude of the error is usually of the order of a few percent for
photons, reaching 20 to 30% only in unusual cases.
Assuming that errors of this magnitude are acceptable, we can use
infinite-medium buildup factors. Equation 8.1-2 must be modified to accept
this compromise. Making the further assumption that the source energy
distribution is divided into groups and that Hj is the mean energy for the jth
group, we can write
423
Sv(r') Ba(~ !liti) exp(-~ !liti) dv

I
(8.1-3)
where t is the thickness of each region and i refers to the number of regions
of different composition interposed in the line-of-sight path x between a
source point and any particular point of energy deposition. The sums over j
imply that we set the path length equal to the cumulative sum of mean free
paths in each material to the point at which the heating rate is to be
obtained.
The buildup factor can be selected according to several rules of thumb.
For homogeneous or near-homogeneous mixtures (regions significantly
smaller than the mean free path), the so-called equivalent Z (atomic number)
method' has been used with good effect. In this method the equivalent
atomic number is obtained by a recipe wirh a sound theoretical basis.
Briefly, this method involves comparing the shape of the curve obtained by
plotting the total attenuation coefficient per electron of the mixture as a
function of energy to the corresponding curves for the individual elements
until a reasonable match is found for energies below source energy. To be
completely equivalent, the ratio of total to scattering cross section should be
the same in the two cases, which usually occurs to a good approximation
when the shapes are matched.
Extension of the buildup-factor method to structures with multiple
layers leans more heavily on rules of thumb and physical intuition. When the
outermost layer exceeds two or three mean free paths, the buildup factor of
the outermost material is generally recommended, but based on the total
number of mean free paths along the line of sight through all materials. This
seems a reasonable procedure since the gamma spectrum would readjust to
the new medium and tend to approach the spectrum that would exist if the
whole structure consisted of the outermost material.
In the most extreme case where the structure is multilayered and each
layer is less than two mean free paths thick, no clear-cut procedure is
available. Some Monte Carlo calculations have been made on energy
deposition in multilayered shields, but correlations for design use have
generally been lacking. Based on Monte Carlo studies of heat deposition in
one- and two-layer lead slabs in water by a plane monodirectional beam,
Bowman and Trubey' suggested that combinations of the Goldstein-Wilkins
energy-deposition buildup factors could be used in the relation representing
424
the buildup factor for stratified slabs. This effective buildup factor is
expressed by
BaU.!, 11 ,IJ., I,)
Ba , (IJ., I, sec Ii,E o ) B a , (IJ., I, sec Ii,E o )
x exp(-IJ.2 I, sec Ii) + Ba, [(IJ., I,
+ IJ.,( 2 ) sec Ii,E o ] [1 - exp(~IJ.212 sec Ii)] 1(8.1-4)

where Ii is the angle between the direction of the incident gamma ray and
the normal to the slab. Subscripts 1 and 2 refer to the first and second
materials, respectively.
For a double-layer shield whose outer layer is less than two or three
mean free paths or a multilayer shield with the exterior layer more than two
or three mean free paths in thickness, use of Eq. 8.1-4 seems intuitively
justified for most purposes. The angular dependence (Ii now becomes the
angle that each path makes with the normal to the shield) does complicate
the integral when Eq.8.1-4 is substituted into Eq.8.1-3; however, the
numerical integration presents no problem when electronic computers arc
used, and in some cases it would not be too arduous for hand calculations.
For conditions that are significantly different from the preceding cases,
some other method should be used if it is desired to do other than bracket
the answer by using techniques that will give results known to be high or
low. For instance, in a multilayered shield with all layers less than two mean
free paths, a high result is obtained by using the largest individual buildup
factor as the effective buildup factor for the composite shield, and a low
result is obtained when the smallest individual buildup factor is used.
Constants for the energy-deposition buildup factors in the Berger form
are given in Appendix F for water, aluminum iron, tin, and lead. Tobias,
Vondy, and Lietzke s have interpolated the results of Goldstein and Wilkins
to obtain energy-deposition buildup factors for even atomic numbers from 2
to 78 and have obtained coefficients for best fit to the third-degree
polynomial
B(p.x,E o ) = 1 + a(p.x) + b(p.x)2 + c(IJ.x)3
(8.1-5)
Values of a, b, and c for seven energies from 0.5 to 10 MeV are given in
Appendix F, Table FA, and can be used for materials not explicitly given
elsewhere. Because they are regarded as most accurate (Sec. 4.8.1), the
Cap0 6 polynomial expressions with regions of applicability are given in
425
Appendix F, Table F.5. This set includes the same materials as Appendix F,
Table F.3.
Walker and Grotenhuis 7 have examined the effective atomic numbers of
concretes and found them to range from 11 for ordinary concrete to 27 for a
barytes aggregate. Buildup factors for a particular concrete can be obtained
by interpolating available data to the effective atomic number. Trubey'
published coefficients for the Berger form of the buildup factor for ordinary,
magnetite, and barytes concretes as well as for sand, wood, air, and lithium
hydride. These are reproduced in Appendix F, Table F.6.
The buildup-factor method has been widely used and is a versatile tool
for shield design. Calculations with this method have generally checked
experimental measurements to within 35%. It is reasonable to expect similar
accuracy when this method is carefully used in design calculations for shields
or reactor structures that do not involve extreme multilayered construction
with each region less than two mean free paths in thickness. Even for
multilayered construction, however, when the outermost layer is more than
two or three mean free paths in thickness, particularly when the low-z
material follows the high-Z material, the method should be quite good since
the spectrum tends to readjust to the new medium.
(b) Other Methods for Gamma-Ray Heating. Every method devised for
gamma-ray attenuation has been applied to obtain heating rates. Greenhow,
Mueller, and Sabian 9 devised a recipe with some success for applying any
neutron multigroup diffusion code to gamma-ray heating by proper
exchange of cross sections. French' 0 applied differential energy spectra
from the Goldstein-Wilkins moments results to carry the gamma-ray
spectral description through multilayered shields. This gives additional detail
to the results, but it is essentially no more accurate and is considerably more
involved than the buildup-factor techniques based on the same moments
data.
Lathrop' 1 has modified a standard LASL one-dimensional Sn code
(OTF) to include up to 10 moments (P, 0) of the group-to-group scattering
cross sections. A subroutine cross-section code produces scattering cross
sections by performing group averages of the Legendre moments of the
Klein-Nishina differential cross section. At the option of the user,
cross-section data can be flux weighted, source weighted, or unweighted.
Test calculations with this code generally compare favorably with Monte
Carlo calculations. Engle'2 also developed a one-dimensional Sn-p/ code
(ANISN) which in principle can handle any number of moments of the
426
group cross sections; a two-dimensional version (DOT) was developed by

Mynatt, Muckenthaler, and Stevens.! 3
Trubey, Penny, and Lathrop l4 compared the results of an S,. -P3
calculation with Monte Carlo and buildup-factor methods (the latter is
equivalent to the moments method for the homogeneous medium) for a
sphere of iron 70 em in radius containing a unit gamma-ray fission source
distributed uniformly within the inner 25-cm radius. The Monte Carlo results
were obtained with the FMC code. [S No importance sampling was used, and
a total of 5000 case histories was calculated. When the buildup-factor
method was applied, the usual point kernel was integrated over source and
energy with each buildup factor obtained by a Lagrangian interpolation of
the buildup factors in the paper by Goldstein and Wilkins. 2 As shown in
Fig. 8.1, the agreement between the three methods is excellent; in fact, it is
rather surprising that the agreement is good all the way to the edge. The
Monte Carlo results are typical in that the statistics get bad as the distance
from the source increases, and the net result is that the flux is usually
underestimated. At the origin the poor statistics resulted because the scoring
volume was too small. Toward the edge the decrease in the Sn results with
respect to the moments method most probably is a boundary effect. The Sn
calculation assumed a vacuum beyond 70 cm; use of buildup factors implies
an infinite region.
For deep penetrations of the order involved in shielding calculations, it
had been generally felt that a P 3 approximation was insufficient to account
for the anisotropy of gamma scattering. In some preliminary studies by
Lathrop,! I however, the difference in results between the Po and P [
approximation was quite pronounced, but only small improvement resulted
from the use of higher order approximations in place of the P [
approximation.
For multilayered shields it is apparent that Monte Carlo or discreteordinate methods are the best available, and they should be used where
accuracy requirements exceed the capability of buildup factors.
8.1.2 Neutron Heating

Direct heating effects are potentially important for elastic scattering of
fast neutrons in low-z materials. For equal numbers of I-MeV neutrons and
gamma rays, the contribution of elastic neutron scattering to the heating rate
varies widely with the atomic number of the medium. Foderaro, Hoover, and
Marable' show that the ratio of neutron to gamma-ray heating is 5 to 1 in
l(r 4 , - _ . . ,
...,....
. -_ _......
-.-
427
, -_ _- ,
10-6
.,.
0
0-
10- 7
':'
E
0
>
::; 10- 8
Z
i=
in
0
0-
">-
10-9
"az:
Moments method
I:>, Sn
_
Monte Carlo
10- 10
lo-12!!:---~:_-
10
_ _:!::---:!::_--__:l!,_-_
20
30
40
_+_,..._-_+.,.-_~
50
60
70
DISTANCE FROM CENTER, em
Fig. S.l-Comparison of three methods for calculating heating in iron from 25-cm
spherical source. (From Trubey et. al. 14 )
428
water, 6 to 10 in carbon, and 1 to 50 in iron. We can show this by

developing the neutron analog to Eq. 8.1-3. From Sec. 3.2.2, a neutron of
energy E initially and energy E' after an elastic collision is
E'
=E
[A +(A2A+cos1)20+ 1]
2
(8.1-6)
where A is the mass number of the target nucleus and 0 is the scattering
angle (in center-of-mass coordinates). The energy loss is
(8.1-7)
We need the average energy loss. If the average scattering angle is given by
cos 8,
-(2
A)(A + 1)2 (1 - cos 0)
(8.1-8)
f>E(E) = E
We can now write the heating rate caused by elastic scattering in a

material with Ni atoms/cm 3
(8.1-9)
Hn,n = K Ni IE f>E(E) airEl <l>(E) dE

where Hn,n
K
=heating rate (watts cm- 3 )

=conversion constant (K = 1.6 x
10- 13 watt-sec Me\T l

airEl = elastic-scattering cross section of the ith material
<l> = flux density of neutrons in dE about E
E = incident neutron energy
Foderaro et al. 16 use the orthogonality property of Legendre polynomials to show that the average value of the cosine of the scattering angle is
cos 0
=II irE)
(8.1-10)
where II irE) is the first expansion coefficient of the Legendre polynomial

expansion of the differential elastic-scattering cross sections of the ith
429
element. Since these cross sections are commonly expressed in terms of

Legendre polynomial expansions, this is a very convenient result. Values of
I. (E) for selected materials are quoted by Foderaro et al. and are given in
most compilations of differential elastic-scattering distributions. The average
energy loss is then
(8.1-11)
The significant term is Ai/(Ai + 1)2. It tells us that neutron heating
decreases rapidly with the mass of the target nucleus and supports the
conclusion that fast-neutron heating in heavy materials is unimportant in the
presence of comparable numbers of gamma rays. For A < 12, both neutrons
and gamma rays tend to contribute significantly.
A corollary to these conclusions relates to the importance of inelastic
scattering of fast neutrons. Since inelastic-scattering cross sections in the
energy range of interest (E < 6 MeV) are significant only in heavy nuclei,
inelastic neutron scattering is usually unimportant in heat-generation
computations. Even where inelastic scattering is significant, it is generally
less than elastic scattering. Thus where elastic scattering is unimportant for
heating considerations so is inelastic scattering.
Indirect neutron heating through production and absorption of gamma
rays is significant where there is appreciable radiative capture and conceivably where there is appreciable gamma-ray production by inelastic
events.
8.1.3 Charged-Particle Heating
Three nuclei account for most of our interest in neutron-induced
charged-particle heating, 6 Li, lOB, and 14N. Because the cross sections tend
to be in the millibarn range in the MeV region, charged-particle-out reactions
are not important in heating except at thermal-neutron energies. Table 8.1
summarizes the reactions of interest.! 7
Note that three isotopes of uranium are included in Table 8.1. The large
energy release per fission would make these isotopes important heat
generators if they were used in a shield.
For any of these reactions, say (n,a), the heating rate due to neutrons in
energy interval t;.E is simply
H n,. = KN un,. (E) (E) (E + Q) t;.E
(8.1-12)
430
where K = same conversion constant given before (1.6 x 10- 13 watt-sec

MeV-I)
N = number of atoms of target material per cubic centimeter
an,. (E) = (n,a) cross section at energy E
(E) = neutron-flux density at neutron energy E
Q = energy release for the reaction
Table 8.1-CHARGED-PARTICLE REACTIONS IMPORTANT IN HEATINGt
Thermalneutron
Nuclide
6Li
lOB
I4
238U
235U
233
Reaction
(n,,,)
(n,,,)
(n,p)
Fast fission
Thermal fission
Thermal fission
cross section,
950
3840
1.8
0.58 (3 Mev)}
582
527
Energy emitted,
MeV
4.8
2.3,0.48
0.63
Average total energy
from fission,
-190 MeV
tFrom C. N. Klahr, Reactor Handbook, Second Edition, Vol. III, Part B, Shielding,
E. P. Blizard (Ed.), p. 174, Interscience Publishers, a division of John Wiley & Sons, Inc.,
New York, 1962.
8.2 AIR TRANSPORT

Although transport of neutrons and gamma rays through air is not
usually an important consideration in the design of shields for fixed power
plants, it is an important shielding consideration for any portable or mobile
system. Laboratory experiments involving reactors, isotopes, or accelerators
also frequently require some assessment of air transport. Actually two
programs have accounted for most of the research in air transport that has
been done: nuclear-powered aircraft and weapons radiation studies. Interest
from the weapons standpoint is obvious; all the initial radiation from a
nuclear weapon detonated within the atmosphere arrives at a distant receiver
via the atmosphere. In analysis of exposure to fallout radiation, the transport
through air from airborne or ground-deposited activity is the primary
concern. The reasons for interest in air transport from the standpoint or
nuclear aircraft are also evident; weight limitations give additional incentive
to tight design, and transport of radiation by air and structure from reactor
shield to passengers and crew becomes a crucial design factor.
431
The processes involved in air transport arc identical to those in dense

media. The difference that makes air a separate topic is entirely in the
geometry. Relaxation lengths in air are of the order of a hundred meters as
compared with relaxation lengths in dense materials of the order of a tenth
of a meter. Put in a more familiar way, the density of air at standard
conditions is 1.29 mg cm'3; the density of most solids is in the range 1 to
10 g cm'3.
The geometry of interest includes the ground and the cover of air
extending out to large distances in space. The importance of the boundary
effects at the air-ground interface depends on the ratio a/h, the sourcedetector separation divided by the height of source or detector above the
ground. For a/h ;;, 1, the presence of the ground significantly affects the
amount of scattered radiation arriving at the detector. For a/h < 1, the
relative importance of the ground effect depends on the height of the source
and the detector. If the height of either the source or the detector is less
than a mean free path in air, the ground effect is likely to be significant.
Consider the case where both the detector and the source are located in
the upper regions of the atmosphere. Here we have what can be described as
an infinite medium within which neutron and gamma-ray mean free paths
are extremely long and within which the density of material varies
significantly over distances of interest. The density variation must be
included if the accuracy of the air-transport calculation is to be acceptable.
As the source and the detector are lowered in rhe atmosphere toward the
earth's surface, the density increases, and the density gradient must be
considered until the gradient becomes unimportant over several mean-freepath distances. Air-transport calculations in the lower atmosphere at heights
from 3 to 10 mean free paths or so above the ground surface are very similar
to infinite-medium calculations for other materials of uniform density.
As the source and the detector are lowered further toward the ground
surface, the effects of the ground become noticeable, and the air-ground
interface effects become most important when both the source and the
detector arc on the surface.
[f
either the source or the detector is raised
above the surface, the interface effects diminish.

The radiation transport in an air-over-ground geometry is generally
calculated to obtain the radiation distribution in the air. For such problems
the ground is considered as a perturbing medium. The perturbation consists
in absorption and reflection of the incident radiations as well as the
contribution of secondary gamma rays produced by neutron interactions in
the ground. Only for special problems, such as ground activation and heating
432
rates, is there a need for neutron and gamma-ray distributions within the
ground proper.
8.2.1 Infinite Air Medium
As mentioned in the preceding text, the transport of neutrons and
gamma rays in an infinite air medium can be calculated in much the same
manner as an infinite medium of any other material provided the density is
uniform. In fact, probably more infinite-medium calculations have been

performed for air than for any other material.
Air transport has been calculated' with moments, discrete-ordinates, and
Monte Carlo methods in recent years. The advantages of the Monte Carlo
and discrete-ordinate methods in providing angular distributions at the
receiver have resulted in their use almost exclusively. Early definitive
calculations of air scattering of gamma rays were done by Lynch, Benoit,
Johnson, and Zerby' 8 using the Monte Carlo method. Similar calculations
for neutrons were reported by Wells" and Zerby.20
Trubey2' showed that for source-detector separation distances less than
about 30 m a single-scattering calculation gives an excellent approximation
of the intensity at a detector but does not give an adequate representation of
the energy or angular distributions at the receivers.
Figure 8.2 illustrates recent comparative calculations of the neutron dose

as a function of distance from a point isotropic 14-MeV neutron source by
Straker and Gritzner, 22 Keith and Shelton,2 3 and Webster. 24 Straker's
results were obtained by discrete ordinates, and the remainder, by Monte
Carlo. Straker's data arc lower in magnitude because of his use of different
cross-section data in the high-energy region. Webster and Keith and Shelton
used identical cross-section data, and the difference in their calculations at
larger distances is probably due to poor statistics in the Monte Carlo output.
The abscissa and ordinate of this graph deserve mention. By plotting the
dose per source neutron times 41TR 2 (R the source-receiver separation) as the
ordinate, we can remove the inverse-square geometric effect. Using slant
range in density units (g cm- 2 ) for the abscissa, we can avoid dependence on
the air density. This graph is then applicable to any uniform air density, and
the separation (or slant range) in centimeters is obtained by dividing the
slant range in g/cm 2 by the air density in g/cm'. For convenience, Table 8.2
gives the conversion to meters for sea-level air density.
Figure 8.3 shows similar data from Straker and Webster for a point
isotropic fission-neutron source. The differences noted in the 14-MeV source
""-
"-
"- '\
'\
433
\
\
\
\
\
\
\
\
Keith and Shelton
a:
w
o
o
N
a:
10- 10
.!;
10,11 L.._---l_ _..L_ _L_---l-:-_...L_ _L_....l_--,,.,l,,_ _L..I......J
100
200
SLANT RANGE, g/cm 2
Fig. 8.2-Comparison of the neutron dose in infinite air from a 14-MeV neutron source
calculated by Straker,22 by Keith and Shelton,23 and by Wchster. 24
434
g
iii
c: 10-9
[3
cr
f-
::J
cr
::J
Sl
cr
w
o
'"01010
N
0:
.'i
~
~
:-..
Webster
\
\
\
\
10- 11
~_---JL_-l
_ _--L_ _~_ _.L_ _L _ _L_-':J':-_-l_ _--'

100
200
SLANT RANGE, g/cm 2
Fig. 8.3-Comparison of the neutron dose in infinite air from a fission-neutron source as
calculated by Straker 22 and by Webster. 24
SHIELD HEATING. AIR TRANSPORT, ETC.
435
data are seen to wash out for the fission source as a result of the smaller
number of high-energy neutrons in the fission distribution. Some differences
in slope are evident, however. Comparisons of calculated and measured data
for an infinite air medium are shown later in this chapter.
Figure 8.4 compares the results of three calculations for gamma rays in
an infinite ait medium. Two are based on the moments method: a
Table 8.2-AREAL-DENSITY CONVERSION FOR
SEA-LEVEL AIR DENSITY (p = 1.293 mg/cm 3 )
Range, m
Range, g/cm'
50
6.5
77
10.0
12.9
25.9
50.0
100
200
387
500
773
1000
1160
1547
64.6
100.0
129.3
150.0
200.0
calculation for ait by Vogel' 5 in 1963 and a direct application of the 1954
buildup factors for water. The two sets of data are close together except at
0.5 MeV, where the older results are based on an interpolation. The thitd set
of data shown is from a Monte Carlo calculation by Marshall and Wells' 6 in
1966. This set of data is widely used because it is recent and because it is
differential in angle and energy at the receiver, which makes it suitable for
application to shielded detectors.' 7
Unfortunately a doubly differential data set is needed to convey useful
working information: for 10 angle groups and 10 scattered energy groups, a
10-by-10 matrix is requited for each initial energy. Thus one complete data
set includes about 10 3 entries. All this is preliminary to stating that we do
not have the luxury of including such data in this text. We respectfully refer
the reader to the references and information centers.
The effect of ait density on ait scattering is straightforward and simply
related to source-receiver separation distance. Zerby' 8 showed that the flux
density or dose rate from an anisotropic point source in an infinite
homogeneous medium at a given ait density can be transformed to the
436
10- 12
Based on:
-
Goldstein and Wilkins,

moments method
l:i Vogel,26
moments metho~
Marshall and Wells, 6

Monte Carlo
1O.13:0~_ _L_4:\O:-_L_.,jBO~-.L-"12!'O:---J'-''''-;1-!;60;;-_L--'2,JOO;;;---l
SLANT RANGE, g/cm 2
Fig. SA-Comparison of the calculated gammaray dose as a function of range in an

infinite air medium.
437
corresponding quantity at any other air density in the medium:
- p~
D(P2 ,R 2 } - 2' D(p, ,R,)
p,
(8.2-1)
where p is the air density, R is the separation distance, and

(8.2-2)
This relation can be used only when the source and the receiver are at
the same air density.
The question sometimes arises whether or not the scattering in air is
affected by moisture (rain, fog, or water vapor). The possibility is most
simply resolved by considering the relative densities. A cubic meter of air at
standard conditions weighs about 1.3 kg. Saturated water vapor at OC
weighs 5 g/m 3 ; even at 35C (95F) saturated water vapor weighs 40 g/m 3
During a heavy rainstorm of 5 cm per hour, the water density is only about
5 g/m 3 A heavy fog yields a water density in the same range. Thus the water
density never exceeds a few percent of the air density, and such a water
content would have a comparable effect (of no more than a few percent) on
the scattered intensity.
8.2.2 Air-Over-Ground Calculations
Calculations of the neutron and secondary gamma-ray transport in an
air-over-ground geometry are difficult to compare since interface effects
depend on the source and the detector height, the separation distance, and
the composition of the ground used in the calculation.
Probably the first definitive air-over-ground calculation for gamma rays
to reach the open literature was the Monte Carlo study of Berger. 29 The
point isotropic source was 1.28-MeV gamma rays, and the results were given
in terms of boundary correction factors, which made them convenient to
apply to fallout radiation studies. Air-over-concrete calculations for point
monoenergetic and monodirectional gamma-ray 30 and neutron 31 sources
were reported by Wells. These data were to be applied for the nuclear
aircraft with a reactor 4 m above a concrete ramp.
438
REAcrOR SHIELDING FOR NUCLEAR ENGINEERS
-~
~
N
'5
10-9
Z'
oa:
I::J
u
a:
::J
'"a:w
0-
Discrete ordinate
lBCI 10-10
::;
::;
"a:
10,11~_.L_-L _ _L-~--L_ _L_.L_-,J,,-_L_.L--L---,J.
50
100
150
SLANT RANGE, g/cm 2
Fig. 8.S-Comparison of the neutron dose vs. range as reported by Straker22 and by
Marcum 32 for a 14-MeV type source in an air-over-ground geometry.
A comparison of more recent results is shown in Fig, 8,5, where the

discrete-ordinate data of Straker" are compared to the Monte Carlo data of
Marcum" fot an identical neutron source composed largely of 14-MeV
neutrons. Differences in the data are probably due to cross sections used for
nitrogen and oxygen since several new sets were published in the intervening
- - Infinite air medium
~,
....
c
""~
....
E
u
....
10.9
aC<
t-:::>
w
Z
:::>
10. 10
'"ac
z
aC<
t--
:::>
w
15.2-m source height
" "-
FiSSton-neutron)',
source
'"C<
w
,......... ..... .... .....

.... , , ,
, , ....'
, '. /',
'
14-MeV source_J<
OJ
C<
.....
':'
z'
439
C<
10. 11
,, , ,
,, , '
,,, '
,,
'"
'"
100
Fig.8.6-Comparison of the neutron dose as a function of range for a fission-neutron

source and a 14-MeV neutron source in an infinite air medium and for a source height of
15.2 m. (From Straker.")
3 years. Both sets of data were for detectors at the ground, but the Straker
set is for a 15-m source height and the Marcum set for a ground-level source.
This difference does not account for the differences in the results.
8.2.3 Air-Ground Interface Effects
Figure 8.6 compares the neutron dose in infinite air and in an
air-over-ground geometry for both a high-energy 14-MeV neutron source and
a fission-neutron source. These data are from the same set by Straker"
quoted previously. The neutron dose is appreciably lower for the air-over-
440
10-9
-- ........
g
,
E
u
2'
"'"=>
-- .....
............
.............
....
..... ....
t-
w
2
.... ....
'" ~'5.2-m source""""'" .........

'"
10. 10
..... .....
height
"
""
14-MeV source
.......
~------ " .... :--.
......
..... ....
.....
'"=>
"
"'"
'"
W
'"
""<:
""
"
"
Fission
""source
"
10- 11
:;;
:;;
<:
"
"
'"~
lO'12~_ _--:;;~_-"7.!:;;---~;;----;;;';::;---;;';;;;---,;t;;;-..J
100
300
500
700
900
1100
1300
SLANT RANGE, m
Fig. 8.7 -Comparison of the secondary gamma dose from a fission-neutron source and a
14-MeV neutron source in an infinite air medium and for a source height of 15.2 m.
(From Straker. 22 )
ground geometry than for infinite air, and the effect appears to be somewhat
dependent on the neutron source energy. This effect would be much more
evident at larger ranges. Figure 8.7 shows the effects of the interface on the
secondary gamma-ray dose for the same conditions as those given for the
neutron dose. Here the effects are highly dependent on both the source
height and the neutron-energy distribution.
Some indication of the accuracy of the infinite-air and air-over-ground
transport data can be obtained from Figs. 8.8 and 8.9, which compare
441
lO- Br-'-"""-""'---'--'--"'--'T"-..,..---r---,r--,--,
I
Infinite air medium
c
g,
~
'E 10-9
u
:g
o
a:
~
Air-over-ground geometry
8.2-m source height

O.9-m detector height
- Calculated (discrete ordinate datal

Measured
1O-10':-_......._...J._ _L-_*_...J._ _L-_......._~,....._L--.J......"....J.-_J

o
50
100
150
SLANT RANGE,
g/cm2
Fig. 8.B-Comparison of the measured 33 and calculated22 neutron dose in an infinite air
medium and in an air..over-ground geometry resulting from the HENRE neutron source.
Straker's data and measurements made during Operation HENRE 33 at the

Nevada Test Site. Operation HENRE (High Energy Neutron Reaction
Experiment) was designed to provide a high-intensity source of 14-MeV
neutrons from an accelerator using the D- T reaction. The accelerator was
mounted on a 457-m tower and could be raised to approximately 366 m to
approximate an infinite air medium. Neutron and secondary gamma-ray
doses were measured as a function of range for the soutce and the detector
at 336 m (simulating an infinite air medium) and for the source at an 8.2-m
height and the detector on the ground surface. 34 35 Results of the
measurements are compared with calculated values in Figs. 8.8 and 8.9. The
442
Calculated :
Infinite air
easured (HENRE):
336-m source
and receiver height
z
o
a:
Calculated:
15.2-m source height
Surface receiver
w
Z
w
u
a:
::>
lillO- IO
Measured (HENRE):
8.2-m source height
O.9-m receiver height
..
a:
w
w
'"
g
:IE
'"a:
lO-l1L_ _....I..
20
L _ _J.._ _...l
40
60
SO
.l.,-_ _.,J.,_...J
100
120
SLANT RANGE, g/cm2
Fig. 8.9-Comparison of the measured 33 and calculated 22 secondary gamma dose in an

infinite air medium and in an air-over-ground geometry resulting from the HENRE
neutron source.
vertical bars describing the measurements represent uncertainties in the

measurements caused by shadowing of the neutron source by components of
the accelerator with a consequent departure from isotropy; the actual source
intensity was therefore uncertain.
Limited success has been achieved in developing theoretical models to
describe the effects of the air-ground interface. One such model, called the
First-Last collision Model, devised by Frcnch 36 has been reasonably
successful in describing interface effects on the integrated neutron dose. The
mo(fel was originally conceived as a possible means of adjusting infinite-airmedium transport data to account for the perturbation resulting from an
interface. The model presupposes that interface effects operate primarily on
443
the first collision of neutrons near the source and on the last collision of
neutrons near the detector. Perturbation fractions are calculated, which
represent the effects of the interface on the number of first and last
collisions, as the source, or detector, or both approach the ground surface.
The model accounts only for the scattered-neutron component and is used
only for separation distances where multiple scattering predominates.
Results with the first-last collision model also indicate that the interface
effects are limited to those neutrons which arrive at the detector from the
lower 2rr hemisphere below the detector.
An experiment for testing this last conclusion was devised and carried
out during Operation HENRE. The 14-MeV neutron source was fixed at a
height of 34 m and the detector height was varied from 0.2 to 21 m. A 2rr
shield was sandwiched between two detectors to separate the neutron dose
arriving from the upper 2rr hemisphere from that arriving from the lower. A
third detector was positioned on a boom attached to the shield to measure
the total free-field dose as a function of distance above the ground. The
entire assembly was suspended at a horizontal distance of 305 m from the
source tower. More details of the experimental arrangement are given in
Sec. 9.3.1.
Measurements were made on all three detectors as a function of distance
above the ground. Appropriate corrections were made to the measured data
to account for backscattering of the neutrons by the shield and for ground
shadowing by the shield.
Figure 8.10 compares the experimental results and the values calculated
from the model. The free-field measured and calculated neutron doses were
normalized at the maximum height of 21 m since the absolute source
strength could not be established with sufficient accuracy. In general, the
agreement is excellent, and the experimental results can be considered a
reasonable verification of the model.
8.3 SHIELD MATERIALS
The selection of materials is obviously an important part of reactor
shield technology. Materials selection and a related task, optimization,
together offer the greatest potential for weight or cost savings. Nevertheless,
materials selection has not been reduced to an analytical approach and has
received rather less attention by shielding investigators than other more
elegant aspects of shield design. Of course, there has been an active
nuclear-data collection program, and extensive properties handbooks are
444
4,.--,.....---r---,---,---,----,-----,.---,--,--,..--.,
Measured
"Last collision" calculation
Free-field detector
o
0
0
Top detector
10- 6
8
w
Bottom detector
'"
o
o
z
,,
oa:
....
::>
w
:""-CaICulation with
:
,,,
I
,,
"Ground shadowing"
._-----305m
2 '
,VToPdetector
Free-field detector_~21r shield
~ '\. Bottom detector
AIr
1O-7~-+_+-_;_----i:--t<_-*-+.-~;_---,!o-~-_:i.
10
12
HEIGHT (HI, m
14
16
18
20
22
Fig. S.IO-Comparison of measured and calculated scattered fast-neutron dose near

air-ground interface 305 m from 14-MeV neutron source. [From R. L. French and L. G.
Mooney, Nucl. Sci. Eng., 43: 279 (1971).]
available. 3 ? Virtually every accelerator laboratory has been engaged in

measuring some interaction cross sections, most research reactors have been
used to measure attenuation constants, and numerous national and international organizations are engaged in collecting, compiling, and evaluating the
resulting data. But all this effort has been devoted to nuclear data; relatively
little effort has been spent relating these results to other physical and
chemical properties important to the selection of shield materials.
Perhaps the deficiency arises because the materials task requires a variety
of skills: nuclear interaction, heat transfer, chemical interaction, structural
characteristics, other physical properties, and economics must all be applied
445
in a thoroughgoing analysis. It is evident that the selection of shield materials

deserves additional attention.
8.3.1 Considerations in Materials Selection
The primary nuclear properties to be considered are obviously neutron
and gamma-ray attenuation. As a rule, good gamma-ray attenuators also
produce secondary gamma rays from neutron inelastic scattering and
radiative capture; thus the production of secondaries is also a consideration.
For shields requiring several orders of attenuation, volume and mass affect
materials choices. Structural strength at operating temperatures is also
involved. Some low-strength materials having excellent attenuation properties may be rejected because the structure necessary to support them
offsets their primary advantage.
Inner layers of the shield usually require some thermal protection; thus
cooling must be considered. Good gamma-ray attenuators are heavy metals
and are usually the highest density material in the shield. To conserve weight
and cost, these materials are usually placed nearest the core, which gives rise
to a cooling requirement.
Different design objectives lead to the selection of different materials.

For a fixed power-reactor system, cost is a primary consideration; thus the
local concrete aggregate becomes a baseline for comparison with other

materials. Higher attenuation substitutes may be considered only if some
cost saving results.
For mobile systems, total weight becomes the primary consideration,
and more exotic materials can be considered. We will discuss materials in
both the exotic and conventional categories.
In a recent Russian book on nuclear power plants, Komarovskii 38 has
summarized in wry fashion the contradictory properties required of
biological shield materials. We quote directly from the translation:
In general, protective materials used for biological (and radiation) shielding
must have the following properties and ensure the following:
1. Maximum possible density (i.e., maximum specific or bulk weight) to
ensure maximum attenuation of gamma radiation from the core and for the
absorption of the secondary gamma radiation caused by the capture of neutrons
by the material of the shield. Materials of higher density also slow down fast
neutrons efficiently. transforming their energy into epithermal and thermal
neutrons which are then more easily absorbed. (This statement evidently refers to
the effect of inelastic scattering in heavy materials.)
2. Constant density and homogeneity of the shielding material (along all or
part of the shielding).
446
3. A definite, economically justified content of hydrogen (usually contained

in chemically bound water) for an efficient slowing down of epithermal neutrons
to the thermal level. It is desirable for the hydrogen content of the shielding
material to be independent of the temperature of the material.
4. Minimum possible energy of capture gamma radiation forming on absorp~
tion and slowing down neutrons.
5. Satisfactory resistance of the material to radiation effects.
6. Low residual radioactivity, i.e., minimum amount of long-living radioisotopes which might accumulate in the shielding under the effect of radiation.
7. Mechanical strength (primarily compressive strength) and three-dimensional
stability of the structure. The latter requirement may be explained by the fact
that the shielding against external radiation is in most cases also the carrying
structure of the nuclear plant (cover, walls of the reactor tank adjoining the
shielding, etc.); the shielding must not have continuous seams and joints or local
caverns, holes, or traversing cracks through which radiation might easily
penetrate.
8. Relatively low modulus of elasticity thus permitting reduction of tensile
stresses in the outer zone of the shielding as a result of its unilateral heating
(heating from one side).
9. Maximum possible heat conductivity so as to reduce the temperature
gradient over the thickness of the shielding and thus the tensile stresses in the
outer zone of the shielding.
10. Minimum thermal expansion of the material, also enhancing reduction of
thermal stresses.
11. Minimum thermal hydration during hardening (of concrete).
12. Minimum contraction (shrinkage) during the construction of the shielding
and later, since upon contraction the various apertures, pipes, and conduits in the
shielding may become displaced.
13. Refractoriness and thermal stability, and if the material catches fire, the
combustion rate should be as low as possible with the flame tending to
self-quenching as the fire source is removed. Particular demands as to refractoriness are put on materials used to shield the reactor core directly. The shielding
materials should not melt even at emergency temperature in the reactor. This is of
particular importance for fast-neutron reactors.
14. Waterproofness.
15. Gasproofness.
16. Safe (in terms of effects of the evolution of gases from the material during
its heating).
17. Chemical inertness and first of all resistance to the chemical action of the
coolant (this refers particularly to sodium-cooled reactors in which the shielding
should not include water-metal mixes).
18. Noncorrosive with respect to metal (steel, lead, aluminum, etc.).
19. Easy construction as well as easy repair and if necessary, replacement.
20. Low cost.
447
In nature there are no materials which satisfy all these requirements. Artificial
materials (such as concrete) must therefore be used, or in some cases a shield
consisting of layers of different materials must be constructed. In selecting the
proper protective materials allowance must be made for the ratio of neutron-flux
intensity to gamma-ray intensity.
Many of the above requirements are conflicting. When using materials of high
density, for example. the requirement of minimum cost cannot be satisfied. High
hydrogen content conflicts with the requirement of maximum content of heavy
elements. [n this case a compromise must be found by using a mixture of heavy
and hydrogen-containing compounds (e.g., concrete) or, which is less frequently
done, by constructing the shield in alternating layers of heavy and light materials.
In selecting the construction materials for the shielding all the engineering and
economic advantages and disadvantages of the various materials must be weighed
carefully and the operating conditions of the particular nuclear plant must be
taken into account. Preference should therefore be given to materials of suitable
properties. When constructing new premises for a reactor plant, it is usually
possible to create a shielding of large volume but of less heavy materials. On the
other hand, when building a reactor in already existing premises the thickness of
the material often has to be reduced due to restricted access conditions, and in
this case more expensive materials of higher density have to be used.
Now let us examine the ptopetties of some materials actually in use for
the shielding of mobile and stationary systems and compare some of their
characteristics with the preceding requirements.
8.3.2 shield Materials for Stationary Reactor Systems
Concrete is the most commonly used radiation shielding material for
fixed reactors and accelerators. Since concrete is a mixture of elements, the
variation in composition is large. Therefore, when referring to concrete as a
shielding material, we should state the specific composition in order for its
characteristics to have meaning. Concretes are generally referred to as
ordinary or heavy. Ordinary (portland) concrete has a density range from 2.2
to 2.4 g/cm' and is largely composed of oxygen and silicon or calcium (or
both), depending on the aggregates (granite, sandstone, or limestone). Heavy
concretes with densities ranging from 3 to 6 g/cm' are those in which the
aggregate consists of iron or iron are, baryte (barium sulfate), steel shot,
punchings, sheared bars, or other metal additives. There is obviously an
added cost factor for heavy concretes.
Other materials that have been used in large quantity for shielding are
compacted soil, iron, lead, and water. Experience has shown, however, that
448
heavy-concrete shielding has many advantages over these materials. 37 ,39 ,4'
Ir is also evident that these materials satisfy few of the preceding desired
properties,
Probably more than 90% of the stationary reactor systems designed for
power production are shielded with concrete, As an illustration, of
twenty-three test-reactor shields described in a ] 968 survey published by the
Radiation Shielding Information Center ,41 six used both heavy and ordinary
concrete, eleven used only ordinary concrete, and five used only heavy
concrete. Only one shield is described without concrete. Fourteen contained

water: eleven were swimming pool reactors, two were boiling-water reactors,
and one was a homogeneous solution. Three others included heavy water.
Other materials mentioned were iron, steel, wet sand, boral, graphite,
borated graphite, polyethylene, earth, and boric acid. Most of the materials
used in the experimental reactor shields, other than concrete, such as iron,
lead, and graphite, individually have poor structural or radiation shielding

properties but, when combined in laminated configurations, result in
extremely good shields.
Let us look at rhese materials in more detail. To compare nuclear
properties of materials, we need a set of standards for comparison. Despite
the universal use of concrete, we shall use lead for gamma rays and water for
neutrons as standards for comparison of attenuation properties. Both arc
frequently used and are regarded as excellent attenuators. Both are pure
substances (and relatively free from impurities in practice) and so do not
present the multiplicity (and variability) of constituents with which we must
contend with concrete.
(a) Water. Because of the success of the swimming pool reactor as a
research tool and of its derivative, the pressurized water reactor, as a pioneer
power source, water layers are found in reactor-shield analyses almost as
often as concrete. Water is an excellent neutron attenuatar because of its
large hydrogen content. It is an inferior attenuator of gamma rays because of
its low electron density. Since oxygen has no thermal-neutron capture and
the 330-mb thermal-neutron capture cross section of hydrogen gives only
2.2-MeV gamma rays, water is relatively low in secondary gamma-ray
production.
Considerations when water is used as a shield material include containment, temperature control, corrosion, and purification (demineralization).
(b) Lead. Best of the conventional materials for gamma-ray attenuation,
lead is exceeded only by uranium. Lead bricks are to be found in almost
449
every radiation laboratory as portable shield bricks. Lead suffers creep at

high temperatures (above 260C) and must be constrained to avoid
downward flow.
Lead shot makes good filler for cavities in plugs and doors in shields. If
provision is made for extracting the filler in place, it becomes a great
convenience should it become necessary to dismantle or remove the door.
Lead has relatively poor neutron properties and produces 7 A-Me V gamma
rays on neutron capture. Impurities in commereiallead can cause significant
activation problems. Reference 37 gives additional physical properties.

(c) Graphite. A good neutron moderator and reflector material, graphite
is used extensively in reactor design. Since it is available in very pure form, it
exhibits few secondary gamma rays. Like water, it is a poor gamma-ray
attenuator. Besides its use as a reflector-moderator, it is also found in
thermal shields and occasionally as a primary shield material (see the
discussion of the Enrico Fermi power plant in Chap. 10). The excellent
high-temperature properties of graphite are additional reasons for its
application.
(d) Iron. Iron is almost always present in the reactor structure as steel
and is used for thermal shields and pressure vessels. Iron produces a number
of secondary gamma rays up to 10 MeV, most of them at 7.6 MeV. Not
surprisingly, it is intermediate in gamma-ray attenuation between water,
carbon, and boron at the low end and tungsten, lead, and uranium at the
high end. Neutron attenuation per unit thickness of iron is good; it is
intermediate between the same materials mentioned for gamma rays. The
isotope s 8 Fe is easily activated by thermal neutrons producing s 9 Fe, with a
59-day half-life, which emits a 1.5-MeV gamma ray. This activation of iron
may limit accessibility to a reactor after shutdown. 38 All steels embrittle
after exposure to thermal-neutron fluence of the order of 10 19 to 10 2
neutrons/cm 2 , showing hardness increases from 50 to 100%, depending on
the alloy and heat treatment. (See Chap. 23 of Ref. 37.)
(e) Boron. This element is used principally to absorb thermal neutrons
because of the cross section of 3840 b for the
B(n,o</'l reaction. The
alphas are easily absorbed (producing a heating problem for large incident
flux densities); the gamma ray at 0.5 MeV should be considered in borated
water and polyethylene. Boron is available in a number of useful forms; as
boric oxide (B 2 0 3 ), as boron carbide (B.C), as elemental boron, and as a
sandwich material called bora!' Boral consists of an aluminum-B.C mixture
1.
450
clad with pure aluminum. Pure boron has also been used as an additive to
polyethylene and graphite to add neutron absorption properties to good
neutron attenuators. It has also been added to some steels ro reduce
secondary gamma-ray productions and activation problems of iron.
(f) Concrete. Since concrete is a mixture of materials, the mixture can
be tailored to provide optimum structural and shielding properties. Table 8.3
gives the elemental composition of ordinary portland concrete and several
concretes fabricated for improved shielding properties. 37 Table 8.3 contains
only a small sample of the many custom-designed concretes available.
References 37, 42, and 43 contain information on mixing proportions,
placement techniques, costs, etc. There is no reported standard concrete

composition since variation in normal concretes is due primarily to the
aggregate. One can, however, use the ordinary portland concrete as a basis
for comparing the improved structural and shielding properties of heavy

concretes.
The ferrophosphorus concrere is of high density and has excellent

bonding qualities. The aggregate is ferro phosphorus ore. The iron-limonite
is a heavy concrete containing the iron-ore limonite and steel punchings as
aggregates. Limonite is composed of goethite and hematite, both ironbearing ores. This concrete is similar to that used to shield the Brookhaven
graphite reactor.
The iron-portland and several baryte concretes were all developed for
shielding use. The iron-portland concrete is of high density. Lumnite is a
cement (as is portland) which is better described as calcium aluminate
hydraulic cement. Lumnite has a high aluminum and calcium content.
Colemanite, used as an aggregate, is hydrated calcium borate with a high
fixed-water and boron content. Baryte is a barium sulfate ore used as an
aggregate in heavy concretes. The baryte concretes contain boron for
neutron absorption and hydrogen for neutron attenuation, and the baryte
ore serves as a good gamma-ray attenuator. The densities vary from 3 to
3.2 glcm', which is lower than iron-aggregate concretes.

Neutron and gamma-ray attenuation per unit mass vary only slightly
from one of these mixes to the next. The principal advantage in the heavier
mixes lies in the smaller volume needed for a given total attenuation.
8.3.3 shield Materials for Mobile Reactor Systems
Weight rather than cost is generally the principal design criteria in the
application to mobile reactor systems. The emphasis, therefore, shifts to the
Table 8.3-ELEMENTAL COMPOSITION OF CONCRETES!

Composition, wt.%
Lumnite-
Element
oetite
Ironlimonite
portland
1.4
1.0
50.5
0.3
61.2
0.5
72.1
0.05
87.50
0.33
52.9
0.2
4.4
32.0
0.6
0.7
10.4
0.2
4.2
18.0
0.2
6.1
5.82
0.13
3.96
Portland
concrete
Mag-
Iron-
Colemanitebaryte
portland-
frits~
Lumnitecolemanite-
colemanite-
baryte
baryte
baryte
2.19
0.56
1.04
33.80
0.23
6.26
40.13
3.31
1.21
0.10
0.64
0.02
3.07
1.09
0.88
36.95
0.14
5.48
38.59
0.96
0.11
1.87
1.10
1.02
36.98
0.20
7.67
38.03
1.49
0.11
1.76
0.12
1.32
0.04
Boron
Ferrophosphorus
concrete
en
Iron
Hydrogen
Boron
Oxygen
Magnesium
Calcium
Barium
Silicon
Sodium
Potassium
Aluminum
Manganese
Chlorine
Carbon
Phosphorus
33.7
1.6
1.3
3.4
3.5
3.4
1.4
0.91
2.9
0.07
0.4
0.5
1.6
0.33
0.35
1.03
0.85
0.98
34.89
0.22
8.46
40.70
1.76
0.11
0.61
0.01
:I:
;;
t"'
"
>
:I:
t"
::l
[>
>
;a..,
~Z
en
."
.::j
0.1
Sulfur
0.17
1.07
Zinc
Fluorine
Titanium
2.8
19.7
0.1
0.05
9.63
9.15
0.66
0.23
9.06
8.97
1.27
0.071
3.1
3.1
..,t"
Density,
g/cm'
2.3
3.45
4.8
4.27
5.8
3.2
3.1
tFrom H. E. Hungerford, Reactor Materials, Vol. I, Materials, C. R. Tipton,

Publishers, a division of John Wiley & Sons, Inc., New York, 1960.
Jr.
(Ed.), p. 1086, Interscience
<.n
~
452
improved shielding provided by the so-called exotic materials used in layered

combination.
The approach in designing minimum-weight shields for reactor systems

usually is to select the most efficient of the neurron shielding materials
compatible with the design and to combine this with the most efficient of
the gamma-ray shielding materials rather than to select a single bulk
attenuator medium. These two component materials are usually layered in
some manner that yield, minimum weight by providing the proper balance
between primary and secondary radiations and geometric effects on material
volume.
The efficiency of a neutron attenuator can be correlated to its hydrogen

density, and the efficiency of a gamma-ray attenuator can be correlated to
the total density of the material. Thus minimum-weight shield component
materials are such combinations as lead plus polyethylene, tungsten (wolfram) alloys plus lithium hydride, or depleted uranium plus titanium hydride.
Shields of this nature placed in proximity to a propulsion reactor
invaria.bly create a thermal problem because of the high-energy absorption.
Internal temperature rise can be limited by cooling the materials with a
forced coolant flow between thin layers of the material or through internal
cooling channels. Some of the high-density gamma-ray shielding materials,
such as wolfram and depleted uranium, have a relatively high thermal
conductivity and can tolerate a fairly high operating temperature if clad to
protect against oxidation and corrosion. In contrast, most materials that are
high in hydrogen content have low thermal conductivity and cannot tolerate
high operating temperatures without melring or decomposing.
Although submarines were the first application for nuclear propulsion,
the application to surface ships is also well known. Two such examples are
the U. S. merchant ship N.S. Savannah and the Russian icebreaker Lenin.
The reactor shield of the N.S. Savannah is composed of light water, lead,
polyethylene, standard concrete, and heavy concrete (described in
Chap. 10). The reactor shields of the Lenin are composed of water, steel, and
heavy concrete. 41
In addition to mobile reactor systems, other mobile or portable sources

of radiation require shielding, e.g., neutron sources fabricated from
plutonium-beryllium or polonium-beryllium and radioisotopes (usually
gamma-ray emitters) resulting from high-intensity neutron exposures in
reactors.
Shipping and storage casks must be provided for sources of this type, and
they are usually constructed of lead for gamma-ray shielding and contain
453
paraffin or orher hydrogenous material for neutron shielding. These sources

are usually small and present few design problems. An exception to this is
found in spent-fuel-element shipping casks where large burn ups have been
experienced. 44 These elements contain sizeable quantities of 242 em,
24 4Cm, and 241 Am, all of which fission spontaneously, giving. rise to a
fission-neutron distribution (as well as the accompanying prompt gamma-ray
spectrum) in addition to the long-term fission-product gamma rays. Shield
design of shipping casks for these fuel elements presents nearly all the
problems of a reactor shield: radiation heating (forced cooling is necessary),
good neutron and gamma-ray shielding, and minimum size and weight
(usually limited by the largest available truck size for transport).
Weight-limited shields are also used for isotope power supplies, which use
the heat generated by the decay of radioactive nuclides to produce
electricity. Such sources have been used to power orbiting satellites and
experimental equipment placed on the moon. Shields for these systems must
obviously be efficient.
As mentioned, among the top contenders in materials for mobile reactor
systems are lead, polyethylene, tungsten alloys, lithium hydride, depleted

uranium (238), and titanium hydride. The properties of some of these
materials arc given in the following paragraphs. Some physical properties of
the metals discussed arc summarized in Table 8.4 and of the hydrogenous
materials, in Table 8.5.
(a) Polyethylene. This plastic is a pure hydrocarbon (CH 2 )n with a
density of 0.92 g/cm' which actually contains 18% (by volume) more
hydrogen than does water (both contain more hydrogen per unit volume
than liquid hydrogen). Water contains 6.7 X 10 22 hydrogen atoms/cm', and
polyethylene has about 8 X 10 22 hydrogen atoms/cm' . Polyethylene does
not activate, but it does soften at 110C, and it supports combustion. A
more-dense version of polyethylene has a higher softening point (200C) and
slightly higher density (0.96 g/cm'). Either material suffers radiation damage
and must be replaced before decomposition reduces its effectiveness
(probably well above 1011 ergs/g energy deposition).'" As mentioned under
the discussion on boron, polyethylene can be borated to produce the added
benefit of high neutron absorption. Linear thermal expansion is large, and
thermal conductivity is very small; both propenies must be borne in mind.
(b) Tungsten Alloys. Tungsten powder blended with binder metals, such
as nickel, copper, and iron, has been sintered to form very dense
machineable alloys with densities varying from 16.5 to 18.5 g/cm'. The
...'"...
Table 8.4-PHYSICAL PROPERTIES OF DENSE SHIELD MATERIALS

[ron
(stainless steels)
Density, g/cm 3
Specific heat,
c.l g-l deg-'

Melting point, c
Thermal conductivity,
cal sec- t cm- t deg- t
Coefficient of thermal
expansion, deg- I
Tensile strength,
kg/cm 2
Lead
Tungsten
Uranium
Concretes
7.9 to 8.0
1.34
19.3
18.5 to 19.04
2.5 to 5.0
0.12
1370 to 1425
0.0309
327.4
6.0
3410
0.02 to 0.04
857
0.16 to 0.2
0.03 to 0.05
0.083
0.40 to 0.46
0.6 to 1.0
(6to 11)10-'
(14 to 20)10-6
29.1 (10-6 )
(4.4 to 5.0)10- 6
(8 to 10)10- 6
(6 to 10)10-6
(4 to 6)10'
160 to 180
(8.1)10'
(7.7 to 9.1)10'
Table 8.S-PHYSICAL PROPERTIES OF NEUTRON SHIELD MATERIALS
Graphite
Density, g/cm 3
Sublimation,oC
Polyethylene
Lithium
Titanium
Zirconium
hydride
hydride
hydride
1.5 to 1.9
36S0
0.92 to 0.96
0.82
3.9
5.6
Sublimes
110 to 200
680
400 (decomposes)
430 (decomposes)
(3.5 to 8)10- 6
(16 to 18)10-'
0.1 to 0.3
(6 to 8)10--
0.2 to 0.5
0.55
Melting (softening)
point,OC
Coefficient of thermal
expansion linear,
deg- 1
Thermal conductivity,
cal sec- t em-I deg- t
Specific heat,
cal g-l deg- t
455
more common of these have a density of about ] 6.9 to 17.1 g/cm 3

Fabrication is limited by currently available equipment to pieces with a
maximum dimension of about] m. Although this material has good strength
properties, it is still likely to be held in place by a steel framework because
of the limit on piece size. Tungsten falls just under lead in gamma-ray
attenuation per unit mass. It is much better than lead per unit thickness for
neutron attenuation and worse for secondary production.
(c) Depleted Uranium. The separation of uranium isotopes to produce

235 U-enriehed material for reactor and weapons uses has reacted a supply of
depleted uranium that is essentially pure 23 8 U. This material has a high
density (18.7 to 19.08 g/em 3 ) and good strength properties. It may be cast,
rolled, forged, extruded, and machined (with environmental precautions).
Uranium oxidizes in air to form first a brown and then a black surface
layer. In moist air corrosion will continue beyond this stage as the coating
layer spalls off and a new layer forms. The metal from this process can be
protected by cladding or alloying or both. Alloying with a 2% molybdenum
has proven satisfactory from the corrosion standpoint and also serves to
improve the mechanical properties of the material. There is a slight activity

associated with depleted uranium, but this does not severely restrict its use.
We should note, however, thai depleted uranium always has a residual
amount of 23 s U that should not be neglected in estimating radiation
hcat:ng, secondary gamma-ray production, and activation. Precautions arc
required in machining the metal because of its pyrophoricity and the hazards
of internal toxicity and internal radiation exposure.
It is unlikely that any structural support or framework would be

required for 238 U shields in that they can be manufactured in about any size
and geometry and they possess strength properties on a par with mild steel.
Uranium is the best attenuator for gamma rays on a mass absorption
basis as well as on a thickness basis, and it provides roughly the same neutron
attenuation as lead.
(d) Lithium Hydride. Lithium hydride (LiH), the best known neutron
attenuator,46 contains approximately 12.6 wt.% H 2 , has a density of 0.78
g/cm 3 , and has a melting point of 683.5C. It is difficult to fabricate,
normally requires surface protection at all temperatures, combines actively
with water, and requires containment, venting, and structural support.
Weight-conscious shield designers like to recommend its use; mechanical

designers shudder at the thought. The structure and cost penalties in
applying this material should be analyzed carefully in considering its use.
456
Lithium has no capture gamma ray. Much consideration, including experimentation work, has been given to this material for reactor shields in
space applications.
(e) Titanium Hydride. Titanium hydride (TiH z ), with a density of 3.76
glcm' and a maximum useful temperature of approximately 600C, has a
maximum Hz content of 3.9 wt.%. This hydride is self-supporting. It needs
no surface protection in air but will evolve hydrogen at high temperatures.
8.3.4 Comparison of Attenuation Properties
The efficiency of various materials for the attenuation of gamma rays
increases with the average atomic number of the constituents. Figure 8.11
shows various material thicknesses required to flcduce gamma-ray int~nsities
by a factor of 10 as a function of the incident gamma-ray energies. These are
narrow-beam conditions. When extrapolated to broad-beam conditions (to
include buildup), the separation between the lines would become even
greater because the more-dense materials inhibit buildup.
Neutron attenuation is greater in low-z materials, particularly those
containing large quantities of hydrogen. Figure 8.12 shows the neutron-dose
attenuation in infinite media of normal concrete,47 hydrous iron-ore
concretc,47 water,47 a simple hydrocarbon and boron mixturc,47 and
lithium hydride 48 for a point isotropic source of fission neutrons. The

concrete compositions are given in Table 8.6 as types A and H. The lithium
hydride data of Kam and Clark' 8 in Fig. 8.12 have been normalized to the
data of Nichols' 7 for the other materials at R = 0 em.
Figure 8.13 shows the capture-gamma-ray dose from the fission-neutron
source in water, concrete, and the hydrocarbon and boron mixture 4 'in
Fig. 8.12. Concrete does not thermalize neutrons as rapidly as water, and it
is also a better gamma-ray attenuator. The effect of the boron is evident in
the lower curve.
The effect of water content in concrete on the neutron-dose attenuation
can be seen in Fig. 8.14 for concretes A, B, C, and 0. 47 The compositions

are given in Table 8.6. Concrete type A is called magnetite A in Volume 1Il
of the Reactor Handbook.' 9 Concretes B, C, and 0 are the same concrete
with three-fourths, one-half, and one-fourth the water content of concrete A, respectively. The effect of decreasing water content on the dose
rates, all shown as solid lines in the figure, is very pronounced. The limonite
concrete (E) has a relatively low density but a high water content; dose rates
457
200
100
80
60
40
g c
E
0
20
2,
"'"'w
Z
'"
U
,0
'i:
....
.:r.'"
w
>-
-'
....
z
w
....
Sintered tungsten alloy

(p' 171
10
E. MeV
Fig. 8.11-Narrow-beam gamma-ray absorption for various materials.
458
10- 9
"ce
~
c 10.10
~
N
'eu
W
:g
"~
10-
11
CH 2 + 5 wt. %
a:
boron
(p = 1.0)
t:;
w
Z
N
a:
i;
10- 12
Lithium
hydride
(p =
0.75)
1O-14~_...J.._-:~_..J.._-:~_..J..._-:~_..J.._-:!~_..J.._""'~_..J..->~
20
40
60
80
100
120
DISTANCE (Rl,cm
Fig. B.12-Neutron dose vs. depth in the indicated materials fission-neutron source.
(Based on data from Nichols47 and Kam and Clark.48 )
Table 8.6-ELEMENTAL DENSITY (g/cm') OF CONCRETES ASSUMED FOR SHIELD ANALYSISt
'":r:
tTi
Type of concrete
iron orc
Hydrous
iron orc
Limonite
Ferrophosphateserpentine
G
Magnetiteserpentine
0.0108
1.218
1.960
0.108
0.128
0.00538
1.176
1.960
0.108
0.128
0.0370
1.113
1.189
0.165
0.227
0.0208
0.7805
1.355
1.044
0.196
0.0122
1.180
1.181
0.457
0.458
A
Hydrous
Hydrous
C
Hydrous
Element:!:
iron ore
iron orc
~Cr,Mn, V
0.0215
1.304
1.960
0.108
0.128
0.0160
1.261
1.960
0.108
0.128
H
0
.!. AI. Mg. s. P
~Ti
t'"
J:;,
Total
3.522
47
3.473
3.425
3.377
2.731
3.396
tFrom Nichols.
:j:Cross sections for the underlined elements were used in the fastneutron dose calculation.
3.288
H
Normal
0.0200
1.116
0.491
0.612
0.118
2.357
:r:
tn
:>'
>-l
~
Z
'"
C3
.~
tn
>-l
r>
460
'E
u
ui
V>
"
"'"
"''""
Water
Ip
1.01
'"::>
t
'"
10-11
u
N
'"4i
Normal
concrete
Ip
2.361
CH 2 + 5 wt. % boron
Ip
1.01
lO-12,=-_-'-_-,I;,-,_.l-__~_--,o<-_-.L._~~_.l--,,*,,_-'-_~
20
80
DISTANCE (Rl, em
Fig. 8.13-Capture-gamma-ray dose for the indicated materials fission-neutron source.
are approximately the same as those for concrete A. Concretes F and G are
fictitious mixtures of aggregates.
Figure 8.15 shows the effect, from an analysis by Lahti,S 0 of placing
various thicknesses of tungsten between the fission-neutron source and the
lithium hydride medium. The data show the neutron dose as a function of
distance in natural lithium hydride after first passing through the indicated
thicknesses of tungsten. Even though tungsten is a heavy material, it provides
good neutron atten uation per unit thickness.
461
'"E
0
10. 2
'0
~
0
g
,
.s
10...3
E
E 10-
e
Eo:
'"0.,:
10... 5
10-6
107L_---'L.._--l._ _..L_ _..J..._ _.l.-__.-JL.._...:il---.J

o
w
~
W
W
100
1W
lW
DISTANCE IR), em
Fig. 8.14-Fast-neutron dose rate in various concretes as a function of the distance from
a unit point isotropic fission source of 235 U. (From Nichols. 4 7)
Hydrogen
Concrete type
Hydrous iron orc

Hydrous iron ore
Hydrous iron ore
Hydrous iron ore
Limonite
Ferrophosphateserpentine
G Magnetiteserpentine
A
B
C
D
E
F
Density,
g/cm 3
content,
3.52
3.47
3.42
3.38
2.73
0.022
0.016
0.011
0.0054
0.037
3.40
0.021
3.29
0.020
g/cm 3
462

10- 5
N
\
,\
E
u
";"u
10. 6
.
0
10- 7
";"
"
10.8
'"
0
"
X
10- 9
a:
Tungsten
thickness
, \
'~
It!.
\'\
Lithium hydride
'\'
'\\
10.5
\:\
i= 10. 10
\,\\
"
:::>
zw
\\\
\'\\\
\
10- 11
'"
"z0
a:
>-
:::>
lithium hydride
\'~
"
followed by
hydride
followed bV
hydride
followed by
'\\\
15.5
Tungsten
lithium
Tungsten
lithium
Tungsten
\\\
" \
>>-
Material
om
\' \
'\:\
10- 12
PENETRATION DEPTH IN LITHIUM HYDRIDE (RI, em
Fig.8.15-Discrete-.-ordinates calculation of fission neutrons in tungsten followed by
lithium hydride (Fwm Lahti.'")
8.4 SHIELD OPTIMIZATION

The problem of finding the optimum materials arrangement for a shield
that satisfies a given set of design criteria is the most intriguing in the area of
reactor shielding. It has challenged many investigators!-6o and, to this
writing, a general solution has not been found and may not be. Assuredly, a
few workable schemes have been developed, and each study has contributed
SHIELD HEATING, AIR TRANSPORT, ETC,
463
some insight to the subject. A few have published evidence (parameter

studies or claimed weight savings) intended to substantiate their methods,
The complex of phenomena involved in expressing the performance of the
shield invariably requires simplifying or limiting assumptions that detract
from the result. The problcm intrigues the shield designer for several reasons,
The classical mathematical analyses of Lagrange and Euler give us the
guidelines for deriving the optimum conditions, Variational calculus and the
relations for determining stationary values of an integral are the usual
starting point, t Since shield weight is usually a significant fraction of the
total system weight, the potential for gain seems large. Even in systems
where weight is not important, optimization with respect to cost is
important and provides ample incentive for pursuing the problem,
Most frequently optimization is with respect to weig;ht. Other conditions
that must he met, such as volume restrictions, dose-rate requirements, or
number of shield layers, are called constraints. These constraints, together
with the boundary conditions, help determine the parameters arising from
the optimization procedure.
The optimization proceeds in at lcast two phases. The first is to apply
the necessary conditions for a minimum value of the weight function. The
analogue in elementary calculus is to find the stationary values of the
function y(x). The function has a stationary or extremum value (both
include maxima and minima) at x = a if y'(a) = 0. To find the value a, one
simply takes the derivative of y with respect to x, sets it equal to zero, and
solves for a.
The second phase is to determine the sufficient conditions for a
minimum as opposed to a maximum or a local minimum (saddle point). This
second phase usually involves proof of the existence of the minimum. To
complete the analogy with differential calculus, the sufficient condition that
y'(a) is a minimum is that y"(a) > [for a maximum, that y"(a) < OJ and
that y'(x) is continuous in the neighborhood of x = a.
The variational approach to shield optimization involves the generalization of these simple conditions to apply to functions of several
variables with accessory constraint conditions. The variational problem is to
find a function such that a related definite integral has a minimum value.
Troubetzko y S 6 provides a recent example of optirTlization work done in
this country in a method called shield synthesis. His objective was to devise a
tVariational calculus, Lagrange's method of undetermined multipliers, and Euler's equation are
found in texts on mathematical physics, for example, ivlorgenau and Murph y 6 I or Butkov. 6 2
464
design technique that yields the selection and arrangement of materials

giving minimum weight for a spherical shield given reactor power and total
dose. He uses simplified models for neutron and gamma-ray attenuation and
secondary gamma-ray production. Application of the variational conditions
to the model leads to restrictions on the choice of possible materials and the
relation of materials in adjacent layers. He concludes that the innermost and
outermost of n arbitrary concentric layers must be pure materials; any other
layer may contain a radius-dependent mixture of no more than two
materials. Regions adjacent to such mixtures must be pure materials identical
to one of the two components of the mixture. Troubetzkoy states that the
approximations due to the simplified attenuation model can be offset by
performing an exact numerical analysis of the shield (presumably by Monte
Carlo or discrete ordinates) to obtain its performance to a better accuracy.
He reports that shield synthesis gives considerable weight savings over more
conventional designs.
This approach yields a procedure for greatly narrowing the choice of
materials and mixtures that can be used. The approximation due to
assumptions of simplified, manageable attenuation models does not affect
the accuracy of the design of the shield, but it probably does affect the
determination of the optimum. However, studies have shown that these
minima are usually rather broad; thus, if one is indeed near the actual
optimum, most of the potential weight saving has been realized.
So argue the Russian authors Suvorov and Fedorenko" 0 who have
recently reported a somewhat different approach to optimization of a
reactor shield. They also use a variational approach but determine the
optimum variations of the location of heterogeneous layers with constant
concentrations of material within each layer. Theirs is also a one-dimensional

approach (using a cylindrical shield), but they use linear perturbation
theory. By altering the number of layers and the choice of materials, they
are able to compare min ima so derived to observe the effect of these choices.
It is noteworthy that their test for having reached an optimum set of
thickness is that comparable variations in each layer give equal variations in
the dose outside the shield. This technique enables them to show the effect
of adding boron to an iron-water (or lead-water) shield either as a
homogeneous mixture or constrained as additional layers. They also show
the advanrage of a ten-layer two-material shield over a five-layer shield of the

same materials. This technique uses a five-group neutron attenuation in a PI
approximation, but exponential gamma-ray attenuation using infinitc-
medium buildup factors. Neutron-produced secondaries arc added to the
465
primary gamma-ray distribution. Weight savings of the order of 20 to 40%

are shown over straightforward designs. From the published description of
the technique, it is evident that a great deal of high-speed computer time is
required for applications of this method.
The Troubetzkoy method clearly gives more guidance in the selection of
materials, but the Suvorov-Fedorenko method would seem to give more
realistic results.
Sheffield 53 and Miller and Cranford 54 in an earlier study used gradient
nonlinear programming to obtain minimum weights of divided crew
shield-reactor shield geometries appropriate to aircraft or rockets. [n this
case the dose in the crew compartment is the primary constraint, and the
effect of air transport from reactor shield to crew compartment must be
included in the optimization procedures. This method places emphasis on
the shape of the reactor shield rather than on the number of layers or
composition. It is interesting that the optimization procedure in one phase
on the computer is to test at each step for the most effective sector on the
reactor or crew compartment to place a given increment of shield material.
The increment is then placed where it gives the greatest reduction in dose.
Aronson, Klahr, Steinberg, and Held" developed a procedure for
optimizing the placement of shadow shields (shield splitting) in space
vehicles. In a paper given at a conference in Harwell, Enginol S 8 describes an
interesting variational technique that can use any available method for
computing the attenuation through the shield and optimizes with respect to
cost or weight or any linear combination of parameters.
Another recent program, ASOP, performs detailed one-dimensional

discrete-ordinate transport calculations with ANISN (described in Chap. 4)
for each step in the iterative optimization. 6 3 This feature distinguishes
ASOP from its forerunners since others involve a simplified transport
approximation. This treatment leads to significant improvements in the
optimization technique since weight savings of 25% are estimated."'
REFERENCES
1. H. C. Claiborne, Survey of Methods for Calculating Gamma-Ray Heating, USAEC
Report ORNL-RSIC-8, Oak Ridge National Laboratory, June 1965.
2. H. Goldstein and J. E. Wilkins, Jr0' Calculations of the Penetrations of Gamma Rays,
Final Report, USAEC Report NYO-3075 (NDA-15C-41), Nuclear Development
Associates, Inc., June 30, 1954.
466
REACTOR SHIELDiNG FOR NUCLEAR ENGINEERS
3. H. Goldstein, Fundamental Aspects of Reactor Shielding, p. 221, Addison-Wesley

Publishing Company, Inc., Reading, Mass., 1959 (out of print); now available from
Johnson Reprint Corporation, New York.
4. L. A. Bowman and D. K. Trubey, Deposition of Gamma-Ray Heating in Stratified
Lead and Water Slabs, USAEC Report CF-58-7-99, Oak Ridge National Laboratory,
July 28, 1958.
5. M. L. Tobias, D. R. Vondy, and M. P. Lietzke, Nightmare--an IBM 7090 Code for
the Calculation of Gamma Heating in Cylindrical Geometry, USAEC Report
ORNL-3198, Oak Ridge National Laboratory, Feb. 26, 1962.
6. M. A. Capo, Polynomial Approximation of Gamma-Ray Buildup Factors for a Point
Isotropic Source, USAEC Report APEX-SIO, General Electric Company, August
1959.
7. R. L. Walker and M. Grotenhuis, A Summary of Shielding Constants for Concrete,
USAEC Report ANL-6443, Argonne National Laboratory, November 1961.
8. D. K. Trubey, Gamma-Ray Buildup Factor Coefficients for Concrete and Other
Materials, Nucl. App/. Technol., 9: 439 (170).
9. C. R. Greenhaw, G. O. Mueller, and G. E. Sabian, Use of Diffusion Theory Codes to
Predict Gamma Heating, USAEC Report KAPL-M-DNA-5, Knolls Atomic Power
Laboratory, 1963.
10. R. L. French, Reactor Shield Heating Calculation:; from Gamma Rays, USAEC file
number NP-9354(FZM-I076), Convair, June 1958.
11. K. D. Lathrop, DTF-]V, A Fortran-IV Program for Solving the Multigroup Transport
Equation with Anisotropic Scattering, USAEC Report LA-3373, Los Alamos
Scientific Laboratory, July 15, 1965.
12. W. W. Engle, Jr., A Users Manual for ANISN, A One-Dimensional Discrete Ordinates
Transport Code with Anisotropic Scattering, USAEC Report K-1693, Union Carbide
Corporation, Mar. 30, 1967.
13. F. R. Mynatt, F. J. Muckenthaler, and P. N. Stevens, Development of
Two-Dimensional Di~\crete Ordinates Transport Theory for Radiation Shielding,
USAEC Report CTC-INF-52, Union Carbide Corporation, August 1969.
14. D. K. Trubey, S. K. Penny, and K. D. Lathrop, A Comparison of Three Methods Used
to Calculate Gamma-Ray Transport in Iron, USAEC Report ORNL-RSIC-9, Oak
Ridge National Laboratory, 1965.
15.J. J. Loeshler and J. E. MacDonald, Flexible Monte Carlo Programs FMC-N and
FMC-G, USAEC Report APEX-706, General Electric Company, Apr. 28, 1961.
16.A. Foderaro, L.]. Hoover, and J. H. Marable, Heat Generation by Neutrons, in
Engineering Compendium on Radiation Shielding, Sec. 7.2, R. G. Jaeger (Ed.),
Springer-Verlag, New York, 1968.
17. C. N. Klahr, Heat Generation in Shields, in Reactor Handbook, 2nd ed., Vol. In,
Part B, Shielding, E. P. Blizard (Ed.), Interscience publishers, a division of John
18.R. E. Lynch, J. W. Benoit, W. P. Johnson, :md C. D. Zerby, A Monte Carlo
Calculation of Air-Scattered Gamma Rays, USAEC Report ORNL-2292 (Vol. I), Oak
Ridge National Laboratory, Sept. 10, 1958.
19. M. B. Wells, Studies in Shielding-III. Monte Carlo Calculations of Neutron Scattering
in Air, Report NARF-57-49-T (FZK-9~120), Convair, Nov. 4, 1957.
467
20. C. D. Zerby, A Monte Carlo Calculation of Air-Scattered Neutrons, USAEC Report

ORNL-2277, Oak Ridge National Labocatocy, May 24, 1957.
21. D. K. Trubey, The Single-Scattering Approximation to the Solution of the
Gamma-Ray Air-Scattering Problem, Nucl. Sci. Eng., 10: 102-116 (1961).
22. E. A. Straker and M. L. Gritzncf, Neutron and Secondary Gamma-Ray Transport in
Infinite Homogeneous Air, USAEC Report ORNL-4464, Oak Ridge National
Laboratory, December 1969.
23. J. R. Keith and F. H. shelton, lA-MeV Neutron Transport Results, Report
KN-774-68-1, Kaman Aircraft Corporation, September 1968.
24. W. M. Webster, Neutron Air Transport Calculations, USAEC Report UCRL-S0570,
Lawrence Radiation Laboratory, January 1969.
25. R. E. Vogel, A Fortran II Program for the Calculation of the Transport of Gamma
Rays by the Method of Moments, Report KN~63~76(R), Kaman Aircraft Corporation,
Jnne 1, 1963.
26. J. D. Marshall and M. B. Wells, The Effect of Cutoff Energy on Monte Carlo
Calculated Gamma-Ray Dose Rates in Air, USAEC Report RRA-M67, Radiation
Research Associates, Inc., Oct. 31,1966.
27. N. M. Schaeffer, Energy and Angle Distributions in Air for Gamma Rays, in
Engineering Compendium on Radiation Shielding, Vol. I, Sec. 4.5.2.1, R. G. Jaeger
(Ed.), Spdnge,-Vedag, New Yack, 1968.
28. C. D. Zerby, Radiation Flux Transformation as a Function of Density of an Infinite
Medium with Anisotropic Point Sources, USAEC Report ORNL-2100, Oak Ridge
National Laboratory, Oct. 9, 1956.
29. M. J. Berger, Calculation of Energy Dissipation by Gamma Radiation Near the
Inte,face Between T"" Media,]. Appl. Phys., 28, 1502 (1957).
30. M. B. Wells, Air and Concrete Scattering of Gamma Rays, Report
NARF-59-11 T(MR-N-229), Convai" MaL 20, 1959.
31. M. B. Wells, Air and Concrete Scattering of Neutrons. Monte Carlo Calculations,
Repoct NARF-58-15T(MR-N-197), Convai" ApL 2, 1958.
32.J. I. Marcum, Monte Carlo Calculations of the Transport of 14~MeV Neutrons in the
Atmosphere, Report RM-3531~PR, RAND Corporation, 1963.
33. F. F. Haywood, T. G. Provenzano, and J. A. Auxier (Eds.), Operations
Plan-Opecation HENRE, USAEC Repo't CEX-65.03, Oak Ridge National
Laboratory, 1965.
34. A. E. Fritzsche, N. E. Lorimier, and Z. G. Burson, Measured High-Altitude Neutron
and Gamma Dose Distributions Due to a 14-MeV Neutron Source, USAEC Report
EGG-1l83-1438, EG&G, May 1969.
35. A. E. Fritzsche, N. E. Lorimier, and Z. G. Burson, Measured Low-Altitude Neutron
and Gamma Dose Distributions Due to a 14-MeV Neutron Source, USAEC Report
EGG-1183-1449, EG&G, Aug. 21, 1969.
36. R. L. French, A Fiest-Last Collision Model of the Air/Ground (nte,face Effects on
Fast-Neutron Distributions,Nucl. Sci. Eng., 19: 151~157 (1964).
37. H. E. Hungerford, The Nuclear, Physical and Mechanical Properties of Shielding
Materials, in Reactor Handbook, Vol. I, Materials, C. R. Tipton, Jr. (Ed.), Chap. 51,
468
38. A. N. Komarovskii, Design of Nuclear Plants, translated from Storoitel'stvo

Yadcrnykh Ustanovok, Atomizdat, Moscow, 1965 (Russian), USAEC Report
AEC-te-6722,1965.
39. H. S. Davis, How to Choose and Place Mixes for High-Density Concrete Reactor
Shields, Nucleonics, 13: 60 (June 1955).
40. R. L. Walker and M. Grotenhuis, Concrete Shielding Constants, Nucleonics, 20: 141
(1962).
41. Oak Ridge National Laboratory, Compilation of Data on Experimental Shielding
Facilities and Tests of Shields of Operating Reactors, USAEC Report ORNL~RSIC~24,
November 1968.
42. H. M. Glen, Materials of Biological Shielding, American Society of Mechanical

Engineers Paper No. 57-NESC-51, March 1957.
43. W. Q. Hullings, Materials Data for Nuclear Systems, Report NARF~61-24T
(FZK-9-162), Gene"l Dynamics/Fact Woeth, July 1961.
44. F. P. Roberts and H. H. Van Tuyl, Chemical Analysis of Spent Power Reactor Fuels,
USAEC Repoct BNWL-SA-503, Battelle-Nmthwest, Dec. 27, 1965.
45. A. Shapiro, Radiation Chemistry of Polymeric Systems, Vol. 15, High Polymer Series,
46. R. Van Houten and W. G. Baxter, High Temperature~High Hydrogen Content Shield
Materials, Transactions of the American Nuclear Society Topical Meeting, Nuclear
Materials for Space Application, Cincinnati, Ohio, April 1963.
47.J. P. Nichols, Calculations of Concrete Shields for Fission Sources, USAEC Report
ORNL-TM-1167, Oak Ridge National Labomtmy,June 17, 1965.
48. F. B. F. Kam and F. H. S. Clark, Fission Neutron Attenuation and Gamma~Ray
Buildup Factms foe Lithium Hydcide, Nucl. Appl. Technol., 3: 433-435 Ouly 1967).
49. E. P. Blizacd (Ed.), Reactor Handbook, 2nd ed., Vol. III, Pact B, Shielding,
Intcrscience Publishers, a division of John Wiley & Sons, Inc., New York, 1962.
50. Gerald P. Lahti, Fission Neutron Attenuation in Lithium~6, Natural Lithium Hydride,
and Tungsten, Report NASA~TN-D-4684, National Aeronautics and Space
Administration, 1968.
51. H. Goldstein and E. P. Blizard, A Criterion for the Experimental Optimization of
Two-Component Unit Shields, USAEC Report TID~5018, Nuclear Development
Associates, Inc., May 10, 1951.
52. H. Hurwitz, Jr., Note on a Theory of Minimum Weight Shields, USAEC Report
KAPL-1441, Knolls Atomic Power Laboratory, 1959.
53. R. D. Sheffield, Shield System Optimization-Gradient Nonlinear Programming,
Repoet NARF-57-62T(MR-N-207), Convaie, 1957.
54. R.
A. Miller and W. Cranford, Shield System Optimization, Report
NARF-58-24T(MR-N-207), Convaie, June 16,1958.
55. R. Aeonson, C. Klahe, H. Steinbeeg, and K. Held, Effect of Shield Splitting and
Placement
m
Nuclear-Propelled
Space
Vehicles,
Report
N-65-22849
(NASA-CR-62456; TRG-136-FR), Technical Reseac<h Geoup, Inc., Mac. 1, 1961.
56. E. S. Troubctzkoy, Minimum Weight Shield Synthesis, USAEC Report
UNC-5017(Pt. A), United Nudeac Coepocation, Oct. 15, 1962.
469
57. S. Sasse, Weight Optimization for Radiation Shields of Large Nuclear Reactors,
USAEC Report ORNL-tr-1525, translated from Report ABS-THH-1023 (German),
Oak Ridge National Laboratory, 1966.
58. T. B. Enginol, A Procedure for Shield Optimization, in Proceedings of the Conference
on the Physics Problems of Reactor Shielding, British Report AERE-R-5773 (Vol. 5),
pp. 1269-1287, May 1968.
59. G. P. Lahti and P. F. Hermann, Comparison of Tungsten and Depleted Uranium in
Minimum-Weight. Layered Shields for a Space Power Reactor. Report
NASA-TM~X~ 1874 (N69~35817). National Aeronautics and Space Administration,
1969.
60. A. P. Suvorov and R. P. Fedorenko, Selection of Optimum Metal-Water Shielding for
Reactors, Problems in the Physics of Reactor Shielding, No.3. D. L. Broder, A. P.
Suvorov, and S..G. Tsypin (Eds.), Report JPRS-49745, Joint Publications Research
Service, Superintendent of Documents, U. S. Government Printing Office,
Washington, 1970.
61. H. Morgcnau and G. M. Murphy, The Mathe"",tics of Physics and Chemistry, 2nd ed.,
D. Van Nostrand Company, Inc., Princeton, N.J., 1956.
62. E. Butkov, Mathematical Physics, Addison-Wesley Publishing Company, Inc.,
Reading, Mass., 1968.
63. W. W. Engle, Jr., A Users Manual for ASOP ANISN Shield Optimization Program.
USAEC Report CTC-INF-941, Union Carbide Corporation, Sept. 8, 1969.
64. W. W. Engle, Jr., and F. R. Mynatt, Development of One-Dimensional Shield
Optimization Techniques, in Neutron Physics Division Annual Progress Report for
Period Ending May 31, 1968, USAEC Report ORNL-4280, Oak Ridge National
Laboratory, 1968.
Experimental Shielding
L. G. MOONEY
and S. T. FRIEDMAN
Mensuration marks the first and last steps in executing a shield system. Not
that the shield designer starts his task with a dosimeter, but he does start
with a measured fission spectrum and a large body of measured cross
sections. The final step in shield design consists in a radiation survey of the
completed system to verify that the design objectives have been met.
Furthermore, all the steps between the first and last are based in some
manner on previous measurements, either by direct input of empirical data
or by use of an analytical method that has withstood the test of
experimental verification. So it is with any technology; no uniqueness claim
can be made for shielding, only that this technology is relatively young and
still has very close ties with experiment. Our dependence on measurement is
evident; key experimental results have been quoted in each topic where they
substantiate a method or provide needed input. In this chapter the
presentation of shielding data is incidental; our purpose rather is to convey
something of the flavor of experimental shielding by describing a selection of
measurements and facilities chosen to represent the range of these
investigations.
Radiation detectors are the tools of this trade, and our discussion will
begin with a brief description of them. Shielding experiments may
approximately be classified in three categories: materials properties, phenomenological studies, and shield-system tests. This is an arbitrary, nonexclusive classification. Some experiments could obviously be classified in
more than one of the three. Investigations of shield materials include those
studies whose objective is to characterize the transport properties of bulk
materials or combinations of materials. These include slab penetration
measurements and cross-section studies. Phenomenological studies involve a
particular geometry; they include air-transport, air-ground interface, albedo,
and duct-penetration studies. Shield-system studies include mock-up experi-
471
472
ments and proof tests. The system test, or mock-up, is always connected to a
particular design; the other two types of experiments usually have more
general applicability. Shield-system tests are discussed in Chap. 10 (Shield
Design), where the tests are described in the context of their related designs.
It is evident that an experiment in anyone of our three classes might
serve several purposes. As already stated, some quantities have not been
predicted and have only been obtained by measurement. Some geometries
are sufficiently complex to defy accurate analytical description, and some
experiments are performed to verify a calculation or a design method. These
various purposes will be clarified by the description of experiments to
follow.
9.1 DETECTORS FOR SHIELDING EXPERIMENTS
Measurements have meaning only to the extent that detector outputs can
be interpreted. Thus it is imperative that the user be familiar with the
response, intensity range of operation, general characteristics, and limitations
of the detector used. In the following sections, we discuss the detectors
commonly used in shielding measurements. References 1 through 6 contain
more complete descriptions.
9.1.1 Active Neutron Detectors
Two types of detectors are used to determine the neutron environment:
(1) active electronic detectors, which provide data directly while the neutron
source is operating, and (2) passive detectors, which are installed before an
experiment, exposed to neutrons, then retrieved and read after the source
has ceased operating.
The Hurst fast-neutron dosimeter is a proportional counter consisting of
a cylindrical chamber lined with polyethylene and containing a hydrogenous
gas, usually methane. The counter is operated with a pulse-height weighting
circuit. Since the proton-recoil pulse-height distribution depends on
incident-neutron energy, proper pulse weighting produces a total count
proportional to fast-neutron kerma in the counter material. A low-energypulse bias voltage (corresponding to 100- to 200-keV neutron energy) is
necessary to reject gamma-ray pulses in a mixed neutron and gamma-ray
field. The bias succeeds in rejecting gamma-ray pulses below 10 to 20 Rlhr
exposures. Above this level, gamma-ray sensitivity must be corrected. Small
neutron sources are used for pretest and posttest calibration. Neutron
EXPERIMENTAL SHIELDING
473
senSitiVity ranges from 0.5 mrad/hr to 0.5 rad/hr. Special pulse-counting

procedures have been used to extend sensitivity to lower intensities.
Fission chambers are ionization chambers coated with 237 N p, 238 U,

232Th, 239pu, or 23SU. The first four are used for counters which respond
only to neutrons whose energy is above their fast-fission threshold (0.6, 1.5,
1.75, and 0.01 MeV, respectively) and which measure the total flux above
that threshold. There are two problems: thermal-neutron reactions in the
other contaminating isotopes normally found with these isotopes (even a few
parts per million of 23 s U in 23 8 U can be troublesome) and with gamma-ray
sensitivity. Frequently, data are obtained with bare and cadmium-covered
23S U chambers to provide an indication of the thermal-neutron flux
(actually subcadmium) rather than an integral over all energies. Fission
chambers can be compensated for gamma rays by incorporating two regions:
one containing no fissionable material and thus responding only to gamma
rays and the other containing fissionable material and responding to both
neutrons and gamma rays. The difference in output of the two regions is
therefore proportional to the neutron fluence.
Detectors containing boron either in the form of a gas (usually boron
trifluoride) or coated on the walls of the chamber are widely used. A
matched set of bare and cadmium-covered detectors provides subcadmium
(En < 0.4 eV) or thermal-neutron flux via the 1 B(n,( reaction. Gamma-ray
compensation also can be used with these detectors. Sensitivity is varied
by
varying the gas pressure in the initial filling, by varying the enrichment of the
boron isotope, by varying coating thickness, or by varying the size of the
detector. Boron detectors can be operated in either the pulse mode or the
current mode with the latter having the advantage of giving essentially
instantaneous readout. The current mode is particularly helpful when a
spatial traverse is being made since detector output can be monitored with a
chart recorder showing detector output vs. a position variable. A considerable variety of boron chambers are commercially available.
Helium-3-filled detectors take advantage of the large (n,p) cross section

of 3He to measure neutron flux. Fairly elaborate electronics are required to
convert the output to a flux, but the detector is quite sensitive.
A number of recently developed liquid scintillators, such as NE-211 ~nd
213, can be used to provide neutron spectra if appropriate shielding and
collimation are provided around the detector as well as appropriate
electronics and response functions for converting the output to a true
spectrum.
474
Much interest has developed in recent years in the use of solid-state

detectors, such as lithium-drifted silicon and lithium-drifted germanium, for
measurements of neutron spectra, flux, and dose. These detectors generally
have good resolutions but low efficiency. They make use of such neutron
reactions as 3He(1l,p), 1 H(1l,n)l H, 6 Li(Il,OI), and 1 0B(1l,OI). The addition of
drifted lithium to silicon and germanium increases the detection efficiency
for high-energy neutrons. Solid-state detectors are sensitive to light, X rays,
gamma rays, and charged particles; thus they must be adequately shielded
and compensated.
9.1.2 Passive Neutron Detectors

A number of passive detectors, nonelectronic devices, arc commercially
available for the measurement of neutron intensities. These detectors can be
placed within a shield, exposed to neutrons, recovered, and later read to

determine the neutron intensity at the time of exposure. These devices have
the advantage of small size, and they can be located in otherwise

difficult-to-monitor situations. Among them are the following:
Small thin specimens (foils) of a great variety of materials become
radioactive when exposed to neutrons. If the composition and mass of the
foil are known, as well as the half-life, cross section, decay energy of the
radioactive species, the irradiation interval, and the specific activity of the
foil (disintegrations per second per gram), one can determine the neutron
flux. Some foils respond primarily to low-energy neutrons (gold and
dysprosium), some to resonance-energy neutrons, and some only to neutrons
above a threshold energy. Foils necessarily must be of high purity and have
half-lives neither too long nor too short. In general, they are reusable. Many
of the fission foils must be enclosed in boron to eliminate thermal-neutroninduced response.
Several p-n junction (solid-state) devices have been developed in which
neutrons change the electrical characteristics in a measurable and re-
producible fashion. Most of these are cumulative (i.e., they cannot be reset
to zero) and so can be reused only by determining the difference in total
dose registered by them before and after exposure. These devices are almost
as small as foils, but they can be stored for some time before readout.
A relatively new device is the damage track-length detector. 2 This device
makes use of the fact that chemical etching of a plastic material penetrated
by fission fragments greatly enlarges the tracks produced by the fragments so
that their lengths can be determined and the number of fissions occurring in
475
foils of fissionable isotopes having a variety of thresholds-can be found; this,

in turn, provides the neutron fluence. These plastic track-l;ngth detectors are
one to two orders of magnitude more sensitive than fission foils and have a
long shelf life.
Nuclear e/llll [siollS are still occasionally used for determining neutron
spectra external to radiation shields. Collimation must be used along with
suppression of the gamma-ray field at the emulsion to reduce gamma-ray
fogging.
9.1.3 Active Gamma-Ray Detectors

Both active and passive detectors have been used to obtain experimental
gamma-ray data of interest in reactor shielding. A number of ionization
chambers have been used, varying in size and sensitivity as well as in the
maximum temperatures to which they may be exposed, their neutron
responsiveness, and the dynamic range over which they can be used. These
devices are usually calibrated with standard isotope sources, such as 6 0 Co
and I 37 es. Ionization chambers mounted on traversing devices can be read
continuously and thus can rna p an entire region since the position variable
can be one of the variables on a chart recorder and the dose rate the other.
A,lthracene scintillation dosimeters (ASD) have often been used for
measuring gamma-ray dose rates since the output of a photomultiplier tube
attached to the scintillation crystal is proportional to the gamma-ray dose
rate at the crystal. Neutron .:aptures in the detector can provide a spurious
response, and temperature sensitivity is also a problem.
The widespread use of scintillation detectors as gamma-ray spectrometers
has been made possible by the availability of large single crystals of NaI(TI)
and the development of end-window photomultiplier tubes with high
resolution. Except for edge effects created when secondary photons from
pair production escape the crystal, essentially all the incident gamma-ray
energy is converted to light via the photoelectric effect, Compton effect, or
pair-production process, and the light pulse energy is proportional to the
primary gamma-ray energy.
A number of high-resolution solid-state gamma-ray spectrometers have
been developed which can be used in some situations; however, neutron
sensitivity can be a very serious problem. Of particular importance is
lithium-drifted germanium. Resolution at high energies is much better with
these devices than with the scintillation spectrometers, but more elaborate
spectrum unfolding techniques and electronics are required.
476
9.1.4 Passive Gamma-Ray Detectors

One of the earliest gamma-ray detectors was photographic film. It is still
used since it is relatively inexpensive and can be read automatically.
Sensitivity can be varied over a wide range. Most emulsions exhibit neutron
sensitivity, but this can sometimes be compensated. Also film does not have
a spectral response approximating tissue. Various absorbers can help reduce
this difficulty.
Chemical indicators of ionization have also been used as gamma-ray
detectors in shielding experiments, although these have largely been replaced
by easier-ta-read chemically activated glass detectors and the newer
thermo luminescent detectors. The latter devices (usually called TLD's) come
in a variety of sizes and shapes, including powder, micro and mini chips,
rods, and hollow tubes. Most TLD's were developed and calibrated for
situations in which only gamma rays would be present and where great
accuracy was not required. Automatic or semiautomatic readers have
become available, although careful calibration is required. The detectors
must be heated before reuse; some are temperature-sensitive in the ambient
range and lose some of the dose indication if an extended time period
intervenes between exposure and readout.
9.1.5 Interpretations of Detector Output
Considerable caution must be exercised when detectors developed for a
particular set of circumstances are used under conditions that are quite
different. For example, small thermoluminescent gamma-;ay dosimeters
were developed initially for use in pure gamma-ray fields produced by
isotope sources or accelerators wherein the energy of the gamma rays was
usually well below 2 MeV. Some care is required in obtaining reasonably
precise data in a mixed field of neutrons and both high- and low-energy
gamma rays. The lithium in LiF TLD's responds to neurrons to an extent
determined primarily by the concentration of 6 Li, which has a high
neutron-capture cross section, and the ratio of the neutron to gamma-ray
intensity. Manganese-activated calcium fluoride used in a TLD may also have
a significant response to neutrons as a result of captures in manganese, which
has a high capture cross section for resonance-energy neutrons. Detectors of
CaF are also frequently exposed in small tantalum covers, which serve to
flatten the energy response of the TLD to make it similar to that of tissue.
These covers present no particular problem in a pure gamma-ray field similar
to that in which the detectors are calibrated; however, tantalum has very
477
high thermal- and resonance-neutron absorption cross sections. The detectors, when exposed to a mixed field of neutrons and gamma rays, may
therefore be responding in part to the gamma rays produced in the covers by
either neutron capture or subsequent decay of the activated tantalum.
For high-energy gamma rays, the efficiency of the TLD, particularly the
very small micro dosimeter, may be reduced with increased energy because
the range of the electrons produced by the gamma rays may be greater than
the dimensions of the detector.
There have also been instances in which the preamplifier attached to a
gamma-ray detector was more sensitive to radiation than the detector itself
and produced a signal even when the detector was disconnected. The long
cables required for remotely controlled experiments can also significantly
change detector-response systems. Calibrations should be made with the
detector attached to the cable to be used in the field.
Thermal-neutron detectors, calibrated in a thermal column where the
ratio of the number of thermal neutrons to the number of epithermal
neutrons may be 40/60, may give quite misleading thermal fluxes in a
situation in which there are few thermal neutrons and many epithermal ones,
such as in a thick region of aluminum outside a highly enriched

intermediate-spectrum reactor.
In short. detectors almost invariably provide data, but frequently

considerable effort is required to ensure that the data are related to the
quantity one is trying to determine.
9.2 SHIELI).MATERIAL MEASUREMENTS
The following experiments are only a sample of many performed in
recent years to characterize the transport properties of bulk materials or
combinations of materials. They were selected as well-designed experiments
that are representative of those performed in materials studies. Experiments
are discussed for reactors, accelerators, and fixed sources. The purpose and
plan of the investigations are emphasized, but sample data are included to
indicate the utility of the results. A recent survey of reactor shielding
facilities has been reported by the Radiation Shielding Information Center. 7
9.2.1
Reactors
An extensive series of slab-penetration measurements were performed in
a specially designed facility called the Outside Test Tank (OTT). These
478
measurements compared the relative effectiveness of various arrangements of
a given set of materials and of different sets of materials. The facility was
also used for testing mock-ups of designs for various shields. The tests were
performed at the Nuclear Aerospace Research Facility at General Dynamics,
Fort Worth, Texas, and are described by Belcher and Zoller. 8 The source of
radiation was the Ground Test Reactor (GTR) which fits into the moderator
tank shown in Fig. 9.1. The OTT provided an unusual capabiliry for the
testing of a wide variety of shield configurations of interest in the
development of minimum-weight airborne or space nuclear systems since
earlier measurements required that slabs and detectors be submerged under

water or oil, and it permitted measurements in air without extraneous
liquid layers. The power level of the reactor used in the OTT was high
enough (up to 500 kw) that active dosimetry measurements could be made
in air to distances of 30 m or more from the reactor. The shields tested were
sufficiently thin compared to the water and lead surrounding the reactor to
ensure that only radiation penetrating the shields reached the external
detectors with no contribution from air or ground scattering. The tank itself
was 4 m high by 4.5 m in diameter and could be rotated remotely. Two
independent dosimeter traversing mechanisms were also remotely controlled
from an underground control foom.
Several hundred shield configurations consisting of either laminated slabs

or boxes containing semiporous shields were tested in the shield compartment adjacent to the reactor moderator tank. A few of these shields were
mock-ups of proposed shield designs, but the great majority were part of a
systematic study of the effects on external dose rates of varying the quantity
or position of a given material either by itself or within an array of another
material. Typically, one might have 2 to 10 em of gamma shielding, such as
tungsten, steel, or depleted uranium, in an array of neutron shielding, such as
lithium hydride, beryllium, or beryllium oxide, with or without borallayers
adjacent to the higher density materials. Borated stainless steel, zirconium
hydride, lead, and boron carbide were also tested. A number of simulated
ducts were investigated, as were various thicknesses of neutron shields with
no gamma-ray shielding. Samples of a great many different materials used in

aircraft turbines and structures were activated in various neutron spectra
obtained by changing the composition of the materials between the samples
and the reactor to give a solid basis for estimates of component activation.
Many measurements made at distances of 3 to 30 m from various reactor
shield configurations are reported by Friedman. 9 Figure 9.2 is typical of the

data obtained. The gamma-ray dose rates at 2.7 and 15.2 m from the reactor
~
l'jl
...
i,,
~,
!,
,,,
i,
Fig. 9.1-0utside Test Tank. (From Belcher and Zoller. 8 )
479
480
10' .....---...,....---.....,.----.....----.10-'
Tank
1.3 em boral
GTRtl . '"
D
E
u
E
u
:::;
;:
'"
,,; ,.;
:I
Ii
wall
E
u
Detector
"> -..
0
,I
2.7 m, 15.2 m
Oem tun!J9ten
10-2
2.5 em tungsten
,,
5.1 em tungsten at 2.7 m
, ...
...
'"1................. -- -- -----=-+-----10.2 em tungsten at 15.2 m
. 10.2 em tungsten at 2.7 m
10-2 L
...L
5
....L
10
L _ _--1 W 4
15
20
THICKNESS OF LiH BEFORE TUNGSTEN, em
Fig. 9.2-Gamma dose rate in air vs. tungsten location in lithium hydride. (From
Friedman. 9 )
481
center are plotted as a function of both the thickness of a tungsten alloy and
as a function of the amount of LiH between the reactor and the tungsten.
Secondary gamma rays were shown to be of considerable importance.
Secondary gamma-ray production decreased as the amount of neutron
shielding between the tungsten and the core was increased. However, for
most situations increasing the inner radius of heavy gamma-ray shielding
material would also increase shield weight. The addition of boron to the
heavy material or the use of thin protective layers of boral was expected to
reduce the number of thermal-neutron captures in the gamma shield
significantly. The number of captures of epithermal or resonance-energy
neutrons should not be reduced significantly, and the production of gamma
rays by inelastic scattering of fast neutrons (E > 200 ke V) should not be
affected. The data shown in Fig. 9.3 were obtained in an attempt to
differentiate between these various sources of gamma rays. The data showed
that adding boron to stainless steel is more effective in suppressing secondary
gamma-ray production than placing boral between the steel and the reactor,
even with a good moderating material in front of the steel. These same
experiments also provided neutron-flux density and gamma-ray dose rate
within the shields for comparison with predictions as well as gamma-ray
spectra outside the tank.
Foil exposures between the slabs for the same configurations as shown in
Fig.9.2 were made as a function of tungsten thickness with 27 Al
(E> 8.6 MeV), 32 S (E > 2.9 MeV), and 6 3CU (E < 0.4 eV) foils. Measurements were made for four thicknesses of tungsten, 2.5, 5.1, 7.6, and
10.2 em, inserted where the tungsten layer is shown in Fig. 9.2. Foils were
placed between slabs separated by gaskets. The configuration was compressed to eliminate the space at the tungsten location and expanded as the
tungsten slabs were inserted.
Foil activities (counts min- 1 g-l watt-I) are shown in Fig. 9.4 at reactor
shutdown as a function of tungsten thickness for the copper foils in terms of
cadmium differences. Note that for a tungsten thickness of 0 the LiH slabs
are moved left to the BeO and boral slabs. As the tungsten slabs are inserted,
the LiH slabs are moved to the right. An effect on the thermal activity is
evident. As the LiH is replaced by tungsten, a valley begins to form,
which indicates the much higher absorption cross section for tungsten.
The valley deepens and widens as the tungsten thickness is increased. The
sulfur and aluminum activities (not shown) are reduced by a factor of 2 in
the LiH region when LiH is replaced by 10 em of tungsten. This reduction
482
'.
E"
on
~ 10- 1
10-3 ...
a:
a:
'"o
'"g
""
'"
""
Stainless
steel
Borated
stainless
j
I
'"
r13 em b0'", I
~'il
:I:
::;
5 Il E
.-,,,, 0
!U) . . .: e
~'"
E
0
....
~1> Detector
--,
1O- 2 L
lL
}-4---2.7 Or 15.2 m--+-t
..J 10-4
Fig. 9.3-Gamma dose rates in air behind stainless steel and borated stainless steel.
(From Friedman. 9 )
demonstrates the effectiveness of adding a high-7- inelastic scatterer to a

low-z elastic scatterer.
Verbinski, Bokhari, Courtney, and Whitesides' 0 describe a reactor

experiment that yielded measurements in bulk water shields at Oak Ridge
National Laboratory (ORNL). The source of neutrons was the Bulk
Shielding Reactor I (BSR-I) described by the Radiation Shielding Information Center. 7 This was a series of measurements of the spectral and spatial
details of the fast-neutron flux density penetrating water shields. The
purpose was to obtain data for evaluating neutron-transport codes that
Symbol
[;.
102
.,
.,~
.,
~
483
Tungsten
thickness,
em
0
2.5
5.1
7.6
10.2
10'
e
0
0
u
>:
!::
>
;::
10
"
10- 2 L----I-------4-----------_---J
BeD AND
BORAL RANGE
Fig. 9.4-Measured 63 Cu activity (cadmium differences) as a function of foil location

for the indicated thicknesses of tungsten.
484
predict the variation of flux density with energy, angle, and penetration
depth in a shield. Two basically different shielding codes were compared
with the measurements. Two different types of experiments were performed
to provide a broad basis for evaluating the codes. In one a large source with
adjacent shield was used; in the other a small source well-separated from the
shielding slab roughly approximated a good-geometry neutron-scattering
arrangement. This discussion is concerned with the first experiment: the
large source with adjacent shield. The small-source experiment will be

discussed later. The first experiment was performed at ORNL and consisted
in the mapping of the absolute spectral intensity in the water shield of a
pool-type reactor (the BSR-I) as a function of both angle and distance from
the source. In conjunction with these measurements, both the absolute
reactor power and the distribution of the neutron-source strength throughout the reactor were measured. The results of the reactor-power calibration
were then used as input to the neutron-transport codes so that the codes
could be evaluated for the complete task of predicting the neutron-transport

pattern, beginning with an absolute fission-source distribution and ending
with spectral intensities at a large number of points and directions in the
shield.
The fuel loading of the BSR-I used for the absolute measurements of
spectral intensity is shown in Fig. 9.5. It is a water-reflected rectangular
configuration (38 by 38 by 60 em) of aluminum-clad elements with two
additional elements in the rear which were required for criticality. For each
spectral measurement the relative power of the reactor was determined to
within 2% by means of two sulfur foils (actually pellets) located 35 cm

above and 38 cm east of the reactor centerline. The sulfur activation was
later related to the absolute total reactor power as well as to the absolute
neutron-source density throughout the reactor volume. The absolute power
was determined to an accuracy of about 5%.
A shielded diode spectrometer, designed especially for these measurements, was positioned to measure the spectral intensity at a distance r from
the face of the reactor, along the center line of the reactor midplane, and at
an angle
with respect to the center line as shown in Fig. 9.5. The
spectrometer consists of a very thin layer of 6LiF between, but not in
contact with, two silicon-gold diodes. Neutrons traversing the spectrometer
collimator strike the 6 Li foil, and the alpha particle and triton emitted in the
resulting 6Li(n,o<)T reaction (Q; 4.78 MeV) are detected in coincidence in
the two diodes. Shadow shielding in the collimator protects the diodes from
-----...
BSR LOADING 96A """
,-
.........
----
ALUMINUM
./ I 20 em
".,.,....
SULFUR
I.
20 em
.I
FOILS~
CONTROL RODS EQUAL TO

ONE-HALF FUEL ELEMENT
TUNGSTEN LIN ING
l~~i~~SILICON DIODES
~
7 UF LAVER
6Li~;~~~~~
6 UF LAYER
LEAD
6Li~""'---'
of
Fig.9.S-Experimental arrangement showing relative pOSItiOns

reactor and spectrometer for measuring neutron
spectra in BSR-I water shield. The dashed circle represents the spherical source region used for the calculations, for
which the distance r is measured from the surface ofthe sphere. [From V. V. Verbinski, M. S. Bokhari, J. C. Courtney,
and G. E. Whitesides, Nuclear Science and Engineering, 27(2): 284'(1967).]
...
00
'"
486
direct reactor radiations. Background measurements are made by substituting a 7 LiF foil for the sensitive 6 LiF element.
The spectrometer has operated satisfactorily in gamma-ray fluxes from
the BSR-I as high as 10 6 R/hr at the point in the water shield where the
neutron flux was being sampled.
The calculations corresponding to this experiment were performed with
the NIOBE'l and the DTK codes (DTK is a version of DTF,' 2 which is
discussed in Chap. 4, Sec. 4.4). Limitations on the geometries that could be
used with these codes decreed that the reactor core be simulated by a sphere.
The radius of the spherical source region was taken as 26 em and is shown by
the dashed circle in Fig. 9.5, all values of r in the calculations being given in
terms of distance from the surface of this source sphere. The spectral
intensity of neutrons was measured at angles (II) of 0, 41 and 52 with
respect to the extended midplane center line of the reactor. The distance r
from the reactor face was varied in increments of 10 em out to a distance of
50 cm. All measured (and calculated) fluxes are given for a total reactor
power of 1 watt.
The spectral intensities obtained at II = 0 are shown in Fig. 2.19, along
with predictions of the NIOBE transport code. (In Fig. 2.19 and succeeding
figures in this chapter, J1 = cos II.) It is evident that the code predicts the
absolute magnitude with good accuracy and reproduces some of the general
features of spectral shape. The strong dip in the spectra at 3.5 MeV is caused
by a broad peak in the oxygen cross section. The measured spectra suggest
that a similar dip may be present in the 20- to 50-cm spectra but at an
energy about 0.5 MeV higher. However, the present model of the spectrometer was designed to provide only moderate energy resolution in order to
achieve good sensitivity. Therefore the discussion of detailed spectral shapes
will be deferred to the section on LINAC measurements (Sec. 9.2.2).
The 0 measurements are compared with the predictions of the DTK
transport code in Fig. 9.6. The agreement between the DTK calculations and
measurements near the source region is about the same as that with the
NIOBE calculations. The dashed lines between 5 and 7.5 MeV show the
improvement in the predictions obtained when the calculation was repeated
with updated cross sections.
The measurements for II = 52 were made at distances of 10, 20, 30, and
40 cm from the face of the reactor. The results are compared in Figs. 9.7 and
9.8 with the predictions of the NIOBE and DTK codes, respectively. Overall
agreement appears to be fair except that at r = 10 and 20 cm the
experimental results are higher than both calculations at high energies, where
487
5
.~
2
.~
1'00
r- Oem
I~
/ ' Exp
10 ej, L'
00
L, ~o
,'I'
Ie 104
5 1--.
20 em
'~
~E
103
1::::i
30 em
.'
40cm
,...r-;
I '1=1--
.-~
.:::I .
50cm
0
0
...L
I'
----.
11 .",L.:.....:
,
I--0
110
I~
L-
f--r
,.".
f-----'
0
r-..
r,
1='-
10
,," ,
;.-DTK
5
~
10"
678
10
11
12
NEUTRON ENERGY, MeV
Fig. 9.6-Measured and DTK-calculated neutron spectral intensities in the BSR-I water
shield at 0. For the calculation the value of 1J is 0.978 rather than 1. [From V. V.
Verbinski, M. S. Bokhari, J. C. Courtney, and G. E. Whitesides, Nuclear Science and
Engineering, 27(2): 286 (1967).]
the neutrons are more penetrating. This must be caused by uncollided plus
small-angie-scattered neutrons coming from the corner of the reactor, which
is within view of the spectrometer collimator at r = 10 cm and almost within
view at r = 20 em. This corner does not, of course, exist for the spherical
source used for the calculations. At r = 30 and 40 cm, the difference in the
calculational and experimental source geometries is relatively unimportant
since the corner of the reactor is farther from the line of sight of the
488
I---~
o
o
o
I&-~~
..
30 em
~.~
,,~
'" lo'_=:-'~
'- ...
----~~40cm__::_.
11>
-.j
.,
'eo
'"
>-
--
---I
---t---t~
10-2 L_l-_l-_l-_L---_L---_L----JL----J_---J_-l_-l_---J
o
2
4
6
8
10
12
NEUTRON ENERGY, MeV
Fig. 9. 7 ~Measured and NIOBE-calculated neutron spectral intensities in the BSR-I water
shield at 52, [From V. V. Verbinski. M. S. Bokhari, ]. C. Courtney, and G. E.
Whitesides, Nuclear Science and Engineering, 27(2): 287 (1967).]
489
...
c
2
1O-2: - _ - , - _ - , - _ - ,_ _,--_-,-_-,-_-,_-,,-_..L.._...J.._-'_-'
o
2345678910
11
12
NEUTRON ENERGY, MeV
Fig. 9.8-Measured and DTK-calculated neutron spectral intensities in the BSR-I water
shield at 52. For the calculation the value of fJ. is 0.650 rather than 0.617. (From V. V.
Verbinski, M. S. Bokhari, J. C. Courtney, and G. E. Whitesides, Nuclear Science and
Engineering, 27(2): 287 (1967).]
490
spectrometer collimator. Much better overall agreement with the experiment

is realized here except that the codes predict fluxes that are somewhat low at
low neutron energies. Note that at 52 the transport calculation using
updated oxygen total cross sections (dashed lines) produced little, if any,
overall improvement in the 5- to 7.5-MeV region where the older
cross-section values are evidently too high.
Measured and calculated spectral intensities were also obtained for
IJ = 41 at r = 20 and 40 cm, but, since they show the same general features
as those obtained for IJ = 52, no further comparisons are given here.
An additional reactor experiment, described by Clifford, Muckenthaler,
Maerker, Straker, and Mynatt' 3 and by Clifford, Straker, Muckenthaler, Verbinski, Freestone, Henry, and Burrus, l 4 was performed to evaluate total cross
sections of several materials over the energy range of interest in reactor
shielding. The experiment was carried out at the Oak Ridge National
Laboratory Tower Shielding Facility (TSF). This investigation was performed with the reactor on the ground and has since been called the
broomstick experiment because of the long, thin samples (Fig. 9.9).
The measurements consisted of spectra of the uncollided flux of
neutrons transmitted through thick samples of shielding materials. Because
the samples were quite thick, the transmitted spectra were influenced most
strongly by the minima in the total cross sections. The spectrum incident on
the sample and the transmitted spectra were measured with an NE-213
neutron spectrometer covering the energy range from 0.8 to 11 MeV. The
total cross sections were evaluated by a direct comparison of calculated and
experimental data. The uncollided spectra were calculated with the most
detailed cross-section data available and then folded with the response
function of the detector. Although the energy resolution of the spectrometer
is rather broad, this technique still allows a determination of the errors in the
very narrow minima in the total cross section since the total number of
transmitted neutrons is measured to within a few percent. Because the
geometry of the experiment was very good, no correction for multiple
scattering was required; therefore the calculations were quite simple and
straightforward, involving only the exponential attenuation of the incident
spectra using fine-mesh data for the total cross section. Experiments were
done to demonstrate that multiple scattering was indeed negligible. The
sample diameter was varied by a factor of 2 with no measurable change in
the transmitted spectra.
Good agreement between the measured and calculated spectra indicates
that the valleys are adequately represented in a cross-section set. Where
491
1-----~-15.2 m---~-+---
H2 0
L1THIATEQ
PARAFFIN
STYROFOAM
H 2 0 + LEAD
NOT TO SCALE. ALL DIMENSIONS IN CENTIMETERS EXCEPT
AS NOTED. BEAM CENTER LINE, -2 M ABOVE CONCRETE PAD.
Fig. 9.9-Arrangement of broomstick experiments. 13 14
disagreement exists, detailed information about the specific region of

disagreement is not possible, unless the minima are well separated, because
of the poor energy resolution of the spectral measurements. Where a number
of minima are close together, only the general energy range in which the
disagreement exists can be determined. Because of this these measurements
can in no way be described as a measurement of the total cross section but
simply as an experimental evaluation of the existing cross-section data.
The spectrometer uses a 5- by 5-cm liquid organic scintillator (NE-213)
and has an efficiency that ranges from 0.15 to 0.4 for neutrons in the MeV
energy range. Thus it can be used in flux densities that are several orders of
magnitude lower in intensity than those measurable with other spectrometers, all of which are limited by low efficiencies (Hr' to 1(j"6). The validity
of the experimental technique was checked by measurements on thick
samples of carbon, lead, and uranium, which could be easily compared with
calculations since the total cross sections for these elements are well known.
So that the ratio of the uncollided flux at the detector to the collided
flux could be maximized, the detector was located at a distance of 30 m
from the shield of the reactor and the sample was placed at the approximate
midpoint between the detector and the reactor shield. The arrangement is
shown in Fig. 9.9. Most of the samples were cylinders that were positioned
with their axes coincident with the axis of the neutron beam. Oxygen and
nitrogen samples in liquid form were contained in either Dewar vessels or
styrofoam-insulated brass cans.,
Because neutrons scattered by the air and gtound become important
when the scattering sample is thick, both the detector and the neutron beam
were shielded and very tightly collimated. The neutron beam incident on the
492
sample passed through two collimators: one which was adjacent to the
reactor shield and another which was adjacent to the sample position and
limited the beam to 9 em in diameter. These collimators also served to
reduce the air-scattered background at the detector without producing any
observable change in the measured spectra incident to the samples.
Measurements were made through some eighteen different materials. We
show only the results for iron, which are particularly interesting because of
considerable uncertainty that existed at the time in the fine structure of the
iron cross section.
Since the experiment was primarily one of cross-section evaluation, data
from the OSR,' 5 KFK,' 6 BNL,l 7 and ENDF IB' 8 files were all examined.
Data from the OSR Monte Carlo cross-section library and the ENDF IB
(Evaluated Nuclear Data File) were considered the best for iron from the
experimental results. Figure 9.10 shows comparisons of the measured and
calculated energy spectra from the 05R and ENDF/B cross-section data.
Good agreement is noted in the data for energies above 2.0 MeV, but both
sets of cross sections are too high for lower energies.
This technique of evaluating minima in total cross sections has been very
useful in indicating which evaluated set of total cross sections is adequate for
shielding calculations and thereby permitting an efficient updating of
cross-section libraries by updating only those cross sections which are not
adequate. It also provided information about the energy regions in which
new cross-section measurements should be made.
In the foregoing sample of reactor shielding experiments, we have
omitted many performed with the reactors mentioned as well as many more
with other research reactors. A recent survey 7 of shield facilities gives a
complete list of shielding reactors and the types of investigations performed.
The Lid Tank Shielding Facility, although no longer used in shielding
investigations, deserves mention. Originally developed at ORNL, later
duplicated elsewhere, the lid tank was successfully used for measuring
removal cross sections and other materials attenuation parameters. A thermal
column from a reactor core (at ORNL, the X-10 reactor) extended to a plate
of fissile material attached to the wall of a tank. The fission plate (or source
plate) produced a plane source of fast neutrons with an effective power of a
few watts. Since the fission plate is removed some distance from the reactor,
the background radiation is greatly reduced. The tank could be filled with
water, oil, or a solution of interest, and materials arrangements could be
lowered into the tank adjacent to the source. All the neutron removal cross
sections given in Chaps. 4 and 6 were derived from lid tank measurements.
493
10'
5
.'" .
f r\.
,
o
-i"
8 in. iron
<'z
'0
Measured
Calculated (0 5R)
==
--- Calculated tENDF/B}

~
'~
\
,
--
.....
~lrn
-- ~
no:
-11 1
,
2
10
11
NEUTRON ENERGY, MeV
Fig. 9.10-Iron penf~tration results. (From Clifford et ai. 13)
Although the lid tank ptoduced many useful measutements, limitations in

intensity and flexibility caused it to be superseded by the teactor facilities
now
In
use.
9.2.2 Accelerators
The development of accelerators and generators capable of ptoducing
high yields of neutrons and gamma rays has provided the shielding
experimentalist with opportunities for measurements previously impossible
in reactor radiation fields. These soutces, coupled with detectors of good
sensitivity and resolution, permit cross-section measurements and evalua-
tions, gamma-ray energy-spectra determinations, basic shield-material evaluations, and other useful measurements.
Experiments with accelerators generally exhibit a flavor different ftom

those with reactors. Better-defined geometry is usually possible, the soutce is
pure neutton, gamma, or charged-particle without the complex backgtound
494
REACTOR SHIELDING FOR NUCLEAR ENGiNEERS
associated with a reactor. of greatest importance, however, is that with some

accelerators the energy of the source radiation can be selected either directly
or by time-of-flight measurement.
In the discussion of the broomstick experiment of Clifford ot al., 13.14

we mentioned that appreciable uncertainties were found in the reported
cross sections for iron at energies below 2.5 Me V. Those differences were
attributed to poor resolution in the valleys of the total cross section. An

experiment with similar purpose but entirely different approach was
petfotmed at Gulf General Atomic (GGA) by Cerbone, Miller, and Profio, 19
who used the linear accelerator (LINAC) with a time-of-flight method' 0.2 I
as a source of neutrons.
The geometry comprised a 76.30-cm-diameter iron sphere (Fig. 9.11), at

the center of which was located the uranium target for photoneutron production by bremsstrahlung from incident electrons. In Fig. 9.11 the electronbeam tube is the 9.s-cm-diameter port hole, port A is for monitor foils, and
the stepped 10.02- to 12.7-cm-diameter port is for the 0 measurements. The
through port at R = 23 em is for off-zero internal angular-flux measurements.
The neutron spectrum and angular distribution
afC
measured
by extracting a
collimated beam from the experimental assembly and sorting the counts with
their time of arrival at a detector placed 50 m distant. Time-zero is defined
by pulsing the neutron source. In the iron sphere, the time to slow down and
migrate to the beam extraction point is small compared to the corresponding
50-m flight time and can be neglected.

The spectrum was measured from 0.5 to 15 MeV with NE-211 organic
scintillators and from 0.6 keV to 0.5 MeV with an NE-908 lithium glass
scintillator. Spectra were measured for several thicknesses of iron at several
angles off the target-detector axis. Figure 9.12 shows the 0 measured
spectra at radius values of 19.05 and 38.1 em (surface). The fine structure of
the iron cross section is evident in the spectral changes. Since this is a good
geometry experiment with fair resolution, the rcsults arc adequate for
evaluating iron cross sections. Discrete-ordinate calculations of the spectra
for conditions specified in Fig. 9.12 were made with the IDF code (a version
of DTF-IV 2 2) using 29 energy groups. Group cross sections were prepared
from the ENDF/B library with careful attention given to minima in the
cross-section structure. An asymmetrical S 16 quadrature and a P3 crosssection expansion were used. Wide discrepancies with the measurements
were found. possible causes were errors in the inelastic and elastic cross
sections or the result of skipping valleys and neglecting self-shielding of the
495
76.30-crn sphere diameter (approximate)
9.53-cm diameter
5.08-cm diameter_
l:--H-+-~
10.02-cm diameter
12.7-cm diameter
Fig. 9.11-lron sphere for LINAC experiment. (From Cerbone et al,19)
resonances, which often happen when too few energy groups are used to
describe complex fine-structure cross sections.
A second calculation was made using 49 energy groups. Figure 9.12
shows results of this calculation compared with measurements for R = 19.05
and 38.1 em. Agreement is still quite poor, which strongly suggests that
errors exist in the ENDF/B point cross-section data for iron.
Thus the results of Clifford et al. 1 3.14 and Cerbone et al. 19 agreed to
the extent that reported values of the iron cross section appeared to be
entirely inadequate for shielding calculations.
Carlson, Cerbone, and Willoughb y 2 3 state:
... recently. much concern has been expressed about the adequacy of the
measurements of the total-neutron cross section of iron. Numerous examples of
496

10- 1 ,-...,.--,--,-.,..,-...,.-.......,..,.---r--r-...rT--,--. .-r-rr-..,
19.05 em
,,
,i
:1 /,
'-,'
I,
"
"
L__ J
I ' :
I,I.. J
1"'"
r~',
r
_J
l
r.
I"~,
, . . " ,.J
II
"
','
"I:"
--"
rr
n
t'Jl
'I
o.rl-'
s'
I
...... -~~-~~
/[~
'v' ,:'",':,
, ,
fl"
,: V
},
lJ
::::~l,.:
"
" :1
f"l
:1 ::
I
:,1'
U
...
I,.
'I
,J'
-:11
:: Ui
'"
t.! l
L,
,I'
I
I
~"I
," I
I I
I
: 1"1
'::
,
I
I
\
"
\
I
~.,
II
''
L,
L.J
....
\
\
,,
!L._, \ ,
, ,
:. ,,
,
,,
R=38.1 em
Hr" ':--'-_.1-.L..J.-':---'_...........l.-L.'-::-L--'--'-.l...L:-........---"LJ-'-L-..J
103
,0"
to'
to"
10'
ENERGY, eV
Fig. 9.12~Comparison between 49-group IDF calculated (A',B') and measured (A,B) 0
neutron angular spectra for iron at R = 19.05 (A',A) and 38.1 em (B',B). (From Cerbone
et al. 19)
the questionable state of this cross section are evident. For some energy
regions,24 there are large discrepancies in the measurements and energy scales;
also considerably more structure than was previously seen has been observed in a
recent set of measurements made by Cierjacks, Forti, Kopsch, Kropp, Nebe, and
Unseld. 25
Many measurements have been made of this cross section which is of prime
importance for shielding applications. However, in almost all cases the resolution
was inadequate. For iron, the level spacing and level widths in the MeV neutron
energy region are such that a multilevel fit for resonance parameters is extremely
difficult; therefore, it is essential that measurements be made with very high
resolution if the cross sections are to be used for shielding applications. It thus
497
became clear that a new set of iron cross-section data taken with high-energy
resolution and covering a large energy range was necessary.
For these reasons the iron total-neutron cross sections have been measured
with high-energy resolution for the energy interval from 0.5 to 9 MeV.
The iron cross-section measurements were made by Carlson et ai. 2 3 at

GGA with the same accelerator used by Cerbone et al. These transmission
measurements were also made using time-of-flight (TOF) techniques but
with a 226.75-m flight path. An overall plan of the geometry is shown in
Fig. 9.13. High-energy neutrons are produced by electron bombardment (the
bremsstrahlung produces photoneutrons) of a spherical uranium target
having a radius of 3.8 em. The neutrons travel through an evacuated drift
tube to a defining collimator at the 100-m station. This collimator restricts
the size of the neutron beam to a circular area having a radius of 6.3 em. In
front of the collimator are two 3 He detectors that are used to monitor the
neutron flux and a 3.1-cm-thick uranium filter that reduces the effect of the
gamma flash on the neutron detector. The transmission samples and the
sample changer are located behind the collimator. With the sample changer
the samples can be placed accurately in a position such that the detector (at
the 220-m station) is completely shadowed from the source. This mechanism
is operated remotely from the data-taking area. The monitor counting rates
were measured with and without the samples in place to ensure that
neutrons scattered from the samples did not affect the 3 He monitors. No
detectable difference was observed. The transmitted neutrons enter an
evacuated drift tube that connects the 100-m and the 220-m flight-path
stations. At the 220-m station the neutrons are detected in a 5-cm-thick by
2.5-cm-radius cylindrical NE-211 scintillator coupled to a 56 AVP photomultiplier tube. The incident-neutron flux is perpendicular to the axis of the
scintillator. The detector is surrounded by a lead shield that reduces the
URANIUM FILTER
e-
'He DETECTORS
EVACUATED
BEAM TUBE
ZANIUM TARGET
~
0
~COLLIMATOR
~I
SAMPLE
LEAD SHIELO"""
EVACUATED}e
BEAM TUBE
NEUTRON DETECTOR
lOQ-m STATION
"""""
220-m STATION
Fig. 9.13-Diagram of flight path and experimental setup used in the total-cross-sec:tion
measurements of Carlson et al. 2 3
498
amb ient background. The contribution to the counting rate from neutrons
that scatter off the detector into the lead shield and then back again into the
detector was found to be negligible.
The iron sample was composed of standard Armco iron (99.94% iron).
The measured density agreed with rhe accepted value for the density of iron.
The sample thickness was 0.386 nuclei/b.
Although measurements were made from 0.5 to 9.0 MeV, only a limited
portion will be shown here. The complete set is given in Ref. 23.
Figure 9.14 shows the measurements of Carlson ef al. 23 from 0.7 to
1.0 MeV. The curve is merely a guide obtained by connecting the points. The
lneasurement of the cross section in this energy region where the cross
section varies sharply with energy requires the utmost in energy resolution.
Figure 9.15 compares the data with the data of Cierjacks ef al. 2 s and of
Schwartz, Schrack, and Heaton. 2 6 of the available iron total-cross-section
measurements, those of Cierjacks ef al. agree best with Carlson's data with
respect to the presence of structure and the neutron-energy scale. Carlson's
measurements, which have a somewhat smaller energy spread than those of
Cierjacks (0.035 nsec/m vs. 0.043 nsec/m), resolve the structure somewhat
better. The measurements of Schwartz ef al. are in agreement with Carlson's
data; however, a considerable amount of structure has been unresolved in
these data, which were obtained with a resolution of 0.3 nsec/m.
The iron total cross sections obtained in the study by Carlson ef al. 2 3
have provided much needed information about the cross-section valleys. It
had previously been conjectured that deeper valleys are needed in the total
cross section to obtain good agreement between the broomstick transmission
measurements of Clifford ef al. [3. [4 and calculations based on the total

cross sections. The latest measurements show a trend toward deeper valleys
than had been previously observed, which is primarily (it is believed) the
result of better resolution.
In our discussion of the transport measurements in water reported by

Verbinski ef al., [0 we mentioned that a small-source experiment with good
geometry was also performed. The source was the LINAC described
previously. In these measurements a slab shield was positioned at a distance
of 40 em from the source, which resulted in a nearly-good-geometry
scattering configuration. Consequently the effects of the sharp resonallces in
the oxygen cross section on the spectral shapes could be observed and
compared with predictions of shielding codes that have a fine energy mesh.
The spectral measurements did not yield information on the attenuation of
501
flux, but the activation of sulfur pellets placed on both the soutce side and
the exit side of the slab provided a measure of the attenuation.
The physical atrangement for the measurements with the pulsed-neutron
source is shown in Fig. 9.16. The LlNAC provided short bursts of electrons
(from 10 nsec for thin shields at 0 to 50 nsec for the thickest shield at 30)
FASTNEUTRON
DETECTOR
50 m 114 m VACUUM,
3~
m AIR, 5 em LEAD)
Fig. 9.16-Experimental arrangement at LINAC for measuring spectra of neutrons

leaking from a water slab at 0. For 30 measurements, the slab-source configuration was
rotated about the point "0." The positions of the sulfur pellets used for transmission
measurements are also shown. l From V. V. Verbinski, M. S. Bokhari, J. C. Courtney, and
G. E. Whitesides, Nuclear Science and Engineering, 27(2): 289 (1967).]
which irradiated a small lead target (5-cm-diameter by 2-(111 disk) and thus
produced bremsstrahlung radiation which, in turn, produced prompt
photoneutrons with a near-fission spectrum. The target was centered in a
20-cm lead housing, which was needed to suppress the bremsstrahlung

radiation that would otherwise have incapacitated the neutron detector.
Some of the neutrons escaping the lead housing penetrated the water shield
and passed down a 50-m flight path to a 5- by 5-in. encapsulated NE-213
liquid scintillator neutron detector. Figure 9.16 shows the arrangement for
measuring the 0 (straight-through) spectrum for a water slab of thickness T.
For some of the measurements, the neutron source and water slab were
502
rotated 30 about the point 0 shown on the exit side of the slab. The sulfut
pellets shown were used for measuring the neutron attenuation.
The configutation shown in Fig. 9.16 was calculated with both the
NIOBE and the DTK codes. Since the NIOBE code cannot use a smallsoutce
geometry, the neutron source for the calculation was taken to be a
20-em-diameter sphere with constant source density having the same
spectrum as that measuted with the water slab shown in Fig. 9.16 removed.
Unfortunately, in this measutement not all the soutce could be viewed
because of the limitation of the 7.62cmdiameter collimator shown in
Fig. 9.16. Comparisons of measutements with the NIOBE and DTK
calculations using the same falsesource geometry were equally poor.
Additional DTK calculations were performed for which the source was
assumed to be a 5cmdiameter sphere, which was a better simulation of the
experimental source. The lead housing around this small soutce was treated
as additional shielding. The spectrum used in these calculations, shown as the
top curve in Fig. 9.17, was the same as that measuted with the lead housing
and water slab removed. The two lower curves in Fig. 9.17 compare the
resulting calculated spectra with the measuted 0 spectra for slab thicknesses
of T; 20 and 40 em. The calculated results at 0 are very similar to the
NIOBE results l 0 except for the dashedline sections at about 5 to 7.5 MeV,
which were obtained with updated oxygen cross sections. The calculations
for the 30 configuration are compared with the measuted spectra in
Fig. 9.18. Here the calculated spectra are quite different from the NIOBE
results, the overall agreement in Fig. 9.18 being much better. From this it is
concluded that the falsesource configuration used for NIOBE does not
adequately approximate the experimental geometry and that reasonable care
must be exercised in reproducing the soutce geometry in calculations that
are being compared with small-source measutements of fast neutrons.
The solid line in Fig. 9.17, which represents the initial smallsoutce DTK
calculations, falls short of reproducing the broad transmission peak observed
in the measuted leakage spectra at 0. Since the geometry for this
measurement roughly approximates the goodgeometry configutations used
in measuring total neutron cross sections, the cross sections used in the
calculations became suspect. Cross sections for the calculated solid lines in
Fig. 9.17 were taken from the 1958 version 27 ofBNL325. The dashed lines
were calculated using the updated cross sections 28 reported in the 1964
Supplement No.2 of BNL-325. Much improvement is noted when the newer
cross sections are used.
..,
c c
g:6
~
-'N
~
5
C
- --14-
20 em
:-~
5
10
-Experiment
--+.
DTK
s.. 1tr~
<>
"
503
f-----
~-
0"1 ,..!
:=
=-
40 em
-c:_,;; _
_.._- t-
00
-=--
--
2 --
-~
-T
103
0
10
NEUTRON ENERGY. MeV
11
12
13
"
15
Fig.9.17-Measured and DTK-calculated spectra of neutrons leaking from 20- and

40-cm-thick water slabs at 0. The calculational source spectrum, shown at the top of the
figure, was taken as an isotropic source of 3.9 X 10 12 neutrons at the center of the lead
housing shown in Fig. 9.16. Experimental spectra have been individually normalized to
the DTK code results. [From V. V. Verbinski, M. S. Bokhari, J, C. Courtney. and G. E.
Whitesides, Nuclear Science and Engineering, 27(2): 291 (1967). ]
9.2.3 Fixed Sources

Fixed isotope sources provide an excellent tool fot the measurement of
materials if the whole reactor energy distribution is not needed. The work of
Alberg, Beck, O'Brien, and McLaughlin 29 provides a case in point. Their
measurements with 137 Cs gamma rays provide differential energy and angle
spectra from a point isotropic source in an effectively infinite medium of
water for gamma-ray penetrations of one, two, three, and four mean free
paths at 15 intervals. The objective of this experiment was to obtain a

definitive set of differential distributions in an infinite medium which could
be used as a standard for comparison with transport predictions.
Figure 9.19 shows the experimental arrangement. The detector was a
5.08- by S.08-cm NaI(Tl) crystal, and an EMI-9SS2 photomultiplier tube and
504
~
~
~ "--
-.:'
.-r=20Cm'llOS
000
F ...
T'" 10 em
--
..>
T'" 30 em
--::J
:;
g,
c
'""
10'
F=
r==
:.-.::=:: f-----__
f-----
'---
l ....
~--
-~
- ---
L:
---==j
t
.-
'--
"----
--+--~--+------1-~-I-~l----
'::::0.._
"1
L:....
t:
--
---. ------.-
456
10
11
12
NEUTRON ENERGY, MeV
Fig. 9.18-Measured and DTK-calculated spectra of neutrons leaking from 10-, 20~, and
30-cm-thick water slabs at 30, Experimental spectra have been individually normalized
to the DTK code results. [From V. V. Verbinski, M. S. Bokhari, J. C. Courtney, and G. E.
Whitesides, Nuclear Science and Engineering, 27(2): 292 (1967).]
REAR APERTURE
Fig.9.19-Schematic diagram of point source 137 Cs differential measurements. The

distance from the back edge of the rear aperture to the front edge of the forward aperture
of the collimator is 80 em; to the end of the methacrylate plastic probe, it is 163 em. The
aperture diameters are 1.0 em. [From M. Alberg. H. Beck, K. O'Brien, and J. E.
McLaughlin, Nuclear Science and Engineering, 30(1): 67 (1967).1
505
a 512-channel analyzer were used. About 700 liters of distilled water was used
to fill a 48- by 97- by 145-cm methacrylate plastic tank. This size was chosen
so that both the source and the detector would always be at least two mean
free paths from the edge of the box. The 1 37 Cs source was calibrated to
5%.
The detector assembly was located outside the medium to minimize
displacement effects. A probe and collimator were used that were based on a
design by Skarsgard and Johns. 30 The defining apertures of the collimator
were located outside the tank. A hollow plastic probe of 2.5 em inside
diameter was inserted into the tank and attached to the front end of the
collimator with a vacuum and pressuretight seal. The path between the
methacrylate plastic window in front of the crystal and the end of the probe
was then evacuated. Pulse-height distributions corresponding to various
angles between the source and the detector for a given penetration were
obtained by moving the source around a semicircle of fixed radius located on

the midplane of the tank. The collimator was equipped with a shutter so that
background readings could be taken at each angle with the source in
position.
The solid angle subtended by the crystal was determined by the method
described by Skarsgard and Johns. 30 In this scheme a sphere is constructed
whose center lies on the collimator axis at the end window of the probe and
whose radius is just large enough to include all points on the window which
subtend a finite solid angle at the crystal. It is convenient to regard this
sphere as the source of scattered gamma rays, even though most of the
radiation seen by the crystal will have been scattered through the sphere
from other points in the medium. A cross section is taken through the sphere
at the window, and the solid angle seen by each differential element on this
cross section is calculated. An integration over these elements yields the total
solid angle seen by the crystal.
Eight sources were used to calibrate the spectrometer: 95 2r_ 9 5Nb
(750keV), 137CS (662keV), 7Be (477keV), 113Sn (393keV), SlCr
(322 keY), 47SC (160 keY), '09Cd (88 keY), and 8SSr (514 keY). Onequarter of the multichannel analyzer, 128 channels, was used to obtain a
standard pulse-height distribution for each source. The eight standard
pulse-height distributions were fitted with energy-dependent functions
describing the photopeak, escape peak, Compton tail, and backscatter peak.
These functions yielded interpolated pulse-height distributions for energies
506
lying in the range covered by the standard sources. When suitably normalized
the set of 124 interpolated pulse-height distributions equally spaced over the
interval 0.0198 to 0.9382 MeV formed a 124 by 124 response matrix for the
spectrometer.
Pulse-height distributions were measured for separations between the

source and the detector which correspond to onc, two, three, and four mean
free paths of water. The angle between the source and the detector was
varied in 15 intervals. The pulse-height distributions were unfolded to yield
the differential energy and angle distributions for the various separation
distances.
These spectra are normalized to a source strength of 1 photon/sec. The

differential energy and angle spectrum I represents the energy carried by
photons of energy E(MeV steradian- 1 ), which in unit time cross a unit area
whose normal is in the angular direction specified. The ordinate is a product
of I and the factor 41fT 2 ello r, where r is the penetration depth in centimeters
and /10 is the absorption coefficient at the source energy. These units
not only facilitate comparison among spectra taken at different penetration depths but also allow comparison of spectra taken in different media.
Marshall and Wells 31 have reported comparisons of their Monte Carlo
calculated
1 37 Cs
transport
data in
infinite air with these measured
infinite-water-media data. They made comparisons of the data differential in

energy and angle as well as integrated over angle. The double-differential
data compared favorably in shape. Energy spectra integrated over angle for
the four distances are shown in Fig. 9.20, for which the measured data have
been divided by a factor of 1.2. Agreement in shape and relative magnitude
is good.
In addition to the measurements, Alberg et al. 2 9 carried out multigroup
transport calculations of differential energy spectra in water for a 137 Cs
source using the latest values of the attenuation coefficients for water. 32
The spherical-harmonics technique 33 (described in Sec. 4.3) was used, in

which the transport equation is solved directly by expanding the angular flux
in polynomials, and a matrix equation is used to solve the resulting set of
multigroup equations. For these calculation expansions in Legendre polynomials up to degree seven were used. Comparisons of the calculated and
measured energy spectra integrated over angle for the four mean-free-path
values are shown in Fig. 9.21. The overall agreement is good. The point
scatter in the experimental curves at higher energies is caused by poor
counting statistics.
10
507
8
6
'b
"-
...J
U.
>-
"w
a:
zw
10
8
6
4
10- 1 L _ - l_ _-'-_ _..J.._ _,-L-_----:.L_-='':-_--'
0.1
0.2
0.3
0.4
0.5
0.6
0.7
ENERGY lEI. MeV
Fig. 9.20-Point isotropic

Wells." )
137 Cs
source total-energy spectra. (From Marshall and
9.3 PHENOMENOLOGICAL MEASUREMENTS

Phenomenological problems are usually geometric in nature, Le., they are
characterized by a particular geometry. Under this general heading we
include air transport, air-over-ground interface effects, albedos, and duct
penetration.
9.3.1 Air-Transport and Air-Ground Interface Experiments
In Sees. 8.2.2 and 8.2.3 we described a series of air-transport
measurements made on a tower in Nevada with a 14-MeV neutron source. It
508
Mean
free
10
.' .' " ..
,~
,p
paths
'
~-"""':'c::..
..,-..:..:::.
... '~--..:.~
...~.- 4
"
~:.::.
1x
....
:l
.oJ
"'~""
LL
>to
'.
'
'.
a:
w
Z
','
'.
",
Multigroup transport
calculation
Experimental
0.1
0.2
0,3
0.4
0.5
0.6
0.7
ENERGY lEI, MeV
Fig. 9.21-Comparison of the experimentally determined differential energy spectra (10 )

in water with multigroup transport calculations. The.co is obtained from an integration of
the differential energy and angle spectrum lover all angleR, [From M. Alberg, H. Beck,
K. O'Brien, and J. E. McLaughlin, Nuclear Science and Engineering, 30(1): 71 (1967).]
is interesting to look briefly at other measurements in this series, which

involved both a reactor and a 60 Co source. There were several purposes for
this series of measurements. The primary objective was to obtain definitive

air-transport data at large distances (>1 km) for the source energies used.
Source height and detector height were varied to obtain the effect of the
air-ground interface. Although performed primarily to support nuclear-
509
weapons-effects studies, measurements such as these are frequently useful in

safety analysis of reactors.
Neutron and gamma-ray measurements were made at the Nevada Test
Site (NTS) during Operation BREN (Bare Reactor Experiment Nevada),
which is described by Sanders, Haywood, Lundin, Gilley, Cheka, and
Ward 34 and by Auxier, Haywood, and Gilley. J5 The BREN facility
consisted of a 465-m-high steel tower (Fig. 9.22). Two radiation sources
were supported by an aluminum hoist car that could be moved along the
face of the tower. Tbe two sources were the ORNL Health Physics Research
Reactor (HPRR)36 (shown on the tower in Fig. 9.22) and a 60Co source
whose effective activity was approximately 800 Ci. The HPRR is a small fast
reactor that produces a nearly isotropic angular distribution of fission
neutrons and is neither moderated nor cooled.
From the extensive array of measurements made during Operation
BREN, we show only dose fatc vs. distance for neutrons, secondary gamma
rays, and 60Co gamma rays reported by Haywood, Auxier, and Loy.3?
Figure 9.23 shows the neutron dose as a function of slant range for
source heights of 8.2, 91, 343, and 457 m and a detector height of 0 m.
Since radiation entering the ground beneath the source is partially absorbed,
the dose at a fixed distance away should increase as the source is raised
because of less ground absorption. Figure 9.23 shows that the neutron dose
does indeed increase with source height.
Figure 9.24 shows the measured gamma-ray dose as a function of slant
range for reactor heights of 8.2,91, and 343 m and a detector height of 0 m.
The gamma-ray dose is seen to increase also as the source height increases,
but tbe mechanism responsible for this is not necessarily tbe same as that for
the neutron-dose increase. Some of the gamma-ray dose is caused by
secondary gamma rays resulting from neutron interactions, a portion of
which come from tbe ground.
Figure 9.25 shows the gamma-ray dose from the 60Co source as a
function of slant range for source heights of 8.2, 91, and 343 In and a
detector height of 0 In. The behavior is identical to tbat for neutrons in that
the 60Co dose increases with source height.
A differential measure of the air-ground interface effects on the neutron
dose reported by French and Mooney3 B was based on data obtained during
the 14-MeV measurements. described in Sees. 8.2.2 and 8.2.3 to test the
concepts of the first-last collision model. The model was tested (in relation
to last scatters only) by placing a slab of polyethylene on a hoisting
mechanism to separate the neutron dose arriving at a detector into fractions
511
107 r - - - - , - - - - , - - - - , - - - - - , - - - - , - - - - , - ,
Do HR
.~
8.2 m, HD = 0
91 m, HD
HR
'"
HR '" 343 m, Hv = 0
X H R '" 457 m, H o '" 0
Air density, 1.02 g/liter
.c
~
1()" ' - -_ _' - -_ _---"'--_ _-'---_ _-'---_ _- ' -_ _-'-..J
200
400
600
800
1000
1200
SLANT RANGE, m
Fig. 9.23-~Neutron dose times slant range squared (D X R 2 ) as a function of slant range
(R) for a detector height (H D ) of 0 m and various reactor heights (H R ). (From Haywood
etal. 37 )
from the upper and lower hemispheres. The slab assembly was designed to
move from the ground surface to a height of 70 ft above the ground.
Hurst-type fast-neutron dosimeters were mounted on the top and bottom of
the slab and on a boom for measurements of the free-field dose.
The experimental setup was located at a horizontal distance of 1000 ft
from the base of the tower supporting the 14-MeV accelerator. The source
was fixed at a height of 112 ft above tho ground for the entire test.
512

106 ~--~--~----r---~---.----r--"""--,
o HR
"" 8.2 m,
t::. HR
HD "" 0
91 m, HD = 0
o HR
""
343 m, Ho "" 0
Air density, 1.02 g/liter
,-
1:
105
a:
'x"
Q
10"
L-_~
_ _--'-_ _--'-_ _
200
400
600
_ _--'-_ _....L-J
.L-_~
800
1000
1200
1400
SLANT RANGE, m
Fig. 9.24-Gamma-ray dose times slant range squared (D X R 2 ) as a function of slant

range (R) for a detector height (HD) of 0 m and several reactor heights (H R ). (From
Haywood et ai. 3 7 )
Figure 9.26 shows the experimental setup looking toward the BREN
tower. The lower detector is attached to the bottom center of the slab, the
free-field detector is on the boom extending from the slab toward the tower,
and the top detector (not visible) is mounted atop the slab. Fast-neutrondose measurements were made on the three detectors for slab heights from
0.23 to 21 m. The uncollided dose, seen only by the top and free-field
detectors, was calculated and subtracted from the measured doses. The
uncollided dose is not affected by the interface and is therefore ignored in
the first-last collision model. In addition to functioning as a 211 shield, the
slab ba.ckscatters neutrons into the detector and reduces the neutron flux at
513
Hs = 8.2 m, HD = 0
6 Hs = 91 m, H D = 0
o Hs = 343 m,HD =0
Air density. 1.01 g/liter
,OB
HI'L---.,-L---J':---:-l-:----'------,J
200
400
600
800
, 000
SLANT RANGE, m
Fig. 9.25--Gamma-ray dose times slant range squared (D X R 2 ) as a function of slant

range (R) for a detector height (Hv) of 0 m and various 60 CO source heights (HS)' (From
Haywood et al. 3 7)
the ground in an area dependent on the slab height. These effects were
calculated and subtracted from the measured dose. Results were shown in
Sec. 8.2.3.
To make quantitative comparisons of the calculated and measured
interface effects, the calculated total last-collision dose was normalized to
the measured free-field detector dose at a height of 21 m. This height was
selected for the normalization because the interface effect should be least at
this height. Comparison of the normalized results shows the extent to which
the theoretical model estimates the fraction of the 21-m free-field dose
observed at lower heights.
Examination of the scattered doses shown in Fig. 8.10 shows that the
dose at the top detector is relatively independent of height and that the dose
at the bottom detector increases by a factor of more than 2 as the height is
increased from 0.23 to 21 m.
514
fig. 9.26-Arrangement for 2rr shield looking toward the neutron source. (From French
and Mooney.38)
9.3.2 Duct Penetration

The streaming of neutrons and gamma rays through entrancewaysl ducts,
vents, cable-route openings, portholes, and other openings has always been a
major problem in shield design. Many of the efforts at devising methods for
calculating streaming are reviewed in Chap. 7. Clifford et al. 1 3 report the
comparison of measurements and calculations of neutron streaming through
multilegged concrere ducts from an experiment performed at the Tower

Shielding Facility at ORNL. Their results have already been described in
Chap. 7, but it is interesting to see the experimental arrangement for a
one-bend duct (Fig. 9.27). A 7.6-cm collimated beam of neutrons was
incident to the opening of the O.91-m-square duct at an angle of 45. The
SIS
Fig. 9.27-Photograph of experiment for duct with one bend. (From Clifford et al. 13)
source of neutrons was the Tower Shielding Reactor II. The fast-neutron
spectrum of the beam was measured by Bokhari, Verbinski, and Todd'9
using the shielded diode spectrometer described previously. The slow-neutron
spectra were obtained from measurements with a boron-shielded BF, detector 40 having lOB filters with 23 different thicknesses.
Three duct configurations were involved in the test: a straight duct, a
duct with one bend, and a duct with two bends. Principal results have been
shown in Figs. 7.52 through 7.54. The measured absolute energy spectrum
of the beam was used for the source spectrum in the calculations. In general,
the agreement with calculations averages better than 20% through five orders
of fast-neutron dose attenuation. Calculations were made of the thermal flux
resulting from all source neutrons with energies above cadmium cutoff to
test the accuracy of the slowing-down model used in the code. In general,
the agreement is also within 20% in the second and third legs and within 5%
for most of the straight ducts. For the three-legged duct, it was necessary to
follow approximately 18 wall backscatterings to obtain 90% of the flux
contribution in the middle of the second leg and 25 wall backscatterings in
the middle of the third leg.
516
Even though we have described widely diffetent shield experiments, we

have not done justice to the diversity of arrangements in the litetature. The
variety of experiments performed is as great as the variety of ptoblems the
shield designet has faced since radiation soutces were first used. Sources of
radiation being measured have included reactots of all kinds (powet,
research, pulsed, and critical assemblies), as well as source plates, isotopes,
neutron generators, and a variety of accelerators. Environments in which
reactor sources and detectots have been placed range ftom the bottom of
swimming pool facilities to towers 200 and 1500 ft high, to aircraft flying at
35,000 ft, and to satellites in orbit. Detectors have been located in aircraft
crew compartments, underground shelters, spacecraft payloads, ducts,
armored tanks, submarines, and surface vessels.
REFERENCES
1. W. J. Price, Nuclear Radiation Detection, 2nd ed., McGraw-Hill Book Company, Inc.,
New York, 1964.
2. C. M. Huddleston (Ed.), Detectors for Shielding Research, in Proceedings of the
Session on Detectors for Shielding Research, held at American Nuclear Society.
Toronto, Canada, June 1968, Report ANS-SD-IO.
3. J. Greenborg (Ed.), Proceedings of the Special Session on Fast Neutron Spectroscopy,
Presented at the 1964 Winter Meeting, San Francisco, Calif., Dec. 1, 1964, Report
ANS-SD-2, American Nuclear Society, February 1965.
4. H. Murphy, Summary of Neutron and Gamma Dosimetry Techniques, Report
AFWL-TR-66-111, Air Force Weapons Laboratory, September 1967.
5. D. D. Glover, Experimental Reactor Analysis and Radiation Measurements, McGrawHill Book Company, Inc., New York, 1965.
6. S. T. Friedman, J. Rathbun, and R. Kaufmann, Use of the Transport Code DTK for
Shielding Design and Analysis, Trans. Amer. Nucl. Soc., 7: 2 (December 1964).
7. Oak Ridge National Laboratory, Compilation of Data on Experimental Shielding
Facilities and Tests of Shields of Operating Reactors, USAEC Report ORNL-RSIC-24,
November 1968.
8. J. A. Belcher and L. K. Zoller, The Outside Test Tank and Associated Hardware,
USAEC Report XDC-59-11-154, General Electric Company, Nov. 6, 1959.
9. S. T. Friedman, Shielding Tests Performed in the General Electric Outside Test Tank
in the Period March-June 1958, USAEC Report APEX-581, General Electric
Company, February 1961.
10. V. V. Verbinski, M. S. Bokhari.]. C. Courtney, and G. E. Whitesides, Measurements
and Calculations of the Spectral and Spatial Details of the Fast-Neutron Flux in Water
Shields, Nuc1. Sci. Eng., 27(2), 283-298 (1967).
11. D. Yetman, B. Eisenman, and G. Rabinowitz, Description of Input Preparation and
Operating Procedures for 9-NIOBE and IBM-7090 Code, in Advanced Shield
Calculational Techniques, I, Report UNUCOR-631, United Nuclear Corporation,
517
1963; S. Preiser; G. Rabinowitz, and E. deDufour, A Program for the Numerical

Integration of the Boltzmann Transport Equation-NIOBE, in Advanced Shield
Calculational Techniques, Il, Report UNUCOR~632. United Nuclear Corporation,
1963.
12. D. K. Trubey and B. F. Maskewitz (Eds.), A Review of the Discrete Ordinates Sn
Method for Radiation Transport Calculations, USAEC Report ORNL~RSIC-19, Oak
Ridge National Laboratory, March 1968.
13. C. E. Clifford, F.]. Muckenthaler, R. E. Maerker, E. A. Straker, and F. R. Mynatt,
Neutron Transport Studies in Ducts and Simple Shields, in Proceedings of the
Conference on the Physics Problems of Reactor Shielding held at the Atomic Energy
Research Establishment, Harwell, Sept. 26-29,1967, British Report AERE-R-5773
(Vol. 4), p. 937, 1968.
14. C. E. Clifford, E. A. Straker, F. J. Muckenthaler, V. V. Verbinski, R. M. Freestone,
K. M. Henry, and W. R. Burrus, Measurements of the Spectra of Uncollided Fission
Neutrons Transmitted Through Thick Samples of Nitrogen, Oxygen, Carbon, and
Lead, Investigation of the Minima in Total Cross Sections, NucL Sci. Eng., 27(2):
299-307 (1967).
15. D. C. Irving, 05R Cross Section Library, Memo 2, Code Package CCC-17, Oak Ridge
National Laboratory, 1965.
16. S. Cierjacks, P. Forti, D. Kopsch, L. Kropp, and H. Unseld, A Novel Method for Very
High Resolution Cross-Section Measurements, German Report KFK-453, 1966.
17. J. R. Stehn, M. D. Goldberg, B. A. Magurno, and R. Wiener-Chasman, Neutron Cross
Sections, Vol. I, 2nd ed., Suppl. 2, USAEC Report BNL-325, Brookhaven National
Laboratory, 1964.
18. Brookhaven National Laboratory, ENDF/B Tape 102, Cross Section Library.
19. R. J. Cerbone, J. E. Miller, and E. A. ProGo, Angular Neutron and Gamma-Ray
Spectrum Measurements in a Bulk Iron Assembly, USAEC Report GA-9149(Vol. I),
Gulf General Atomic, Mar. 27, 1969.
20. A. E. Profio, H. Antunez, and T. Williamson, Measurements of the Neutron Angular
Flux Spectrum in Graphite, Report GA-7470, General Dynamics Corporation,
Oct. 30, 1966.
21. A. E. ProGo and R. J. Cerbone, Measurements of Neutron Angular Flux Spectrum in
Water, Report GA-8355 (DASA Report 2056), General Dynamics Corporation, 1967.
22. K. D. Lathrop, DTF-IV, A FORTRAN-IV Program for Solving the Multigronp
Transport Equation with Anisotropic Scattering, USAEC Report LA-3373, Los
Alamos Scientific Laboratory, July 15, 1965.
23. A. D. Carlson, R. J. Cerbone, and D. F. Willoughby, Measurement of Neutron
Penetration Standards, Vol. II, High Resolution Measurements of the Total Neutron
Cross Section of Nitrogen and Iron, Report GA-9149 (DASA Report 2289), Gulf
General Atomic, March 1969.
24. M. D. Goldberg, S. F. Mughabghab, B. A. Magurno, and V. M. May, Neutron Cross
Sections, Vol. IIA, Z = 21 to 40, USAEC Report BNL-325, 2nd ed., Brookhaven
25. S. Cierjacks, P. Forti, D. Kopsch, L. Kropp,]. Nebe, and H. Unseld, High Resolution
Total Neutron Cross Sections Between 0.5~30 MeV, German Report EURFNR-591
[KFK-1000; EUR-3963; EANDC(EJ-11l-U], June 1968.
518
26. R. B. Schwartz, R. A. Schrack, and H. T. Heaton II, Total Neutron Cross Sections of
Carbon, Iron and Lead in the Me V Region, in Neutron Cross Sections and
Technology, Proceedings of a Conference, Washington, D. C., Mar. 4-7,1968, D. T.
Goldman (Ed.), NBS-SPEC. PUBL.-299, Vol. 2, p.771, Supecintendent of Documents, U. S. Government Printing Office, Washington, D. c., 1968.
27. D. J. Hughes and R. B. Schwartz, Neutron Cross Sections, USAEC Report BNL-325.
2nd ed., Brookhaven National Laboratory, 1958.
28. J. R. Stehn, Murrey D. Goldberg. Benjamin A. Magurno. and Renate Wiener-Chasman,
Neutron Cross Sections, USAEC Report BNL-325. Sup pI. No.2, Brookhaven
29. M. Alberg, H. Beck, K. O'Brien, and ]. E. McLaughlin, An Investigation of
Differential Energy and Angle Spectra for 137CS Gamma Rays in Water and Air,
NucL Sci. Eng., 30: 65 (1967).
30. L. D. Skarsgard and H. E. Johns, Spectral Flux Density of Scattered and Primary
Radiation Genetated at 250 kV, Radiat. Res., 14: 231 (1961).
31. J. D. Marshall and M. B. Wells, A Comparison of Monte Carlo Calculated and
Experimentally Determined Energy and Angle Spectra for 137 Cs, USAEC Report
RRAM74, Radiation Research Associates, Inc., June 1967.
32. J. H. Hubbell and M. J. Berger, Photon Attenuation and Energy Transfer Coefficients,
Tabulations and Discussion, Report NBS-8681, National Bureau of Standards,
May 10, 1965.
33. W. D. Lanning, Application of the Spherical Harmonics Technique to Problems in
Gamma Ttansport, Nucl. Sci. E,!g., 15: 259 (1963).
34. F. W. Sanders, F. F. Haywood, M. I. Lundin, L. W. Gilley, ]. S. Cheka, and D. R.
Ward, Operation Plan and Hazards Report-Operation BREN, USAEC Report
CEX-62.02, Oak Ridge National Laboratory, January 1962.
35. J. A. Auxier, F. F. Haywood, and L. W. Gilley, General Correlative StudiesOperation BREN, USAEC Repott CEX-62.03, Oak Ridge National Labotatoty, 1963.
36. F. F. Haywood, Spatial Dose Distribution in AirOver-Ground Geometry, Health
Phys., 11(3): 185-192 (Match 1965).
37. F. F. Haywood, J. A. Auxier, and E. T. Loy, An Experimental Investigation of the
Spatial Distribution of Dose in an Air-Over-Ground Geometry, USAEC Report
CEX-62.14, October 1964.
38. R. L. French and L. G. Mooney, Differential Measurements of Fast-Neutron
Air-Ground Interface Effects, Nucl. Sci. Eng., 43: 273-280 (1971).
39. M. S. Bokhari, V. V. Verbinski, and H. A. Todd, Measurement of the Fast-Neutron
Beam Spectrum from the Tower Shielding Reactor, II, in Neutron Physics Division
Annual Progress Report for Period Ending August 1, 1963, USAEC Report
ORNL-3499 (Vol. 1), pp. 105-108, Oak Ridge National Laboratory, 1963.
40. T. V. Blosser and R. M. Freestone, Jr., Development of an Epithermal Neutron
Spectrometer. in Neutron Physics Division Annual Progress Report for Period Ending
August 1, 1965, USAEC Report ORNL-3858 (Vol. I), Oak Ridge National Laboratory, 1965.
Shield Design
H. C. CLAIBORNE and N. M. SCHAEFFER
10
To integrate the numerous design considerations in an attempt to gain a

perspective of the whole process of shield engineering, we will proceed in
two different ways. First, from an entirely philosophical standpoint, we will
describe a set of iterations in a shield design based on the experience of
Hungerford 1 in the course of developing the shield system for the Enrico
Fermi Nuclear Plant. His recollection of the tasks is both interesting and
instructive since it is based on actual design experience and has subsequently
been refined by his lectures in radiation shielding at Purdue University..
Dr. Hungerford has also formulated six principles of reactor-shield design;
these are given below. A second way to obtain a perspective of the design
process is to examine the shields of several operating reactors and to review
the approaches used by their designers. We will discuss several widely
differing types; particular attention will be given to special problems
encountered in each case. We have limited the selection of shield design for
this discussion to cases now in the literature, particularly those which
include publication of measureme nts on shield performance as well as design
description. Proof of shield performance is the final step in the design
process; a design is incomplete without it. This limitation means generally
that systems available to us for this discussion were completed in the
mid-sixties. This, in turn, implies shield analyses that date from the late
fifties and early sixties. Consequently the analysis methods used in the work
we will study are not as advanced as the latest ones described in Chaps. 4 and
6. There is an 8- to 10-year period in the cycle from first concept to start-up
tests, and shield technology has changed markedly in the interval.
To return to our first viewpoint, the philosophical viewpoint, consider
the following shield design principles that are based on experience developed
over the years. They are guides to be followed in the conception and
development of a reactor plant and shield system.
519
520
Hungerford's six principles of shield design as stated in Ref. 1 are:

1. Reactor-Shield Unity: A shield is an integral part of a reactor system and
must he designed at the same time as, and as an entity with, the overall reactor
system.
2. Shield Integrity: Adjacent parts of a shield, having the same design criteria,
must be designed with equal performance characteristics.

a. Two adjoining parts of a shield which are made of different materials
must have the same radiation attenuation characteristics.
b. Shields around voids, gaps, and irregularities must be designed to give the
same overall attenuation of the radiation as the bulk shield.
3. Shield Safety: Becaus~ the reactor shield is a safety device and must be
considered as a part of the safety system of the reactor, there can he no
compromise with expediency in its design.
a. Sloppy design and construction techniques in sensitive areas which may
affect the safety of the reactor cannot be tolerated.
b. Use of cheap materials and other cost-cutting practices may be more
costly in the long run, and moreover may jeopardize the safety of the
reactor.
4. Shield Accommodation: A shield should be adapted to provide for the
mechanical requirements of the reactor, its supporting structure, and its
component systems, without sacrificing the principles of reactor-shield unity,
shield in tegrity, or shield safety.
5. Shield Economy: The best possible shield should be designed at the lowest
possible cost, consistent with the overall reactor design, without sacrificing safety,
integrity, or performance.
a. An economical shield is a must in the design of large nuclear power
plants for the production of electric energy.
b. This principle is subservient to the fust four principles.
c. An application of this principle is the concept of shield optimization,
whereby a shield may be optimized according to some parameter such as
weight, size, cost, or some nuclear property, etc.
6. Shield Simplicity: A shield should be designed to be as simple in
configuration as possible, with the minimum number of voids, ducts, and cutouts
for the reactor components and auxiliary systems, consistent with the principle of
shield accommodation.
a. An application of this principle would be to allow all pipes and cables
serving the reactor to pass through the shield in one penetration of the
shield (or as few as possible).
b. Simpler shields will have more satisfactory performance, less maintenance, and will cost less.
10.1 ITERATIONS IN THE SHIELD DESIGN

The job of creating a reactor-shield design proceeds in three distinct
phases: (1) the development of a preliminary conceptual design; (2) the
SHIELD DESIGN
521
correlation of the preliminary conceptual design with the mechanical design

to obtain a final conceptual design; and (3) the translation of the final
conceptual design into a detailed engineering design. Each of these steps has
a distinct subset of problems. The development of the conceptual design
includes the consideration of the selection and availability of materials and
problems with the use of these materials in the shield. Let us follow in detail
the development of the design and later of the shield construction to see
what types of problems may arise in each of these phases.
10.1.1 Preliminary Conceptual Design
The preliminary conceptual shield design begins with a rough sketch of
the plant layout showing where the reactor and every auxiliary system and
component is to be placed. It is important here for the shield designer to
have a voice in the plant layout since this will affect the whole design and
future operation of the plant. From this sketch the designation of all areas to
be shielded and their interrelation is made. Once this is done the shield
design criteria for each area with respect to personnel protection must be
worked out. Next, rough sketches of each shield or shield area must be made
so that the mechanical requirements as well as the shield requirements can be
considered and developed. Close coordination with the mechanical design
people is a necessity.
In most reactor design and construction projects, a time schedule for
developing each design will have be~n set up, hopefully enough ahead of the
construction schedule to give adequate time for the development of the
design. Probably the shield designs for several areas will have to be developed
simultaneously.
(aJ Preliminary Selection of shield Materials. Once the type of shielding
and overall requirements of the shield for each area have been worked out, a
preliminary selection of shield materials is made. The properties of the
selected shield materials are investigated thoroughly with re~pect to the
anticipated environment to ensure that the materials will be able to stand up
to the environment and perform reliably over the lifetime of the plant. The
shields, like much of the permanent structure of the plant, are not amenable
to replacement during plant operation.
(bJ Shield Calculations. Only after the shield materials have been
selected, or the choice is at least narrowed to two or three, should the shield
calculations begin in earnest. The model for the shield calculations can be set
up. The preliminary design sketches may contain several different configura-
522
tions based pardy on the types of materials chosen and pardy on economics
and other considerations. Rough calculations should be made for each shield
region to help narrow down the choice of materials by evaluating shield
performance of each of the selected materials. Considerations should include
preliminary calculations of radiation attenuation, nuclear heating, radiation
effects, temperature rises, and internal stress generation. The next set of
calculations should include a parametric study to determine the best
arrangement of the shield materials with respect to the overall mechanical
design and arrangement of the structure.
(c) Final Materials Selection. Firm and final selection of each shield
material should await the results of definite tests on samples of the material
exposed to a heat and radiation environment equivalent to plant operating
conditions and total exposure equivalent to the plant lifetime if this
information is not already known. One of the biggest mistakes that can be
made in shield design is to choose an untested material that is acceptable on
paper but whose integrity and reliability are questionable.
Once the shield material for a given area has been tentatively selected,
the simple conceptual design of the bulk shield can be developed from the
results of the calculations already at hand. The problems associated with this
stage of the design are the difficulties in selecting the shield material, ill
developing the calculational model, or in executing the calculations.
10.1.2 Detailed Conceptual Design
The next step is to divide the shield areas into a number of regions to
check deviations from the model of the initial calculations. An example
would be that the mechanical design calls for a thinner or thicker shield in
some particular area, with perhaps a change or addition of materials. shield
thicknesses for each of these regions should be calculated separately. The
results of all these calculations are then brought together and assembled into
a preliminary conceptual design.
At this point the design problems begin in earnest since the preliminary
conceptual design must be compared with the mechanical design and the two
fitted together. The preliminary design, as we have stated, includes only the
most important shield regions and perhaps only those at the most important
elevations or cuts through the shield. The detailed conceptual design must
show clearly at every elevation and cut through the shield the exact material
to be used and the broad features and configuration of the material in every
area of the shield. The main penetrations through the shield and the
SHIELD DESIGN
523
shielding around them must be shown. Special calculations must be

considered for the areas of penetration and areas of shielding around
mechanical objects attached to the reactor vessel. Such items as support
structure, coolant lines, detectors, experimental tubes, pipe hangers, cable
trays, sheaths containing thermocouples, or other types of sensors are all
examples ofobjects that the shield must be designed around.
(a) Special Calculations-Radiation Streaming. To calculate radiation
streaming, one must divide the shield into a large number of special areas. In
each area judgment must be passed as to the allowable leakage radiation.
Calculations must be performed to determine the proper amount of shield
material. Many of the calculations will involve streaming in voids, gaps,
circular and rectangular ducts, and annuli. These areas in a shield present the
greatest challenges because one must not only create calculational models to
simplify a complex mechanical design but must also exercise good judgment
in estimating the streaming and in interpreting the results.
(b) Nuclear Heating. Another problem of great concern during this
design stage is nuclear heating. Earlier analysiS of materials should indicate
those regions where there might be excessive heating in the bulk shield. If
detailed calculations in any given area show that there is indeed a heating
problem, then proper steps must be taken to change the material, add
energy-absorbing (thermal) shields, or call for a shield coolant system
(whichever is most appropriate). The concern here is therefore not with the
shield material but with heating or radiation effects on structural material,
equipment, or system components located in each region. If the heating
induces excessive temperature rises or if the neutron fluence is above the
radiation-damage limit, additional coolant or shield material may be needed.
If the problem in a given region is too severe or if there is no room to add
any more material, more drastic action must be taken, e.g., change the
mechanical design or reevaluate the shielding material initially selected. This
may mean upgrading the shielding properties of the material in this region, if
possible, by adding more absorber into the shield material (such as increasing
the boron content of borated graphite) or by replacing it with a more
efficient material. Any of these actions, of course, may have economic
reverberations.
(c) Economic Considerations. Once the various shield regions have been
analyzed, the materials chosen, and the thicknesses and configuration for the
regions decided on, an economic analysis should be performed. Grades or
524
quality of the material must be specified. If the calculated costs exceed the
budget allowance, the shield should be reviewed again to determine if there
are areas that could be simplified or if the material could be changed.
Since all budgets are tight (by definition), there will be a great
temptation during this stage of the design to try to dictate upper limits of
allowed costs and, by inference, the quality or grade of material to be used.
Certainly, the use of an expensive material cannot be justified if there is a
less expensive one that will do the same job and positively stand up to the
anticipated environment during the plant lifetime. On the other hand, if
there are circumstances that call for a certain material because it is the only
safe and reliable material for the job, then no substitutions can be
considered. Decisions on questions of safety vs. economy must obviously
favor safety.
(d) Completed Conceptual Design. When all aspects of the shield design
have been evaluated, the materials selected, and the necessary modifications
made in the preliminary design, the detailed conceptual design is complete.
Documentation should be in the form of a report with drawings at every
important elevation and penetration through the shield showing for each
shielded area the basic configuration and materials.
10.1.3 Final Engineering Design

The conceptual design must be translated into engineering drawings.
Detailed drawings of the reactor and auxiliary systems and components must
be analyzed and redrawn to show the proper shield modifications. It is here
that all the fine detail of the shield structure comes to light.
For instance, the reactor is built at room temperature, but it may be
operated at 540C. The reactor, its support structure, the coolant piping,
any auxiliary gas coolant lines or cover-gas lines, all mechanical appendages,
sensing devices, etc., will have to operate at temperatures at or near the
operational temperature of the reactor.
The mechanical system will have been designed to flex, shift, and move
with temperature to relieve the stresses on the system. The coolant pumps
and the motion of the coolant flowing within the system will create
vibrations and other movements of the system components. The shield
cannot be erected around the reactor and its systems without regard for
these expansions and movements. The engineering design therefore must
include allowances for expansion and motion.
SHIELD DESIGN
525
The drawings must show or indicate clearly how close to the vessel and
its systems the shield can be erected, and all clearances must be detailed. It
rnay even be necessary at this late date to rerun shield calculations, taking
into account the clearances, and perhaps to specify changes in the shielding
material.
One of the problems arising at this stage of the design is the possible
conflict thai may arise between the shielding requirements and the
mechanical requirements. A cooperative effort between the shield designer
and the mechanical engineer is necessary for a satisfactory compromise
solution to such problems. If this type of cooperation is not achieved before
the construction phase, costly rework may result.
The engineering design, besides reflecting the requirements of the

conceptual design, must also take into consideration the requirements for
field erection. Therefore on each drawing all necessary instructions and notes
must be carefully made in language intelligible to construction crews. The
completed engineering design will be a series of elaborate drawings showing
all the requisite detail for proper erection.
(a) Field Drawings. The architect-engineer delivers a set of engineering
drawings to the plant constructors, who copy each drawing and prepare field
drawings by adding pertine';t information for field workers, e.g., welding
instructions, construction tolerances, and erection sequence. It would seem
that the shield design job should be complete by the time the design reaches
this stage, but this is not the case, especially in the matters of construction
tolerances and erection sequences. For example,
if a shield is to be erected
out of a series of graphite blocks that are graded in boron content, the
manufacturer must color-code his blocks or identify the percent boration by
some other easily recognized means, and the color code or identification
instruction must appear on the field drawings.
(b) Construction Follow-Up. Each of the color-coded graphite blocks
will have the specified dimensions and a dimensional tolerance in accordance
with the material specifications given to the manufacturer. The erection
crews will be expected to lay these blocks to form the shield with a specified
construction tolerance. In most jobs the erection tolerances are generous to
keep down the cost of construction. The designer's task here is to ascertain
that the construction clearances and the manufacturer's dimensional
tolerances when taken together do not lead to unexpected streaming paths
through the shield. Even here he should be prepared to rerun some shielding
calculations covering the affected region as a check to determine the effect
526
of excessive clearances if these have not already been included in his previous
calculations. As a result of these checks, the designer should be prepared to
take one of the following actions: (1) modify the configuration of the
blocks, (2) issue orders that gaps greater than a given length and width be
filled, or (3) issue orders that overall erection tolerances be cut to more
severe limits. He should consult and obtain the cooperation of the field
engineer to ensure that the erection crew will follow the spirit as well as the
specification of the changes.
(c) Construction with Concrete. In the pouring and curing of concrete,
simple field procedures can be set up to ensure cleanliness of the aggregate,
proper concrete mixing ratios, and uniformity of density. The poured
material must be worked by hand or with mechanical vibrators to ensure
closing of voids, proper joining at the boundaries between one pour and
another, and setting and curing of the completed pour. For poured shields,
the allowed clearances must include the forming tolerances. Normally these
cannot be considered to be less than 1 in. Therefore concrete to be poured
around ducts and penetrations should have permanent imbedded steel
sleeves, and the tolerances or clearance limitations will therefore be on the
sleeve itself rather than on the concrete; the concrete must be vibrated to
ensure close joining between the sleeve and the concrete shield.
(d) Testing Vendor-Supplied Materials. Quality control of all materials
is essential during erection to ensure that materials received from the vendor
are within the limits specified. If borated material is being used, for instance,
the material must have at least the minimum boron content specified since
that figure determines the proper shield thicknesses. With graphite, tests of
conductivity and other properties must be run to ensure that the product is
really graphite and not (for example) graphite particles bonded together with
a cement. An impurity analysis should be run on samples of each material to
ensure that the impurity level is within specifications, especially wherever
impurities can be a detriment to shield performance. Impurities in graphite,
for instance, accelerate its oxidation by hundreds of times at temperatures
exceeding 800F (the oxidation threshold).
(e) Field Testing and Inspection. A shield cannot be erected successfully without proper liaison with the erection crew. The shield is a
three-dimensional structure to be erected from two-dimensional drawings. In
some instances the blueprints are very difficult to interpret to obtain a
three-dimensional concept. Thus someone knowledgeable with the shield
design must be available to interpret the drawings. Inspectors are necessary
SHIELD DESIGN
527
at all phases in the erection of a shield to ensure that tolerances are held and
directions are followed.
Radiation surveys of the completed shield system serve as the proof-test.
For power reactors surveys are usually conducted first at low power, then at
one or more intermediate levels, and finally at full power. Detailed surveys
are made in each instance to ensure that design criteria are met. Results of
several of these surveys are discussed later in this chapter.
10.2 FAST BREEDER: ENRICO FERMI

The Enrico Fermi reactor is a prototype of the sodium-cooled fast
breeder concept. Shielding of a fast reactor system poses some special
problems that thermal systems do not have:
1. Leakage of fast neutrons into the shield is high because fast-neutron
cores operate at high power density with relatively thin breeder blankets.
2. The neutron leakage spectrum is harder, it peaks between 50 and 500
keY.
3. The absence of a thermal-neutron peak makes the secondary gammaray problem more complex because of the increased importance of
resonance-neutron absorption.
4. The system components are more subject to radiation damage.
5. Moderating materials are not usually allowed within the reactor vessel
because of safety considerations.
6. Liquid-metal coolants reach a high specific activity, which poses the
double problem of shielding the coolant system and contending with the
high temperatures.
An unusually detailed shield test program was carried out on the Enrico
Fermi reactor because of its prototype nature. Aside from obtaining
operating data to ensure that the shield system would permit safe operation
up to the maximum design power level of 500 Mw(t), the test program was
also aimed at getting sufficiently accurate data to make meaningful
comparisons between calculations and measurements. Because of the
detailed test program and the unusually complete documentation, we will
describe the design features of this system in detail and limit discussion of
other reactor systems to their unusual features. The following discussion of
the Enrico Fermi shield system is taken from the shield design report by
Beaudry, Hungerford, Montey, and Chaltron' and the report on the shield
test program by Beaudry. 3
528
The plant shielding can be classified in two main categories: the shield
system for the reactor and primary coolant systems in the reactor building
and the shield system for the service and fuel-handling systems in the various
auxiliary buildings. The shield for the reactor is discussed in detail in
Sec. 10.2.3. A general description of the plant is given in the next section.
General information on the shield design of fast reactor systems is given by
Hungerford, 4 including brief descriptions of the shield systems of other fast
reactors: Experimental Breeder Reactor No.1 (EBR-1),s Experimental
Breeder Reactor No.2 (EBR-2),6 Rapsodie,7 Dounreay,8 and the BR-S. 9
10.2.1 The Reactor Plant
The Enrico Fermi Atomic Power Plant reactor! 0 is a sodium-cooled fast
breeder designed for a maximum power of 430 Mw(t). A plot plan showing
the various supporting facilities is shown in Fig. 10.1, and an elevation of the
reactor containment vessel is given in Fig. 10.2.
The reactor vessel and its associated structures are shown in perspective
in Fig. 10.3. The stainless-steel reactor vessel is composed of four parts:
lower reactor vessel, transfer-rotor container, upper reactor vessel, and
rotating shield plug. The entire vessel is located in the lower compartment of
the reactor building, below grade, in a nitrogen atmosphere within the
primary shield tank.
The reactor vessel contains the reactor core, which is composed of fuel
and the inner-radial-blanket and outer-radial-blanket subassemblies.
Liquid-sodium coolant flows into the reactor vessel through three equispaced
3S-cm-diameter nozzles. The sodium flows upward through the
outer-radial-blanket subassemblies and through the core and
inner-radial-blanket subassemblies. Heat is extracted from each subassembly
as the coolant flows upward into the upper sodium plenum. Sodium leaves
the upper reactor vessel through three equispaced 76-cm-diameter nozzles.
The hold-down mechanism, which is mounted on the rotating shield plug,
is used to prevent the core and inner-radial-blanket subassemblies from
being lifted by the rapid flow of sodium through the core.
The reactor vessel is contained within a carbon-steel liner that is leaktight
up to the coolant outlet lines and designed to carry any sodium leakage to
detectors located at the bottom. All this is contained within the primary
shield tank, the space between the vessel wall and shield-tank wall being
filled by the primary shield. The construction of the primary shield tank is
such that an adequate height of sodium over the core will be maintained
should the reactor vessel leak.
530
t------~=-::22~M.::;:====+------.
27-TON
11 M
TROLLEY
CRANE
MACHINERY
ll-0ME
PRIMARY
SOOIUM
PUMP
OPERATING
FLOOR
EOUIPMENT OOOR
GRAOE
10 M
SECONOARY SOOIUM
PIPING
180 M ELEV.
EARTH
VALVE11=t1t1-~Ii-RIl~~Vt~\~~
THROTTLE
E.M. FLOWMETERS
IHX
'. ,
10 M
SECONOARY
SOOIUM
II 0
SHIELD
WALL -~~tJ:::::;i~1:;:~~:ffL~~=:j~t=rt~
1.5 M
SECONDARY
PUMP OISCHARGE_!1;!-+++--I'l1"1I
GALLERY
PRIMARY
SHIELD TANK
0.76-M
PIPE
5.5 M
Fig. 10.2-Elevation of reactor containment vessel. (From Beaudry et al. 2 )
The machinery dome, which is attached to the top flange of the primary
shield tank, consists of two parts: (1) the lower machinery dome, which
forms a sealed enclosure over the top of the primary shield tank, and (2) the
upper machinery dome, which contains the aluminum energy absorber, the
control mechanisms, and other plug-mounted equipment.
The reactor is an assembly of 870 removable and, to an extent,
interchangeable units, all assembled on a square lattice unit spacing of 6.83
em and contained in the lower reactor vessel.
The fuel subassemblies, which also contain the upper and lower axial
blankets, occupy a region (roughly cylindrical) in the center of the lattice.
The entire core region is about 79 cm in diameter and 79 cm in height, and
~ach axial blanket section is 43 cm in height. Together the core and the
radial and axial blanket regions approximate a cylinder 203 cm in diameter
and 178 cm in height.
SHIELD DESIGN
531
MACHINERY DOME
OFFSET HANDLING
CABLE
MECHANISM
CASK CAR
CONTROL-DRIVE
MECHANISMS
HOLD-DOWN
MECHANISM
,.
.f _(
_--}-''-'r--ROTATING SHIELOPLUG ASSEMBLY

THERMAL SHIELD
EXIT PIPE
SCHEMATIC
-- COOLANT OUT
HOLD-DOWN DEViCE
TRANSFER ROTOR
ltJ-COOLANT IN
~"'HI--+--":-UPPERREACTOR
CORE
TRANSFER ROTOR
CONTAINER ---ti~*AXIAL
NOT
~;4I-Il-r-RADIAl
BLANKET----+\}~~~:JJ'g\i
GRAPHITE
SHIELDING
VESSEL
BLANKET
LOWER REACTOR
VESSEL
THERMAL SHIELD
SHOWNL-":_-l~i=i~~~nlri~~~~33'
SUPPORT
PLATES
PRIMARY SHIELD
TANK
Fig. lO.3-Perspective view of reactor. (From Beaudry et al. 2
The reactor is controlled by two operating rods and eight safety rods.
Fission counters and ionization chambers for monitoring the neutron flux
are located in the six neutron-counter tubes embedded in the graphite shield
that surrounds the reactor vessel. An Sb-Be neutron source is located at the
core-blanket interface to maintain a sufficient flux for control purposes
during reactor start-up ann in the shutdown condition.
532
The heat-removal system comprises three primary coolant loops and
three secondary coolant loops. Heat removed from the reactor core and
blanket by the primary sodium coolant is transferred to the secondary
sodium coolant in three parallel intermediate heat exchangers located in the
reactor building. The secondary sodium system transfers the heat to the
steam and feedwater system in three once-through-type steam generators.
A primary sodium service system, located in the shielded sodium service
building, stores and purifies the highly radioactive primary coolant. The
major components of this system are three storage tanks, a cold trap, a
plugging indicator, and a primary-system overflow tank.
A separate secondary sodium system (for which shielding is unnecessary)
is located in the steam-generator building. Both systems are blanketed with
argon gas.
10.2.2 Shield Design Criteria
The design criteria for the shield system include limitations on the
biological dose, radiation damage to materials, and radiation heating.
Maintenance philosophy and safety were also considered in establishing these
criteria.
(aJ Biological Dose. A maximum dose rate from both neutrons and
gamma rays was set at 0.75 mrem/hr for all unlimited access areas. Based on
a 40-hr work week, this is 30% of the average weekly exposure permitted by
federal regulations I 1 governing workers in nuclear facilities.
Many areas of the reactor plant do not need to be occupied on an
unlimited basis; consequently the shields for certain areas of the plant were
designed for only limited access, i.e., access controlled and limited to
specified time periods. The design criterion for these areas was based on
estimates of the required working times.
(b) Radiation Damage. Radiation-damage criteria were set for the
stainless-steel reactor vessel, the graphite primary shields, and the various
organic materials, such as gaskets and diaphragms. For concrete, it developed
that radiation heating effects were more severe than damage effects, and
consequently heating dictated the exposure limitation.
On the basis of an expected 20-year plant life, a maximum fast-neutron
fluence of 10 22 neutrons/cm 2 (estimates indicated this value to be
conservative) was used as the fast-neutron exposure limit of the stainlesssteel reactor vessel. This limitation was a principal factor in determining the
thickness of the thermal shielding within the reactor vessel.
SHIELD DESIGN
533
Energy storage through lattice distortion in graphite (Wigner effect)12

becomes a problem at room temperature for fluences of the order of
2 X 10 21 neutrons7cm 2 Studies of the Wigner effect in the graphite primary
shield indicated that setting a minimum operating temperature of 127C in
the graphite would anneal out the energy storage for the expected fluences
of 10 18 to 4 X 10 21 neutrons/cm 2 .
Most of the organic materials subject to radiation damage are in the
auxiliary equipment, where the radiation sources are gamma rays from the
spentfuel assemblies and the primary sodium coolant. A dose limit of 108
rads over the plant life was set for organics except for a few cases that
required more stringent limits, such as Teflon at high temperatures.
(c) Radiation Heating Limitations. Concrete shields were limited to an
incident-radiation energy flux 13 of 4 X 10 10 MeV cm-2 sec- 1 This limit
ensured that the temperature rise due to nuclear heating within the shield
would not exceed S.SoC. Local rises greater than this value can create large
gradients that may cause thermal stresses that result in cracking, spalling, and
void formation.
Temperatures in excess of 93C cause concrete to rapidly lose its bound
water and hence decrease its neutron-attenuation properties. Therefore as a
design criterion the maximum ambient temperature around the primary
neutron shields of ordinary concrete was limited to 82C with an internal
maximum of BsDe.
The maximum allowable temperature gradient in all shields was limited

to 73C/m to prevent excessive thermal stresses. Concrete used only for
gamma-ray shielding was allowed a temperature limit of 182C. The
temperature limit for serpentine concrete shielding was set at 445C.
(d) Remote-Maintenance Consideration. Maintenance considerations
were important in setting requirements for the shield system. These
considerations had a strong influence on certain of the design details of the
shields. Some areas in which inspection is necessary require shielded manhole
access or other means of entrance through or around shields. Shields must
allow the removal and handling of activated equipment by remote means. So
that induced activity in these limited-access areas could be kept at a
minimum, the neutron-flux-density limit was set at 10 4 neutrons cm- z sec-I,
low enough to prevent significant activation of the equipment.
The Fermi plant philosophy on remote transfer and maintenance of very
large radioactive equipment was that the reactor would be shut down. The
plant would be evacuated or all personnel sent to designated safe areas
534
except those necessary for the plant safety and those involved in the remote
operations. This philosophy avoided the requirement for shielding all work
areas, offices, etc. Remote maintenance of large equipment is an operation
apart from the routine affairs of the plant and is necessary only a few times
during the plant lifetime; therefore the expense of shielding every area was
unwarranted.
(e) Safety Design Standards. Precaution was taken in the design of the
Fermi shields to ensure that the plant would be safe and also free from
defects leading to unsafe conditions. Each shield design was analyzed from a
safety point of view, including the designs for all penetrations for piping,
cables, and ducts, to ensure that in the event of any predictable accident
condition the plant would be safe.
The graphite for the primary shield was specified as nuclear-grade
graphite with acceptable physical properties between predetermined ranges.
Dimensional tolerances of the graphite blocks were limited to about 2% to
keep the total void volume in the primary shield at a predetermined value.
Gaps between blocks of adjacent layers or rows were offset to minimize
radiation streaming. All graphite within the rotating plug was canned in
rectangular steel containers; weep holes were provided in the tops of the
containers to allow gases to escape, and room was allowed within the cans
for expansion under radiation.
Concrete for shields had to meet density and strength standards, and no
defects between various pours of the concrete were allowed. Curing periods
up to 28 days were imposed on neutron-shield concrete to ensure proper
hydration. All concrete and other shield structures underwent rigid
inspection during erection. All concrete shields and other concrete in the
compartments below the operating floor (which house the primary coolant
system) were lined with a layer of steel to prevent spallation of the concrete
and load shifting should a sodium leak occur in the primary sodium system.
10.2.3 Reactor-Shield Systems
The reactor shielding can be divided into three main sections: (1) the
primary shield, located within the primary shield tank; (2) a secondary
concrete shield wall that surrounds the primary shield tank; and (3) a
biological shield that forms the operating floor and also surrounds the
containment building. Elevation and plan views of these shields are shown in
Figs. 10.4 and 10.5.
SHIELD DESIGN
535
SQUARE CANS FILLED WITH

BORATED GRAPHITE
HOLD-DOWN AND CONTROLROD CHANNELS

MACHINERY DOME
[
'"'22:~;;;;;~:;R~O~T~A~TING SHIELD PLUG
E.
'0... : .(J.~.~o.
,j..... , .:
STEEL SHOT
_0
:::::..... .... "QQ.

'::::::::-"
CONCRETE
BIOLOGICAl-
o.O~
'~~~I-~~~Cs.~C:;~"~Q.~'~SHIELD
FLOOR
STEEL GAMMA-RAY
SHIELD
15-CM
OVERFLOW LINE
=~-r,./ 7&-eM
SODIUM
OUTLET LINE
15-CM BORATEDGRAPHITE LINERS

INSULATION
Rs:A~::R_
iEicl"l..1I
INLET
: LINES
~EEL
THERMAL SHIELD
AND REACTORVESSEL WALL
SEAL PLATE
GRAPHITE THERMALCONDUCTION LAYER
c::J
BORATED
GRAPHITE
"'..;..... PLAIN GRAPHITE
Fig. lOA-Primary shield system. (From Beaudry et al. 2 )
(a) Primary Neutron Shield. The primary shield system has three main
parts: (1) a stainless-steel thermal shield within the reactor vessel, (2) a
graphite and borated-graphite neutron shield outside the reactor vessel, and
(3) a steel and graphite rotating-plug shield that fits into the top of the
reactor vessel and acts as a biological shield. The upper part of the primary
tank extends around the rotating plug and is also a biological shield
containing steel as well as graphite.
SHIELD DESIGN
537
The thermal shield is located between the breeder blanket and the vessel
wall (Fig. 10.4) and consists of 304 stainless-steel bars and plates. Sodium
coolant flows in the spaces between adjacent layers to remove the heat
generated within the steel. The thermal shield is 30 em thick in the lower
region and 15 em thick in the upper region. The functions of the thermal
shield are: (1) attenuate gamma rays and reduce the heat load to the reactor
vessel, (2) prevent radiation damage to the reactor vessel by lowering the
energy of the high-energy neutrons by inelastic scattering, (3) act as a
reflector and reduce the core leakage, and (4) protect the reactor vessel from
thermal shocks caused by temperature transients in the coolant.
The graphite shield is composed of layers of graphite and borated
graphite (5%). The main portion is located outside the pressure vessel but
inside the primary shield tank (Fig. 10.4). In conjunction with the thermal
shield, the graphite shield is designed principally to reduce the neutron
leakage to levels low enough so that material outside this shield will not be
damaged or highly activated. The problems and costs of erecting the graphite
shield led to the use of two block sizes, 33 by 33 by 10 em and 30.5 by 15.2
by 7.6 em (with a 2% dimensional tolerance). All blocks were canned and
stacked in staggered arrays. They were fitted together as tightly as possible,
and in areas where it was necessary to maintain a rigid structure, the blocks
were fastened together.
Owing to the eccentricity of the reactor vessel with respect to the
primary shield tank, the thicknesses of the graphite and borated graphite
vary greatly, being about 86 em wide at the narrowest part and about
152 em wide at the maximum. At an elevation corresponding to the central
plane of the core, the minimum distance between the core center and the
outer edge of the primary shield is 2.3 m and the maximum distance is about
3.8 m. The basic conceptual design specified a IS-em-thick layer of 5%
borated graphite next to the reactor vessel followed by a 7.6-cm-thick layer
of insulation. Adjacent to the insulation is a region of plain graphite of
varying thickness. A final 15.2-cm layer of 5% borated graphite is located
next to the inner surface of the primary shield tank. This design was
arranged to provide a predetermined amount of neutron attenuation, which
was established by the magnitude of the neutron flux leaving the reactor
vessel and by the required neutron-flux levels on the outside of the primary
shield.
The original concept as described had to be modified in many details as
design progressed to accommodate the necessary structural members,
instrumentation, piping, and auxiliary systems. A cross-sectional view of the
538
actual shield at the approximate elevation of the reactor center is shown in

Fig. 10.6. The designs of the reactor vessel, the associated sodium piping, and
the primary-shield-tank arrangement were frozen at an early stage; these
represented the best mechanical arrangement. Many large components were
already fixed in position. In addition, many items were installed during the
detailed shield design and construction, such as gas lines, instrumentation
HEATER-CABLE
RACEWAY
BORATED-GRAPHITE PATCH
5-CM PURGE-GAS INLET
LOOP 3
35.S-GM Na PIPE
LOOP 1
15.2-CM Na PIPE
HEATER-CABLE
RACEWAY
lOOP 3
15.2-eM Na PIPE
lOOP 1
35.S-eM Na PIPE
DETECTING-TUBE
COOLING DUCT
LOOP 2
35.5-eM Na PIPE
PRIMARY SHIELD
TANK
azm 5% BORATED GRAPHITE

IIIIIBiIB 7% BORATED GRAPHITE
c::::J
PLAIN GRAPHITE
Fig. lO.6-Plan view of graphite-shield layout at 172-m elevation. (From Beaudry

et al!)
tubes, heaters and cables, strain gages, thermocouples, pipe hangers and
supports, and cable trays. Where possible, with these late design additions,
advantage was taken to arrange components conveniently for shielding
reasons. Such changes included the bending and routing of cables and gas
lines, alterations in structural details, reduction in size, changes in gap
size, etc.
The presence of objects inside the primary shield tank caused local
shielding problems by decreasing the amount of available shield material and
creating streaming paths owing to the gaps that were allowed around these
objects to accommodate thermal expansion and erection tolerances. It is
SHIELD DESIGN
539
estimated that at least 500 separate regions occurred for which shielding had
to be specially designed.
Shielding of the primary sodium lines within the tank was very difficult
since their large volume, combined with the additional insulation and
expansion gaps, resulted in very large holes in the shield. A patch of extra
borated graphite was installed outside the primary shield tank (but inside the
secondary shield wall) to compensate for loss of shielding from voids.
Borated-calcium "donut" shields were hung directly on the upper end of the
sodium pipe and large blocks of steel were installed opposite the lower elbow
to reduce neutron streaming along the pipes within the shield. Each of the
six neutron instrument tubes created a 25-cm-diameter hole in the shield,
which resulted in excess streaming up to the operating floor. As compensation, large donuts of borated graphite were fastened to the upper ends of the
tubes, and special blocks of graphite were cut to fit inside and around the
electrical cables within the tubes.
Although gamma rays were of little consequence in most areas of the
graphite shield, the gamma rays formed by neutron capture in the upper
portions of the reactor vessel and thermal shield were found to be of great
importance at the top of the primary shield near the rotating plug. The
entire upper section of the shield waS filled with a mixture of steel shot and
borated graphite (installed in special compartments) to attenuate the gamma
rays to a level of 0.75 mrem/hr at the top of the shield (the operating floor).
In certain regions where the shield was very thin, the boration was increased
to 7%. The complexity of the upper shield design is illustrated in Fig. 10.4.
The primary shield was designed so that it would not require any
auxiliary or internal cooling. About 85% of the total nuclear heat in the
shield (approximately 77 kw at 500 Mw reactor power) is generated in the
inner layer of borated graphite. The location of the insulation is such that
the thermal gradient is toward the reactor vessel; thus most of the heat
generated is returned to the vessel and carried away by the sodium coolant.
Very little heat is conducted through the insulation to the outer graphite
regions. The heat generated in the outer regions is about 12 kw. It is
conducted to the outside of the primary shield and removed by the subfloor
ventilation system.
The rotating-plug shield (see Fig. 10.4) serves both as a biological shield
for the operating floor and as a means of access to the reactor vessel. The
plug is 3.7 m thick, and its cylindrical surface has two steps or offsets. The
top of the plug is slightly above the level of the operating floor; the bottom
is located in the upper part of the sodium pool so that nuclear heat
540
generated in the lower part of the plug will be removed by the primary
coolant. The plug was designed to attenuate neutrons, primary sodium
gamma rays, and gamma rays produced by neutron absorption in the steel of
the plug to levels corresponding to an overall dose rate of 0.75 mrem/hr.
The shell of the plug is made of stainless steel. A 45-cm layer of stainless
steel is located at the bottom of the plug. This layer moderates the higher
energy neutrons at the top of the sodium pool. Above the stainless steel are
six 28-cm-thick layers of plain graphite capped by one 28-cm-thick layer of
1.5 wt.% borated graphite. Each of the graphite layers is separated from the
other by a 5-cm thickness of steel. The 1.8-m zone of plain graphite acts as a
thermal column, which moderates the neutrons so that they are captured in
the final layer of borated material. Directly above the layer of borated
graphite is a 30-cm layer of carbon-steel gamma-ray shielding, which
attenuates the high-energy gamma rays produced by neutron capture in steel
and stainless steel. A 30-cm layer of stainless-steel insulating material and a
28-cm cover plate of solid carbon steel complete the shield. The layer of
insulation enables the plug to be maintained at a relatively high temperature
without overheating the steel cover plate, which bears the weight of the plug.
The graphite used was in the form' of canned 7.6- by 7.6- by 28-cm blocks.
Any large gaps or spaces between cans were filled with thin wedges of steel.
Boron-steel castings were used in regions close to the access ports for the
hold-down mechanism and offset handling mechanism to depress the
neutron-streaming flux in these areas.
The rotating plug contains a large number of penetrations and access
passages to the interior of the vessel. Among these are the large bathtubshaped access passages for the fuel-handling mechanism and the large circular
penetration for the core hold-down device and for the control- and
safety-rod channels. Other smaller penetrations include access ports (plugs
that can be removed to allow access to the interior of the vessel), the sweep
mechanism (an instrument for ensuring that all fuel subassemblies are
correctly seated), and surveillance tubes (small penetrations that allow the
introduction of experimental gear or materials within the reactor). Shielding
for the fuel-handling mechanism matches as nearly as possible the shielding
in the main part of the plug. The very complicated annular void area
between the plug and this shield required the construction of special shield
rings and the minimizing of clearances. Patch shielding of steel and graphite
was used above the plug in this area to reduce the streaming to acceptable
levels. The hold-down mechanism also contains the control- and safety-rod
channels. The hollow stainless-steel control- and safety-rod guide tubes
SHIELD DESIGN
541
contain special shields of stainless steel and boron steel. Areas between the
channels were partially packed with a special powdered shielding mix
containing iron powder (34%), graphite powder (57%), and powdered boron
carbide (9%) compacted to a density of 1.8 g/cm3.
The core and blanket assemblies are transported into and out of the
reactor through the exit pipe (see Fig. 10.3), which is an 18-cm-diam~ter
tube that extends downward from the operating floor to the transfer-rotor
container. Subassemblies stored there during reactor operation undergo
fission and contribute to the dose in the operating floor. Gamma rays
reaching this area during operation come from the sodium coolant and from
neutron capture in upper portions of the reactor vessel. As the fuel
assemblies are raised upward through the tube, their fission-product gamma
rays become increasingly important. Figure 10.7 shows the design that was
evolved to cope with these radiation sources.
A shield plug, about 2.1 m long, fits into the top of the exit port, which
is recessed about 20 em into the operating floor. The plug remains in place
except during fuel transfer. Shielding within the plug consists of 1.5 m of
1.5% borated graphite capped by 0.6 m of steel. The shielding outside the
plug consists of borated graphite, steel, lead, serpentine concrete, dry-pack
serpentine, and ordinary concrete. Vertical streaming is minimized by a
series of sliding borated-graphite rings located between the inner exit-port
tube and an outer tube that is part of the primary-shield-tank structure. The
rings are placed in alternating layers; the smaller rings are attached to the
inner tube, and the larger rings are attached to the outer tube. Additional
steel and lead shielding is located both above and partially around these
disks, and a steel-and-graphite plug is tightly fitted inside the inner tube.
Since the upper part of the exit tube must be heated during fuel transfer
to prevent the sodium from freezing and clogging the valving, dry-packed
serpentine aggregate is used for shielding in this area. A 0.6-m removable
plug located on top of the exit port completes the special shield.
(b) Secondary shield Wall. The function of the secondary shield wall
(Fig. 10.5) is to reduce the neutron-leakage flux coming from the primary
shield tank to an average value of not more than 1 X 104 neutrons cm-2
sec- t within the equipment compartment, a level that will not appreciably
activate the secondary coolant or the primary coolant equipment.
The secondary shield wall is composed of prepacked ordinary concrete
with a density of 2.3 g/cm 3 and it has a lining of 1.3-cm steel plate on each
side. The concrete thickness is mainly 76 em, but it widens to 100 em near
542
TOP
REMOVABLE
MACHINERY DOME
eLL'''-~
BORATED GRAPHITE
SEAL
OPERATING FLOOR
EQUIPMENT
HANDLING lUG
DECK
LEAD
SERVICE GAS
LINES IN TUNNEl-"
'
'. ,~I~m~cuwl1Y/'~""
CONCRETE~~
DRY-PACK
' :
~ . '. :.'
"
.'''~ ..
SERPENTINE
:.'~1rJII
SERPENTINE
:':, '.~. ,:'
CONCRETE
~
. :(
.~
::"
BOXOUT FO~RI"="~'
;
",..
GAS LINE
GRAPHITE
BORATED-GRAPHITE
DONUTS
:~
EMERGENCY- '.'.":
COOLING LINE
/"Il.::lPi~~lfil
PRESSURE-DETECTOR TUBING
EXIT-PORT LINER TUBE

STEEL GAMMARAY SHIELD
EXIT-TUBE PLUG
PRIMARY SHIELD-
TANK WAll
Fig. 10.7 -Fuel exit-port assembly. (From Beaudry et al. 2 )
the location of the transfer rotor. The increase in thickness here is necessary
because when the transfer rotor is filled with spent subassemblies the
neutron leakage from the primary shield in this area is of higher magnitude
than the design leakage value.
A prepacked construction method whereby the aggregate is laid in first
and the cement pressure-grouted into the aggregate was used to ensure
uniformity in the density and Water content of the completed shield. Heavy
reinforcing steel bars on a lS-cm mesh formed two planes of reinforcement
for the secondary shield.
A S-cm-thick steel thermal shield was placed around the wall in the
high-gamma-flux areas to prevent excess heating and drying out of the
SHIELD DESIGN
--PRIMARY
SHIELD TANK
STEEL SLEEVE
PIPE
"n~ ~ i V
.~
..
\-
"II
..
'I
~.d
'
It
)1', J CONCRETE ..~, I,

":, , . , .... J 0 I,
J
'\
,_ "
1\1
543
r;;::-COOLING GAP
!IOLOGICAL SHIELD
SMALL AIR HOLE
GAP FOR COOLING
(,.
INSUL:A~T~'O~N~'\~~'~"~-~~.~~~i .~.~ .~.~(.~ . ~\~~~t~~~~~~~~~~

{
"CONTAINMENT'AIR GAP
PIPE...!'
--h---1~-GAP - - - - - - - -
(fl
I-CONTAINMENT
Fig. lO.8-Schematic view of typical sodium-pipe penetration through secondary shield

wall. (From Beaudry et al. 2 )
concrete by gamma rays emitted from the sodium. Cooling was accomplished by ducting the lower building ventilation system to the 12.7-cm gap
between the wall and the thermal shield.
Special shielding problems were created by the various penetrations of
the secondary shield wall for the primary coolant inlet pipes, primary
sodium overflow pipe, gas lines, and electric cables. Electric- and gas-line
penetrations were kept small; they pierced the wall at angles close to
tangential with the primary shield tank. The 76-cm-diameter sodium exit
lines pass under the shield wall and require special shielding within the
reactor compartment and in the areas where they penetrate the floor of the
equipment compartment. Because of mechanical considerations, the sodium
inlet lines were required to pass through the shield wall in directions nearly
perpendicular to the wall's surface.
The design of a typical sodium-pipe penetration through the secondary
shield wall is shown in Fig. 10.8. The penetration contains not only the
primary sodium pipe but also its outer containment pipe and an annular
cooling gap. Cooling is required within the penetration to remove the nuclear
heat deposited in the steel through the pipe insulation so that the concrete
will not become overheated. Neutron streaming through the annular void
544
regions was minimized by placing donut shields around the primary sodium
pipe.
Shielding for the primary sodium overflow line, which is a normally
empty 15.2-cm pipe, was handled by passing the pipe through the shield at a
45 0 angle with the horizontal. This arrangement ensured that rhe gamma
radiation passing through the penetration would be scattered at least once
through an angle more than 90 0 , which reduces the intensity by more than
an order of magnitude.
(c) Operating Floor and Radial Biological shield. The biological shield
for the operating room, i.e., the floor (see Fig. 10.4), and the radial
biological shield surrounding the lower section of the containment building
are designed to allow continuous access during full-power operation. The
radial shield is ordinary concrete with a thickness of 2.1 m. The floor shield
has a total thickness of 1.5 m (concrete and steel) in all areas except under
the cask-car tracks, where the thickness is 2.1 m to support the 180-ton cask
car. Magnetite aggregate with a density of 5.1 g/cm 3 was added to the
concrete to maintain a minimum dry density of 2.4 g/cm 3 in the
operating-floor concrete. As a result of detailed studies of the possible
radiation paths, the steel plate on the underside of the floor varies in
thickness from 8.9 to 25 em. The final design thicknesses at each location
are shown in Fig. 10.9. Concrete with serpentine aggregate (density, 2.1
g/cm 3 ) was used for high-temperature areas around the floor plugs for the
primary heat exchange and the sodium pumps. The reduced density in the
areas where serpentine concrete was used required that the thickness of the
steel in the shield be correspondingly increased.
There were numerous penetrations of the operating floor for the reactor
nuclear instrumentation, primary sodium pumps, and heat exchangers and
for electric cables, gas lines, and other miscellaneous equipment. Each
penetration required a separate shield design, which was handled in a manner
similar to the previously described special shield designs.
(d) Intermediate-Heat-Exchanger and Sodium-Pump shield Plugs. The
outer shells of the primary sodium pump and intermediate heat exchangers
rise through the operating-floor shield. The shells are stepped within the
floor shield, and removable stepped shield plugs fit into the shells at the
floor level.
The equipment is designed so that there is a double annular gap around
each of the plugs, which allows clearance for expansion caused by heat and
permits removal for maintenance. A sketch of the shield design around a
SHIELD DESIGN
545
1.5-M FLOOR THiCKNESS

EVERYWHERE EXCEPT AREAS
INDICATED AS 2.1 M
SERPENTINE CONCRETE
ABOVE THESE AREAS
DECAY TANK
NO. 3
IHX
2.1-M THICKNESS
WITHIN THIS
PRIMARY SODIUM
PUMP
DOTTED
BOUNDARY
DECAY TANK
NO. 4
CASK-CAR
TRACK
OVERFLOW
PUMPS
DECAY TANK NO.2
PRIMARY SODIUM
PUMP
DECAY TANK NO.1
IHX
THICKNESS OF PLATE. eM
fZ125
1!1!114.6
rjlB.4
.,3.3
1II1B
fZI10.B
rj 17. 1
015.9
B.9
~CONCRETE
PLUG
Fig. 10.9-Plan view of operating floor showing steel requirements beneath the concrete.
(From Beaudry et al. 2 )
heat exchanger is shown in Fig. 10.10. Since the offsets in the floor
penetrations were fixed before a detailed analysis of the necessary shielding
was made, a dual-purpose shielding material was needed to reduce tv proper
levels both the neutron and gamma radiation levels within the gap" and
annuli around the plug shields. Boron steel (containing 1% boron) was
chosen for this purpose and was installed below the stepped annuli around
the intermediate heat exchanger. Plain carbon steel was used around the
sodium-pump floor penetratlons.
Because of high temperatures in this area, each of the three intermediate-heat-exchanger plugs and three pump plugs were filled with
dry-packed serpentine. A 15.2-cm layer of steel shot was added to each
pump plug for additional gamma-ray shielding.
546
OUTER SHELL
OF IHX
~ DRY-PACKED SERPENTINE ROCK
f:';':7:~:~ SERPENTINE CONCRETE

h:>.~~:::A ORDINARY PORTLAND CONCRETE
STEEL
Fig. to.10-Sketch of intermediate-heat-exchanger shield plug. (From Beaudry et al.
(e) Primary Sodium-Pipe shields. Special considerations were given to

the problem of neutron leakage from the 76-cm-diameter primary sodium
pipes. A calcium borate mineral with the approximate formula 2CaO'
3B 2 0, . 2H 2 0 in an asbestos matrix (trade name Sigma K) with a density
of 1.1 glcm' was chosen as shielding for these lines. This material contains
12 wt.% boron and withstands temperatures up to 1l00C in high-radiation
environments.
The problem of designing a proper pipe shield was complicated because
the pipes were surrounded by 0.9-m-diameter secondary containment pipes
for protection against a sodium leak. Other factors in the design of the
shields for these pipes were the type and design of insulation, heaters, and
construction gaps. Allowances for pipe movements due to thermal expansions had to be considered. The thickness of the calcium borate around the
three 76-cm pipe loops varied from 23 to 28 em. The design of these shields
is shown in Fig. 10.11. Four donut or ring shields of calcium borate were
SHIELD DESIGN
547
91-CM CONTAINMENT
CALCIUM BORATE
SHIELDING MATERIAL
HANGER
INSULATION
Fig. lO.II-Calcium borate shield for primary sodium pipe. (From Beaudry et al. 2 )
installed directly on the pipe and on its containment to prevent neutron

streaming through the annuli around the pipes.
(f) Shield Systems for the Auxiliary Facilities. The auxiliary facilities
requiring shielding are the fuel and repair building, the sodium service
building and its associated pipe tunnel, the inert- and waste-gas building, the
secondary sodium galleries, and rhe control building.
In these areas shielding against gamma rays only was required. Most of
the shields were of ordinary concrete, but steel shot, steel plate, sand, gravel,
and lead were used in special situations. Design criteria and construction
considerations for these auxiliary facilities parallel those of the primary
reactor shields and will not be described here.
548
10.2.4 shield Costs

The type of material was an important consideration for each shield, and
the cost was highly influential in the final selection.
The main shield materials were ordinary concrete, steel, stainless steel,
lead, and graphite. The more unusual or exotic materials used were
(1) magnetite and serpentine aggregates in concrete, by themselves, or in

mixes in a dry-packed condition; (2) steel shot; (3) boron steel; (4) boron
carbide; (5) borated graphite (boron carbide in graphite); (6) calcium borate;
and (7) uranium (depleted). Each material was selected on its merits for the
intended use. Earth, sand, and gravel were used to advantage.
The total cost of all shields ($5,581,300) was approximately 5.5% of the
entire cost of the project. Of this amount, $1,000,000, or 18%, was for
research and development. The remainder ($4,581,300) was for the
construction of the shields. The shield research and development cost was
slightly more than 3% of the research and development cost for the entire
plant.
Table 10.1 shows a breakdown of the shield costs for the Fermi plant.
10.2.5 Calculational Techniques
The calculation of neutron attenuation presented the greatest uncertain-
ties in the shield design and followed the availability of improved transport
techniques. The original attenuation calculations were done before 1960
with removal theory. The design of the secondary shield wall is an example
of the application of removal theory and shows the conservatism incorporated. When the thicknesses of concrete required to reduce the flux in the
equipment compartment to 10 4 neutrons cm- 2 sec- t were determined,
calculations along various paths (see Fig. 10.12) indicated concrete thicknesses ranging from 81 em at 0 0 (opposite the thin primary shield tank) to a
0
maximum of 100 em at 180 (opposite the transfer rotor). An average
thickness of 91 em was chosen for the final design except for the north wall
opposite the transfer rotor, which was set at 100 em. In the final design
stages, the AIM-6 diffusion-theory code' 4 was primarily used with some
checks being with the NIOBE' 5 numerical integration, the TDC' 6 (twodimensional Sn-P,), the DTK' 7 (one-dimensional Sn-P,), and the
CRAM'S (one- and two-dimensional diffusion theory) machine codes.
Monte Carlo or the removal-diffusion techniques, which are used in design
work today, were not used since the state of the art did not appear
sufficiently advanced at the time.
SHIELD DESIGN
549
Table 10.1-APPROXIMATE SHIELD COSTS, ENRICO FERMI

ATOMIC POWER PLANTt
Area
Reactor building
Concrete, poured
Concrete, prepacked
Steel plates and sleeves
Primary-shield-tank shell
Primary-shield-tank graphite
Reactor-vessel thermal
shield (stainless steel)
Offset-handling-mechanism
shielding
Rotating-plug shielding
Sodium-pump plug shields (3)
Intermediate-heat-exchanger
plug shields (3)
Calcium borate pipe shields (3)
Total reactor building
Auxiliary facilities
Concrete
Steel
Total auxiliary shields
Plant total
Cost
Item
Total material and erection

Research and development
Grand total shield costs
45,000+
113,200
294,000
110,000
2,200,000
137,500
26,200
235,000
234,000
395,000
280,000
----$4,070,800
398,500n
112,000
--$510,500
4,581,300
1,000,000**
$5,581,300
t From Beaudry et al. 2

+Includes forming and reinforcing steel.
About $1,750,000 of this amount is for the removal of subgrade
graphite and its replacement with high-quality material. This procedure
was very much more expensive than the original installation.
~ Includes prorated additional structural costs owing to the presence of
the shield.
* *Prorated from total research and development costs.
Generally the AIM-6 diffusion code was used with the I8-group
neutron-cross-section set of Mills l9 which was revised and updated.
Occasionally the newer CRAM diffusion code was used, in both one and two
dimensions, with a special 24-group set of cross sections. Little attempt was
made to investigate the effect of different cross-section sets or the
energy-group structure. It was felt that errors due to the calculational
methods or uncertainties inherent in the mock-up of the physical problem
550
would override errors due to the nuclear cross sections since diffusion theory
uses the approximation of isotropic scattering. It should be pointed out,

however, that, for deep-penetration calculations with the modern transport
codes that allow for anisotropic scattering, the result is very sensitive to the
nuclear cross section.
Most of the calculations were one-dimensional, but in later work some
two-dimensional CRAM calculations were made. It was not possible to
mock-up the entire primary shield in a single two-dimensional problem
because of the large size and asymmetry of the system. A simplified
computer model used in cylindrical geometry calculations is shown in
Fig. 10.13. The corresponding regional descriptions and material densities
are shown in Table 10.2. This model represents a section of the primary
shield from the core center upward to the rotating plug and outward to the
primary shield tank through the narrowest part of the graphite shield. The
safety rods are assumed to be in their raised positions above the subassembly
handling heads.
Neutron streaming through the various shield penetrations, such as pipes,
ducts, and offset plugs, was a potential hazard that required considerable
design study. In general, the streaming equations for the various geometries
were used, but they were modified with the Fisher 20 formula to include the
effects of scattering. These streaming equations for the various geometries
are tabulated by Hungerford. 4 Gamma-ray attenuation was calculated with
point kernels and ray-tracing techniques as described by Rockwell. 2 1
Some of these techniques were programmed for computer calculations
for more detailed analysis, such as determining the optimum steel and
concrete requirements of the operating floor at various locations, as shown
in Fig. 10.9.
For this study the below-floor primary system was divided into
radioactive sources of several different geometries. A computer program was
written which calculated the radiation levels at various positions on the
operating floor through the shield. The program calculated the contribution
from each source and summed overall sources at each point on the floor. The
points were located on a grid system of radial lines marked every 10
between 0 and 360 starting at north and circumferential lines that were
about 0.91 m apart. The results of the computer program were used to
determine the final balance between concrete thickness and steel thickness in
each broad area.
552
Ii
VERTICAL CORE
V
l
20
20
38
3 5.6
I
I
163.7
~6~
20.1
:......
17
reENTER LINE OF
TEMPORARY
INSTRUMENT
THIMBLE
fta:7
-
l19
15
17
34.3
4 5.6
16
12 11
3.4
9
5
3 8.8 2
13
73.6~
43.2
1
35.6
-.l
10.8
22
18
21
21
10
8
4
3
19.0
13
6
1
r-
6r
14 15
f-
ALL DIMENSIONS I N CENTIMETERS
NOTE:
REFER TO
TA8LE 10.2
HORIZONTAL
CORE
15.2
28
59
I I I
I20.8 I 26.715.215.2
73.6
-li
15.2
Fig. to.I3-Primary shield: Simplified layout of zones for computer problems. See
Table 10.2 for key to description and composition of reactor zones. (From Beaudry.3)
SHIELD DESIGN
553
Table to.I-REGIONAL DESCRImON AND CQMPOSmON OF REACI'OR ZONES FOR COMPlTI'ATIONSt

Number dena.tea fen itotope., 102 4 acom./em'
ZoM
ZoneRame
numMf
,
1
3
4
Cn~
Control- and safety-rod region

End gap
~wblanlet
S.fuy~
6
7
Radial blanke~
,,
,.
11
"13
14
15
16
17
18
19
20
""
region of blanket
AllW gap
Handling hud.
Safety rod. and hold-down
Hold-down (outft" region)

Hold-down (inner region)
Safety roth (poison lection)
Steel ihicld bt."

Steel thermal. .Meld
nlU
Mo
7z
F.
0.00360
0.00215
0.009632
0.00616-4
0.0035
0.00223
0.00196
0.00125
0.00171
0.OO39
0.OO25
0.01101
0.00705
0.0008
0.00051
0.0148
0.0141
0.017
0.014
0.0194
0.0106
0.0136
0.0179
0.0139
0.0148
0.OO66
0.019
0.0013
0.0158
0.017
0.017
0.018
0.00043
0.0092
0.0189
0.0189
0.0172
0.0223
0.023
0.0219
0.0711
0.0648
0.0172
0.0161
0.0034
0.00584
0.0236
0.0632
0.0020
0.0085
0.0098
0.085
0.0063
0.023
0.0198
0.0009
U5U
Sodium plenum
Steel thermal.Meld
HoId.4own column
Safety rod. (poison tection)
Safety roo. (poison I"don)
Steel of rotatina plUS
Borated graphite
Plain I'aphite
N.
Natural
Boron
0.0012
0.0104
0.0104
0.0104
0.0104
0.0045
0.0769
0.0803
tFrom Beaudry.]
10.2.6 Comparison of Measurements and Calculations

The shield test program for the Enrico Fermi reactor plant was unusually
complete and detailed because of the prototype nature of the reactor. The
test program principally covered regions inside the reactor building, and the
major emphasis was on neutron measurements, although the initial emphasis
was in measuring the induced 24 Na activity because of its importance in the
overall plant design.
Neutron measurements were made throughout the reactor vessel and the
primary shield tank. Threshold detectors and 2 3 S U and 2 38 U fIssion foils
were used in the core and axial and radial blankets. Neutron fluxes were
measured in the upper sodium pool up to the bottom of the rotating plug.
Vertical 23 S U fission distributions were obtained at the side of the pool in a
surveillance tube near the reactor-vessel wall. A temporary instrument
thimble located near the center of the pool was used to obtain vertical
traverses (234 U , 235 U , 236 U , 238 U , 239 Pu , and 240 Pu) with both foils and
absolute fission counters. In the primary graphite shield, both 235 U foils and
various neutron detectors were used. Vertical distributions were obtained
from the flex-leg regions (LRV) up to the machinery deck (in the
surveillance tubes) or up to the operating floor (in the neutron-counter
554
tubes). The neutron-flux levels in the reactor compartment (between the

primary shield tank and the secondary shield wall) were measured at 33
locations with 23 s U foils and at one azimuthal position with a fission
counter and a spherical moderating dosimeter.
At various times after criticality, radiation surveys were made throughout the entire above-floor area of the reactor building, including the rotating
plug and machinery dome, and throughout the various service buildings.
The gamma-ray surveys were conducted with area survey monitors. The
neutron surveys used portable proton-recoil counters, BF 3 counters, and a
set of multisize spherical polyethylene dosimeters. In general, the detecting
instruments were capable of measuring radiation levels above 0.01 mR/hr for
gamma rays and 0.02 to 0.05 mrem/hr for neutrons.
Although a large number of measurements were made and compared
with calculations, we will summarize them with a very general comparison in
the following section and with a few detailed comparisons subsequently.
(a) Overall Comparisons. No shield regions were found to be underdesigned, but some areas were found to be significantly overdesigned, which
was to be expected. A summary of the principal test results is shown in
Table 10.3. The original design calculations for the 22 Na and 41 A activities
T.bl.l0.3-SUMMARY OF PRINCIPAL RESULTS OF THE ENRICO FERMI

SHIELD TEST PROGRAMl
Sodium-24 activity
Sodium-22 activity
Argon-41 activity (STP)
Secondary 24 Na activity
Radiation streaming
Component activation in
reactor vessel
Uraniurn-235 fission rates:
Sodium pool
Transfer rotor
Radial blanket and reflector

Neutron-counter tubes
Foil-sample tubes
Reactor compartment
Equipment compartment
tFrom Beaudry.3
Measured,
,..Ci cm- 3 Mw(tfl
Original design
calculation,
~Ci em -3 Mw( t 1
Most recent
calculation,
IlCi em-3 Mw(tft
67
0.004
0.0017
<3 X10"
100
0.034
0.3
4 X 10-'
76
0.006
0.0014
No radiation measured above reactor-building floor at 110 Mw(t}.

Measured and calculated dose rates agree within a factor of 2.
Measured values
Measured values
Measured values
Measured values
Measured values
Measured values
Measured values
within 50% of calculated values.

are less than 1/10 of calculated values.
higher than calculated by 50% (in reflector).
within a factor of 3 of calculated values.
within a factor of 2 of calculated values.
less than original calculations.
less than original calculations.
555
SHIELD DESIGN
were extremely conservative, but later more careful calculations with better
data produced good agreement. In general, with the complexities involved,
the overall agreement can be considered good. In spite of concern about
radiation streaming, nothing significant was detected on the operating floor.
The principal reason for overestimating streaming was undoubtedly due to
the original use of estimates of the radiation levels underneath the rotating
plug which were too high. In the early calculations of neutron streaming, a
neutron-flux value of 2 X 10' 3 neutrons cm-2 sec-' under the plug was used,
which is equivalent to 4 X 10' 0 neutrons cm-2 sec-I Mw-'. The best present
estimate of the neutron flux at the base of the plug is about 4 X 10 7
neutrons cm- 2 sec- 1 Mw- 1 , a factor of 1000 lower. Secondary reasons were:
(1) the use of conservative streaming formulas and (2) the assumption of
gaps larger than those which actually existed.
(b) Measurements in the Temporary Instrument Thimble. Vertical flux
238 U, 239 PU, and 235 U fission counters were made through
the temporary instrument thimble (see Fig. 10.14). This instrument thimble,
which extended down the vertical axis of the reactor core, replaced the
traverses with
HOLD-DOWN AND SAFETY RODS\

FUEL-HANDLING MECHANISM ~
OPERATING
FLOOR
FUEL
EXIT TUBE
REACTOR
VESSEL
HOLD-DOWN
MECHANISM
TEMPORARY
INSTRUMENT
THIMBLE
ONCRETESHIELD WALL
Fig.lO.14-Elevation of primary shield system showing detector locations. (From

Beaudry.3 )
556

10',,'..,.,,-,--.,--,..--...,--,.--,--r--,.---,.--,.--,----,
(\~
'C(
'"
10
\00
"
'.... , 0
u
~
10"
ui
t-
v;
10"
~
~
t-
=>
--'
'"
,, 0
',0 0
.........
"
"
0
,
0
............ 000
a:
z
o Measured. absolute fission counter

- - Calculated, CRAM: 2-D
\0
10
10'
... ...
o
o
o
o
o
<Xl
, 06~_;~___;t;;-...I...;-!;;;---';'*"-;*:::-;;:t;_---;;;;,;-'*",~,*;-_f,;;;o 40 80 120 '60 200 240 280

400 440 .....
VERTICAL DISTANCE ABOVE CORE MIDPLANE, em
Fig. 10.15-Plutonium~239 fission distribution in the temporary instrument thimble.

(From Beaudry.')
central safety rod. The thimble, which was 5 em in diameter, was sealed
against sodium at the bottom and open to the containment building at the
top. Comparisons between the measurements and two-dimensional CRAM
calculations for the 23 9 Pu and 23 5 U counters are shown in Figs. 10.15 and
10.16, respectively. All measurements and calculations were normalized to
the same value at the core center. In the CRAM calculations the 24-group set
previously mentioned was contracted to 9 neutron groups. The sharp dip in
data at about the 400-cm distance is caused by a piece of steel that
completely surrounds the upper part of the thimble. The agreement is
generally good, with 2' s U measurements being in better agreement with the
calculations than the 2'9 Pu data. This probably results because the 2'9 Pu
cross sections in the thermal and near-thermal ranges cannot be as well
represented as 2'5 U with the small number of neutron groups that were
used. It is probable that, if the calculations were carried beyond the 280-cm
mark, the agreement would become poor since diffusion theory would
become an inadequate model at deeper penetrations, particularly around a
heavy absorber such as steel.
SHIELD DESIGN
557
o Measured, absolute fission counter

o Measured, 235U pins
- - Calculated, CRAM: 2-D
'~
~
10"
ui
!;:
IX'
o 10"
uw
I::l
-'
~ 107
a>
<{
106'~_~~~"""';:,:;;-,-:~---,=---,~_~,-;;:!;;,-=;-~;-*,,,......_-l
40
80
120
160
200
240
280
320
360
400
440
Fig. 10.16-Uranium-235 fission rate in the temporary instrument thimble. (From

Beaudry.')
A large number of one-dimensional calculations were made with the

AIM-6 and CRAM codes, and a number of cross-section sets were used.
Figure 10.17 shows the results of some of these one-dimensional calculations
in comparison with the measurements (curve B) and the two-dimensional
CRAM results (curve A), the measurements and CRAM results being the
normalized values from Fig. 10.16. In the 18-group AIM-6 calculation
(curve E), the handling head and hold-down regions were considered to be
described simply by a 20-cm-thick section of steel and sodium, followed
directly by a large thickness of pure sodium. This hold-down region is much
thinner than the actual hold-down region, and no attempt was made to
include the safety rods. This is the type of problem mock-up that had
originally been used in the majority of the shielding problems run for the
shielding design of the upper vessel regions (e.g., the rotating plug, shot
tanks, and upper reactor vessel). Figure 10.17 shows that the use of this
simplified type of geometry is quite conservative. The lack of agreement
between this calculation and the measurements appears to be caused more
by an inadequate geometrical mock-up than by the use of spherical
558
100.000""";:---,----,----,------,----,----,
AIM l8-group spherical
calculation: simplified
hold-down region
NIOBE SO-group spherical
I
~',:!f~~::_._+~,,_I__calculiltion: simplified
hold-down region
lO,Doot-
~
'"z
o
iji
::J
'"
Two-dimensional
CRAM calculation
I!l
N
calculation: complex
hold-down region
(no boron rods)
1.000
u:
~
AIM l8-group spherical
100
19
9'OUP~ 1
+-----"""L-"-.-+--J
\,-+
AIM la-group spherical
calculation: complex
hold-down region
(with boron rods)
f-----+----+---+---
80
120
160
200
240
280
~-----1
320
360
400
440
Fig.lO.17-Uranium-235 fission-rate traverse in the temporary instrument thimble.
(From Beaudry')
geometry. This is more evident from curves C and D of Fig. 10.17, which
were obtained from spherical 18-group AIM calculations. In each of these
and in the two-dimensional CRAM problem (curve A), however, the same
detailed hold-down configuration was used. The geometric model used for
calculating curve D included the safety rods with their full boron content,
whereas the geometric mock-up for curve C omitted the boron altogether.
The strong poisoning effect of these safety rods accounts for the much lower
values of curve D, and this is apparent for the entire distance from the axial
blankets out to the rotating plug. Although the fission-rate data shown in
curve C are most different from the measured data in the region of the safety
rods, they match the measured response (curve B) farther out in the sodium
pool (all the way to the rotating plug).
A more complete study of the differences obtained from various
cross-sectional sets also involved such items as mesh spacing, fluxconvergence criteria, group-collapsing schemes, and possible differences
between the AIM-6 and the CRAM codes. Spherical geometry was used. As
tentatively found by the study, there were essentially no differences between
SHIELD DESIGN
559
the results from the AIM-6 and the CRAM codes per se. However, the use of
the various cross-sectional sets did result in great differences in the 2 3 s U
fission response. Again these differences first became pronounced above the
axial blanket, where the twin effects of spectrum softening and the presence
of large amounts of steel are encountered.
All these problems have clearly demonstrated that a one-dimensional
problem mock-up can be quite incorrect in those regions of the hold-down
where there are abrupt changes in absorbing materials. The general shielding
calculations (as exemplified by curve E of Fig. 10.17) used in the basic shield
design appear to be on the conservative side even without considering any
additional correction factors. Nate, however, that this entire discussion has
been based primarily on the 235 U fission rate, which is mainly an index of
thermal and near-thermal fluxes.
(c) Reactor-Compartment Measurements. The neutron flux in the reactor compartment next to the secondary shield wall is the source term in the
design of that shield. A number of measurements with a fission counter and
a spherical dosimeter were made directly underneath manhole No.2 (see
Fig. 10.5). All the measurements by the two different instruments were in
close agreement, but no detailed comparisons with calculations were made.
In a number of regions of the reactor compartment, neutron-flux
measurements were made with 23 s U foils. When the measured foil counts
were converted to fluxes on the inside of the secondary shield wall with an
adjusted AIM-6 spectra, all were within the maximum design value of 10 8
neutrons cm-2 sec-I.
10.3 FAST BREEDER: DOUNREAY FAST REACTOR

This British power plant is contemporary with the Enrico Fermi plant,
and the two plants share a number of design features. The Dounreay Fast
Reactor (DFR) is a sodium-potassium-cooled fast breeder system; the
reactor shield contains mostly graphite (pure and borated) and steel. The
rotating top shield is designed to provide access channels for charging the
core and breeder regions as well as eighteen 5-cm penetrations for control
purposes. Figures 10.18 and 10.19 show the general arrangement and
materials details of the rotating top shield. The removable top shield
(Fig. 10.18) was not in place for the experiments we will describe. A
material called Resilon, which is a hydrogenous resin-bonded glass fiber, is
used in the top shield.
560
REMOVABLE CONCRETE
BLOCKS
REMOVABLE TOP SHIELD

MILD STeEL AND JABROC
CORE CHARGE PLUG
CONCRETE-VAULT
ROOF
ROTATING SHIELDS
LAYERS OF STeEL AND RESILQN
ALUMINUM-FOIL INSULATION
MILD STeEL
BORATED GRAPHITE
VESSEL.SUPl'ORT-tt:;:~~
ROLLER
~~~H--VESSEL
t
BORATED GRAPHITE
MILD STeEL
'i OF
CORE
STAINLESS-STEEL PUNCHINGS
PUREGRAPHITE ZONE
FOR ION CHAMBERS
BORATED GRAPHITE
VESSEL SUPl'ORTS
(NITROGEN COOLED)
Fig. to.IS-General arrangement of the DFR shield. (From Adamson et al.
22
Adamson, Judd, and McNair22 report a study of the neutron flux in the
top shield of DFR to test the validity of removal-diffusion theory in
calculating the bulk shield attenuation for a fast reacror shield. Although the
method had been applied successfully to large thermal reactors where the
principal attenuating medium is concrete, it had not been applied to a
relatively small fast reactor where the medium is principally borated
graphite. We will describe some of their results for comparison with the
SHIELD DESIGN
561
CORE-CHARGE-HOLE LINER
(WHICH TAKES EXPERIMENTAL
SHIELD PLUG)
CONTROL-ROD
FOLLOWER
INNER ROTATING SHIELD
LABYRINTH SEAL
ROLLERS
SUPPORTING
INNER ROTATING
SHIELD
RESILON AND STEEL
~~~~
LAMINAT'ONS,~~~~~~~ii
ALUMINUMFOIL
INSULATION
-.:'
'
..
ILL4-M ~
"
PLUG
LINNEA ROTATING SHIELD
' - - - - - OUTER ROTATING SHIELD

~ STAINLESS
~STEEL
~MILD
~STEEL
BORATED
[J
-. . ' GRAPHITE
0.5
M
Fig. lO.19-Details of the rotating shields. (From Adamson et aL 22)
562
corresponding work shown in the preceding sections on the Enrico Fermi
plant.
10.3.1 Calculational Model for Bulk Shield
The models of the core, blanket, and top shield used for the calculation
are shown in Fig. 10.20. The Spinney removal-diffusion method and
CENTIMETERS
ABOVE CORE
CENTER PLANE
VOID
373.0------ - - - - - - - STEEL AND RESILON
6
337.0------ - - - - - - - 5
MILD STEEL
290.8
SHIELD AXIS'-...I4'
4 1
BORATED GRAPHITE
1 :-STAINLESS STEEL
153.5=====~L~=======
146.9
I
3
127.0
107.0
COOLANT
10
BLANKET FUEL-
ELEMENT TOPS
RADIAL BLANKET
CORE FUEL-
61.2
40.9
27.0
fJ84=:;;j----t-AXIAL
7
4----,
ELEMENT TOPS
lCORE CENTER PLANE
BLANKET
CORE
,
28.5
,
68.5 CENTIMETERS FROM AXIS
Fig. lO.20-Model of reactor and shield used for RASH D calculations. (From Adamson
et al.")
variations, including RASH E, are described in Sec. 4.9. At the time of this
study (1964), the previous version, RASH D, was in use. The removal source
is in effect the density of first collisions of neutrons born in fission. A
discrete-ordinate calculation, TDC,' 6 was used to define the boundary
condition flux densities at the core-shield interface. The shield comprised
regions 1 through 6 in Fig. 10.20. The one-dimensional RASH D calculation
was done in infinite-slab geometry.
SHIELD DESIGN
563
10.3.2 Measurements
An experimental shield plug was designed to fit within close tolerance
into the core charge hole of the top shield (Fig. 10.19) so that foils and
fission chambers could. be mounted at appropriate positions in the shield
plug and positioned in the core charge hole within the shield. Close-fitting
blocks of shield material were assembled within the plug to match adjacent
materials in the surrounding shield. Detectors are summarized in Table lOA.
Table lOA-DETAILS OF MEASURED REACTIONSt
Reaction
27 Al(n,O!)24 Na
32 S (n,p)32p
232
Th(n,f)
237 Np(n,f)
55
Mn(n, 1/ 6 Mn
197 Au(n, ')')198
Au
Energy of neutron
flux involved
Means of
measurement
Fast flux, threshold

about 9 MeV
Activation of aluminum metal

Activation of elemental sulfur
Thorium fission
about 3 MeV
about 1.5 Me V
about 0.5 MeV
Thermal and epithermal flux
4.geV
counter
Neptunium fission
counter
Activation of manganese foils, unprotected and in
cadmium capsules
Gold sandwich foils
t From Adamson et al. 2 2
Results for the 32 S measurements are given in Fig. 10.21, and for the
ss Mn, in Fig. 10.22. We have picked these two since the 32 S results are
typical of the fast-neutron fluxes and the s 5 Mn of the thermal and
epithermal fluxes. The 32 S results (as well as 27 AI, not shown) compare very
well with the removal-diffusion approximation. Results of some earlier
measurements reported by Smith 23 are also shown in Fig. 10.22, and they
match the present data rather well.
Agreement between the low-energy measurements with the removaldiffusion calculations is poor, l'articularly in the upper regions of the shield
and in air above the shield. This indicates an appreciable flux of neutrons
above the top of the shield which have not diffused through the shield but
which streamed through the various ducts formed by control penetrations
and are reflected back from above.
564
Experimental
X Experimental, results quoted from Ref. 23
RASH 0 calculation
_.- Calculated contribution from streaming
107
106
::;;
I-
<l:
106
a:
:>
u.
oJ
:>
10
CIl
oJ
<l:
a:
.I,-, \
103
:>
~
Z
102
u.
0
"
.I
10
"
'\.
\
CIl
..
a:
w
CIl
10. 1
;::
U
<l:
w
10- 2
a:
10-3
Core top plate
Coolant
10"
0.3% borated graphite
Stainless steel
10- 5
50
100
150
200
250
300
350
DISTANCE ABOVE CORE CENTER PLANE, em
Fig. lO.21-Rate of 32 S(n,p)32 P reaction. (From Adamson et al. 22 )
400
565
SHIELD DESIGN
--,~
10"
"'q,:
10"
,
0
10'
10"
I-
"
lOS
"o
103
"'w
:l
la"
:"l
u.
"
10'
10
o
Experimental
Experimental, detailed scan in resilon and
steel region
- - Experimental, results quoted from Ref. 23
RASH D calculation
- - RASH 0 calculation with streaming contribution
~.-
Core top plate
50
100
Stainless steel
150
200
250
300
350
500
DISTANCE ABOVE CORE CENTER PLANE, em
Fig. lO.22-Rate of 5 5 Mn(n,'Y)S6 Mn reaction. (From Adamson et al. 2 2)
10.3.3 Effect of Streaming

Adamson et aL 2 2 attempted to estimate the streaming through the 18
control ducts and the breeder charge hole (Fig. 10.19). Each duct was
represented as a sink (negative source) at the end of the duct and a line of
point sources along the duct axis. Age theory was used to obtain the
distribution from each source in the surrounding medium. Since this is
566
applicable only in infinite homogeneous media, it was not intended to give a

reliable estimate at interfaces between different shield regions and at the top
of the duct. Thus the streaming estimate cannot be expected to give the
characteristic hump at the graphite-steel or steel-Resilon interfaces.
Nevertheless, the streaming estimate when added to the RASH D bulk shield
result does agree well with the 5 5 Mn measurements (and those for the gold
resonance at 4.9 eV, not shown). In fact, Adamson et al. insist that the
agreement for 55 Mn is fortuitous because of the imprecise nature of the
streaming estimate. It does confirm, however, that streaming through the
ducts vitiates the otherwise acceptable bulk shielding properties of the top
shield. Had the attenuation of the top shield been of the same order as that
of the streaming neutrons, it would have been no less effective and perhaps
more economical.
10.4 HEAVY WATER, NATURAL URANIUM: AGESTA
The Agesta reactor is the first power-reactor prototype to be built in

Sweden, and it has some extremely interesting features. It is also the first
pressurized heavy-water natural-uranium reactor in the world. Most of the
power produced is used for heating in a suburb of Stockholm. This reactor type is particularly important in countries not possessing uraniumenrichment facilities since it uses natural uranium. The demand for D 2 0 is
expected to grow rapidly because of the desirability of this system in many
countries. Comparisons of measured and predicted radiation levels at Agesta
are reported by Aalto, Sandlin, and Krell. 2 4
10.4.1 Description of Reactor and Shield
The reactor with cooling system, service facilities, and auxiliary
equipment is installed in a series of underground caverns excavated in a rock
outcrop, as shown in Fig. 10.23. The rock medium provides a great deal of
natural shielding and containment. The design philosophy that led to the
layout shown was to provide for full access to as much of the plant as
possible even during full power. This led to the compartmented arrangement
where systems are separated into functional units isolated by shielding from
nearby units and accessible by shielded passageways. In such an excavated
plaht, there is an obvious premium on space; thus magnetite concrete was
used in the main shields to conserve space.
567
SHIELD DESIGN
ROOM
FUEL~ASSEMBl Y AND
CONTROLROD STORAGE
=
=
REACTOR
VESSEL
SWITCHGEAR
DO
TRANSPORT
AIR LOCK
000
TURBINE HALL
<I(lt>
}
~------~OF
ROCK OUTCROP
Fig. lO.23-Agesta plant containment layout showing the relative positions of the major
components in the reactor part of the plant. (From Aalto et al. 24)
A cross section of the reactor and the primary shields is shown in

Fig. 10.24. The pressure vessel has an inside diameter of 4.55 m and a height
of 4.68 m. The core has a diameter of about 3.6 m and a height of 3.0 m.
The reflector layer of D 2 0 surrounding the core is about 30 em thick.
Beyond the reflector are two steel thermal shields (5 and 10 em) separated
by cooling spaces. The steel pressure vessel has a wall thickness of 7 em.
The radial shield consists of a 2.2-m layer of magnetite concrete.
10.4.2 Calculational Model
Aalto et al. 24 mention three sets of calculations: a first set of simple
hand calculations in 1958 with which the shield dimensions were set; a
second set using the British RASH removal-diffusion method several years
later; and a set in 1955 based on the NRN method of Hjiirne and
Leimdorfer. 2 s The latter is also a variation of the Spinney method and is
described in Sec. 4.9.2.
568
,-----1"------:
I
IL------I
CONTOUR OF
I
I
CONTROL-ROD
'
!-MECHANISMS lair + Fel-!
""
!I
.
I
II
LID BOX
(H 20 + FeJ
1-----~T~!.~f.T.Q.!!.ll
I
I
"I
I
I ., .
I
~ro":~~:",
SHIELDS
t !?"'~~r
I .
LTHERMAL
1,------,--,-;
- -,
~
it
r=='
,Fe + 020
i,
----- ------- '9''f/'
- .'
'I
--- -'-i
<II
,f
E
,9
~'
-ri
.'-.
330m
",
A.
Fig.lO.24-A simplified vertical cross section of the Agesta reactor and the primary
shields. (From Aalto etal. 24 )
From the geometry of the shield shown in Fig. 10.24 and our recent
review of the Dounreay analysis, it will become apparent that attenuation in
the radial direction is governed entirely by massive unbroken layers, and the
removal-diffusion method should be adequate. 1n the axial direction,
however, the large annular duct surrounding the pressure vessel should be
569
SHIELD DESIGN
Rows 1 + 2+ 3
M
1.00
Rows 1 +2
Row 1
a:
'oa:"
~
;;
:!<
-:: 0.50
CENTIMETERS
Fig. lO.25-Ratio of the radial thermal fluxes from individual fuel channels. (From
Aalto et at. 24 )
important in determining the radiation levels above the reactor in the region
of the control-rod drives.
Radial NRN predictions were made for the ftrst, second, and third rows
of fuel elements nearest the shield to illustrate the diminishing influence of
elements deeper in the core. These are shown for the thermal flux densities
in Fig. 10.25 in terms of the ratio of the thermal flux density contributed by
each row to the total at a distant point in the radial shield. The relative
positions of the elements and their power contribution are shown in
Fig. 10.26. The removal source from the third row of elements adds only 4%
to the flux density and the second row contributes only 15%, even though
rows 3 and 2 account for about 5X and 2X, respectively, the power in the
first row. These results show clearly that the true power-density profile,
particularly the density at the outermost regions of the core, is extremely
important in obtaining an accurate calculation of the radiation penetrating
570

ROW 3
P '" 6.71 Mw
,0
0.505
,0
0.596
ROW 2
p '" 2.6 Mw
,0 ,0
1.055
x
0.538
Row 1
P=l.39Mw
10 0.493
10 0.414
10
1.'07
-z
,0 ,0 ,0
1.03'
,0
0.956
0.544
0.421
,m
0.382
0.518
x""
z '" 0
1[]-I-I
27 em
RELATIVE POWER IN z-DIRECTION
10 l[I)-1[]- - -
RELATIVE POWER IN x-DIRECTION
~804
0.434
0.340
~27 en: I
10
'L.!.J
0.382
0.705
,0
0.411
,0
0.343
Fig. lO.26-Core configurations used when calculating the removal source in the radial
NRN prediction. (From Aalto et al. 24 )
the shield. This is well known as an important source of error in shielding

calculations.
The actual power distribution in this case did deviate from the one on
which the shield predictions were based, and it affected comparisons with
attenuation measurements accordingly.
10.4.3 Measurements
This study was performed as a shield survey during commissioning of the
plant. No special provisions had been made for locating detectors within the
fixed shield; thus measureme nts were made only on available surfaces. This
limitation accounts for the paucity of intermediate values in the radiation
levels to be discussed.
SHIELD DESIGN
571
The range of magnitude of fluxes to be measured is worth noting: 10' o.

Since reactor power could be varied from 10 kw to 50 Mw, a range of
5 X 10 3 , the detector sensitivity range had to be sufficient to cover the
remainder, about 10 6 The limitation to available surfaces meant that the
measurements were made mainly at the extremes of this range.
For high neutron-flux densities, thermal and epithermal measurements
were made by double-foil (gold-copper) and cadmium-ratio (manganese)
methods. Sulfur, with its 3-MeV threshold, was used for fast-neutron-flux
measurements. The low fluxes outside the primary shields were measured
with boron counters.
Gamma-ray measurements were made with glass dosimeters, ion chambers, and Geiger-Mueller detectors.
Calibration foils on the outer face of the tank at the core midline
(along line R in Fig. 10.24) served as power-level monitors. They also served as
an automatic correction for the variation of neutron attenuation in the radial
direction because of varying moderator temperature. Neutrons degrade to
the temperature of the moderator, and a change from 28C to 220C alters
the observed "thermal" flux by a factor of 3.
Power distribution varied 20% ftom the mean in the radial direction
and was about 30% lower at the outer radius of the core than had been
predicted. The axial distribution was asymmetrical; typically of reactors with
top-mounted control rods, the maximum in the flux occurs well below the
midplane of the core. The power level at the top of the core was 20% lower
than predicted. These deviations translate directly to the measurements.
One additional source of error is the detailed structure near the detector.
In this instance some thermal insulation and steel conduit used to support
the detector were not included in the predictions, but they may have
affected the thermal flux measurement by a factor of 2. The predicted radial
thermal flux densities are shown in Fig. 10.27. Only one measurement is
shown (at the edge of the tank). The intensity beyond the concrete was
below the minimum observable. The epithermal fluxes are similar and are
not shown.
The axial variation of the thermal flux density is shown in Fig. 10.28 tor
two vertical lines: AI and A 2 in Fig. 10.24. The calculations (solid and
dashed curves) apply to AI and do not include an estimate of the streaming
around the sides of the tank; this is shown separately, and obviously it
controls the intensities above the tank in the same manner as was shown for
Dounreay. The dotted lines in Fig. 10.28 connect the measurements.
572

1014 ,
,--_ _- ,
...,-_---..,,-,
\
... ,
' \ - RASH B, P cone. :::
3.5(g/cm 31
\
\
NRN,p cone. = 3.8(g/cm 3 j
_ _, , 1 05
\,\
,,
\
"
Magnetite concrete
150
200
Minimum
\\11" '"'
r
..
250
300
Z, em
Fig.l0.27-Predicted and observed radial thermal fluxes (line R, Fig. 10.24). (From
Aalto et al.")
Aalto et al. 2 4 do not describe their method for calculating the streaming
(which agrees very well with the observed intensities). However, their
mention of multiple reflections would indicate an albedo calculation
possibly coupled with Monte Carlo (see Sec. 7.10.2).
Gamma-ray dose rates are summarized in Table 10.5, and sulfur
fast-neutron reaction rates (proportional to flux densities), in Table 10.6.
The gamma-ray dose rate outside the pressure vessel is two to three times the
predicted value, probably because of activation of the nearby steel from
previous high-power runs. Since the shield design is very conservative, the
only critical neutron intensities are those in the vicinity of the control-rod
drives because these levels control servicing procedures. However, the
original streaming estimate was in excellent agreement with the measurements; thus no additional problems developed.
Aalto et al.' 4 conclude that, although local flux predictions often differ
from observations by a factor of 2, these are easily accounted by the error
sources mentioned. They also point out that the complexities of a power
574

Table 10.5-AGESTA GAMMARAY DOSE RATESt
Measuring line, def.
inition of point
Coordinate,
em
Predicted,
mremjhr
Measured,
mremjhr
)*
7.5 X 10'
Axial, upward (A,)

Control-rod mechanisms
Below the concrete
lid
Hall loo,
Radial (R)
Outside pressure
vessel
Outside biological
shield
Axial, downward
Not shown in figures
200 to 570
(-10
610 to 760
(-10)
1.5 to 3.0
810
<J:J.7
0.2
54
3 X 10 7
8 X 10 7
290
0.Q4
<0.1
Level +7.1
center line
2 m above
1.1 X 10'
2.3 X 10'
plane +7.1
tFrom Aalto et al. 24
+Not comparable to the measured value because in the latter
streaming effects dominate.
plant coupled with the limitations on measurements make such a survey an

inferior means for studying the accuracy of shielding prediction methods. In
such a conservatively designed plant, only a gross design error resulting in
excessively high fluxes would draw any attention. Thus the Agesta survey
was a successful design verification that confirmed the original analysis in a
highly satisfactory way.
10.5 BOILING WATER: PATHFINDER

Designed as a demonstration boiling-water-reactor plant, the Pathfinder
plant was built near Sioux Falls. S. D., to supply 66 Mw(e) to the area power
network. Start-up was in March 1964, and the reactor was shut down in
September 1967 for repair of a condenser-tube break. Bjeldanes 26 reports
that subsequent inspections in the reactor vessel disclosed failures in the
steam separators and a baffle in the fuel hold-down structure. After various
considerations of required repairs, plant size, and operating costs, the reactor
was decommissioned, and the plant was later converted to a fossile-fuel
system.
SHIELD DESIGN
575
Table 10.6-AGESTA FAST-NEUTRON FLUXESt
Measuring line, def

inition of point
Coordinate,
em
Predicted by
NRN
Measured
Axial, upward
(A , andA 2 )
Bottom of the lid
box
60
1.45 X 10"
(1.4 1.3) X 10"
610 to 760
610 to 760
(1 X 1O-5 ):j:
(1 X 10"):1:
5 X10-2
(5 to 13) X 10-2
44
2.85 X 10"
(3.1 1.3) X 10"
54
7.27 X 10 5
(1.95 0.10) X 105
290
3.08 X 10"
<5 X 10-3
Below the concrete
lid:
A,
A2
Radial (R)
Thermal shield
Outside the pressure
vessel
Outside the biologi-
cal shield
Axial, downward
Below pressure vessel, center line
(3.1 0.2) XI0 2

24
tFrom Aalto et al.

Fast-neutron fluxes are given as S(n,p) reaction rates. To
obtain > 1 MeV for fission spectrum, multiply by (9 X 102 ).
~Not comparable to the measured value in which streaming effects dominate.
The following description of the design is taken from a shield-analysis

report by Holl and Stephen? 7 The plant consists of four connected
buildings: the reactor building, the turbine and control building, the
fuel-handling and waste-disposal building, and the water-treatment building.
A plan view of the operating floor is shown in Fig. 10.29. This plant layout
evidently resulted from many iterations and reflects a good deal of planning
for management of radiation problems. Each building may be characterized
by its principal radiation source. The water-treatment building was to
contain unused water and have no radiation sources. The reactor and
fuel-handling buildings are controlled by active and spent fuel; however, the
fuel-handling building is also affected by activated corrosion products (called
crud by the authors) in the used elements and equipment. Radiation levels in
the turbine building were to be controlled by gaseous activation products
from (n,p) reactions in oxygen of the water; levels were underpredicted here,
576
and this created the only unforeseen radiation problem in the plant. We will
return to this problem in the discussion of the survey measurements.
Figure 10.30 is an elevation of the plant showing the containment shell
of the reactor building with the adjacent structures. The core is located
below grade to take advantage of earth Hll as shielding. The primary
HYDROGEN
HOUSE
TURBINE BUILDING
CONTROL
ROOM
FUELHANDLING
BRIDGE '-J=~"/111!1
WATER-TREATMENT
BUILDING
REACTOR
BUILDING
Fig. lO.29-Pathfinder operating floor. (From Holl and Stephen.' 7)
biological shield is ordinary concrete about 3 m thick. A clearance space of

about 1 ft between the reactor vessel and the shield provides a water-cooling
space. Penetrations for reactor controls and instruments are shadowed by
91-cm-thick concrete plugs. Above the reactor a 5.5-m pool of water
attenuates radiation to the operating floor.
The turbine is recessed into the operating floor. Original plans called for
restricting access (limiting working times) near the turbine. The control
room and administrative offices are adjacent to the turbine room.
578
10.5.1 Calculations
As with other systems described, the initial shield analysis for Pathfinder
was done by hand from attenuation curves. The detailed calculations were
done with fast-neutron removal theory and gamma-ray linear-absorption
buildup factors. A point-kernel machine program was used that could
adequately handle the large number of regions, materials, and compositions
encountered in the design. Duct scattering was done with the SimonClifford equations (Sec. 7.10.2).
The list of radiation sources that must be considered in a boiling-water
system is impressive. Gamma-ray sources in the core include those from
prompt fission; delayed fission; capture

35 U, 238 U, zirconium in
fuel-element cladding, and hydrogen); and inelastic scattering. Decay gamma
rays from the shutdown core must be considered as a function of core
history and time after shutdown for maintenance planning.
Gamma rays from other materials include captures in iron and various
activation products from the crud (16 N, 17 N, and 19o), and potential
activities from
17 N
neutrons.
The crud activity is a result of the high water velocity, which erodes
activated species from the core and carries them throughout the system. The
activity consists of 59pe, 60Co, 56Mn, Ster, 9SZr, and 6sNi. Various
assumptions were made for the buildup and plate-out of the crud in
calculating the total gross activity.
Two (n,p) reactions and related captures deserve further mention:
16 O(n,p) 16 N with a 7.3-sec half-life,
16N .... 16 0+{3+oy (6 MeV)
80% include the gamma ray. The other reaction sequence, 170(n,p)17N, has
an 8-MeV threshold and a very low cross section (7 Ilb), but
17N .... {3+ 170 .... 160+n
with a 4-sec half-life; the beta has an energy of 3.7 MeV, and the neutron, 1
MeV. The neutron can be troublesome since the nitrogen is carried by the
water and steam to all parts of the system.
There is also a capture reaction in 180, producing 19o, which has a
29-sec half-life,
180+n .... 19o .... 19p+{3+oy (1.4 MeV)
SHIELD DESIGN
579
Calculated concentrations of '6N, !7N, and [90 in the water and steam
loops at full power were 7.4 X 10 9 , 1 X 10', and 4.3 X 10' atoms per
pound of water or steam, respectively. Thus the dominant activity is [ 6N.
Predictions were made for a detailed schedule of dose points and
multiplied by a factor of 3 to include a safety margin.
10.5.2 Survey Measurements
Lewellen, Lacy, and Rab y 2' report the results of radiation surveys made
at 3, 20, 40, 60, 80, and 90% of full power. The large number of
intermediate power levels was apparently run because some of the radiation
levels were not expected to increase linearly with reactor power. Standard
portable survey meters were used: a beta-gamma Gieger-tube probe and a
BF 3 neutron detector.
Results of the radiation survey can be summarized as follows: Although
many of the survey measurements differed somewhat from the calculations,
no discrepancies related to radiation directly from the core. Normal
overdesign (a factor of about 3 in intensity) of the primary shield prevented
any radiation problems in the reactor building. The only serious discrepancy
occurred in the turbine building, where intensities varied from 3 to 10 times
the predicted levels near the turbine itself. Accordingly, a concrete wall
2.4 m high by 9.1 m long was erected on the operating floor between the
turbine and the administrative section of the building. The addition is shown
in Fig. 10.31; the added attenuation in the direction of the administrative
side ranged from 2.5 to 5, which reduced the radiation to desired levels.
10.6 SHIP PROPULSION: N.S.
SA VANNAH
Important as safety is for fixed plants, it is doubly so for propulsion

systems. A fixed-plant location is chosen for remoteness from high
population density; a ship must berth at any deep-water harbor, frequently
in a city's high-density area.
The reactor plant of the N.S. Savannah was selected as an example of the
pressurized-water type because safety considerations, weight, and space
limitations on the shield presented a more difficult design problem than
land-based power plants. In addition, a detailed set of measurements was
made on the finished design under operating conditions. The following
discussion of the reactor plant is mostly taken from the Smith and Turner
shield design summary reports 29 '3 0 and the shield test report by Blizard,
580

NOT TO SCALE
D
H
--------
~:
EXIST. WALL
EXITS
TURBINE BUILDING, OPERATING FLOOR

- - - FENCE
===
CONCRETE-WALL ADDITION
Fig. 10.31-Shield wall and fence additions. (From Lewellen et .1.'")
Blosser, and Freestone."

mid-fifties.
Bear in mind that the shield design dates from the
10.6.1 Description of Ship, Reactor, and Main Shielding
The N.S. Savannah is a single-screw, combination passenger-cargo ship

with an overall length of 181 m. She has a molded beam of 24 m and draws,
when fully loaded, 9 m, displacing 20,000 tons. The vessel has three
complete decks, A, B, and C (from top down), with platform decks at
various levels. A promenade deck extends over A deck for about one-third of
the ship's length; the forward end being slightly forward of amidships. The
SHIELD DESIGN
581
promenade deck is devoted exclusively to public rooms and spaces; all

passenger staterooms are on A deck. The passenger dining room and some of
the crew spaces are on B deck. The remainder of the crew spaces are on C
deck. Below on the smaller D deck various machinery and equipment are
located.
The Savannah reactor is the pressurized-water type, fueled with slightly
enriched (4.4% 23 s U) U0 2 . The fuel is in the form of l.08-cm-diameter
pellets, clad in 0.089-cm-thick stainless steel, contained in bundles of
stainless-steel tubes. The core is 229 cm long overall, with an active fuel
length of 168 cm and an equivalent core diameter of 158 cm. Typically, an
initial loading contains 312.4 kg of 23SU and 6787.5 kg of 238U. The
predicted lifetime of such a core is 52,200 Mwd (>750 days at full power).
The primary shield for the Savannah reactor consists of an annulus of
light water, 84 cm thick, contained in the primary shield tank surrounding
the reactor. The water annulus is supplemented by lead in thicknesses from
2.5 to 10 cm at the primary-shield-tank outer wall. Outside the primary
shield tank are boilers, pumps, demineralizers, and other parts of the
steam- production system. The core, pressure vessel, primary shield, and
steam-production system are enclosed in the containment vessel, a 10.7-mdiameter 6.3-cm-thick steel cylinder with hemispherical ends. The total
length of the containment vessel is 15.4 m. Figure 10.32 shows a model of
the containment vessel and the relative elevations of the various decks.
The secondary shield is outside the containment vessel. Above the
midplane of the containment vessel, the secondary shield consists of a layer
of lead, varying between 12.7 and 15.2 cm in thickness, and a minimum of
20 cm of polyethylene. Below the midplane of the containment vessel, the
shield is primarily concrete in thicknesses ranging from 84 to 122 cm. Fore
and aft, on the hemispherical ends, ilmenite concrete is used; ordinary
concrete is used at the sides. Collision mats consisting of alternate layers of
redwood and steel replace some of the concrete on the sides.
Water-storage tanks, approximately l.5 m deep, extend the secondary
shield below the containme nt vessel from the tank tops to the hull. A
schematic diagram of the shield is shown in Fig. 10.33.
10.6.2 Shielding for Refueling and Control-Rod Maintenance
Special provisions for refueling and servicing are worth noting. Before
entrance can be gained to the reactor after shutdown for refueling and
control-rod maintenance, the insulation covering, pressure-vessel head,
control-rod assembly, flow-baffle assembly and upper grid assembly must be
removed. The pressure-vessel head is secured by studs and a seal weld.
582
Fig.l0.32-The New York Shipbuilding Corporation's model of the shield of the N.S.
Savannah. (From Blizard et ai. 31 )
The stud-removal operation takes place with the teactor water level up in
the head. The water level is lowered before the seal-weld cutting operation
begins. Additional water in the upper extension of the shield tank reduces
dose rates from core fission products and activated internals to 1 to 2
mR/hr. Distance attenuation from primary-loop corrosion and fissionproduct activity further reduces these dose-rate contributions.
Before the dismantled pressure-vessel head is removed, a permanent tank
above the pressure-vessel flange is filled with transitory shield water to limit
dose rates from the reactor internals and spent core. (These are exposed
when the head is removed.) The head is then removed, and a telescopic
internals cask is lowered onto the pressure-vessel flange. The upper
flow-baffle assembly is then withdrawn into the extended cask. Figure 10.34
is a schematic view of the internals cask in place showing the upper
583
SHIELD DESIGN
79 CM TO
CENTER OF CORE
III CORE 'EDGE II

9.5 eM OF H20 P '" 0.85.7 eM OF H20 p '"
12.0 eM OF H20 p = 0.8_
h.L
O.8q;22?2??22222?2222222':~:~ ~~ g~;:
_2.5 eM OF Fe
THERMAL
SHIELD
7.6 eM OF H20 p = O.8~_15.2 eM OF Fe

10 CM OF AIR
PRESSURE VESSEL
_1.0 eM OF Fe
INSULATION -
PRIMARY
SHIELD
TANK
::
....... 1.5 eM OF Fe
;-7.6 eM OF Pb
(TYPICAL)
2.7 TO 5.2 M OF SPACE

6.3 eM OF Fe
WITH EQUIPMENT
6.3 CM QF Fe
~
i. :. '.
"'~"' .. '. \.~~~i:~NMENT
.q-.
~~~~~gARY
15.2 CM OF Pb
15.2 CM OF POLYETHYLENE
84 TO 122 CM OF CONCRETE
BELOW CENTER LINE

OF VESSEL
ABOVE CENTER LINE

OF VESSEL
Fig. lO.33-Schematic cross section at the midplane of the N.S. Savannah reactor shield.
(From Blizard et. al. 31 )
flow-baffle assembly withdrawn. The transitory shield water is also shown.

The cask is telescoped to remove the more highly activated grid assembly.
After the flow baffle and grid assembly have been removed, a shield plug,
or manipulator (equivalent to about 51 cm of cast iron, for gamma-ray
attenuation), is placed on the pressure-vessel flange, and the transitory water
level is lowered. The refueling cask, with its control-rod extension, is
positioned on a port in the shield plug, and the port is positioned over the
fuel element or control rod to be removed. Figure 10.35 is a schematic
diagram of the shield plug and refueling cask in place.
10.6.3 shield Design Criteria
The design requirements for the shielding on the N.S. Savannah were
chosen to ensure that the passengers would receive no more radiation than is
allowed to the general public, even though they were to live continuously on
584
11.4CMOFPb
10.55 M
TRANSITORY
SHIELD WATER
TANK
10 M
WATER-lEVEL
ELEVATION
ELEVATION. 9.44 M
UPPER FLOW
BAFFLE
2.5 eM OF Fe
6.3 eM OF Pb
250-CM DIAMETER
2.8 eM
ELEVATION, 9.0 M
TOOL
ATTACHING
POSITION
UPPER GRID
CORE
PRIMARY
SHIELD
TANK
Fig. 10.34-Telescopic internals cask. (From Smith and Turner. 29 )
board ship with the reactor constantly at full power. The requirements are
also that the crew of the vessel receive no more radiation than workers in
atomic energy installations, even though they remain on board ship all the
time, also with the reactor constantly at full power. The third requirement
for the shielding is that stevedores handling cargo will receive no more
radiation than that permitted for the general public, even though they
continuously handle cargo in the cargo spaces with the reactor at one-fifth
full power. These requirements have dictated dose rates that are extremely
low in passenger spaces and well below ordinary permissible levels in crew
and cargo spaces owing to the presumption that these spaces are occupied
continuously with the reactor at full power.
..
585
SHIELD DESIGN
12.7 eM OF Pb
CONTROL-ROD
EXTENSION
25.4 eM OF Pb~',""'I~
FUElELEMENT
CASK
5.1 eM OF Fe
51 eM EQUIVALENT
CAST IRON
tWATER
LEVEL
PRESSURE
VESSEL
PRIMARV
SHIELD
TANK
COR E
AFT
LEAD
FORE
Fig, lO,35-Shield plug and refueling cask, (From Smith and Turner,")
The design dose rates used to ensure that these requirements were met
are summarized in Table 10.7.
10.6.4 Lead-Polyethylene shield Construction
The design of a compact lead-polyethylene shield that can withstand
the design stresses is somewhat unusual; consequently some of the details are
shown here.
586

Table 10.7-DESIGN DOSE RATES FOR N.S. SAVANNAHt
Location
Operating conditions
Type
access
Design dose
rate
Access spaces outside

secondary sh ield wall
(passenger spaces)
Access spaces outside
secondary shield wall
(crew spaces)
Full-power operation
Unrestricted
0.5 rem/year
Full-power operation
Unrestricted
(crew)
5.0 rem/year
Cargo holds
Operation at one-fifth
full power
Reactor operating
Reactor shut down
~2 hr or longer.
No fission products
in primary loop
Reactor shut down
3 days or longer;
refueling operations
Unrestricted
500 mrem/year
None permitted
Limited
200 mrem/hr
Limited
200 mrem/hr
Inside containment
Inside containment
Locations occupied during

fuel-transfer operations
for transient
conditions,
10 mrem/hr
for continuing
conditions
t From Blizard et al. 3 I
Polyethylene shielding is made from sheets; differences in shield

construction consist mostly in variations in the method of supporting and
covering the laminations after they have been placed.
Two properties of polyethylene which are of importance in shield
0
fabrication are its low softening point (180 to 260 F) and its relative
chemical inertness. The first of these permits the slabs to be shaped with
relative ease in the field. The second makes the use of solvent cements
generally impractical. Instead of cement, corrugated nails are staggered in
alternate layers to fabricate the shield. Sheet size and gauge were chosen for
maximum ease of fabrication, and the edges of the polyethylene plates were
staggered slightly to prevent streaming.
In the secondary shield the polyethylene was applied over the lead on
top of the containment vessel. Rib plate stiffeners and intercostals welded to
the shielded surface (containment vessel) were used both to strengthen the
shield and to provide a means of attaching the cover plate over the
polyethylene. This cover plate was thin steel and served to protect the
polyethylene surface from the hazards of combustion and physical damage.
587
SHIELD DESIGN
COVER PLATE
CORRUGATED NAILS
EXPANSION GAPS
FOR POLYETHYLENE
LEAD
"--''-L..L...'-LL..L..L...L.L.l...L.L..L--L-"--,'-LL..L...<~:..LL...<'-L..L...r-CONT
AI NMENT VESSE L
EDGE BOND
ANGLE IRON BOLTED TO RIB AND COVER PLATES
EDGE BOND
BACKING PLATE FOR LEAD

LEAD EDGE BONDED TO STEEL
BACKING PLATE
FILLET WELD
1lI!!!1Il;><t-STEEL-TO-STEEL WELD
LEAD BURNED TO LEAD
(NO BOND TO STEEL REQUIRED)
Fig. lO.36-Typicallead-polyethylene shield construction details. (From Smith.'o)
It also provided a convenient surface on which to hang pipes, electrical

connections, etc. Typical lead-polyethylene shield-arrangement details are
sketched in Fig. 10.36. Note the gap spacing shown for polyethylene in these
sketches; inadequate allowance for expansion could cause serious structural
damage and buckling in the shield.
10.6.5 Attenuation Calculations
The attenuation calculations for the primary shield and the secondary
shield are discussed separately in the following sections since different
techniques were used.
588
(a) Primary Shield. Two principal methods of calculating neutron

attenuation were used. In all cases conservative assumptions were made since
it was recognized that the calculational models were relatively crude and
uncertainties existed in the basic cross-section data.
One method involved prediction of the neutron fluxes in the reflector,
thermal shields, and pressure vessel with a computer code utilizing two-group
diffusion theory. The calculated fast and thermal flux profiles are shown in
Fig. 10.37 along with the fast flux distribution through the shield tank,
which was calculated with the method described at the end of this section.
14
12
7u
.,.,
10
':'
E
u
~
~
0
8
~
""~
-e:
'0-'"
,;
~ fii:"
>
t
~
f-
oN
00
. ,;"
0
~
~
<t
;t
60
80
100
120
140
160
180
DISTANCE FROM REACTOR CENTER LINE, em
Fig. lO.37-Neutron-1ux profile radial center line, 69 Mw. (From Smith and Turner. 29 )
Diffusion theory is inadequate for the deep penetration through the

primary shield tank. For this calculation an empirical kernel was used for the
attenuation of fast neutrons by water along with the removal-cross-section
concept to account for the additional attenuation due to the steel. The
kernel was obtained by fitting the fast-flux results of the moments-method
calculations for a point fission source in an infinite medium of water by the
following equation:
</J =_1_
41TT2
(e-0 '2 3r + 0.0276 e-o.os'r)
(10.6-1)
589
SHIELD DESIGN
where r (in em) is the penetration in water. Basic to the use of Eq. 10.6-1
was the assumption that the energy spectrum approached an equilibrium
shape in water for penetrations greater than 40 em and that r was the
equivalent amount of water (actual water plus equivalent steel).
The attenuation of gamma rays was calculated with the kernel method
with infinite-medium buildup factors.
For the radial attenuation of core gamma rays (fission plus capture and
inelastics born in the core), the self-shielding factors for uniform cylindrical
sources were used. Since self-absorption factors have not been determined in
."
8
7
g'"
~
ii
e
8
l'
ci
~~
<4
'"
I~
80
c
N
;\
;-
"- .tl
100
120
:i!
.fi
.t~
~~
..
ill' fc
B
~
140
il
~
~
I------i 3
160
160
I
200
I
220
12.2 rem/hr
240
r, em
Fig. lO.38-Primary gamma dose rate, radial center line, 69 Mw. (From Smith and
Tumer. 29 )
general for radiation from the end of a cylinder, this was handled by
representing a 1-cm-thick disk of core source by an isotropic disk source and
interpreting numerically along the core axis. The radial profile of the core
gamma rays is shown in Fig. 10.38.
Calculating the secondary gamma flu:x: through a series of water and steel
slab shields involved an e:x:pression that described the rate of production of
gamma rays by capture of thermal neutrons and attenuation in the source
slabs (slab assumption was used since the radius of curvature is large
compared to the thickness of the steel) and shield materials on either side.
An e:x:ponential neutron removal was assumed in each source slab, and the
590
10
9
8
~
7
E
6g
ci
5 E 0N
"
0
-'
;;
J:
"""
"
.~
>
:E
.0
J:
J2
~
J:
,E<
8.6 rem/hr
0: J:
80::'-:C'OO!:":':1!::20:-:'~40:-:':-:80::-:C180=-:::200=-:::2:':20:-:2~40~2::80::-:::280=-:::300:::-''';3~20::'''34:7:0"-::380=-:::380=-~400
DISTANCE ALONG AXIS FROM CORE CENTER, em
Fig. lO.39-Secondary gamma dose rate through bottom shield axis. (From Smith and
Turner. 29 )
number of source slabs was chosen such that the calculated thermal flux
profile through the thermal shield could be matched closely by a simple
exponential in each source slab. The results of these calculations for
secondary gamma rays are shown in Fig. 10.39.
(b) Secondary Shield. In calculating fluxes or dose rates for points
outside the secondary shield, integration over the original isotropic source
volume becomes unreliable because of the implications of the combined
geometries of the primary and secondary shields. For this situation the
pressure-vessel top and the primary shield tank top and outer walls were
chosen as surface sources for the subsequent secondary-shield attenuation
calculations.
The problem then resolves itself into establishing an angular flux density
on these surfaces for both neutrons and gamma rays and then numerically
integrating over source surfaces with due regard to the directional orientation of the flux. It was assumed that both neutron- and gamma-ray flux
densit,ies could be represented by a cosine distribu tion, flux density on the
surfaces being calculated by point-kernel integration over the volume sources
contained within the primary shield tank. With surface source corrected for
the intensity in the particular direction, the kernel calculations for
SHIELD DESIGN
591
attenuation through the secondary shield become similar to that for the
primary shield calculations with the exception of a correction for the iron
removal cross section. The minimum thickness of 15.2 em on the secondary
shield surface was not sufficient to produce an equilibrium spectrum in
hydrogen. Data from iron-water mock-ups at ORNL were analyzed to
provide an estimate of the effect of nonequilibrium spectral conditions in
hydrogenous material following metal to obtain the necessary corrections to
the removal cross sections.
10.6.6 Measurements
Accurate measurements of the low dose rates specified in the design were
a difficult task. For this purpose a mobile laboratory was developed. 31
Three types of dosimeters were used in the Savannah survey. All gamma-ray
dose-rate measureme nts were obtained with an anthracene scintillation
detector in which the sensitive element is a 2.5-cm-diameter 1.3-cm-thick
anthracene crystal. The choice of anthracene was made because the energy
lost within anthracene by a secondary electron resulting from a given-energy
gamma-ray photon is very nearly proportional to the (computed) singlecollision dose over the energy range from 100 ke V to 10 MeV. The
instrument has a sensitivity of about 13,300 counts min- 1 mrem- 1 hr.
Thermal-neutron dose rates (proportional to flux density) were measured
with a l-in.-diameter BF 3 proportional counter having an active length of
12 in. In the construction of this counter, high sensitivity was achieved by

filling the tube with 96% 'OB-enriched BF 3 at the relatively high pressure of
76 em Hg. This pressure, essentially 1 atm, results in a sensitivity of
approximately 7400 counts min-' mrad-' (tissue) hr- 1 and also makes the
counter more resistant to accidental damage during use.
Fast-neutron dose rates were measured with a modified long counter that
uses the BF 3 tube of the thermal-flux detector as the sensitive element
within a 20-cm-diameter 30-cm-long cadmium-covered polyethylene cylinder. The polyethylene thicknesses at the end and side of the cylinder had
been adjusted by experiments with monoenergetic neutrons from a Van de
Graaff accelerator to give a counter response proportional to the multicollision tissue-dose curve to within 10% over the energy range from 0.025 to
approximately 10 MeV. The sensitivity of this instrument is approximately
7800 counts min-' mrad- ' (tissue) hr-'.
The instrumentation described was carefully calibrated against known
sources and fluxes and compared to standard instrumentation in a series of
experiments at the ORNL Bulk Shielding Facility (BSF). The BSF
592
experiments used an idealized mock-up of the SalJannah's shield, thus

ensuring that the comparisons would be valid for the energy spectrum that
would be encountered aboard the SalJannah.
The experimental procedures used in the SalJannah survey were
developed and tested in a series of studies of land-based reactors made during
the months preceding this study. By the time the mobile laboratory was
lifted aboard the SalJannah, methods and procedures were firm, and a group
of trained and experienced personnel was available to act as a nucleus for the
large group required for the survey. Before the survey started, the secondary
shield was completely marked with a rectangular 0.9-m grid (see Fig. 10.32).
The general procedure for the survey consisted in obtaining a 1-min or
O.S-min count, depending on counting statistics, in the center of each grid
with both the thermal-neutron and gamma-ray dosimeters. After the count
the detector was slowly swept over the area within the grid to detect any
small area of unusual leakage. The entire shield was covered with both
thermal-neutron and gamma-ray dosimeters. Fast-neutron dose rates were
measured at carefully chosen locations over the shield, which permitted a
good evaluation of the ratio of fast to thermal neutrons over the shield.
10.6.7 Comparison of Measurements and Calculations
The reactor was held at approximately its full power of 69 Mw for the
entire duration of the detailed shield survey. The dose rates recorded thus
represent the maximum to be expected. A detailed comparison of the
measurements and calculations is shown in Table 10.8, which has been
abridged from Blizard et al. 3 1 The point numbers identifying the data
positions were selected by Smith and Turner 29 for later comparison with
measurements. The locations on each deck are shown in Fig. 10.40, with
point numbers corresponding to those of Table 10.8.
Note that the measured fast-neutron dose rates in Table 10.8 are
reported in absorbed dose units (mrad) and are converted to dose equivalent
(mrem) for comparison with calculations by multiplying by a relative
biological equivalence factor varying from 8 to 9 depending on energy
spectrum (or total shield thickness). These units and their relations are
discussed in Sec. 3.3.4.
1n general, the actual dose rates appear to be considerably lower than the
predicted dose rates, which reflects the intended conservatism in the design.
Spot checks of the comparisons show overestimates of the gamma-ray dose
rate and underestimates of the neutron dose rate.
SHIELD DESIGN
Tabl~
593
to.8-COMPARISON OF MEASURED DOSE RATES WITH CALCULATED

DOSE RATES AT SPECIAL POINTSt
Observed dose nile
Neutron
Point
number
Babcock & wilcox 1144-1 calculated dOle I'lIte
Fast,
Thermal,
Gamma,
Total,
Neutron,
mrad/hr
mrem/hr
mrem/hr
mTern/hr
mrem/hr
Gamma,
mrem/hr
Tot,I,
mremlhr
Prom",ade Deck
1.50 X 10-3
0.988
0.976
1.65
45
46
47
1.31 X 10""
1.16
0.711
0.925 X 10-'2
2.23 X 10"
1.0 X 10"1
41.9 X 10-2
4.29 X 10"1
1.42
0.569
2.28
0
0.7
'.5
17.9
0.95
1.86
1.0
2.4
11.0
8.4
11.0
1.20
1.08
'.3
8.'
8.'
1.27
1.30
1.4
19.2
t9.1
12.7
14.6
15.2
27.2
8.'
'.0
17.0
18.3
12.4
2.06
2.05
1.43
2.28
2.52
2.00
1.75
2.02
3.19
1.08
1.40
2.20
2.73
2.14
23.5 X 10"';1
'.5
10.8
24.9
1.8
33.7
5.67 X 10"
0.95
1.12
5.85
0.18
3.37
20.2
2.02
'.7
5.'
8.5
15.6
69.4
0.83
0.65
1.01
1.79
7.20
'.5
2.'
2.'
121.4
'.2
1.04
0.48
0.41
14.99
0.83
1.52
1.14
1.51
0.412
72
1.40
0.897
0.845
1.09
1.43
0.570
2.09
0.369
0.232
0.318
1.09
1.53
0.624
0.360
0.823
0.68
3.43
2
3
3.32
1.98
2.83
4.17
2.68
2.76
2.26
8.84
1.911
1.55
9.75
2.35
1.86
3.58
4.33
1.34
3.45
4.19
2.41
1.26
1.25
0.803
0.792
0.697
0,422
1.17
0.232
48
54
5.
57
71
1.42
2.58
2.39
2.75
1.42
1.10
1.76
1.91
I.,
1.26
A Deck
14
17
22
23
2.
803
804
6.46
7.07
3.97
3.47
4.17
2.26
2.79
'.30
3.84
2.93
2.05
l.70
1.72
1.12
0.694
2,45
1.26
5.27
7.93
6.17
5.32
4.01
9.32
3.37
2,46
9.09
4.52
2.86
I.,
I.,
I.'
I.'
I.'
2.'
5.0
'.7
I.,
5.0
5.0
'.0
'.0
21.4
23.8
18.1
CDeck
252
253
2'1
32.
325
'"
"2
.13
7.76 X 10-3
2.25
4.56
5.37
1.87
2.94
3.15
2.33
12.77 X 10'"
2.76
1.66
10.90
2.49
2.64
2.47
2.25
13.78 X 10"';1
2.19
0.820
15.79
4.53
2.13
3.46
13.2
20.59 X 10"'1
4.16
5.76
20.52
6.17
4.69
6.20
15.25
33.2 X 10"';1
0
0.'
33.6
0
0
DDe<:k
27'
278
332
813
81'
10.51
1.41
1.38
2.02
1.13
0.781
0.400
0.517
1.01
0.370
0.644
0.706
0.777
2.24
1.63
0.465
0.317
0.1$0
0.486
0.317
0.343
0.338
0.273
0.685
0.288
9.69
1.92
2.02
3.99
2.61
I.'
0.'
I.,
2.3
2.'
Tank top
297
307
33.
640
.41
0.390
0.338
0.910
1.24
0.754
t Adapted from Blizard d al.3 I
0.684
0.532
1.06
1.76
0.939
I.,
0.'
1.5
28.5
2.1
594
203
".

71
46 ~47 ~1
79
53
48
_:4.1 ,_., 49_ _
_86
52
82
85
204
3.4. 5. 6
7.
IDOCT~R
23
14
201\
FAN ROOM
1,2
802 803 __ ~
REACTOR .,_.
.601
-.~ DISPENSARY 13 SPAce
LAB. I ~I'
830
B4
\.
".
57
56
50
12
11
l-
~
A DECK
PROMENADE DECK
FROZEN
FOOOS--L,'L,--i
27:;1--.__--"-
WINE
IIIIlI[
613
332
611
276
612
REACTOR SPACE
>w
613
~r-=~=::;---;=====r-:;;;;;;;~
~
PASSENGER
fiSH
BAGGAGE
o DECK
C DECK
297
334
DRY STORES
8S6~ STEVEDORE SPACE
L.L:J lIIIIII L
307
64.
REACTOR SPAce
- ~ _ . _ - -
~
~
>wz
ji,
" I
u
<
STEWARD'S
STORES
r=-
ITIIIII
<
641 u
"
w<
z
HOLD
Fig. lOAO-Special point locations aboard the N.S. Savannah. (From Blizard et al.
31
SHIELD DESIGN
595
The measurements showed, however, that the dose rates did not exceed
tolerance limits in areas where passenger access is permitted. The design dose
rate was exceeded only in a small region on D deck, to which only the crew
is permitted access and which can easily be controlled.
10.7 SPACE POWER: SNAP-lOA FLIGHT TEST
The Snapshot-1 flight test, which took place in April 1965, is the only
known space light of a reactor. Strictly speaking, this was an experiment,
and it represents a test system in development rather than a finished design
application. However, the magnitude of the program, the prelight design
development effort, and the postlight analysis give us the same pattern of
input that we have sought for the previous discussion of reactor-shield
applications in this chapter. We should, nevertheless, bear in mind that the
Space Nuclear Auxiliary Power (SNAP)-10A is not a finished product but an
intermediate development model.
The SNAP-lOA reactor is described by Glasstone 32 as a very small
system, being only 35 cm in diameter and about 46 cm long. It is cooled
with NaK (sodium-potassium) and controlled by rotating beryllium
relector material in sernicylindrical drums toward the core. The core is
about 23 cm in diameter and 39.5 cm long. Fuel elements are enriched
uranium and zirconium hydride clad in Hastelloy with hydrogen serving as a
moderator. The choice of materials was based largely on the need for
compact design and high-temperature operation. The NaK coolant leaves the
core at 545C and is pumped to a thermoelectric converter for production of
electricity. The reactor operated at about 40 kw(t) and produced a little over
0.5 kw of electric power.
The SNAP-10A-Agena spacecraft was launched into a near-polar orbit
Apr. 3, 1965. After 6 days of monitoring and manual control of the reactor
by ground commands, the reactor was switched to automatic temperaturecoefficient control. Subsequent operation was automatic without further
ground command. The reactor was to have operated for at least 6 months,
but on May 16 loss of signals from the spacecraft indicated that the reactor
had shut down. The failure was apparently not in the reactor itself but in a
voltage regulator.
10.7.1 Shield Analysis
Belcher, Flynn, and Thomson" summarized an extensive prelight
shield analysis to ensure that the shield met its design objectives. These
596
objectives were that the 1-year gamma and fast-neutron levels not exceed
107 R or 10. 2 neutrons/cm 2 , respectively, across a 152-cm-diameter plane
located 5.3 m below the reactor core. This plane is called the reference-dose
plane, as shown in Fig. 10.41. The weight objective for the shield was
102 kg, including container and structure, which was certainly a stringent
restriction.
The gamma-ray- and neutron-level specifications were intended to limit
radiation effects to solid-state electronics equipment. Belcher et al. 3 3
estimate an approximate gross leakage of 108 Rand 10' 6 neutrons/cm 2
and a corresponding attenuation requirement of 104 for fast neutrons and
10 for gamma rays. They chose LiH as the shield material for its high
neutron-shielding efficiency (light weight) and high-temperature properties,
estimating that about 60 em would be required.
They also point out, as is evident in Fig. 10.41, that the control drums
and sides of the reactor were not shadowed by the shield from the
thermoelectric converter and the near part of the Agena stage. Not shown
on the sketch is the NaK piping around the shield. The control drums,
converter, and NaK piping all constitute sources of scattered neutrons in the
near section of the Agena structure. Had the shadow shield been designed
with an annular bulge and shortened to avoid a weight increase, it might have
been more effective. An alternative arrangement within the 102-kg weight
limit would be an interesting design exercise. Belcher et al. 3 3 state that it
was impossible to rectify this condition without exceeding the weight
limitations.
Although the specifications refer to the reference-dose plane, most of the
measurements and predictions use the SNAP-Agena mating plane (also
shown on Fig. 10.41) as a primary reference for radiation intensities. We will
use only the SNAP-Agena plane to avoid confusion.
The core region was divided into an array of effective point sources, and
a ray-tracing technique using removal cross sections determined attenuation
through the core and LiH shield. Attenuation measurements for LiH were
repeated at the Atomics International (AI) Shield Test Facility in Canoga
Park, Calif., and the General Dynamics Aerosystems Test Facility in Fort
Worth, Tex. From the measurements a LiH attenuation coefficient of 0.156
cm-\ for the 0.7 g/cm 3 density was estimated at operating temperature
(-"370C).
As might be suspected from the discussion of the geometry, neutrons
scattered around the shield sufficiently exceeded the number penetrating
SHIELD DESIGN
_~
597
~cm
REACTOA~
CONTROL) DRUMS
122 OUT
RADIATION SHIELD
I "-
"l1?..l..o::!1'_--------.rII--t-i1
.....
,
THERMOELECTRIC
CONVERTER~
'I.
I
SNAP INSTRUMENT
COMPARTMENT,{"
SNAP-AGENA
\=t'r-,=----P-,,<:;;::::1rn='
MATING PLANE
5.34 m
---,..--
AGENA
2.6 m
REFERENCE- --f------;-;';;r----+-<---'--
DOSE PLANE
1.52 m
Fig. 10,41-SNAPI0A ,adiation shield geometry. (From Belcher et
al. 33 )
598
through the shield that the latter component was unimportant in analyzing
the flight-test data.
Gamma-ray exposure through the shield was computed by the same
ray-tracing method as the fast neutrons. The sources considered were fission
gamma rays, captures in the core, and activation from the NaK coplant.
Neutron capture in 2' Na produces 24 Na, which decays by producing two

gamma rays, 2.76 and 1.38 MeV. The decay of 42 K gives a 1.51-MeV gamma
ray about 25% of the time. The moments-method buildup factors
(Sec. 4.8.1) were used. None of the gamma-ray intensities exweded the 10 7
R/year maximum at the SNAP~Agena mating plane. The computed total
consisted mostly of gamma rays from the core.
The principal heating rate in the shield was due to neutron capture in
6Li:
6Li+n --> 'H +C/+ 4.8 MeV

The, alpha has a very short range and may be considered deposited at the
capture site. Neutron kinetic energy in elastic collisions with hydrogen
produces almost as much heat as the 6Li captures. Captures in 7Li (the
principal isotope) produce gamma rays, which give unimportant intensities in
this case. At 40 'kw the heating in the ,hield was computed to be about 160
watts with a maximum power density of 0.3 watt/em', 49% from 6Li(n,C/),
39% from neutron elastic scattering, and 12% from gamma-ray absorption.
10.7.2 Flight-Test Results
Bennett, Krakar, and Kruse 34 report the nuclear environment and
compare it with the analysis just described. We have not described the
method of calculating the important scattering contributions from the
control drum, converter, and reactor sides because the estimates of Belcher
et al.' 3 were admittedly crude and were revised for inclusion with the flight
results of Bennett et al. 3 4 We will discuss these estimates later.
Fast-neutron-flux densities were measured with a 2' 7 Np fission chamber. The fission cross section has a threshold of 0.65 MeV, which makes this
chamber well-suited for fast-neutron fluxes limited by radiation-effects
considerations. Neutrons below the threshold do not significantly contribute
to radiation damage. The counter was enclosed in cadmium to eliminate
thermal-neutron interaction with impurities in the 2' 7 Np.
A similar fission chamber with 2' 5 U was used for thermal-neutron
measurements; actually two sensors were used, one cadmium-wrapped, to
SHIELD DESIGN
599
obtain the thermalflux density by subtraction of the two signals. Gammaray exposures were obtained with a CO 2 -filled ionization chamber.
Additional neutron measurements were made with a radiation-sensitive
transistor, which responds to the cumulative fast-neutron-flux density.

Telemetry and data processing need not concern us in this discussion
except to note they were necessarily elaborate and contributed to
reproducibility within a nominal factor of 1.22 in fast-neutron flux or
gamma-ray exposure.
Table 10.9-SNAPSHOT RADiATIONN-DETECTOR
RESULTS AT SNAP-I0A-AGENA MATING PLANE
Fast-neutron-
flux density,
10 14 neutrons
->
-,
em year
Gamma-ray
exposure,
106 R year- 1
LMSC detectors
RAIl
RAI-2
0.48
1.09
6.3
AI detectors
166
167
0.67
0.63
168
170
0.41
0.42
0.50
171
172
0.52
169
138
0.47
5.6
Table 10.9 gives flux densities for the detectors located on the mating
plane we have used for reference. Each value for Lockheed Missiles and
Space Company (LMSC) detectors represents a mean of 1500 telemetry
samples; for AI detectors each value represents a mean of 10 telemetry
samp les from each of 20 data acquisitions. The former are fission-chamber
measurements and the latter are the solid-state detectors. We have converted
the rate data of Belcher ef al. to 1-year totals.
The flight data prompted an immediate review of the preflight analysis.
The fast-neutron-flux densities.were found to be underestimated by a factor
of 6.5 to 7.8; gamma exposures were within 10% of prediction in this area.
Revised estimates were made of th.e scattering from the thermoelectric
converter-radiator, increasing the scattered component by a factor of
600

1015, ._ _-,-_ _- ,_ _- , _ - - , -_ _-,-_ _--,._ _- ,
;;
10
,-_-,
14
""'
'"ci
RAI-l
/\
.,!:!:.
Thermoelectric
converter -tad iator
>
':'
Reflector
E
0
~
,;
10 '3
:::LCpteflight estimate.
-'
-----
thermoelectric converter
=>
Fine control drum

0
projection, 16
a:
....
rotation
=>
w
,.:.
'"
'"
Reactor core
10
12
,Ol1E:=:;t:=:;r:::::;~-......,b_--:h_--:J:c--:J".-......,~-J
o
20
40
60
80
100
'20
'40
'50
DISTANCE FROM CENTER LINE, em
Fig.lO.42-Revised AI fast-neutron flux at SNAP-IOA-Agena mating plane below

out-rotated control drums. (From Bennett et al. 34)
about 8. The various contributions to the fast-neutron-lux density are

shown in Fig. 10.42. The projection of the shadow shield to this plane of the
vehicle has a 51-em radius. Transmission through the shield is contained in
the component marked reactor core. The rest of the components do not
601
SHIELD DESIGN
FAST-NEUTRONFLUX DENSITY, neutrons cm- 3 year- 1 (E
> 0.65 MeV)
-z
-75
-50
152-em-
-25
diameter /
vehicly
4.7 x 10 13
+v+-+--~-------'~-----+--~r-l~-v
+25
+50
+75
+z
em
Fig.10.43-Polar graph of fast-neutron data at SNAP-Agena mating plane. (From

Bennett et al. 34)
penetrate the shield. The early preflight prediction of scattering from the
thermoelectric converter is shown as a dashed line. The revised estimate of
this component raised the total at the center line by a mctor of 5; near the
skin of the vehicle (76 em), the total was unaffected. Shown in the figure are
two measurements from Table 10.9 that are in reasonable agreement.
This was a complex set of scattering sources, and the predictions were
indeed within the accuracy of the approximate methods used. This is a
clear-cut instance where a Monte Carlo calculation of the scattered
contributions is the only method that could be expected to give the desired
accuracy (presumably within a factor of 2).
The flight-test measurements were compared with the predictions, and a
model was constructed that gave a better estimate of the radiation
environment throughout the vehicle. The polar plot for our plane of
reference is shown in Fig. 10.43. Note that the effect of the control drums is
602
to push the isoflux lines inward toward the center'. Intensities beyond the
mating plane could not be satisfactorily modeled without accounting for the
attenuation contributed by some 19 pieces of equipment aboard the Agena
stage. Accordingly, attenuation measurements made during preflight testing
were used to derive removal cross sections for each piece of equipment, 3 4 A
point-kernel calculation was made from the mating plane to each location to
generate the data to model the radiation environment. The resulting model,
based on hybrid combination of preflight prediction, in-flight measurement,
and empirical adjustment, reproduced measured data within a factor of 2.
REFERENCES
1. H. E. Hungerford, Problems in Design of Fast Reactor Shields, Trans. Amer. Nucl.
Soc., 11: 702 (1968).

2. R. ]. Beaudry, H. E. Hungerford, R. F. Mantey, and W. F. Chaltron, Design of the
Radiation Shield System for the Enrico Fermi Atomic Power Plant, Report
APDA~203. Atomic Power Development Associates, Inc., May 1967.
3. R. ]. Beaudry, Operational Shield Tests for t~e Enrico Fermi Atomic Power Plant,
Report APDA-221, Atomic Power Develop~ent Associates, Inc.,]uly 1968.
4. H. E. Hungerford, Shielding, Chap. 8, in Fast Reactor Technology, Plant Design,], G,
Yevick (Ed.), M.I.T. Press, Cambridge, Mass., 1966.
5. H. V. Lichtenberger, F. W. Thalgott, W. Y. Kato, and M. Novick, Operating
Experience and Experimental Results Obtained from a NaK-Cooled Fast Reactor, in
Proceedings of the First United Nations International Conference on the Peaceful
Uses of Atomic Energy, Geneva, 1955. Vol. 3, p.345, United Nations, New York,
1956.
6.J. L. Koch, H. O. Monson, W. R. Simmons, M. Levenson, F. Verber, E. Hutter,
R.Jaross, T. R. Spalding,]. R. Simanton, and A. Lovoff, Construction and Design of
EBR-II, in Proceedings of the Second United Nations International Conference on the
Peaceful Uses of Atomic Energy, Geneva, 1958, Vol. 9, p. 323, United Nations, New
York,1958.
7. C. P. Zaleski and L. Vautrey, Le Reacteur Rapide Surregenerateur, Rapsodie, Vols. I
and II, French Report CEA-2193, 1962.'
8. H. Cartwright, J. Tatlock, and R. R. Matthews, Dounreay Fast Reactor-Basic
Problems in Design, in Proceedings of the Second United Nations International
Conference on the Peaceful Uses of Atomic En';rgy, Geneva, 1958, Vol. 9, p. 323,
United Nations, New Yark, 1958.
9. A. I. Leipunsky, V. G. Grabin, N. N. Aristarkhov, I. I. Bondarenko, O. D. Kazachovsky, O:L. Lubimtsev, S. A. Pashkov, M. S. Pinkhasik, K. K. Renne, U. Y. Stavissky,
F. A. Ukraintsev, and L. N. Usachev, Experimental Fast Reactors in the Soviet Union,
in Proceedings of the Second United" Nations International Conference on the
Peaceful Uses of Atomic Energy, Geneva, 1958, Vol. 9, p. 348, United Nations, New
York, 1958.
SHIELD DESIGN
603
10. Power Reactor Development Company, Enrico Fermi Atomic Power Plant Technical
Information and Hazards Summary Report, Vol. 3, June 1961.
11, Title 10, Code of Federal Regulations, Chapter 1, Atomic Energy Commission, Part
20, Standard for Protection Against Radiation, in Federal Register, Superintendent of
Documents, U. S. Government Printing Office, Washington. D. C., Nov. 17, 1960.
12. L. F. Zartman and G. B. Greenough (Eds.), Proceedings of US/UK Graphite
Conference, London, Dec. 16-18, 1957, USAEC Report TID-7565, Part I, AEC
Division of Reactor Development and United Kingdom Atomic Energy Authority,
March 1969.
13. R. F. Mantey, Temperature Distribution and Thermal Stresses in a Concrete Shield
Due to Gamma Radiation, Trans. A mer. Nucl. Soc., 1: 66 (1958).
14. H. P. Flatt and D. C. Baller, AIM-5, A Multigroup One-Dimensional Diffusion
/'Equation Code, Report NAA-SR-4694, North American Aviation, Inc., Mar. 1, 1960.
05. S. Preiser, G. Rabinowitz, and E. de Dufour, A Program for the Numerical Integration
of the Boltzmann Equation, Report UNUCOR-632, United Nuclear Corporation,
1960.
16. B. G. Carlson, C. Lee, and W. J. Worlton, DSN and TDC Neutron Transport Codes,
USAEC Report LAMS-2346, Los Alamos' Scientific, Laboratory, 1962.
17. W. J. Worlton, Sn Reactor Code DTK, a Predecessor to DTF-IV, USAEC Report
LA-3373, Los Alamos Scientific Laboratory, 1965.
18. A. Hassitt, A Computer Program to Solve the Multigroup Diffusion Equations, British
Report TRG-Report-229, Mar. 8, 1962.
19. C. B. Mills, Neutron Cross Sections for Fast and Intermediate Nuclea.r Reactors.
USAEC Report LAMS-2255, Los Alamos Scientific Laboratory, 1959.
20. E. Fisher, Streaming of Neutrons in Shields, Nucl. Sci. Eng., 1(3): 222-238 (1956).
21. T. Rockwell 1II (Ed.), Reactor Shielding Design Manual,USAEC Report TID-7004,
1956.
22. J. Adamson, A. M. Judd, and N. I. McNair, Neutron Flux in the Shield of the
Dounreay Fast Reactor,]. Brit. Nucl. Energy Soc., 4: 119-126 (1965).
23. D. C. G. Smith, Physics of the Dounreay Fast Reactor, in Physics of Fast and
Intermediate Reactors, Symposium Proceedings, Vienna, 1961, International Atomic
Energy Agency, Vienna, 1962 (STI/PUB/49).
24. E. Aalto, R. Sandlin, and A. Krell, Some Comparisons of Measured and Predicted
Primary Radiation Levels in the Agesta Power Plant, Swedish Report AE-325, 1968.
25. L. Hjarne and M. Leimdorfer, A Method for Predicting the Penetration and Slowing
Down of Neutrons in Reactor Shields, Nucl. Sci. Eng., 24: 165-174 (1966).
26. M. N. Bjeldanes, Pathfinder, Nucl. News, 13(6): 56 Uune 1970).
27. R. J. Hall and D. W. Stephen, Pathfinder Atomic Power Plant Shielding Analysis,
USAEC Report ACNP-62016, Allis-Chalmers Manufacturing Company, July 15,
1962.
28. J. W. Lewellen, P. S. Laey, and T. M. Raby, Pathfinder Atomic Power Plant. Shielding
and Radiation Survey Measurements, USAEC Report ACNP-67520, Allis-Chalmers
Manufacturing Company, June 1967.
29. W. R. Smith and M. A. Turner, Nuclear Merchant Ship Reactor Shield Design
Summary Report, USAEC Repnrt BAW-1I44-1, Babcock & Wilcox Comp.ny, Aug. 1,
1959.
604
30. W. R. Smith, Nuclear Merchant Ship Reactor, Shield Fabrication for the NMSR,
USAEC Report BAW-Illl, Babcock & Wilcox Company, May 1958.
31. E. P. Blizard, T. V. Blosser, and R. M. Freestone, Jr., Radiation Leakage Survey of the
Shield of the Nuclear Ship Savannah, USAEC Report ORNL-3336. Oak Ridge
National Labotatory, Aug. 29, 1962.
32. Samuel Glasstone, Sourcebook on Atomic Energy, 3rd ed., p. 612, D. Van Nostrand
Co., Inc., Princeton, N. J.,1967.
33. J. A. Belcher, W. A. Flynn, and R. J. Thomson, SNAP lOA Radiation Shield Analysis.
USAEC Report NAA-SR-9647. Atomics International, Oct. 15. 1964.
34. C. M. Bennett, ]. J. Krakar, and W. G. Kruse, SNAP-IDA Flight Vehicle Nuclear
Environment and Radiation Effects, USAEC Report TID-22418(LMSC-A764793),
Lockheed Missiles and Space Company, 1965.
Gamma Rays from Inelastic

Neutron Scattering and Fission
A~pendix A
Table A.I gives representative data on important elements in the neutronenergy region of interest. Table A.2 gives gamma-ray energy release vs. time
after fission.
Table A.I-GAMMA RAYS FROM INELASTIC NEUTRON SCATIERING t t
---
Energy of
prominent
gamma ray$,
3
5
Element
Li
"8
M,V
0.478
0.478 (n,a)
0.717
1.02
do/dfl.(90o), mb/ sr
En,MeV
Average
anisotropy,
W(O")/W{90")
1.0
1.0j
1..
15
21.3
l.ot
1.5
2.
8.'
0.3
13.0
3.4
3.0
.5
4.
3.5
8.2
3.8
1.1
<0.1
<0.1
0.1
0.2
4.0
0.2
0.3
44
45
5
39
47
10
38
64
9.5
50
2
15
36
7.
9.9
19.5
8.0
16.8
3.4
13.6
8.8
18.6
5.1
14.6
3.9
12.0
0.5
3.37 (n.p)
C
2.13
4.4
1.6
2.31
4.90
1.0
2.0
22
10.6
2.3
1.43
"8
5.0
5.4
1.2
0.7
0.'
0.5
0.3
5.3
2.15
2.86
5
6
7
4.0
0.5
5.10
II
23Na
0.110
0.198
0.44
0.63
12
24 Mg
1.64
1.37
1.81
13
27 AI
26Mg
0.84
1.01
1.72
2.21
14
2ISi
18
"",.
2.73
3.00
(doublet)
1.77
0.68
1.08
1.45
1.77
2.53
1.0
1.0
25
66
42
1.4
1.4
1.0
1.0
44
45
18.1
I..
7.0
17.0
2.4
8.0
1.0
1.0
1.2
1.45
1.0
0.9
1.3
1.8
1.6
32
6.0
6.4
41
30
39
9.6
7.6 10.8
5.2
8.0
7.2
54
61
6.
2.2
6.5
9.2
5.0
3.8 lOA
(Table continues on next page.)
605
606
APPENDIX A
Table A.l-(Continued}
-----_ _---..
Energy of
Element
prominent
Average
gamma rays,
anisotropy,
M,V
W(0")IW(90")
19
2.52
2.81
20
c.
0.51
.--_._-
daldS1:(90o), mb/sr
E",McV
1.0
1.5
2.0
3.0
5
1.5
3.05
3.5
4.0
5.0
7
7
3
05
10
10
7
6.9
15.2
4.6
3.7
5.8
2.1
2.6
0.5
(Nat.)
0.768
0.9
1.158
3.730
24
52 Cr
3.900
1.43
2.37
2.65
2.4
1.7
1.4
1.3
40
2.77
26
s6Fe
2.96
3.16
0.845
1.24
UI
2.12
2.30
2.60
3.45
2.76
3.60
26
54Fe
1.41
1.13
1.56
2.96
1.75
3.16
1.93
60
3
4
1.8
1.6
1.4
1.5
1.1
1.0
0.9
21.1
35.5
83.0
86.0
9.7
12.3
40
6
4
4
I4
5
89.2
IDA
11.3
8.4
4.8
1.08
1.27
65.1
13.0
1.2
1.25
>1.3
1.55
<1.2
1.48
1.7
61.5
17.3
4.6
6.1
1.6
5.2
27
SIlCo
1.10
1.19
1.29
1,48
1.0
1.0
1.0
1.0
1.74
2.10
2.50
2.80
39
"V
0.71
1.31
1.51
1.62
1.75
1.95
1.98
2.10
2.36
2.56
2.70
2.86
3.05
3.49
3.85
0.75
1.1
0.85
0.85
1.25
0.75
5.6
12.6
2.7
6.4
107.6
17.2
14.9
12.1
6.8
9.2
4.2
3.2
2.8
69.5
22.8
8.3
7.7
2.5
8.1
4.6
11.2
9.3
10.1
32.4
31.7
33.8
11.7
12.6
13.0
15.3
8.2
3.9
5.5
5.7
27.8
7.5
-22.6
5.4
3.9
15.2
11.6
32.8
10.9
3.6
27.2
11.9
34.7
19.2
15.7
13.3
33.4
6.7
4.0
5.7
8.5
35.0
8.2
29.6
5.4
22.4
15.8
7.0
0.75
80.3
6.9
5.6
4.1
8.3
7.1
4.0
4.5
1.6
1.8
29.4
18.2
14.6
35.5
16.9
15.3
14.2
37.9
8.1
7.2
3.3
5.6
6.0
31.8
4.7
27.7
2.5
3.0
7.7
7.0
2.42
2.66
2.85
3.37
0.55
0.68
lOA
11.6
14.5
10.0
16.4
2.7
18.7
lOA
5.2
5.4
7.6
24.6
7.9
20.2
4.8
4.5
4.5
2.6
7.5
2.7
9.0
12.7
2.'
3.2
607
APPENDIX A
Table A.l-(Continued)
Energy of
prominent
gamma rays.
X
40
40
Element
90Zr
92 Zr
M,V
daldn(90o}, rob/sr
Average
anisotropy,
W(OoJ/W('O")
En,MeV
1.0
0.43
0.57
0.89
2.18
0.46
1.5
2.0
0.57
10
4
1
80
18
20
100
1.12
1.41
1.84
2.16
82
206Pb
20?Pb
0.54
0.80
0.57
0.894
1.08
1.77
2.12
2.64
82
20llPb
3.5
10
10
14
7
50
10
23
130
25
10
8
5
'7
175
137
13
0.93
82
3.0
0.7 (4 MeV)
1.2 (4 MeV)
1.0 (4 MeV)
0.82 (4 MeV)
1.0 (4 MeV)
1.0 (4 MeV)
J.32 (4 MeV)
13
18
43
42
14
51
10
20
130
20
10
8
4
.5
118
74
43
10
14
8
2
0.58
0.76
0.86
2.61
1.75
40
.,
I.
14
30
8
7
55
4.0
5.0
48
48
117
t82
150
.3
25
44
2.
22
I.
48
7
87
tFrom I. L. Morgan. Gamma Rays from Inelastic Neutron Scattering, in Engineering Compendium on
&diation Shielding, Vol. I, R. G. Jaeger (Ed.), Sec, 2.3.2.2, Springer-Verlag, New York, 1968.
jdaldn (55").
Table A.2-GAMMARAY ENERGY RELEASE VS. TIME AFTER FISSIONt

Time
'00
200
400
100' 10'
200' 10'
400' 10'
100' 10'
200' 10'
400' 10'
700' 10'
100' 10 3
200 103
400 103
100' 104
200 104
400' 104
100' lOS
200' 105
400' 105
100' 10'
239' lOS
159 t0 5
914' 10
421 10
226' to
105' 10
361 . 103
167 103
821 10 1
445' 10'
289' 10 1
118 10 2
525' 10 1
203' 10'
955
447
154
544' ler1
168' to-I
471 10"'
347 lOS
202 lOS
988' 10
403' 10
200' 10
841 10 3
257 103
112'103
484' 10'
218' 10'
124 10'
406' 10'
149' 10'
574
28'
134
466 . 10- 1
201 Hr l
840' 10-'
200' 10"'
'00
200
400
100' 10'
200' 10'
400' 10'
100' 10'
200' 10'
400' 102
700' 10 2
100' t0 3
200' 10 3
400' 10 3
100' 104
200' t0 4
400' 104
100'105
451 lOB
390' lOB
319' lOB
220 lOB
145 108
835' 10'
359' 10'
181 107
930' 10'
512' 10'
335' 10'
138' 10'
619 . 105
240' lOS
113' lOS
529 104
182' 10_4
644' 10 3
199' 10 3
467 10&
382' 10&
296' 10&
192' 10&
120' 108
200' lOS
400' 105
100 10'
558' 102
640'
251'
120'
542'
248'
143'
474
175'
679'
342'
159'
551
238'
995'
237'
10'
10'
t0 7
10'
10'
10'
105
lOS
104
104
10
10 3
103
10 2
102
r,
rT
r,
r,
The Values are for 10' 0 Instantaneous Fissions

634 103
279 10
111'105
474' 10
501 103
654' 10
165' 10
374' 104
3
3
946' 10
328' 10
234' 10
373' 10
3
3
4
901 10 3
249t0
584' 10
139' 10
463 10 3
151 103
429 103
656' 103
301 . 103
644' 10'
229' 10 3
254' 10 3
633' 10'
863' 10'
156' 10'
133' 103
632' 10'
456' to'
235' 10'
310' 10'
1
1
984 10 1
201 10
341' 10'
994' 10
125' 10'
440' 10'
238' 10'
319' 10 1
219' 101
148' 10'
177' 10'
863' 10 1
343' 10'
814
508
563
245' 10 1
376
124' 10'
195
518
1
1
546' 10100' 10
111
401
387
1
1
371 10455
180
163' 10203
1
1
992 10110'10-'
432' 10576' 10- 1
177
275 1(]'"3 140' 10-'
498 10-'
462' 10-'
150'10-'
160 10-'
452' 10-4 488 "10-'
226' 10- 3 155' 10-'
114 '1(f4 271 1<f"
196' 10-'
645' 10-' 174 '10"2
553 10"5
313' 10-3 116'10-4 271 10"3 259' 10"5
349' 105
230 105
121 t0 5
443' 10
224' to
t06' 10
362' 103
128' 103
482' 10'
235' 10'
144' 10'
549' 10 1
The Reactor Operating Time Was I hr at 1 watt

102' 10&
229' 10'
658' 10'
153'10B
126' lOB
598 10'
888 10'
213' 10'
538 . 10'
976' 10'
710' 10'
189' 10'
455 10'
141 10'
612' to'
475' 10'
377 10'
918' to'
348' 10'
336' 10 1
534' lOB
444 lOB
342' tO B
222' lOB
148' 10&
870' ]0'
345' 10'
133' 10'
540' 10'
269' 10'
167' 10'
642' 105
289' 105
119' 105
538' 104
209' 10
589' 10 3
190'10 3
232'10 2 ,
370' 10 1
455' 10'
149 10'
680' 105
393' 105
277'105
206' 105
953' t04
443' 104
131' 104
438' 103
130' 103
178' t0 2
267' 10'
764
137
279 . 10'
127' 107
474' 10'
230' 10'
145'10'
100' 10'
401 105
147' lOS
474' 104
213' 10
117' 10
547 103
183' 10 3
206' 102
321 10'
165 .
605'
247'
112'
498'
250'
592'
230'
645'
193'
511
325'
535
135
307
10'
10'
to'
10'
105
105
to
to
t0 3
103
10 2
101
211 10'
819' 10'
322'10'
108' 10'
357 105
169t05
664 t04
614' 104
458' 104
240' 104
682' 103
165' 10 2
577
10- 1
321
654 10- 1
r,
r.
r.
979 103
794' t0 3
544 103
287 103
156' 103
654' 10'
223' 10'
996' 10'
271 10'
605
230
772 Hr'
437' IIT I
101 10- 1
ltO'lo-'
255' 10- 3
211 10- 3
160' 1(f3
913' 10"4
171 1<f"
276' 10
20t 10
117' 10
391 103
180t03
96t 10'
309' to'
926' 10'
161 10'
232
781 10-'
540' 10-'
477' l(J-1
282'10- 1
133' 10- 1
369' 10-'
903' 10-4
393' 10- 5
243' 10- 5
649' lIT'
119'
778'
344'
618'
204'
647'
969'
212'
301
304'
412'
697'
877'
731
398'
115'
463'
162'
104'
278'
278' 10'
251 10'
212' 10'
148' 10'
939' 10'
500 ]0'
225' 10'
t04' 10'
289 105
667' 104
262' 104
907' 10 3
515' 10 3
119' 10 3
129' 102
302' 10 1
250' 10 1
189' 10 1
108' 10 1
470' 10'
396' 10'
301 10'
187' t0 7
128' 107
115' lOB
829' 10'
487' 10'
200' 10'
102' 10 1
399 10'
828 105
206' 105
307' 104
318 103
439' 102
828 10'
104' 102
0
'"
00
105
10
to
10 3
103
10'
10'
10'
Hr'
10-'
to- 3
tcr 3
10- 3
to- 3
W- 3
10- 5
IIT 5
10- 5
1(f'
'"trt
Z
0
:>-
203
783' 10'
287' to'
920' 105
167' 105
251' 104
894' t0 3
639' 103
565'103
333'103
157' 10 3
437' 102
107' 101
466' 10- 1
287' 10- 1
768' Hr2
865' 10'
470' 101
136' 10'
548' 10- 1
192'IIT I
123' 1IT I
329' 1er2
The Reactor Operating Time Was 10 hr at 1 wan

100
200
400
100' 10 1
200' 10 1
400' 10 1
100' 102
200' 10 2
400' 10'
700' 102
100 103
200' 10 3
400' 10 3
100' 104
200' 104
400 104
100' 105
200 105
400' 10 5
100' 106
722' 10'
658' 10'
580' 10'
463' 10'
366 10'
271 10'
170' 10'
111 10'
676' 10'
408' 10'
281 10'
126' 10'
594' 106
236.10 6
112' 10 6
526' 105
182 105
644 104
199' 10
558' 103
659' 10'
572 10'
481 10'
362' 10'
272' 10'
190 - 10'
111'10'
680' la'
362' 10'
185 10'
115' 10'
421 106
167' 106
668 lOS
339' 105
158 10 5
550' 104
238' 104
995' 103
237 103
780' 10'
686' 10'
577' la'
439' 10'
3J9' 10'
242' 10'
129' 10'
702' 10'
373' 10'
209 10'
138 10'
582' 106
278' 106
117' 10 6
533' lOS
208' 10 5
588 10
189 104
231 103
370' 10'
352 10'
334 10'
306 . 10'
248 . 10'
185 10'
121 10'
672' 106
451 106
325' 106
237 10 6
179' 106
873 lOS
424 lOS
128 105
433' 104
129' 104
177 103
267 . 102
764 10 1
137' 10 1
151' 10'
144 10'
136'10'
122' 10'
107' 10'
846' 10'
481 10'
278 10'
179 - 10'
119 '10'
847' 106
359' 106
139' 10 6
465 10 5
212 10 5
117'105
546' 104
183' 104
205' 103
321 102
199'10'
17110'
141 10'
101 10'
696' 10'
448' 10'
241 10'
139' 10'
744 . 106
350' 106
186' 106
522' lOS
220' lOS
628' 104
190'104
506' 103
324' 102
534' 10'
135' 10'
307
161' 10'
147' 10'
127'10'
995' 10'
793' 10'
577' 10'
301 10'
153 '10'
620 . 106
241' 106
130' 106
645' lOS
612' lOS
453' lOS
238' 10 5
675' 104
164'10 3
577 10'
320' 10 1
654
43610'
407 10'
36410'
28910'
220' 10'
155' 10'
898' 10 6
453' 10 6
143 106
410' lOS
196' 105
851' 104
495' 104
115' 104
126' 10 3
302 102
250 102
189' 102
108'102
203' 10'
670' 10'
592 10'
492 10'
360' 10'
277' 10'
192' 10'
871 106
326' 106
706 iQs
143 . 105
754' 104
635' 104
557' 104
328' 104
155 104
432' 10 3
106' 102
465
287
768 . 10- 1
122 la'
904' 10'
55810'
259' 10'
148' 10'
712' 106
215' 106
672 105
117' lOS
134' 104
220' 10 3
857' 10 2
104' 10 3
857 102
466 102
135' 102
545
192
123
329 10- 1
400' r~
700 102
100 103
200' 103
400' 103
100' 104
200 104
400 10
100 lOS
200 )05
400 105
100 106
930' 10'
866 10'
787 10'
669 10'
569' 10'
469 10'
354 10'
276 10'
205 10'
150'10'
120'10'
732' 10'7
440 '-10'7
203' 10'
103' 10'
S04 106
178-10'
635' 105
197' lOS
556' 10
751 10'
663' 10'
572 10'
452' 10'
360' 10'
275' 10'
188' 10'
134 10'
870' 10'7
562' 10'
413' 10'
220' 10'
121 10'7
583' 106
314 106
151 106
540' lOS
235' 105
990 10
237' 10
886 . 10'
792 10'
683' 10'
544' 10'
442' 108
341' 10'
221 10'
151 10'
103 10'
724 10'7
564' 10'
341' 10'
210' 10'
100' 10'
486' 106
196 106
576 lOS
186' 105
228' 104
370' 103
475' 10'
456' 10'
428 10'
369 10'
305 10'
240 10'
180' 10'7
149' 10'
123' la'
986' 106
816 106
516' 106
297' 106
103' 106
383' 105
119' lOS
170' Hr'
261 101
760' 102
136' 102
209 10'
202' 10'
194' 10'
108' 10'
164'10'
140 10'
100' 10'
748 10'
571 10'7
427' 10'
335' 10'
185' 10'
967' 106
391 106
197' 106
113'106
537' lOS
180' 105
202' 104
321 103
215 10'
187'10'
157'10'
117' 10'
852' 10'
596' 10'7
372 10'
247' 10'7
151 10'7
867 106
575' 106
6
280' 10
152' 106
498' lOS
167 105
462' 10
308' 103
524 102
134' I~
306 101
'"'"
01
The Reactor Operating Time Was 100 hr at 1 watt

100
200
400
100' 10'
200' 10'
400' 10'
100' 102
200 102
175' 10'
161.' 10'
142' 10'
114' 10'
933' 10'
710' 10'
418' 10'
251 10'
139' 10'
879' 106
716' 106
6
614' 10
579 10 6
409' 106
215'106
611 105
149 104
572 102
319' 101
651 10 1
464 - 10'7
435 - 10'7
392' 10'
316' 10'
246' 10'7
179 10'
110' 10'
602 - 106
247' 106
U8' 106
853' 105
567 105
343' 10 5
805 1<t
965' 103
299' 103
249' 103
188' 103
108' 103
202 102
685 10'7
607 10'
506' 10'
374' 10'
291' 10'
205' 10'7
973 106
405' 106
132' 106
703' 105
620' 105
572 lOS
486 lOS
288' 105
140' lOS
391 104
963' 102
462 101
286' 10 1
765
123' 10'
906 10'
559 10'
261 10'7
150' 10'7
725' 10 6
224' 106
717'10'
134' 105
52
><
:
226' 10
107' 104
999 103
103' 104
781 103
422' 10 3
122'10 3
515'10 1
191' 10'
123'10 1
"8
0
'"
'"
Table A.2-(Continued}
r,
Time
100
200
400
100. 10 1
200 10 1
400' 10 1
100. 102
200 102
400 102
700 102
100' 103
200' 103
400 103
100 10
200 10
400 10
100' lOS
200' lOS
400' lOS
100' 106
108' 109
102' 109
938' 10'
820'10'
720' 10'
619' 10'
503'10'
424' 10'
349 10'
290' 10'
255'10'
197' 10'
150' 10'
953' 10'
616 10'
361 la'
148 10'
557' 106
186 10 6
544 lOS
794' 10'
707' 10'
615' 10'
496 . 10'
403' 10'
319'10'
231' 10'
176'10'
129' 10'
966 10'
806' 10'
580' 10'
433' 10'
285' 10'
186' 10'
107' 10'
460' 106
214' 10 6
958 10'
864 10'
755' 10'
61, 10'
514' 10'
413' 10'
292' 10'
221 10'
171 la'
139' 10'
12i 10'
928' 10'
708' 10'
438' 10'
259' 10'
128' to'
480' 106
156' 106
943' 10 5
229 lOS
199' 10 5
366' 104
r,
r,
r,
The Reactor Operating Time Was 1000 hr at 1 '!:itt

553' 10'
535 10'
507 10'
447 10'
383' 10'
317' 10'
257 10'
225' 10'
196' 10'
169 10'
148' 10'
109 10'
748 . 106
362 106
175' 106
675 10 5
121 10 5
217 '10 4
722 103
130' 103
240' 10'
233' lOll
225' 10'
211 10'
194 10'
171 10'
131' la'
105 10'
863' la'
709' la'
607' 10'
431 10'
305 10'
189 . 10'
131 10'
897' 106
455 ]'06
151 106
175'10 5
319' 104
219' 10'
191'10'
161 10'
121 10'
889' 10'
634 10'
409 10'
284' 10'
187' 10'
1"20 . 10'
895 106
554 106
360' 106
162' 106
718' lOS
224 lOS
209' 10
443' 103
126'103
295' 102
201 10'
187' 10'
168 10'
140' 10'
119' 10'
96"8 10'
676 10'
507 10'
392' 10'
336' 10'
316' 10'
291' 10'
260' 10'
179' 10'
950' 106
271 106
696' 104
534 10 3
306 103
624 1.Q2
471 10'
442' 10'
398 10'
322' 10'
253' 10'
186' 10'
116'10'
666' 106
307' 106
175'106
138 10 6
984. 10 5
609 lOS
162 10 5
408' 10
282' 10
238' 10
180' 104
103' 104
193' 103
Definitions of quantities in Table A.2.

Time
D
B
rT
r,
r,
r,
r,
r,
r,
r,
Time after fission or reactor operation (sec)

Disintegration rate (sec-I)
Beta-ray energy release (MeV sec-I)
Gamma-ray energy release (MeV sec-I) for gamma-ray energies above 0.1 MeV
Gamma-ray energy release (MeV sec-I) for gamma-ray energies from 0.1 to 0.4 MeV
Gamma-ray energy release (MeV sec-I) for gamma.:ray energies from 1.8 to 2.2 MeV
Gamma-ray energy release (MeV sec-I) for gamma-ray energies> 2.6 MeV
tData for this table were.obtained from]. F. Perkins, Report RR-TR--63-11, U. S. Army Missile Command.]uly 1963.
704 '.10'
626' 10'
525' 10'
393' 10'
309' 10'
224 10'
116' 10'
590' 106
314' 106
248 106
235' 106
218' 106
186 10 6
119' 10 6
612 10 5
174' 10 5
456' 10 3
437' 102
276'10 2
738' 10 1
123' 10'
907' 10'
560' 10'
261 10'
150 . 10'
730' 10 6
229' 10 6
765' 10 5
182' 10 5
-698' 104
573'104
540 104
495' 104
350' 104
188' 10
556' 103
352 102
184 10~
118' 102
316' 10 1
..,..,
Z
"'
:>-
S2
:><
:>-
N.eutron Fluence-To-Kerma
Conversion Factors
for Standard-Man Model
Appendix Bt
Table B.l presents neutron fluence-to-kerma conversion factors calculated
by Ritts, Solomito, and Stevens for a standard-man model. The neutron
energies range from approximately 0.022 eV to approximately 15 MeV, and
the compositions of the models are shown in Table B.2. Table B.l gives the
conversion factors for the'standard man together with those for the elements
making the greatest contributions in the model. Table B.3 shows the neutron
reactions considered for each element.
t All data in this appendix have been taken from J. J. Ritts, E. Solomito, and P. N. Stevens. Calculations of Neutron Fluence-to-Kerma Factors for the Human Body, USAEC Report ORNL-TM-2079,
Oak Ridge National Laboratory, January 1968.
611
612
APPENDlXB
Table B.l-(Continued)
Energy,
M,V
9.45
9.30
9.15
9.00
8.85
8.70
8.62
8.47
8.32
8.17
8.02
7.87
7.72
7.57
7.42
7.27
7.12
6.97
6.82
6.67
6.52
6.37
6.21
6.06
5.91
5.76
5.61
5.46
5.31
5.16
5.01
4.69
4.50
4.31
4.16
4.01
3.82
3.67
3.52
3.33
3.18
3.03
2.84
2.69
2.50
2.40
2.31
2.21
2.10
2.00
1.91
1.80
1.70
1.59
1.50
1.40
1.31
1.20
Fluence-to-kerma factor!. (erg$D/(neutrc~~s cm-::)

Nitrogen
Total
Hydrogen
Oxygen
Carbon
5.271(-7)
5.222(-7)
5.173(-7)
5.123(-7)
5.048(-7)
5.007(-7)
4.988(-7)
5.036(-7)
5.062(-7)
4.456(-7)
4.443(-7)
4.430(-7)
4.417(-7)
4.410(-7)
4.402(-7)
4.399(-7)
4.391(-7)
4.383(-7)
4.376(-7)
5.40(-8)
5.00(-8)
4.74(-8)
4.78(-8)
4.54(-8)
4.37(-8)
4.01(-8)
3.90(-8)
4.38(-8)
4.24(-8)
2.19(-8)
2.23(-8)
2.14(-8)
1.75(-8)
1.61(-8)
1.56(-8)
1.56(-8)
1.57(-8)
1.66(-8)
2.15(-8)
4.97(-9)
4.91(-9)
4.85(-9)
4.70(-9)
4.64(-9)
4.60(-9)
4.53(-9)
4.40(-9)
4.28(-9)
4.30(-9)
4.977(-7)
5.038(-7)
5.060(-7)
4.935(-7)
4.970(-7)
4.995(-7)
4.773(-7)
4.650(-7)
4.745(-7)
4.707(-7)
4.368(-7)
4.345(-7)
4.319(-7)
4.292(-7)
4.266(-7)
4.239(-7)
4.211(-7)
4.187(-7)
4.173(-7)
4.159(-7)
2.94(-8)
3.81(-8)
3.89(-8)
3.65(-8)
4.11(-8)
5.32(-8)
4.13(-8)
3.13(-8)
4.14(-8)
3.94(-8)
2.63(-8)
2.62(-8)
3.09(-8)
2.34(-8)
2.49(-8)
1.80(-8)
1.08(-8)
1.10(-8)
1.21 (-8)
1.17(-8)
4.74(-9)
4.38(-9)
3.88(-9)
3.84(-9)
3.98(-9)
3.87(-9)
3.52(-9)
3.49(-9)
3.27(-9)
3.27(-9)
4.605(-7)
4.731(-7)
4.611{-7)
4.572(-7)
4.633(-7)
4.457(-7)
4.502(-7)
4.312(-7)
4.294(-7)
4.449(-7)
4.145(-7)
4.130(-7)
4.115(-7)
4.100(-7)
4.077(-7)
4"'8(-7)
4.018(-7)
3.987(-7)
3.956(-7)
3.925(-7)
2.97(-8)
3.89(-8)
2.46(-8)
2.77(-8)
3.73(-8)
2.32(-8)
3.08(-8)
1.47(-8)
1.70(-8)
3.46(-8)
1.25(-8)
1.71(-8)
2.09(-8)
1.53(-8)
1.42(-8)
1.38(-8)
1.37(-8)
3.38(-9)
3.57(-9)
1.33(-8)
3.77(-9)
3.67(-9)
3.51(-9)
3.55(-9)
3.25(-9)
3.39(-9)
4.17(-9)
4.344(-7)
4.180(-7)
4.116(-7)
4.181(-7)
4.245(-7)
4.102(-7)
4.121(-7)
4.099(-7)
4.060(-7)
3.981(-7)
3.727{-7)
3.555(-7)
3.499( -7)
3.396(-7)
3.892(-7)
3.810(-7)
3.760(-7)
3.709(-7)
3.666(-7)
3.623(-7)
3.562(-7)
3.511(-7)
3.459(-7)
3.392(-7)
3.337(-7)
3.280(-7)
3.203(-7)
3.138(-7)
2.76(-8)
1.79(-8)
1.66(-8)
2.91 (-8)
3.37(-8)
2.24(-8)
2.91(-8)
3.05(-8)
3.09(-8)
3.46(-8)
2.06(-8)
1.53(-8)
1.14(-8)
1.10(-8)
1.32(-8)
1.50(-8)
1.42(-8)
1.33(-8)
1.89(-8)
2.05(-8)
2.26(-8)
2.42(-8)
2.45(-8)
4.01(-9)
3.73(-9)
4.34(-9)
4.48(-9)
4.98(-91
4.78(-9)
3.88(-9)
3.67(-9)
4.48(-9)
3.279(-7)
3.150(-7)
3.136(-7)
3.119{-7)
3.057(-7)
2.995(-7)
3.054(-7}
3.000(-7)
2.950(-7)
2.898(-7)
2.835(-7)
2.780(-7)
2.952(-7)
2.883(-7)
2.800{-7)
2.717(-7)
2.644(-7)
2.577(-7)
2.602(-7)
2.426(-7)
2.721(-7)
2.648(-7)
2.585(-7)
2.507(-7)
2.440(-7)
2.370(-7)
2.297(-7)
2.209(-7)
5m8(-7)
1.42(_8)
1.30(~8)
3.65(-9)
2.07(-8)
3.32(-9)
1.54(-8)
3.40(-9)
7.36(-9)
1.08(-8)
1.14(-8)
1.13(-8)
1.60(-8)
1.26(-8)
1.09(-8)
1.03(-8)
9.85(-9)
9.51 (-9)
9.68(-9)
9.16(-9)
2.83(-9)
2.47(-9)
2.01(-9)
1.89(-9)
1.34(-9)
1.07(-9)
1.21(-9)
1.61(_9)
9.62(-10)
9.20(-10)
1.31(-8)
1.31(-8)
1.21(-8)
1.18(-8)
1.15(-8)
1.13(-8)
2.23(-8)
1.39(-8)
8.82(-9)
8.56(-9)
8.36(-9)
8.12(_9)
7.93(-9}
7.73(_9)
7.52(-9)
7.25(-9)
9.82(-10)
1.72(-9)
8.76(-10)
9.84(-10)
9.061-10)
1.66(-9)
6.44(-10)
5.11(-10)
1.01 (-8)
9.44( -9)
Phosphorus
---_.~--_
Sodium
..
_.-
Chlorine
613
APPENDIXB
Table B.l-(Continued)
Energy,
M<V
~_l~~lIce-to-kerma factor, (ergs If I )/(neutron~~~

~------
Total
Hydrogen
Oxygen
Carhon
Nitrogen
1.10
2.371(~7)
8.0(-1)
7.0(-1)
6.0(-1)
5.0(-1)
1.980(-7)
1.857(-7)
7.00(-9)
6.70(-9)
6.35(-9)
5.97(-9)
5.56(-9)
5.07(-9)
4.52(-9)
7.45(-10)
2.432(-7)
2.121(-7)
2.037(-7)
1.939(-7)
1.63(_8)
1.00
9.0(-1)
4.5(-1)
4.0(.1)
3.5(-1)
3.0(-1)
2.5(-1)
2.0(_1)
1.5(-1)
1.0(-1)
7.0(_2)
5.0( ..2)
1.639(-7)
2.90{-10)
2.96(-10)
2.46(-10)
2.13(-10)
1.93(-10)
1.82(-10)
7.30(-9)
4.20(-9)
3.89(-9)
3.51(-9)
3.12(-9)
2.70(-9)
2.23(-9)
1.75(-9)
1.22(-9)
8.76(-10)
6.37(-10)
2.122(-7)
1.719(~7)
1.592(-7)
1.834(-7)
1.730(-7)
1.609(-7)
1.483(-7)
3.21(-8)
1.15(-8)
7.95(-9)
6.61(-9)
5.61(-9)
5.74(-9)
1.223{-7)
1.142(-7)
1.104(-7)
9.649(-8)
8.165(-8)
6.335(-8)
4.973(-8)
3.830(-8)
1.037(-7)
9.1l6{-S)
7.758(-8)
6.056(-8)
4.774(-8)
3.705(-8)
2.560(-8)
1.807(-8)
4.43(-9)
2.98(-9)
2.18(-9)
1.47(-9)
1.03(-9)
7.53(-9)
5.94(-11)
4.43(-11)
2.98(-11)
2.22(-11)
3.92(-10)
2.66(_10)
1.95(-10)
1.34(-10)
9.24(-11)
6.74(-11)
5.32{-11)
3.96(-11)
2.67(-11)
1.99(-11)
4.97(-11)
3.73(-11)
3.14{-1l}
2.29(-11)
1.87(-11)
1.56(-11)
1.44(-11)
1.32(-11)
1.27(-11)
1.30(-11)
2.88(-12)
3.02(-12)
3.12(-12)
3.20(-12)
6.52(-9)
4.73(-9)
3.74(-9)
2.89(-9)
2.16{-9)
1.45(-9)
1.02(-9)
Chlorine
3.39(-10)
3.90(-10)
4.97(-10)
2.80(-10)
5.32(-10)
1.339{-7)
1.18(~8)
Sodium
6.33C-IO)
1.403(-7)
1.318(-7)
1.236( -7)
1.486(-7}
1.90(-8)
Phosphorus
1.47(-10)
1.08(-10)
9.36(-11)
7.62(-11)
3.0(-2)
2.0(-2)
1.5(-2)
1.0(_2)
7.0(-3)
5.0(-3)
4.03(-3)
3.00(-3)
2.01(-3)
1.50(-3)
9.550(-9)
6.776(-9)
4.880(-9)
3.951(-9)
2.964(-9)
2.002(-9)
1.495(-9)
2An{-B)
1.747{-8)
1.316(-8)
9.287(-9)
6.558(-9)
4.832(-9)
3.820(-9)
2.855(-9)
1.928(-9)
1.436(-9)
1.03(-3)
7.68(-4)
5.13(-4)
4.02(-4)
3.10(-4)
2.01(-4)
1.42(-4)
1.006(-4)
7.1(-5)
5.03(-5)
1.030{-9)
7.774(-10)
5.299{-10)
4.244(--10)
3.347(-10)
2.299(-10)
1.764(-10)
1.417(-10)
1.192(-10)
1.071(-10)
9.832(-10)
7.368(-10)
4.925{-10)
3.867(-10)
2.986(-10)
1.939(-10)
1.372{-10}
9.753(-1l )
6.920(-11)
4.94{-1l)
1.52(-11)
1.14(-11)
7.60{-12)
5.96(-12)
4.60(-12)
2.98(-12)
2.11(-12)
1.49(-12)
1.05(-12)
7.46(-13)
1.36(-11)
1.02(-1l )
6.80(_12)
5.34(-12)
4.12{_12}
2.67(-12)
1.88(-12)
1.33{-12)
9.42(-13)
6.67(-13)
1.40(-11)
1.52(-11 )
1.77(-11)
1.95(-11)
2.19(-11)
2.67(-11)
3.15(-11)
3.73(-11)
4.42{-1l)
5.237(-11)
3.25(-12)
3.27(-12)
3.30(--12)
3.31(-12)
3.32{-12)
3.33(-12)
3.33{-12)
3.34(-12)
3.34(-12)
3.34(-12)
3.50(-13)
4.11(-13)
2.72(-13)
1.84(-12)
3.28(-12)
1.97/-12)
5.91(-14)
1.12(-13)
3.87(-13)
9.84(-14)
1.10(-13)
4.11(-5)
3.43(-5)
2.74(-5)
2.05(-5)
1.60(-5)
1.26(-5)
8.0(-6)
4.0(-6)
2.0(_6)
1.0(-6)
1.036(-10)
1.025(-10}
1.033(-10)
1.077(-10)
1.1431-10)
1.230(-10)
1.454(-10)
1.965(-10)
2.723(-10)
3.811(-10)
4.07(-11)
3.43(-11)
2.78(-11)
2.13(-11)
1.71(-11)
1.40{-11)
1.01(-11)
7.20(-12)
6.63(-12)
7.59(-12)
6.09(-13)
5.09(--13)
4.07(-13)
3.05(-13)
2.38{-13)
1.86{-13)
1.19(-13)
5.94(-14)
2.97(-14)
1.49(-14)
5.45(-13)
4.56(-13)
3.64(-13)
2.73{-13)
2.13(-13)
1.67{-13)
1.06(-13}
5.32(-14)
2.66(-14)
1.33(-14}
5.786(-11)
6.323(-11)
7.065(-11)
8.155(-11)
9.226(-11)
1.041(-10)
1.302(-10)
1.836(-10)
2.591 (-10)
3.656(-10)
3.34(-12)
3.34(-12)
3.35{-12)
3.35(-12)
3.35(-12)
3.35(-12)
3.35(-12)
3.35(-12)
3.35(-12)
3.35(-12)
4.54(-13)
4.99(-13)
5.58(-13)
6.44(-13)
7.34(-13)
8.34(-13)
1.04(-12)
1.46(-12)
2.09(-12)
2.97(-12)
1.31(-13)
1.57(-13)
1.95(-13)
2.59(-13)
3.14(-13)
3.79(-13)
5.08(-13)
7.24(-13)
1.03(-12)
1.45(-12)
5.0(-7)
2.5(-7}
1.7(-7)
7.2(-8)
4.5(-8)
2.23(-8)
5.354(-10)
7.529{-10)
1.144(-9)
1.398(-9)
1.766(-9)
2.489(-9)
9.87(-12)
1.35(-1l}
2.03(-11)
2.48(-11)
3.14(-11)
4.18(-11)
5.159(-10)
7.280(-10)
1.109(-9)
1.356(-9)
1.715{-9)
2.424(-9)
3.35(-12)
3.35(-12)
3.35(-12)
3.35(-12)
3.35(-12)
3.35(-12)
4.15(-12)
5.89(-12)
9.00(-12)
1.10(-11)
1.39(-11)
1.83(-11)
2.01(-12)
2.12(-12)
2.12(-12)
2.12(-12'
2.12(-12)
2.12(-12)
1.361(-8)
tThe standard-man mOdel also inclUdes the elements calcium, sulfur, potassium, and magnesium, but their
contributions to the kerma are negligible.
:tRead: 7.079 X 1 (f7, etc.
Table B.2-ELEMENTAL PERCENTAGES AND NUMBER DENSITIES USED IN NEUTRON KERMA CALCULATIONS
Number density. 1024 atoms!g
Percent composition
Standard
Element
H
0
C
N
Co
P
S
K
Na
C1
Mg
man
Lung
Mu.de
Bone
10
60
24
2.9
1.2
1.1
0.24
0.20
0.20
0.20
0.03
10
75
10
2.6
0.011
0.11
0.24
0.20
0.18
0.25
0.0096
10
75
11
2.6
0.0031
0.18
0.24
0.30
0.16
0.18
0.019
7.1
39.5
22.4
4.7
12.37
13.16
0.41
tThe values for
0.08
0.28
Brain
10
71
15
1.9
0.0086
0.34
0.17
0.30
0.18
0.23
0.Q15
Red
Standard
marrow
man
Lung
Muscle
Bone
Brain
marrow
10
74t
12.5t
3.2
5.977(-2J*
2.259(-2)
1.204(-2)
1.25(-3)
1.77(-4)
2.14(-4)
4.51(-4)
3.08(-5)
524(-5)
3.40(-5)
7.43(-6)
6.062(-2)
2.864(-2)
5.09(--3)
1.13(-3)
1.68(-6)
2.17(-5)
4.57(-5)
3.13(-5)
4.78(-5)
4.31(-5)
4.78(-6)
5.977(-2)
2.824(-2)
5.518(-2)
1.118(-3)
4.66(-7)
3.50(-5)
4.51(-5)
4.62(-5)
4.19(-5)
3.06(-5)
4.71(-6)
4.50(-2)
1.49(-2)
1.13(-2)
2.05(-3)
1.87(-3)
2.57(-3)
7.69(-5)
5.977(-2)
2.674(-2)
7.52(-3)
8.17(-4)
1.29(-6)
6.61(-5)
3.19(-5)
4.62(-5)
4.72(-5)
3.91(-5)
3.72(-6)
5.977(-2)
2.7866(-2)
6.270(-3)
1.377(-3)
0.12
oXJJi:n and carbon were increased by 1 and 0.5%, respectively.
tRead: 5.977 X l , etc.
Red
2.08(-5)
6.87(-5)
...~
'gZ"
~
'"
2.039(-5)
615
APPENDIXB
Table B.3-REACTlONS CONSIDERED IN NEUTRON KERMA

CALCULATIONS
Type of reaction
Radiative capture
Elastic scattering
Inelastic scattering
(n,p) reaction
(n,ex) reaction
(n,n 3a:) reaction
(n,2n) reaction
(n,d) reaction
(n,t) reaction
(n,20:) reaction
(n,n?) reaction
(n,n a) reaction
x
x
x
x
xt
x
x
x
x
x
x
x
x
x
x
xt
x
x
x
tIncludes beta decay after reaction.

:j:lncludes positron decay.
Ca
x
x
Na
cl
Mg
x
x
x
xt
xt
x
x
x
x
x
x
x
x
xt
x
x
x
x
x
x
x
x
x
xt
xt
x
x
x
x
xt
Cylindrical Tissue Phantom
Appendix C
Table C.I-MAXIMUM ABSORBED DOSE AND MAXIMUM
DOSE EQUIVALENT FOR MONOENERGETIC NEUTRONS
INCIDENT ON A CYLINDRICAL TISSUE PHANTOMt
Maximum
Effective
MeV
mead
neutron-I cm-2
Maximum
dose
equivalent,:/:
mrem
neutron-I cm-2
Thermal
0.000001
0.00001
0.0001
0.001
0.01
0.1
0.5
1
2.5
5
7
10
14
4.68(-7)
5.89(-7)
5.18( -7)
4.45(-7)
4.32(-7)
4.34(-7)
8.02(-7)
1.81(-6)
3.01(-6)
3.99(-6)
5.72(-6)
5.70(-6)
7.25(-6)
8.31(-6)
1.15(-6)
1.34(-6)
1.21(-6)
1.01(-6)
8.85(-6)
9.92(-6)
4.86(-6)
1.89(-5)
3.26(-5)
3.50(-5)
4.41(-5)
4.03(-5)
4.31(-5)
6.15(-5)
2.46
2.27
2.34
2.27
2.04
2.29
6.06
10.4
10.8
8.77
7.70
7.07
5.94
7.40
Neutron
energy ,
absorbed
dose.*
quality
factor
tBased on data from J. A. Auxier, W. S. Snyder, and T. D.

Jones, Radiation Dosimetry, Vol. I, Chap. 6, F. H. Attix and
W. C. Roesch (Eds.). Academic Press. Inc. New York. 1968.
*Multiply by 3600 to convert to (mrad hr- I )/(neutron
cm-2 sec) or to (mrem hr-1 )/(neutron cm-2 sec).
Read: 4.68 X 10-'. etc.
616
Systems~ Vector
Operations~ and Legendre
Coordinate
Polynomials
Appendix D
0.1
COORDINATE SYSTEMS
The three common spatial coordinate systems and the basic unit vectors of
each are illustrated in Fig. 0.1 and Table 0.1.
z
t __>p
I
I-~
SPHERICAL
CYLINDRICAL
CARTESIAN
Fig. D.l-Principal coordinate systems.
Table D.I-POSITION VARIABLES

System
Independent variables
x,y, z
p,tJ,z
r, 0, t/J
Cartesian
Cylindrical
Spherical
Unit vectors
i, j, k
P, 6, k
T,
0.2 CooRDINATESYSTEM TRANSFORMATION

(a) Cartesian-Cylindrical
x; p cos 0
p; (x 2 + y2)%
617
6, '"
618
APPENDIXD
cos /I = x(x 2 + y2 r~
y=psin/l
ax = ap = cos /I =x(x
ap
ax
2 + y2
r~
~
ap =sin /I =y(x 2 + y2 r ~
ap =ay
ax = -sm
. /I = -y; ax
a/l = -y (2
a/l
x
ay
- = P cos /I
a/l
a/l
=x;=x(x'
~
+ Y ")-1
+ y'
= -p1.sm /I
r =-1p cos /I
1
(b) Cartesian-Spherical
x
y
=T sin /I cos '"

=T sin /I sin'"
sin'" = y(x 2 + y'
z = T cos /I
ax
aT
= (xl + y' + z' )~

cos /I =z(x' + y' + Z2 r ~
T
r~
+ z, r~
=ax
aT =sin /I cos'" =x(x' + y'
=sin /I sin'" =y(x' + y' + z, r~

aT =.!:.
ay
ay
az aT
~
-=-=cos/l=z(x' +y' +z')aT az
ax
a/l
= T cos /I
a/l
-aX = xz(x'
+ y' r~ (x' + y' + z, r 1 = -T cos /I cos '"
ay =T cos /I sin'"
a/l
"
cos'" = xz(x' + y' t~
=yz(x'
+ y'
a/l
2
2
-a
y = yz(x + y' r~ (x + y2
: ; = -T sin /I = _(x 2
+ y2)~
rl;
1
+ z2 r 1 = -T cos /I sin '"
APPENDIXD
Z2 (X 2
3e
3z
r'
+ y2 + Z2
(x 2 + y2) II
-7
3Vi =
_ (x2 +
3x
.
stn
1 .
--stn
r
e stn
.." = -y
3x=
-
3Vi
-1
619
II'
2r'
= _ ysin
Vi)
r ~sin e
3y
3Vi = r sin e cos Vi = x

Vi )
3Vi_(22)_,_1(COS
-xx+y
---3y
r sin e
3z
3Vi
= 3Vi =0
3z
D.3 VECTOR OPERATORS AND FUNCTIONS

In the following, F(u,v,w) is a vector function of the form
F(u,v,w) = F, (u,v,w) u + F 2 (u,v,w)
+ F 3 (u,v,w)
in which Fi(u,v,w) is a scalar function of the three variables u, v, and w,

which have the corresponding basis unit vectors u, v, and w, respectively.
(a) Gradient Operator V
Cartesian: V F(x,y,z) =
(~~ i + ~~ j + ~: k)
(b) Divergence Function V' F

.
3F
3F 2 3F 3
Cartesian' ""F(xy z) = -t + +3z. v
"
3x
3y
.
.
1 (3F t
3F
3F 3 )
Cylmdncal: VF(p.e,z) = - - + - 2 + p p 3p
30
3z
620
APPENDIXD
Sphencal' V-P(r
.
1 aF
1
aF
e '/J) =aF"1
_ + __2 +---~
ar
r ae
r sin e a'/J
"
(c) Laplacian Operator 1/2
(1/2
F = V- VP)
a2 F -a2-F +a2-F
Cartesian: 1/2 P(x,y,z) = --+
ax 2 ay2
az 2
1[a
1 F)
2
a 2F
Cylindrical: 1/2 F(p,e,z) = - - (aF)]
p+ - (a
- - +-P ap
ap
p2 ae 2
az
. I
./.
Sp henca : 1/ F(r,e,,,) - r 2
[a-a;'{( a;:aF)]
2
+,2 sin e
aF)]
aea (.sm eae
(d) Vector Derivative
ap(x,y,z)
at
Similar relations exist for the other coordinate systems.
(e) Divergence Theorem

Let V be a closed and bounded volume as in Pig. D.2 with P a
vector function everywhere continuous and differentiable in V.
Then
Iv V-PdV=!s
P-ndA
where dv = differential-volume element of V

S = surface of volume V
dA = differential-area clement of S
n = outward-directed unit normal of dA
(f) Some Miscellaneous Theorems
(1) P(u-v) = (uF)-v = u-vF
(2)V-uF = F(V-u) + u-W:
(3)V(P I F 2 ) + 1"1 (VF 2 ) + F 2 (VF I )
(4) V 2 (1'1 F 2 ) = PI V 2 F 2 + F 2 V 2 P j + 2(VF j -VF 2 )
APPENDlXD
621
\1
Fig. D.2-Divergence theorem .
'/ /
./
Volume V enclosed
within surface S
(5) u' (v + w) = (u'v) + (u'w)

0.4 DIRAC DELTA FUNCTION
(a) The Dirac delta function, b(x-xo, Y-Yo, 00 0)' is defined such that
b(x-xo, Y-Yo, 0 00)
0; x oFXo, Y oFYo, 0 0 0
=l;x=xo,Y=Yo,'"
It has the property that
X
/ X
_ (;2 000 ji(x,y, 0 00) b(x-xo, Y-Yo, 00 0) d x dYo 00
o+<. /Yo+<,
o f
Yo
= f(xo, Yo, 000)
where ; > 0 for i = 1, 2,
(b) Miscellaneous relations:
(1) b(x-xo) = b(-x-xo)

(2) b(cx-xo) =
Icl- I b(x-xo)
(c is a constant)
(3)b[(x-xd,(x-X2)] =IXI-x21-1 [b(x-xd+b(x-X2)]
(5)~b(x) =.!.dx
(6)
lim
1fk_~
(k cos kx
X
kX)
sin
x2
/"0+<
b'(x--xo)f(x) dx = - ['(x)lx
xo-e
.
622
APPENDIXD
(7) o'(x-xo) = -o'(xo-x)
(8) o(x)
=.!1fk-""oo
lim r
J
(9)
1
o(y)dy=--
(x
Z
J
=+Z
=.!1fk-H'O'
hm (sin kX)
x
cos kx dx
< 0)
(x> 0)
(10) Lt/ln(X) t/lti(y) = o(x-y), where the set t/ln(x) forms an ortho{;ormal set and t/lti(x) is the complex conjugate of t/ln (x).
(11) Let the vector r

Then
orr)
= o(x)
= xi
+ yj + zk where i, j, and k are unit vectors.
o(y) o(z)
-1-1n
(Z11")'
e iQ " dn
and
J.b orr) f(r)
dl = frO)
(a";;; 0";;; b)
= 0 otherwise
(c) The Kronecker delta function is a discrete analog of the Dirac delta
function and is defined for integers m and n by
o(m,n,) = 1
(m = n)
= 0 otherwise
0.5 LEGENDRE POLYNOMIALS

(a) The Legendre polynomials Pn(x) of degree n are defined by the
generating relation
(1- Zxl + 12 r~ =
It can be shown that
Pn(x) =
1:
koO
ak x
1:
n=O
L (__
[n/2J
Pn(x) In
koO
1jk[Zn-Zk]ixn- 2k
Znk!(n - k)!(n - Zk)!
where [n/Z] = n/Z for even nand (n - 1)/Z for odd n.

(b) Po (x)
=1;P1 (x) =x;P2 (x) =(3x 2 -
l)/Z;P,(x)
=(5x 3
3x)/Z
APPENDIXD
623
(c) Miscellaneous relations:

(1)
f'
2
_[ Pm(x) Pn(x) dx = 2n + 1 o(m,n) (orthogonality relation)
1
(2) x Pn (x) = 2n + 1 [(n + I )P n+1 (x) + n Pn+l (x) I

(recursion relation)
(3) Pn(x) = l
1T
r
0
[x + (x 2
1)~ cos eJ de
(integral form)
(4) Pn(x) = (_I)n Pn(-x)

1 d(n)
(5) Pn(x) = -n2I d n (x 2
n. x
.
l)n (Rodrigues formula)
(6) For f(x) piecewise continuous in [-1,1] and having a first

derivative there,
[(x) =
an Pn(x) (expansion theorem)
n=O
1) L, f(y) Pn(y) dy
2n +
where an = ( - 2 -
D.6 ASSOCIATED LEGENDRE POLYNOMIALS

(a) The associated Legendre polynomial of the first kind of degree n
and order m is defined by
(1_/1 2 )m/2 d(n+m)
PW(/1) = - 2 n ,
d/1 n+m
n.
(2) Letting /1 = cos
e,
(0 <;; m <;; n)
624
APPENDIXD
PW(cos 0) = sin m 0
d(m)
d cosm 0
Pn(cos 0)
(4) Two recursion relations are:

1
p.PW(p) = 2n + 1 [(n + m)P:;'~l(p) + (n - m + 1)P:;'+11
pw(p) =
(12~
:T"
[l"ilW (p)
-l'::'~ I
(p)]
(5) The set PW(p) are orthogonal:
-I
2
(n + m)!
PW(p) Pl"(p) dp = 2n + 1 (n _ m)! 6 (n,l)
Let G(O) be a function continuous over the unit sphere and

having second-order derivatives there. If 0 and tf; are the polar
and azimuthal angles, respectively, of 0, then G(O) = G(O,tf;)
may be expanded as
(6)
G(O) =
[an,o Pn(p)
n=O
k=l
(an k cos ktf; + bn k sin ktf;)P,;(p)]

J
where p = cos 0
(dO
=sin 0 dO do/; = -dp dtf;)
(7) Letw = 0'0', where O(O,tf;) andO'(O',tf;') are two unit vectors.
The useful addition theorem is an expansion of the Legendre
polynomial in terms of the associated functions:
APPENDIXD
n
Pn(W)=
I:
m=-n
625
F 2 (n,m)P:;'(IJ)P:;'(IJ')e im (w-w') = Pn(lJ)Pn(JJ')

n
I:
m=l
+2
p2 (n,m) P:;'(JJ) P:;'(JJ') cos m(1ji -lji')
where JJ = cos 0, JJ " = cos 0, and F(n,m) = [(n - m)!]%

(n + m)!
D.7 ASSOCIATED SPHERICAL-HARMONIC FUNCTIONS
(a) With the unit vector fI,(O,Iji) defined as above, the associated spheriical harmonics of degree I and order m are defined by
PI
m
(fI,) = (_I)m F(l,m) eim>J; [Sin
,m
211!
where IJ = cos 0, F(l,m)
=[i~:
d(l+m) (u 2 _ 1)1]
dJJ{I+m)
:?:r,
and (0';; m';; I)

(1) PI,m(fI,) = (_I)m F(l,m) pr(lJ) eim>J;
(2) PI,m(fI,) = (_I)m F(I,m) PY' (IJ) (cos mlji - i sin mlji)
(3) The complex conjugates are thus given by

Pt.m(fI,)
= (_I)m F(l,m) pr(lJ) e-im>J;
,
Pt.m(fI,)
,
=(_l)m
F(I,m) pr(lJ) (cos mlji - i sin mlji)
(4) In the degenerate case where m = 0

P',O (fI,) = P{o (fI,)
=P'l (IJ) =P,(IJ)
Pl,_m(fI.) = (_I)m Pt,m(fI,)
(5) There are two orthogonal properties:
140 P',m(n) Pj,k(fI,) dfl,=
2::
4rr
1 3(1,j) 3(m,k) 3(m,0)

.
f 4n P',m(fI,) Pf,k(fI,) dfl, = 21 + 1 3(I,J) 3(m,k)

(6) The addition theorem can be written as an expansion of Pn(w)
in spherical harmonics:
n
Pn(w) =
I:
m=-n
Pn,m (fI,) Pit,m (fI,')
Exposure Buildup Factors
Appendix E
Table E.1 contains exposure buildup factors. plots for a given initial energy
and penetration depth (/10 r) vs. atomic number can be used for interpolation.
Table E.I-EXPOSURE BUILDUP FACTOR (B,) POINT
ISOTROPIC SOURCE!
!J.o'
Eo.
MeV
Water
0.255
0.5
1.0
2.0
3.0
4.0
6.0
8.0
10.0
3.09
2.52
2.13
1.83
1.69
1.58
1.46
1.38
1.33
7.14
5.14
3.71
2.77
2.42
2.17
1.91
1.74
1.63
Aluminum
0.5
1.0
2.0
3.0
4.0
6.0
8.0
10.0
2.37
2.02
1.75
1.64
1.53
1.42
1.34
1.28
[ron
0.5
1.0
2.0
3.0
4.0
6.0
8.0
10.0
1.98
1.87
1.76
1.55
1.45
1.34
1.27
1.20
Material
10
15
20
23.0 72.9
14.3 38.8
7.68 16.2
4.88
8.46
3.91 6.23
3.34 5.13
2.76 3.99
2.40 3.34
2.19 2.97
166
77.6
27.1
12.4
8.63
6.94
5.18
4.25
3.72
456
178
50.4
19.5
12.8
9.97
7.09
5.66
4.90
982
334
82.2
27.7
17.0
12.9
8.85
6.95
5.98
4.24
3.31
2.61
2.32
2.08
1.85
1.68
1.55
9.47 21.5
6.57 13.1
4.62 8.05
3.78 6.14
3.22 5.01
2.70 4.06
2.37 3.45
2.12 3.01
38.9
21.2
11.9
8.65
6.88
5.49
4.58
3.96
80.8
37.9
18.7
13.0
10.1
7.97
6.56
5.63
141
58.5
26.3
17.7
13.4
10.4
8.52
7.32
3.Q9
2.89
2.43
2.15
1.94
1.72
1.56
1.42
5.98 11.7
5.39 10.2
4.13 7.25
3.51 5.85
3.03 4.91
2.58 4.14
2.23 3.49
1.95 2.99
19.2
16.2
10.9
8.51
7.11
6.02
5.07
4.35
35.4
28.3
17.6
13.5
11.2
9.89
8.50
7.54
55.6
42.7
25.1
19.1
16.0
14.7
13.0
12.4
626
627
APPENDIX E
Table E.1-(Continued)
iJ.o T
Eo.
Material
MeV
Tin
0.5
1.0
2.0
3.0
4.0
6.0
8.0
10.0
1.56
1.64
1.57
1.46
1.38
1.26
1.19
1.14
2.08
2.30
2.17
1.96
1.81
1.57
1.42
1.31
3.09
3.74
3.53
3.13
2.82
2.37
2.05
1.79
4.57
6.17
5.87
5.28
4.82
4.17
3.57
2.99
0.5
1.0
2.0
3.0
4.0
6.0
8.0
10.0
1.28
1.44
1.42
1.36
1.29
1.20
1.14
1.11
1.50
1.83
1.85
1.74
1.62
1.43
1.32
1.25
1.84
2.57
2.72
2.59
2.41
2.07
1.81
1.64
Lead
0.5
1.0
2.0
3.0
4.0
5.1
6.0
8.0
10.0
1.24
1.37
1.39
1.34
1.27
1.21
1.18
1.14
1.11
1.42
1.69
1.76
1.68
1.56
1.46
1.40
1.30
1.23
Uranium
0.5
1.0
2.0
3.0
4.0
6.0
8.0
10.0
1.17
1.31
1.33
1.29
1.24
1.16
1.12
1.09
1.30
1.56
1.64
1.58
1.50
1.36
1.27
1.20
Tungsten
10
15
20
6.04
8.85
8.53
7.91
7.41
6.94
6.19
5.21
8.64
13.7
13.6
13.3
13.2
14.8
15.1
12.5
18.8
19.3
20.1
21.2
29.1
34.0
33.4
2.24
3.62
4.09
4.00
4.03
3.60
3.05
2.62
2.61
4.64
5.27
5.92
6.27
6.29
5.40
4.65
3.12
6.25
8.Q7
9.66
12.0
15.7
15.2
14.0
(7.35)
(10.6)
14.1
20.9
36.3
41.9
39.3
1.69
2.26
2.51
2.43
2.25
2.08
1.97
1.74
1.58
2.00
3.02
3.66
3.75
3.61
3.44
3.34
2.89
2.52
2.27
3.74
4.84
5.30
5.44
5.55
5.69
5.07
4.34
2.65
4.81
6.87
8.44
9.80
11.7
13.8
14.1
12.5
(2.73)
5.86
9.00
12.3
16.3
23.6
32.7
44.6
39.2
1.48
1.98
2.23
2.21
2.09
1.85
1.66
1.51
1.67
2.50
3.09
3.27
3.21
2.96
2.61
2.26
1.85
2.97
3.95
4.51
4.66
4.80
4.36
3.78
2.08
3.67
5.36
6.97
8.01
10.8
11.2
10.5
(6.48)
9.88
12.7
23.0
28.0
28.5
tFrom H. Goldstein and J. E. Wilkins, Jr., Calculations of the Penetrations of

Gamma Rays, Final Report, USAEC Report NYO-3075, Nuclear Development
Associates, Inc., June 30, 1954.
Coefficients for Gamma-Ray

Buildup Factors
Appendix F
Table F.I-COEFFICIENTS fOR LINEAR, QUADRATIC, AND BERGER FORMS OF
EXPOSURE BUILDUP FACTORS FITIED OVER THE RANGE 0 TO 7 MEAN FREE PATHS
FROM POINT ISOTROPIC SOURCE t
Linear
Quadratic
Maximum
error ,t
Material
[:, MeV
A,
A,
Water
0.255
8.6524
4.6800
1.9953
1.0301
0.7397
0.5884
F3.1
-0.2525
0.6684
1.0053
0.8242
0.6962
0.5801
0.4616
0.3782
0.3251
0.5
1
2
3
4
6
Aluminum
0.3406
10
0.2877
0.5
1
2
3
4
6
8
10
Iron
0.5
1
2
3
4
6
8
10
Tin
0.4321
0.5
1
2
3
4
6
8
10
F2.3
40
10
3
1
3
4
4
2.6461
1.6089
0.9686
0.7197
0.5663
0.4334
0.3476
0.2847
Ft.5
1.4283
1.2373
0.8556
0.6691
0.5403
0.4297
0.3391
0.2681
25
20
12
0.5153
0.7199
0.6731
0.5837
0.5146
0.4153
0.3317
0.2550
3
6
30
13
5
3
2
1
2
7
8
8
8
8
11
12
17
17
16
Berger
Maximum
error,
1.0688
0.9316
0.7437
0.6355
0.5284
0.4142
0.3346
0.2715
b
1.4984
0.6750
0.1666
0.0346
0.0073
0.0014
-0.0050
-0.0063
-0.0063
0.2654
Maximum
error,
30
1.7506
0.2609
1.3245
1.0622
0.2078
0.1052
2
2
1
1
1
1
1
0.8093
0.0408
0.6876
0.0125
0.0024
-0.0126
-0.0214
-0.0257
0.5800
0.4655
0.3860
0.3342
%
10
5
3
1
1
1
1
1
1
0.0378
0.0142
0.0064
0.0032
0.0022
0.0022
2
2
2
1
1
1
1
1
1.2435
0.9589
0.7267
0.6294
0.5253
0.4177
0.3371
0.2752
0.0486
0.0227
0.0127
0.0061
0.0050
0.0055
3
3
2
1
1
1
1
1
0.8642
0.8026
0.6526
0.5338
0.4366
0.3237
0.2473
0.1785
0.0949
0.0731
0.0342
0.0228
0.0175
0.0178
0.0154
0.0151
1
1
5
1
1
1
1
1
0.9081
0.8214
0.7020
0.5323
0.4366
0.3271
0.2563
0.1876
0.0752
0.0684
0.0319
0.0384
0.0358
0.0457
0.0464
0.0592
2
2
3
1
1
1
1
1
0.5479
0.6153
0.5455
0.4284
0.3420
0.2082
0.1371
0.0945
-0.0055
0.0176
0.0215
0.0261
0.0290
0.0348
0.0327
0.0270
2
1
1
1
1
2
2
2
0.5608
0.6219
0.5498
0.4379
0.3583
0.2369
0.1692
0.1232
-0.0146
0.0244
0.0338
0.0479
0.0601
0.0925
0.1103
0.1190
1
1
1
1
1
1
1
1
0.1140
628
0.1250
0.0864
629
APPENDIX F
Table F.l--(Conrinued)
-----~uadratic
Linear
--------
Material
Tungstell
Lead
--~
B,McV
0,5
I
2
J
4
6
,
10
o.2On
0.5
1
0.1549
0.2990
0.3796
0.3810
0.3523
0.3219
0.3034
0.2419
0.19.1.3
4
5.1
6
8
10
UI'JmUlll
0.1003
0.3817
0.4376
0,4171
0.4054
0.3363
0.2624
0.5
1
2
3
O.:~692
0.41 (>4
0.351 S
0.2540
0.1435
O,426<:l
17
16
14
JJ
15
15
13
O.IOS;
7
7
J
0.2264
7
l2
n
11
Maximum
error,
-- 0.0133
.0.0076
2
2
1
1
1
2
2
2
0.2938
0.4425
~O.O751
2
1
1
1
1
1
1
1
0.0036
0.01l 0
0.0255
0.0324
0.0281
0.0223
O,(J957
O.074R
8
6
1
5
10
O.30::D
10
8
5
12
0.316<)
0.3010
0.2571
0.2081
0.1621
4
6
8
<,
---- -------_.-
'X,
/1 1
Bergcr
Maximum
<'nor,
,\'l.l:..illt>."tl
error ,:i
0.4172
-0.0255
0.0080
0.3501
0.0295
0.2710
C.1771
0.1245
0.0666
0.1049
0.0974
0.2273
0.361.1
0.3787
0.3164
0.2389
0.1747
0.1 346
0.0894
0.(J642
--0.0122
--0.0105
0.2526
0.3779
0.0001
1
1
1
1
0.1637
0.2990
0.3250
0.27('0
0.21')')
0.1.314
0.08RS
0.06.3J
-0.0098
0.0122
-0.0038
0.0069
0.01}1;
0.0212
0.0201
0.0165
0.0109
0.0191
0.0248
0.0284
0.0257
0.0217
3
3
2
2
1
1
1
1
I
0.1223
0.1238
-0.0848
--0.0403
0.3862
0.0032
0.3267
0.0253
0.2530
0.1936
0.1622
0.1220
0.0939
0.0547
0.0839
0.1027
0.1112
0.1167
0.1825
0.3204
0.3321
0.2814
0.2283
0.1476
0.1081
0.0798
-0.0951
--0.0599
-0.0162
0.0196
0.0458
0.0916
0.1076
0.1163
2
1
1
1
1
1
1
2
2
2
2
]
1
1
1
1
]
----_._-_.- ----_._--------- ------_._-
tFr0m D. K. Trub"y, A Survey "f Enlpiriul FUllctions Used 1',) rit Gamrna-Kav Buildup Factors, USAEC Report
ORNLRSIC-1 0, Oak Ridge National L.. [wratory. Fchnl,"y 1966.
:1:1'3.1 m"allS "factor of 3.1. etC.
Table P.l --COEFFlCliNTS FOR LINEAR, QUADRATIC. AND BERGER FORMS

OF EXPOSURE BUILDUI' FACTORS fiTTED OVER THE RANGE 0 TO:W MEI\N FREE PATHS
FROM POINT ISOTROPIC SOURCES!
------- ..._---_._- Berger
Quadratic
------------ -------_. __._-- - - - -
Linear
Material
W'ltCI
A,
'XI
I'12
F5.6
36.1015
13.0926
3.4 7R8
1.2549
0.7863
0.5951
0.4030
0.3085
0.2584
7
7
E, MeV
0.255
05
1
2
10
F2
25
6
I
Maximum
error,
Maximum
error,
MJ>.il1llll11
crrur 'f
.-1 2
-"12.9947
-0.9744
1.1152
0.9173
0.7218
0.5')07
0.4471
0.3551
0.3002
b
3.0515
0.8743
0.1469
0.0210
0.0040
0.0003
..0.0027
--0.0038
-0.0026
%1
F~
2.5048
F3
6
6
2
1.8035
1.2282
0.1623
0.1224
0.0649
0.8594
0.0240
0.7004
0.5826
0.0074
0.0014
0.4853
-0.0082
-O.ot24
2
2
3
0.3741
0.3206
------
-0.0139
30
25
11
5
2
1
1
2
630
APPENDIX F
Table F.2_(Continued)
Lincar
Maximum
---_.Quadratic
error ,t
Material
Aluminum
Iron
E,MeV
Tungsten
0.3150
0.0883
0.Q197
20
'0
1.4412
1.0831
0.7869
0.6812
0.5503
0.4252
0.0078
3
2
2
0.6504
0.3395
0.0019
0.0021
F1.9
F1.6
10
0.1143
0.0189
0.0225
0.5
1
2
3
4
6
8
,0
0.5090
0.8495
0.8521
0.8509
0.8643
1.0786
1.1907
1.1075
4
18
25
40
F1.58
F2.2
F2.8
F3.0
0.5150
0.6666
0.5826
0.4254
0.2845
-0.1374
-0.4693
_0.6523
-0.0005
0.0114
0.0168
0.0264
0.0360
0.0756
0.1032
0.1094
0.5
1
2
3
4
6
0.1550
0.3382
0.4671
0.5919
0.8102
1.2616
1.3753
1.2730
13
9
0.2206
0.4149
0.4072
0.3255
0.0995
-0.5462
-0.9399
-0.9502
-0.0054
-0.0048
0.0037
0.0165
0.0442
0.1124
0.1439
0.1382
0.1791
0.3133
0.3695
0.2990
0.1449
_0.1480
-0.5070
-1.l40E.
-1.0279
-0.0047
-0.0036
0.0015
O.ot33
0.0306
0.0624
0.1014
0.1589
0.1408
0.1262
0.2556
0.3185
0.2614
0.1621
._0.2492
-0.5357
-0.6560
-0.0037
-0.0053
-0.0022
0.0091
0.0210
0.0658
0.0912
0.0980
0.5
1
2
3
4
5.1
6
B
10
Uranium
0.6696
1.1185
0.8751
2.3773
1.8643
1.1194
0.8446
0.6942
0.6134
0.5245
0,4759
10
Le,d
F3
0.5
0.5
2
3
4
6
8
'0
0.1043
0.2549
0.3947
0.5123
0.6378
0.8560
1.1247
1.4165
1.2370
0.0812
0.1914
0.2838
0.4081
0.4991
0.8088
0.9323
0.9203
A,
5.7374
2.5385
1.1928
0.8061
0.6075
0.4626
0.3697
0.3087
2
3
4
6
Maximum
0.5
1
2
3
4
6
8
10
1
Ti"
A,
Berger
Maximum
error,
F1.9
30
12
6
6
5
5
33
25
25
34
40
50
30
F1.8
F3
F3.5
F3.8
15
"
3
30
F1.6
F2.1
F2.8
F4
F4
""
5
20
43
F2.3
F2.9
F3.3
0.2750
0.0036
0.0023
0.9019
0.0917
0.921"2
0.7423
0.0586
0.5840
0.4605
0.3201
0.2207
0.0234
0.0162
0.0145
0.0182
%!
,,
error,
D
0.0850
0.0535
0.0266
0.0137
0.5343
0.4182
0.0082
0.3366
0.0058
0.2738
0.0074
0.9814
0.0548
0.0460
0.0063
%
12
9
6
3
2
1
1
0.0428
0.0553
7
7
4
4
3
3
2
1
0.5457
0.6378
0.5678
0.4533
0.3700
0.2401
0.1669
0.1190
-0.0063
0.0180
0.0254
0.0388
0.0518
0.0891
0.1145
0.1278
3
3
3
3
3
2
1
5
0.2692
0.4279
0.4163
0.3484
0.2727
0.1704
0.1161
0.0882
-0.0477
_0.0150
0.0070
0.0324
0.0653
0.1160
0.1405
0.1510
5
2
4
2
1
2
6
7
-0.0500
-0.0211
0.0021
0.0279
0.0557
0.0883
0.1143
0.1440
0.1513
5
4
FF-
0.2243
0.3530
0.3791
0.3244
0.2526
0.1904
0.1554
0.1075
0.0824
1
2
4
12
12
5
5
2
2
6
F2
F5
F-
0.1635
0.2991
0.3240
0.2781
0.2273
0.1426
0.1004
0.0721
-0.0606
_0.0385
--0.0084
0.0234
0.0475
0.1011
0.1274
0.1442
5
5
2
1
1
2
5
7
3
7
6
4
2
1
3
6
3
3
3
1
5
40
F3
F12
5
2
3
2
15
F4
FF5
5
2
2
10
40
F4
0.8932
0.7173
0.5571
0,4518
0.3381
0.2603
0.1902
0.0277
0.0261
0.0268
0.0368
,,
tFrom D. K. Trubey, .'\ Survey of Empirical Functions Used To Fit Gamma-Ray Buildup Factors, USAEC Report
ORNL-RSIC-I0, Oak Ridge National Laboratory, Fcbruary 1966.
tFl2 means "factor of 12," etc.
Tahle F.3-BERGER FORMULA CONSTANTS FOR ENERGYDEPOSITION BUILDUP,

POINT ISOTROPIC SOURCE, 20-MEAN-FREE-PATH PITt
Aluminum
Water
Iron
-----
E,McY
0.2S5
0.5
1
2
3
4
6
8
10
Maximum
error, %
2.4039
1.7128
1.2443
0,8829
0.7506
0.5905
0.4550
0.1598
0.1201
0.0660
0.0243
0.0065
0.0012
-0.0079
-0.0121
-0,0133
30
24
12
5
2
1
1
2
2
0.3757
0.3036
1.7237
1.2204
0.8480
0.6746
0.5411
0,3995
0.3079
0.2492
0.0902
0.0553
0.0268
0.0134
0.0079
0,0060
0.0060
0.0074
Maximum
error, %
14
10
6
3
2
1
1
1
1.7887
1.2192
0.7868
0.5847
0.4494
0.2975
0.2076
0.1565
-- -_.
-----------_.
D
0.0550
0.0466
0.0280
0.0255
0.0281
0.0353
0.0420
0,0528
Maximum
error, %
7
8
4
4
2
3
2
1
__ ._-
1.1873
1.0819
0.7490
0.4909
0.3352
0.1931
0.1237
0.0930
Tin
... ---_._---M;J.ximum
D
error. %
-0.0069
0.0134
0,0245
0.0380
0.0534
0.0820
0.1030
0.1148
4
3
4
9
8
2
2
4
Lead
...
_-------------
Maximum
error, %
0.4551
0.7027
0.5510
0.3476
0.2227
0.1646
0.1167
0.0781
0.0588
-0.0753
--0.0278
-0.0014
0.0251
0.0532
0.0818
0.1060
0.1269
0.1327
10
7
3
2
1
3
5
8
15
tFrom D, K. Trubey, A Survey of Empirical Functions Used To Fit Gamma-Ray Buildup Factors, USAEC Report ORNL-RSIC-IO, Oak Ridge National
Laboratory, February 1966.
:
~
~
"'
632
APPENDIX F
Table FA-PARAMETERS a, b, AND c FOR ENERGY-DEPOSITION
BUILDUP F ACTOR IN THE EQUAnON t
B(/lx) = 1.0 + a(/lx) + b(/lx)' + c(/lx)'
FOR A POINT ISOTROPIC SOURCE
Zof
material
Parameters for energy of
0.5 MeV
1 MeV
2 MeV
4MeV
6 MeV
8 MeV
10MeV
0.37
0.312
0.39
0.36
0.34
0.31
0.34
0.32
0.30
0.28
0.26
0.29
0.26
0.24
0.22
0.20
Parameter a
7.5 (H,O)
2
4
6
8
0.840
0.22
0044
0.64
0.83
1.01
1.19
1.32
0.970
0.97
0.98
0.99
1.00
0.830
0.76
0.76
0.76
0.76
0.589
0.65
0.63
0.60
0.58
00460
0.381
0.54
0.51
0045
0042
0040
0.76
0.76
0.76
0.76
0.76
0.56
0.54
0.52
0.51
0043
0041
0049
0.38
0.36
0.34
0048
0046
10
12
14
16
18
1.50
1.01
1.02
1.03
1.04
1.05
20
22
24
26
28
1.55
1.59
1.62
1.64
1.64
1.06
1.07
1.08
1.09
1.10
0.75
0.75
0.75
0.75
0.74
0047
0046
0044
0043
0042
0.33
0.31
0.30
0.28
0.27
0.24
0.22
0.21
0.20
0.19
0.18
0.17
0.16
0.15
0.14
30
32
34
36
38
1.64
1.63
1.61
1.59
1.56
1.11
1.11
1.12
1.12
1.13
0.74
0.74
0.73
0.73
0.72
0041
0.39
0.38
0.37
0.36
0.26
0.25
0.24
0.23
0.22
0.18
0.17
0.17
0.16
0.16
0.14
0.13
0.13
0.13
0.12
40
42
44
46
48
1.53
1.49
1.44
1.39
1.34
1.13
1.13
1.13
1.13
1.12
0.72
0.71
0.71
0.70
0.70
0.36
0.35
0.34
0.33
0.32
0.22
0.21
0.21
0.20
0.20
0.15
0.15
0.15
0.14
0.14
0.12
0.12
0.12
0.11
0.11
50
52
54
56
58
1.29
0.24
1.19
1.15
1.10
1.11
1.10
1.09
1.08
1.06
0.69
0.68
0.67
0.67
0.66
0.32
0.31
0.30
0.30
0.29
0.19
0.19
0.18
0.18
0.17
0.14
0.13
0.13
0.13
0.12
0.11
0.11
0.10
0.10
0.10
60
62
64
66
68
1.05
1.01
0.96
0.91
0.87
1.04
1.02
1.00
0.98
0.96
0.65
0.65
6.64
0.64
0.63
0.29
0.28
0.27
0.27
0.26
0.17
0.16
0.16
0.16
0.15
0.12
0.12
0.11
0.11
0.11
0.10
0.10
0.10
0.09
0.09
1042
633
APPENDIX F
Table F.4-(Continued)
Zof
6 MeV
8 MeV
0.15
0.15
0.14
0.14
0.14
0.11
0.11
0.11
0.10
0.10
0.Q9
0.Q9
0.09
0.09
0.Q9
0.0015
0.05
0.04
0.03
0.015
-0.0035
0.05
0.04
0.03
0.015
-0.005
0.05
0.04
0.03
0.015
-0.0055
0.05
0.04
0.03
0.015
0.047
0.045
0.043
0.041
0.039
0.010
0.005
0.007
0.008
0.011
0.010
0.003
0.005
0.006
0.008
0.010
0.002
0.003
0.004
0.006
0.010
0.001
0.002
0.003
0.005
0.113
0.109
0.105
0.100
0.094
0.038
0.037
0.036
0.036
0.036
0.013
0.015
0.017
0.019
0.021
0.010
0.012
0.014
0.015
0.017
0.008
0.010
0.010
0.011
0.011
0.007
0.009
0.009
0.009
0.009
0.103
0.091
0.070
0.050
0.031
0.088
0.082
0.079
0.067
0.058
0.036
0.036
0.036
0.036
0.036
0.022
0.023
0.023
0.023
0.022
0.018
0.019
0.019
0.019
0.018
0.011
0.010
0.010
0.009
0.009
0.009
0.008
0.008
0.007
0.007
40
42
44
46
48
0.01 4
0.0
-0.013
-0.025
-0.035
0.050
0.041
0.032
0.022
0.013
0.036
0.036
0.036
0.036
0.036
0.022
0.021
0.021
0.020
0.020
0.018
0.017
0.017
0.016
0.016
0.008
0.007
0.006
0.005
0.004
0.006
0.005
0.004
0.003
0.002
50
52
54
56
58
-0.043
-0.048
-0.053
-0.056
-0.058
0.004
-0.005
-0.013
-0.020
-0.027
0.036
0.035
0.032
0.028
0.025
0.019
0.01 8
0.018
0.017
0.017
0.015
0.015
0.014
0.013
0.012
0.004
0.003
0.003
0.002
0.002
0.002
0.001
0.001
0.001
0.0
60
62
64
66
68
-0.060
-0.062
-0.063
-0.064
-0.065
-0.033
-0.037
-0.040
-0.043
-0.045
0.022
0.018
0.015
0.012
0.008
0.016
0.015
0.015
0.014
0.014
0.011
0.010
0.009
0.008
0.007
0.001
0.001
0.001
0.0
-0.001
-0.001
-0.001
-0.002
-0.003
-0.003
material
0.5 MeV
1 MeV
2 MeV
70
72
74
76
78
0.82
0.78
0.73
0.69
0.65
0.94
0.92
0.90
0.87
0.85
0.62
0.62
0.61
0.61
0.60
7.5 (H 2 O)
2
4
6
8
0.600
0.53
0.50
0.464
0.436
0.175
0.126
0.125
0.125
0.124
0.037
0.055
0.053
0.051
0.049
10
12
14
16
18
0.402
0.367
0.336
0.305
0.275
0.123
0.121
0.120
0.118
0.116
20
22
24
26
28
0.244
0.215
0.188
0.160
0.142
30
32
34
36
38
4MeV
0.26
0.25
0.25
0.24
0.24
Parameter b
10MeV
APPENDIX F
634
Table FA--(Continued)
Z of
material
70
72
74
76
78
0.5 MeV
1 MeV
2 MeV
4MeV
-0.066
-0.067
-0.067
-0.068
-0.068
-0.047
-0.048
--0.050
-0.051
-0.052
0.005
0.002
-0.002
--0.004
--0.006
0.013
0.01 3
0.012
.0.011
0.010
6MeV
0.006
0.005
0.004
0.003
0.002
. 8 MeV
10 MeV
-0.001
-0.002
-0.002
-0.003
-0.004
-0.004
-0.005
-0.005
-0.006
-0.006
Parameter c
7.5 (H2 O)
2
4
6
8
0
0
0
0
0
0
0
0
0
0
10
12
14
16
18
0
0
0
0
0
20
22
24
26
28
a
a
0
0
0.0001
0
0
0
0
30
32
34
36
38
a
0
a
a
a
0
a
a
0
0
0
0
0
0
0
0
0
0
0
a
a
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
a
a
a
a
0
a
a
0
0
0
0
0
0
0
-0.0001
0
0.0001
0
0.0001
a
0.0001
0
0
0.0001
0.0002
0.0003
0.0005
0.0008
0.0010
0
0
0.0001
0.0001
0.0002
-0.0002
-0.0002
-0.0003
--0.0004
-0.0005
0.0002
0.0004
0.0005
0.0006
0.0007
0.0002
0.0003
0.0004
0.0005
0.0006
0.0002
0.0003
0.0005
0.0008
0.0010
0.0002
0.0003
0.0005
0.0008
0.0010
40
42
44
46
48
0.0012
0.0014
0.0015
0.0016
0.0017
0.0002
0.0003
0.0004
0.0005
0.0006
-0.0005
--0.0006
-0.0006
-0.0006
-0.0007
0.0008
0.0008
0.0009
0.0009
0.0010
0.0006
0.0007
0.0008
0.0009
0.0016
0.0012
0.0013
0.0014
0.0015
0.0017
0.0012
0.0014
0.0015
0.0016
0_0017
50
52
54
56
58
0.0018
0.0019
0.0020
0.0021
0.0022
0.0007
0.0008
0.0009
0.0010
0.0010
-0.0007
-0.0006
-0.0006
-0.0006
-0.0005
0.0010
0.0010
0.0010
0.0010
0.0010
0.0010
0.0011
0.0012
0.0013
0.0013
0.0017
0.0017
0.0018
0.0018
0.0019
0.0018
0.0019
0.0020
0.0020
0.0020
60
62
64
66
68
0.0023
0.0023
0.0024
0.0025
0.0025
0.0011
0.0012
0.0013
0.0014
0.0015
-0.0004
-0.0004
-0.0003
-0.0002
-0.0001
0.0010
0.0009
0.0009
0.0009
0.0009
0.0014
0.0014
0.0015
0.0015
0.0015
0.0019
0.0020
0.0020
0.0020
0.0020
0.0020
0.0020
0.0020
0.0020
0.0020
APPENDIX F
635
Table F.4-(Continued)
P"rameters for energy of
Z of
0.5 MeV
material
70
72
74
76
78
1 MeV
2 MeV
0.0016
0.0017
0.0018
0.0018
0.0019
0.0026
0.0026
0.0026
0.0027
0.0027
0.0
0.0001
0.0002
0.0003
0.0004
4MeV
6MeV
0.0008
0,0008
0.0016
0.0016
0,0017
0.0017
0.0018
O.OOOS
0.0007
0.0007
8 MeV
10 MeV
0.0020
0.0021
0.0021
0.0021
0.0021
0.0020
0.0020
0.0020
0.0019
0.0019
tFwm M. L. Tobias. D, R. Vondy, and M, P. Lict,ke, NIGHTMARE-An IBM 7090

Code for the Calculation of Gamma Heating in Cylindrical Geometry, USAEC Report
ORNL-3198, Oak Ridge National Labora'o<y, Feb. 26, 1972.
Table F .5.1-~WATER ENERGY-DEPOSITION BUILDUP DATA t

Polynomial coefficients ~i
E,MeV
PI
Po
P,
P,
Oper:1ting Gamma Energies
0.5
1.5
2.5
3.5
4.5
5.5
6.5
7.5
8.5
9.5
9,26231(-1)
9.42191 (-1)
7.79558(--1)
6.44960(-1)
5.48872(-1)
4.78918(-1)
4.26224(-1)
4.85274(-1)
3.52605(-1)
3.25966(-1)
1.00125(0)
9.95465(-1)
9.97165(-1)
9.99216(-1)
1.00080(0)
1.00198(0)
1.00290(0)
1.00361 (0)
1.00419(0)
1.00466(0)
5.06779(-1)
8.54431 (-2)
2.32124(-2)
5.92178(-3)
--6.25612(-4)
-3.53293(-3)
-4.93742(-3)
-5.63875(-3)
-5.98227(-3)
-6.13305(-3)
1.05625(-2)
-1.40297(-3)
-5.09605(-4)
-1.60968(-4)
-1.31755(-5)
5.64411(-5)
9 .13922(-5)
1.09377(-4)
1.18446 (-4 )
1.22589(-4)
Shutdown Gamma Energies
0.4
0.7
1.0
1.5
2.0
2.5
2.8
9.96370(-1)
1.00043(0)
9.96994(-1 )
9.95465(-1)
9.96087(-1)
9.97165(-1 )
9.97823(-1 )
1.19891 (0)
8.86758(--1)
9.59381(-1)
9.42191(-1)
8.62337 (-1)
7.79558(--1)
7.34476(-1)
5.98646(--1)
3.38048(--1)
1.90309(-1)
8.54431 (-2)
4.33880(-2)
2.32124(-2)
1.59527 (-2)
2.58224(-2)
5.62186(-4)
-1.77445(-3)
-1.40297(-3)
-8.56510(-4)
-5.09605(-4)
-3.69642(-4)
tFrom M. A. Capo, Polynomial Approximation of GammaRay

Buildup Factors for a Point Isotropic Sour<.:c, USAEC Report APEX-51O,
General Electric Company, August 1959.
636
APPENDIX F
Table F.5.2-ALUMINUM ENERGYDEPOSITION BUILDUP DATA t

Polynomial coefficients
E,MeV
OJ
~o
Operating Gamma Energies
0.5
1.5
2.5
3.5
4.5
5.5
65
7.5
8.5
9.5
9.993 7 5(--1)
9.89303(-1)
9.96422(-1 )
9.99988(-1)
1.00133(0)
1.00177(0)
1.00183(0)
1.00173(0)
1.00156(0)
1.00137(0)
L30777(0)
8.80067(--l)
7.10143(-1)
5.84845( .1)
4.91214(1)
4.20424(-1)
3.65631 (-1)
3.22193(--1 )
2.87011 (-1)
2.57981(.1)
3.09928( -1)
8.17524(-- 2)
3.04919(-2)
1.31824(-2)
6.60650(-3)
3.'J6087( .. 3)
2.90324( 3)
2.68803(.3)
277497( -.3)
3.03014( -3)
4.17211(-3)
--1.45150(-3)
-6.93062(-4)
.-2.99415(-4 )
-1.34319(-4)
-6.03646(5)
-3.89701 (5)
- 3.25229(-5)
.3.57358(-5)
-4.36250(-5)
0.4
0.7
1.0
1.5
2.0
2.5
2.8
9.96852(-1 )
1.02587(0)
9.95468(.--1 )
9.89303(-I)
9.93089(--] )
9.96422(-1 )
9.97858(--1)
1.47519(0)
1.37596(0)
1.05501 (0)
3.80067( 1)
7.87367( 1)
7.IOI43( -1)
6.68707 (-])
2.95900(1 )
1.64305( I)
1.l3496( I)
8.17524(2)
4.90708( --2)
~).O,i919 (----2)
2.33 4 84(--2)
1.58034(--21
2.71874(3)
-8.37533(-4)
-1.45150(--3)
--1.04754(3)
-6.93062(-4)
5.37691 ( -4)
tFrom M. A. C'po, Polynomial /\pj'u,~:im;l.tion of Gamma-Ray

Buildup Factors for a Point Isotn)t1ic S()ur,~c, USABC Report APEX-SIO,
General Electric Company, August 19S5~,
APPENDIX F
637
Table F.5.3-IRON ENERGYDEPOSITION BUILDUP DATAt

Polynomial coefficients (3j
.---_.-_.~._--_._.
E,MeV
~I
~o
__.,.,-_..
~2
~3
0.5
1.5
2.5
3.5
4.5
5.5
6.5
7.5
8.5
9.5
1.00284(0)
1.00295(0)
1.00009(0)
9.98723(-1)
9.98231(-1)
9.98087(-1)
9.98085(-1)
9.98142(-1)
9.98223(-1)
9.98311(-1)
1.59020(0)
8.66308(-1)
6.37132(-1)
4.91866(-1)
3.92656(-1)
3.21302(-1)
2.67764(-1)
2.26206(-1 )
1.93054(-1)
1.66013(-1)
1.69181(-1)
6.05394(-2)
3.03015(-2)
2.07774(-2)
1.62387(-2)
1.35838(-2)
1.18363(-2)
1.05957(-2)
9.66759(-3)
8.94618(-3)
1.73772(-4)
-7.72469(-4)
-3.77266(-4)
-1.64953(-4)
-3.22198(-5)
5.86485(-5)
1.24756(-4)
1.75005(-4)
2.14486(-4)
2.46324(-4)
0.4
0.7
1.0
1.5
2.0
2.5
2.8
1.00480(0)
9.90035(-1)
1.00093(0)
1.00295(0)
1.00140(0)
1.00009(0)
9.99533(-1)
1.80880(0)
1.34931(0)
1.06991(0)
8.66308(-1)
7.37006(-1)
6.37132(-1)
5.87104(-1)
1.90712(-1 )
1.81651(-1)
1.12726(-1)
6.05394(-2)
4.02749(-2)
3.03015(-2)
2.65084(-2)
3.23403(-4)
-1.20219(-3)
-1.11370(-3)
-7.72469(-4)
-5.39837 (-4)
-3.77266(-4)
-3.01289(-4)
tFrom M. A. Capo, Polynomial Approximation of Gamma-Ray

Buildup Factors for a Point Isotropic Source, USAEC Report APEX~510,
638
APPENDIX F
Table F.5.4-TlN ENERGY-DEPOSITION BUILDUP DATAl

Polynomial coefficients (3i
E.MeV
~o
~l
~2
~3
0.5
1.5
2.5
3.5
4.5
5.5
6.5
7.5
8.5
9.5
1.00259(0)
1.00156(0)
9.97988(-1)
9.95935(-1)
9.96384(~1)
9.97424(-1)
9.97634(-1)
9.97014(-1)
9.95564(-1)
9.93284(-1)
1.25931 (0)
8.65069( ~1)
5.73675(-1)
4.03733(-1)
3.00490(-1)
2.35258(-1)
1.85578(-1)
1.51450(-1)
1.32875(-1)
1.29852(-1)
-3.63953(-2)
3.58198(-2)
2.88829(-2)
1.65694(~2)
1.29627(-2)
1.11403(-2)
8.58655(-3)
5.30151 (-3)
1.28515(--3)
-3.46252(-3)
1.57962(-3)
-6.85867(-4)
3.55381 (-5)
9.51860(-4)
1.20395(-3)
1.30735(-3)
1.43864(-3)
1.59782(-3)
1.78488(-3)
1.99984(-3)
0.4
0.7
1.0
1.5
2.0
2.5
2.8
1.00240(0)
1.00452(0)
1.00398(0)
1.00156(0)
9.99520(-1)
9.97988(-1)
9.97257(-1)
1.30553(0)
1.16701 (0)
1.06706(0)
8.65069(-1)
6.99078(~1 )
5.73675(~1 )
5.13412(-1)
-4.68267(-2)
-2.65822(-2)
1.52859(-2)
3.58198(-2)
3.47173(-2)
2.88829(-2)
2.50130(~2)
1.56464(-3)
2.33401 (-3)
2.68730(~4)
-6.85867(-4)
-4.44094(-4)
3.55381(-5)
3.30280( ~4)

Buildup Factors for a Point Isotropic Source, USAEC Report APEX-S1D,
APPENDIX F
639
Table F.5.5-LEAD ENERGY-DEPOSITION BUILDUP DATAt

Polynomial coefficients (3j
E,MeV
~o
~1
~2
~3
0.5
1.5
2.5
3.5
4.5
5.5
6.5
7.5
8.5
9.5
1.02446(0)
1.00591(0)
1.00341(0)
1.00230(0)
1.00046(0)
9.96828(-1)
9.93954(-1)
9.91842(-])
9.90492(-1)
9.89905(-1)
4.50964(-1)
6.85608(-1)
4.49401(-1)
2.81788(-1)
2.01870(-I)
1.7] 349(-1)
].470]8(-1)
1.28877(-1)
1.1 6925(-I)
1.11164(-1)
-4.10998(-2)
--1.89736(-2)
-2.58366(-4)
8.13507(-3)
7.30625(-3)
4.27843(-4)
-4.98319(-3)
-8.92684(-3)
-1.14031(-2)
-1.24120(-2)
1.48297 (-3)
5.62874(-4)
3.62998(-4)
4.48166(-4)
9.53997(-4)
1.64896(-3)
2.12918(-3)
2.39466(-3)
2.44540(-3)
2.28140(-3)
0.4
0.7
1.0
1.5
2.0
2.5
2.8
1.00415(0)
1.01437(0)
1.00920(0)
1.00591 (0)
1.00435(0)
1.00341 (0)
1.00300(0)
3.79448(-1)
6.15504(-1 )
7.47687(-1)
6.85608(-1 )
5.61596(-1)
4.49401(-1)
3.91690(-])
-3.68912(-2)
-5.27118(-2)
-3.85780(-2)
-1.89736(-2)
-7.47115(-3)
-2.58366(-4)
2.87674(-3)
1.36614(-3)
2.13950(-3)
1.26739(-3)
5.62874(-4)
3.80808(-4 )
3.62998(-4)
3.80735(-4)

Buildup Factors for a Point Isotropic Source, USAEC Report APEX-SIO,
640
APPENDIX F
Table F.6~COEFFICIENTSFOR BERGER BUiLDUP-FACfOR

FUNC'TION FOR POINT ISOTROPIC SOURCE TO 20 MEAN FREE PATHSt
E,MeV
Max.
error, %
Max.
C
error, %
ORDINARY CONCRETE
Exposure:!:
Tissue kerma
0.5
1.0
2.0
3.0
1.3029
1.0914
0.8126
0.6731
0.08610
0.04566
0.01980
0.00942
18
15
9
4
1.3110
1.0221
0.7744
0.6525
0.08073
0.05160
0.02456
0.01258
12
9
5
3
4.0
6.0
8.0
10.0
0.5953
0.4915
0.4164
0.3585
0.00299
-0.00159
-0.00172
-0.00005
2
3
6
9
0.5672
0.4718
0.4203
0.3716
0.00753
0.00198
0.00096
0.00127
2
1
1
1
Energy deposition:!:
Energy deposition
0.5
1.0
2.0
3.0
1.6280
1.2367
0.8850
0.7171
0.09272
0.04795
0.01987
0.00825
17
9
3
4.0
6.0
8.0
10.0
0.6154
0.4882
0.4049
0.1435
0.00199
-0.00205
-,-0.00208
-0.00041
1
3
6
9
21
1.6568
1.1633
0.8203
0.6702
0.08538
0.05325
0.02465
0.01221
13
10
0.5672
0.4583
0.00753
0.00168
0.00084
0.00083
2
1
1
1
0.3985
0.3609
5
3
MAGNETITE CONCRETE
Tissue kerma
Energy deposition
0.5
1.0
2.0
3.0
1.0971
0.9281
0.7318
0.6154
0.06458
0.04768
0.02620
0.01836
8
8
5
4
1.6607
l.J 664
0.8170
0.6386
0.06696
0.04877
0.02591
0.01747
8
8
5
3
4.0
6.0
8.0
10.0
0.5308
0.4345
0.3752
0.3245
0.01513
0.01118
0.01153
0.01199
3
2
2
2
0.5263
0.4224
0.3390
0.3006
0.01428
0.00705
0.01010
0.00999
3
3
2
2
BARYTESCONCRETE
Tissue kerma
Energy deposition
0.5
1.0
2.0
3.0
0.7002
0.7237
0.6270
0.5367
0.01624
0.03007
0.02373
0.02158
1
4
4
4
1.3542
1.1102
0.7672
0.5806
0.01056
0.02792
0.02229
0.01995
4
3
3
3
4.0
6.0
8.0
10.0
0.4645
0.3727
0.3172
0.2667
0.02122
0.02233
0.02588
0.02878
3
3
0.4689
0.3403
0.2787
0.2420
0.01948
0.02045
0.02251
0.02368
2
2
2
1
,
2
APPENDIX F
641
Table F .6-(Continued)
Mal(,
E,MeV
error, %
Max.
error, 'j;,
AIR~
Tissue kcrma
0.5
1.0
2.0
3.0
1.6001
1.1571
0.8363
0.6974
4.0
6.0
8.0
0.6081
0.5146
0.4635
0.4235
10.0
Energy deposition
0.10094
0.05749
18
0.02430
5
3
0.01017
11
0.00324
-0.00316
-0.00362
~O.OO350
0.09920
0.05687
17
0.6872
0.02407
0.01002
5
3
0.6020
0.5080
0.4567
0.4261
0.00323
-0.00289
-0.00349
-0.00333
1.5411
1.1305
0.8257
11
SAND~
Tissue kerma
Energy deposition
0.5
1.0
2.0
3.0
1.4474
1.0876
0.08932
0.05482
0.8077
0.6758
0.02503
0.01239
14
10
5
3
4.0
6.0
8.0
0.5875
0.4886
0.4334
10.0
0.3883
0.00651
0.00091
0.00009
-0.00224
2
1
1
1
1.6641
0.5844
1.1729
0.8343
0.6833
0.4734
0.09348
0.05619
0.02532
0.01212
0.00653
0.00100
0,4175
~O.OOO55
0.3800
--0.00034
16
10
5
3
2
1
1
1
WOODll
Energy deposition
Tissue kerma
0.5
1.0
2.0
3.0
1.6187
1.1676
0.1 0280
19
1.5079
0.05802
11
1.1163
0.8457
0.7051
0.02392
0.00913
5
3
0.8234
0.6949
4.0
6.0
8.0
10.0
0.6149
0.5176
0.4727
0.4328
0.00253
-0.00342
-0.00438
-0.00390
0.6122
0.5186
0.4729
0.4430
0.09921
0.05688
0.02388
0.00915
17
0.00276
-0.00343
~0.00445
-0.00403
11
5
3
1
1
1
LITHIUM HYDRIDE ~
Tissue kenna
0.5
1.0
2.0
3.0
2.0198
1.2652
0.8391
0.6607
0.13997
0.07091
0.02272
0.00546
34
16
6
2
4.0
6.0
8.0
10.0
0.5778
0.4671
0.4047
0.3676
-0.00157
-0.01017
-0.01213
-0.01135
1
2
3
3
tFrom D. K. Trubey, Gamma-Ray Buildup Factor Coefficients for Concretes and

Other Materials, Nllclear Applications and Technology, 9: 441 (September 1970).
:j:Constants based on A. B. Chilton, Nuclear Engineering atld Design, 6: 205 (1967).
Constants based on F. H. Clark and D. K. Trubey, Nuclear Applications, 4: 37
(1968).
~
Francis H. Clark, Gamma-Ray Buildup Factors for Sand, Air, and Wood (Cellulose),
Nuclear Applications, 6: 588 (1969); Francis B. K. Kam and Francis H. S. Clark, Fission
Neutron Attenuation 'nJ Gamma-Ray Buildup 10< Lithium Hydride, Nuclear
Applications, 3: 433 (1967).
GraEhs and Formulas of

Exponential Integral
Functions
Appendix G
Several exponential functions that are extremely useful in shielding
calculations are presented in this appendix. They include the exponential
function e-X and the exponential integral functions En(x), which are
expressed in integral form by
and in series form by
En(x) =
~
m:;:;O
(_ x)m
-,-----'--c_----,---:
(n-l-m)m!
+ ( 1t
xn-,
(n-l)!
(In 'Y x - An)
mo#n-l
-,
whereln-y=0.577216 ... , Al =0, A n =
~ ~,andn=0,1,2,3,. ...
m=l
The recursion relation for the exponential integrals is
In particular, the first-order exponential integral is given by

E,(x)=
oo
642
e7
-dy
APPENDlXG
643
and the series representation by

(_l)n x n
n!n
E,(x) = -In 'Y x-
---
n""O
In some publications the E, (x) function is denoted by -Ei(-x); i.e.,

E, (x) = -Ei(-x). When tables of the exponential function are used, caution
should always be exercised not to confuse -Ei( -x) with -Ei(x).
For negative arguments the E j (x) function is frequently denoted by the
symbol Ei(x), which is defined by
Y
_X
Ei(x)
==
I--
"--- dy
y
and related to E, (-x) by
-Et(x)=Ej(-x) =
J-x
dy
eYy
Graphs of e-X and E,,(x) for 11 = 1,2, and 3 arc presented j in Figs. G.1 to
G.6. The function Ei(x) is plotted 2 in Figs. G.7 to G.10.
Some approximations that are frequently used are
e-x
e-X.
(11;;'1)
- - <E" (x) ,;;; ------:;

x+n
x+n-l
e-x {
En(x) = --x- I
11
11(11 + 1)
-;+~-x"2~
_ "(11 + 1x)(3" +
[
2)] + ... }
E (x) =_ 1 +x +11 e-x

n
- X + (x +11)2
(x> 1)
. _ x(l-+2'+3+.+'"
1
2! 3!
)
Et(x)=e
(x 1;11 =0)
(x> 10)
Reference 3 contains further information abont the exponential integrals.
644
APPENDIXG
2
E,(x)
1<1'
+,
.... -+
"
-t' .
,"
10'
1:
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Fig. G.l- The functions e-x, E, (x), E, (x), and E 3 (x) for x
data tabulated in Trubey ,2)
0 to 0.7. (Plotted from
645
APPENDIXG
lrf'
10-2 LW-ll.l.L-U1LllULl.W-ll.l.Lll.Ll.-"UllWJ-ll.l.Lll.LllnL.LWJ-ll.l.LllLlllL-WJJJJ
o
2 3 4 5 6 7
x
Fig. G.2-The functions e-x, E 1 (x), E 2 (x). and
tabulated in Trubey.')
(x) for x ::: 0 to 5. (Plotted from data
646
APPENDIXG
;,
1
,,
,.
,t
, ,1,
II
1.I.1 pi
!
10....
I' ,,
:: T'
"
1 11
10
Fig. G.3-The functions e-x, E I (x), E 2 (x), and E 3 (x) for x = 3 to 10. (Plotted from data
tabulated in Trubey ,2)
APPENDIXG
647
5. . . . .
5._
10'"
x
Fig. GA-The functions e- , Et{x), E 2 (x), and E 3 (x) for x = 8 to 15. (Plotted from data
tabulated in Trubey.' )
648
APPENDIXG
IEltIHjll1
_.1'
-l
.-x
2_
10-8
!
It!!.
, t'''t1!
10-9
II
I,
lWliWliUl1iLJ:jt;flt'lli:~:l'll~llWllW~illllWllWlilifrlliJ
' !
13
14
15
t :,
16
17
18
19
20
Fig. G.5-The functions e'x, Edx), E, (x), and E 3 (x) for x = 13 to 20.
data tabulated in Trubey.2 )
(Plotted from
649
APPENDIXG
10""
2 _
5_
10- 10
21
20
22
23
24
Fig. G.6-The functions e-X, E, (x), E 2 (x), and E 3 (x) for x = 17 to 24. (Plotted from
data tabulated in Trubey.2)
650
APPENDlXG
-1
E,(xi
+1
+7
-0.1
-0.2
-0.3
-0.4
-0.6
-0.6
-0.7
Fig. G.7-The function E 1 (x) for negative arguments. (Plotted from data in Ref. 1.)
APPENDIX G
-T ut+ I, i
;r; i 'HtH
,
tr
jf
,
-\
:
I
/tf:
I'
[':::'
651
,
t
t,
E,'x)
rr
,I
II
-1
-2
-3
-5
-4
-6
-7
Fig. G.8--The function E 1 (x) for negative arguments. (Plotted from data in ReLl.)
j-
"
I
_10'
I
I
E,'x)
o -
_102
-1
-6
-9
-10
-11
-12
-13
-14
Fig. G.9-The function E 1 (x) for negative arguments. (Plotted from data in Ref, 1.)
652
APPENDlXG
_'0"
_
'i
!:! i
,07
11'1
1'
E,(x)
:: i,
:!::
-10
1:
_,0
", _
I'
.,-
~i+!I-h' ).4,4l.'L
1'+W-H+fj+Jffi+W-J+HfHJffi+++H-HfHI+H+H-HH LHc-I+ti+l
11111 ii, i I'
I I I , I,
-14
-15
-16
-17
-18
-19
-20
-21
Fig. G.I0-The function E 1 (x) for negative arguments. (Plotted from data tabulated in
Ref. I.)
REFERENCES
1. Federal Works Agency, Works Project Administration, Tables of Sine, Cosine, and
Exponential Integrals, Vals. ] and II, prepared under sponsorship of National Bureau of
Standards, Superintendent of Documents, U. S. Government Printing Office, Washington, D. C., 1940.
2. D. K. Ttubey, A Table of Three Exponential Integrals, USAEC Report ORNL-2750,
Oak Ridge National Laboratory, June 18, 1959.
3. M. Abramowitz and I. A. Stegun (Eds.), Handbook of Mathematical Functions with
Formulas, Graphs, and Mathematical Tables, National Bureau of Standards, Applied
Mathematics Series 55, Superintendent of Documents, U. S. Government Printing
Office, Washington, D. C., 1964.
Tables of Attenuation
Functions for Finite
Slab Geometry
Appendix H
This appendix consists of tabulations of attenuation functions for slab
shields and disk or rectangular sources obtained by integrating over a point
kernel of the form given for gamma rays in Sec. 4.8.1. The geometry for the
disk-source configuration is shown in Figs. 4.7 and 4.8. Data are presented in
Tables H.l and H.2, for sources with both isotropic and cosine angular
distributions, respectively, the isotropic-source data being taken from the
work of Hubbell, Bach, and Herbold,l and the cosine-source data, from the
work of Trubey.2 Corresponding data for a rectangular-source configuration
are given in Table H.3, from Hubbell, Bach, and Lamkin,3 and in Table H.4,
from Trubey.2 The geometry for the rectangular source is shown in Fig. 4.9.
The usc of these attenuation functions is explained in Sec. 4.8.1 and
summarized here.
The variables for the disk source (see Tables H.l and H.2) are
r(llt,ro/z,p/ro) = dose rate
G(E) = flux-to-dose-rate conversion factor
S = source strength (particles cm'2 sec"), 41T solid angle

Ilt = optical shield thickness (mean free paths)
rolz
= ratio
of source radius to the source-detector distance
p/ro = distance off axis (units of source radius)

Y = Ilt sec
For an isotropic source the unscattered flux attenuation for the dose ratc
on the disk axis is given by
(Text continues on page 658.)
653
...
0-
Table H.I-UNSCATTERED FLUX DENSITY FROM A DISK SOURCE

WITH AN ISOTROPIC ANGULAR DISTRIBUTlONt
en
41fr/(S G(EJI
z/ro
0.1
0.2
"'
0
0.01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
o.
0.01
0.02
0.05
0.01
0.2
0.5
1.0
2.0
5.0
10.0
0.5
o.
om
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
1.0
om
0.02
0.05
pfro - 0
p!ro ::: 0.2
p!ro = 0.5
p!ro - 0.8
plro - 1.0
p!ro - 1.2
p!ro ::: 1.5
p/ro - 2.0
plro - 5.0
p!ro = 10.0
1.45(1)!
1.39(1)
1.34(1)
1.20(1)
1.01(1)
7.38(0)
3.50(0)
1.38(0)
3.08(-1)
7.22(-3)
2.61{-5)
1.44(1 )
1.38(1)
1.33(1)
1.19(1)
1.00(1)
7.34(0)
3.50(0)
1.38(0)
3.08(-1)
7.22(-3)
2.61(-5)
1.36(1)
1.31(1}
1.26(1 )
1.13(1}
9.57(0)
7.10(0)
3.46(0)
1.38(0)
3.08(-1)
7.21(-3)
2.61 (-5)
1.15(1)
1.1 0(1)
1.06(1 )
9.54(0)
8.11(0)
6.12(0)
3.17(0)
1.32(0)
3.05(-1)
7.21(-3)
2.61(-5)
7.39(0)
7.03(0)
6.71(0)
5.86(0)
4.81(0)
3.45(0)
1.66(0)
6.63(-1)
1.49(-1)
3.54(-3)
1.29(-5)
3.60(0)
3.33(0)
3.08(0)
2.47(0)
1.76(0)
9.94(-1)
2.51(-1)
4.25(-2)
2.11(-3)
1.83(0)
1.62(0)
1.44(0)
1.02(0)
5.99(-1)
2.09(-1)
2.74(-2)
1.03(-3)
9.00(-1)
7.53(-1)
6.31(-1)
3.76(-1)
1.66(-1)
4.56(-2)
3.45(-3)
2.49(-4)
1.28(-1)
7.84(-2)
4.81(-2)
1.13(-2)
3.16(-2)
1.17(-2)
4.31(-3)
1.36(-3)
2.17(-5)
1.34(-4)
1.72(-5)
1.02(1)
9.98(0)
9.74(0)
9.04(0)
8.01(0)
6.35(0)
3.37(0)
1.37(0)
3.08(-1)
7.22(-3)
2.61(-5)
1.01 (1)
9.87(0)
9.62(0)
8.94(0)
7.92(0)
6.28(0)
3.35(0)
1.37(0)
3.08(-1)
7.22(-3)
2.61(-5)
9.42(0)
9.19(0)
8.96(0)
8.32(0)
7.38(0)
5.88(0)
3.19(0)
1.34(0)
3.06(-1)
7.21(-3)
2.61(-5)
7.65(0)
7.44(0)
7.25(0)
6.71(0)
5.93(0)
4.71(0)
2.60(0)
1.14(0)
2.80(-1)
7.08(-3)
2.61(-5)
5.37(0)
5.20(0)
5.04(0)
4.60(0)
3.97(0)
3.17(0)
1.55(0)
6.35(-1)
1.42(-1)
3.48(-3)
1.27(-5)
3.25(0)
3.11(0)
2.98(0)
2.63(0)
2.15(0)
1.49(0)
5.79(-1)
1.67{-1)
2.06(-2)
1.04{-4)
2.7l{-8}
1.77{0)
1.67(0)
1.57(0)
1.31(0)
9.78(-1)
5.66(-1)
L37(-1)
1.73(-2)
5.93(-4)
7.67(-7)
8.90(-1)
8.13(-1)
7,43{-1)
5.70(-1)
3.70{-1)
1.62(-1)
2.19(-2)
1.02(-3)
1.28(-1)
1.00{-1)
7.83(-2)
3.76(-2)
1.12(-2)
3.48(-3)
1.62(-4)
3.16(-2)
1.92(-2)
1.17(-2)
2.63(-3)
1.01(-3)
3.16(-4)
4.90(-6)
5.06(0)
4.98(0)
4.90(0)
4.68(0)
4.34(0)
3.73(0)
2.38(0)
1.16(0)
2.93(-1)
7.21(-3)
2.61(-5)
4.98(0)
4.90(0)
4.82(0)
4.61(0)
4.27(0)
3.66(0)
2.34(0)
1.14(0)
2.90(-1)
7.20(-3)
2.61(-5)
4.53(0)
4.46(0)
4.39(0)
4.19(0)
3.87(0)
3.32(0)
2.11(0)
1.03(0)
2.69(-1)
7.04(-3)
2.61(-5)
3.69(0)
3.62(0)
3.56(0)
3.38(0)
3.11(0)
2.63(0)
1.63(0)
7.78(-1)
2.03(-1 )
5.76(-3)
2.31(-5)
2.96(0)
2.90(0)
2.84(0)
2.69(0)
2.45(0)
2.03(0)
1.21(0)
5.43(-1)
1.32(-1)
3.29(-3)
1.22(-5)
2.26(0)
2.21(0)
2.16(0)
2.02(0)
1.81 (0)
1.47(0)
8.07(-1)
3.26(-1)
6.64(-2)
1.09{-3)
2.29(-6)
1.49(0)
1.45(0)
1,41(0)
1.30(0)
1.13(0)
8.69(-1)
4.0.9(-1)
1.30(-1)
1.73(-2)
9.00(-5)
2.52(-8)
8.26(-1)
7.95(-1)
7.66(-1)
6.84(-1)
5.68(-1)
3.94(-1)
1.38(-1)
2.73(-2)
1.62(-3)
6.03(-7)
1.27(-1)
1.15(-1)
1.04(-1)
7.74(-2)
4.74(-2)
1.79(-2)
1.72(-3)
1.87(-5)
3.15(-2)
2.58(-2)
2.11(-2)
1.16(-2)
4.30(-3)
2.18(0)
2.15(0)
2.13(0)
2.05(0)
2.15(0)
2.12(0)
2.10(0)
2.02(0)
1.99(0)
1.96(0)
1.94(0)
1.87(0)
1.72(0)
1.70(0)
1.68(0)
1.61(0)
1.51 (0)
1.49(0)
1.47(0)
1.41(0)
1.30(0)
1.26(0)
1.20(0)
1.20(0)
1.01{O)
9.95(-1)
9.78(-1)
9.29(-1)
6.66.1- 1)
6.52{-1)
6.38{--'1)
5.99(-1)
1.23{-1 )
1.17(-1)
1.11(-1)
9.57(-2)
6.69(-5)
3.13(-2)
2.83(-2)
2.56(-2)
1.90(-2)
>
."
."
m
Z
S2
><
:r:
0.1
0.2
0.5
1.0
2.0
5.0
2.0
1.20(0)
6.64(-1)
2.05(-1)
6.54(-3)
2.58(-5)
7.01 (.-1)
0.01
0.02
0.05
6.94(-1)
1.90(0)
1.69(0)
US{D)
6.51(-1)
2.01{-I)
6.42(-3)
2.56(-5)
6.96(-1)
6.89{-1)
6.81(-1)
1.76(0)
1.51 (OJ
1.31 (0)
l.12{O)
1.55(0)
1.07(0)
5.85(-1)
1.32(0)
8.96(-1)
4.73(-1)
1.37(-1)
4.21(-3)
1.78(-5)
1.14(0)
9.62(-1)
6.18(-1)
3.02(-1)
1.77{-1)
5.70(-3)
2.41(-5)
6.71{-1}
6.14(-1)
5.93(-1)
5.61(-1 )
5.02(-1)
3.60(-1)
2.06(-1)
4.09{-1)
5.40(-1)
3.91 (-1)
10.0
2.41(-1)
8.36(-2)
3.52(-3)
1.84(-5)
2.28(-1)
7.79(-2)
3.17(-3)
1.61(-5)
1.23(-1)
1.23(-1)
1.22(-1)
1.22{-1)
1.21{-1)
1.22(-1)
1.17(~1)
om
0.1
0.2
0.5
1.0
2.0
5.0
6.60(-1)
6.26(-1)
5.63(-1)
4.90(-1)
3.92(-1)
4.84(-1)
3.86(-1)
5.39(-1 )
4.78(-1)
5.20(-1)
4.60(-1)
4.32(-1)
3.81 (-I)
3.65(-1)
4.90(-1)
4.35(-1)
3.05(-1)
3.SI{-I)
2.96(-1)
2.62{-1)
1.95(-1)
9.74{-Z}
1.10{-1)
1.07(-1)
l.04{-I)
9.60(-2)
8.40(-2)
6.44(-2)
2.91(-2)
7.83(-3)
3.40(-1)
1.70(-1}
1.41(-1 )
4.12(-2)
1.14(-3)
3.45(-6)
2.50(-2)
6.16(-4)
4.77(-4)
2.28{-7)
3.03(-2)
2.88(-2)
2.74(-2)
2.35(-2)
1.83(-2)
1.1 D( -2)
2.41(-3)
2.35( -4)
1.19{~1)
l.17(~1)
1.14(-1)
1.07(-1)
6.28(-2)
2.52(-2)
1.18(-1)
1.12(~1)
1.06{-1)
6.27(-2)
2.51(-2)
1.21(-1)
1.20(-1)
1.18(-1)
1.11e- 1)
1.05(-1)
6.11(~2)
2.41(~2)
."
1.17(-1)
1.16(-1)
l.14(-1}
1.16(-1)
l.14(~1 )
1.11(-1)
1.01(-1)
1.10(-1)
9.96(-2)
735(-2)
4.43(-2)
1.60(-2)
7.66(-4)
4.83(-6)
5.85(-2)
5.46(-2)
2.25(-2)
2.02(-2)
1.61 (~2)
3.13(-2)
3.12(-2)
0.01
0.02
0.05
3.09(-2)
3.09(-2)
3.06(-2)
2.97(-2)
2.83(-2)
2.56(-2)
1.89(-2)
1.15(-2)
t From Hubbell et ai. I
5.52(-1)
5.45(-1)
5.57(-6)
1.29(-2)
1.20(-1)
1.19(-1)
7.43(-2)
4.48(-2)
1.63(-2)
7.87(-4)
5.05(-6)
tRead; 1.45" 10 1 , etc.
7.40(-5)
2.66(-8)
1.21(-1)
10.0
10.0
3.34(-1)
1.89(-1)
8.65(-2)
4.22(-2)
6.51(-4)
1.13(-6)
7.67{-2)
1.79(-3)
6.86(-6)
7.44(-2)
4.49(-2)
1.15(-2)
4.21(-3)
2.08(-4)
1.39(-6)
4.70{-1)
2.35(-1}
1.15{-2)
4.24(-3)
3.96(-4)
2.27(-6)
9.55(-6)
l.1l(~I)
1.89(-2)
5.85(-1)
5.78(-1)
5.58(-1)
5.27(-1)
4.52(-2)
1.02(-2)
1.06(-3)
6.65(-6)
1.24(-5)
1.01(-1)
2.83(-2)
2.56(-2)
5.91 (-1)
7.45(-2)
5.29{-2)
1.64(-2)
7.91(-4)
5.09(-6)
0.1
0.2
0.5
1.0
2.0
5.0
3.87(-1)
1.06(-1)
2.95(-3)
1.14(-5)
5.39(-1}
4.37(-1)
6.09(-2)
2.16(-3)
1.11 (--1)
2.97(-2)
7.56(-1)
8.53(-1)
7.20(-1)
4.37(-1)
1.95(-1)
2.60(-3)
1.01(-1)
3.06(-2)
6.27(-1)
6.20(-1)
6.63(-1)
6.56(-1)
6.35(-1)
6.02(-1)
6.86(-1)
6.65(-1)
6.31(-1)
5.67{-1)
4.l3{-1)
2.44(-1)
8.47(-2)
3.59(-3)
1.89(-5)
0.02
0.05
10.0
1.71 (0)
10.0
0.1
0.2
0.5
1.0
2.0
5.0
5.0
1.93(0)
3.12(-2)
J.09{--2)
3.06(-2)
2.97(-2)
2.82(-2)
2.55(-2)
6.85(-2)
1.72(-3)
1.09(-1)
1.16(-1)
1.13(-1)
1.07(-1)
9.81(-2}
9.66(-2)
7.22(-2)
4.33(-2)
1.56(-2)
7.29(-4)
4.45(-6)
7.10(-2)
6.961\~2)
4.25(-2)
4.14(-2)
1.52(-2)
1.47(~2)
6.97(-4)
4.13(-6)
6.60(-4)
3.78(-6)
1.08(-;1)
1.02(-0
9.20(-')
6.71 (-2)
3.97(-2)
1.39(-2)
5.97(-4)
3.20(-6)
3.11(-2)
3.10{-2}
3.08(-2)
3.08(~2)
3.06(-2)
3.03(-2)
2.94(-2)
2.80(-2)
3.05(~2)
3.04(-2)
2.95(-2)
2.81(-2)
2.54(-2)
1.89(-2)
1.14(-2)
1.88(-2)
4.21(~3)
4.19(-3)
2.08(-4)
1.38(-6)
2.06(-4)
1.37(-6)
4.16(-3)
2.04(-4)
1.34(-6)
1.14(-2)
2.53(-2)
1.87(-2)
1.13(-2)
1.05(-1)
9.50(-2)
9.02(~7)
9.60(-2)
8.62(-2)
6.23(-2)
4.74(~2}
3.10(-2)
1.53(~2)
3.62(-2)
1.23( ~2)
3.76(-3)
4.83(-4)
2.27(-6)
5.71 (-5)
5.99(-8)
3.06(-2)
3.01(-2)
3.03{~2)
2.98(-2)
2.51(-2)
2.48(-2)
2.45(-2)
2.37(-2)
3.02(-2)
3.00(-2)
~.95(-2)
2.93(-2)
2.79(-2)
2.91(-2)
2.76(-2)
2.52(-2)
1.86{ ~2)
2.50(-2)
1.84(-2)
2.86(--2)
2\72(-2)
2.45(-2)
8.26(-3)
2.71(-3)
2.94(-4)
3.77(-7)
~
."
'"
g
><
;I:
1.57(~2)
1.55(-2)
1.53(-2)
1.46(-2)
2.24(-2)
2.00(-2)
1.36(-2)
1.80(-2)
1.43{-2}
7.75{-3}
1.18(~2)
l.12(~2)
1.11(-2)
4.13(-3)
4.09(-3)
3.82(-3)
9.31(-4)
2.01(~4}
1.08(-2)
3.90(-3)
1.82(-4)
8.20(-3)
2.68(-3)
2.04(~4)
4.03(-3)
1.93(-4)
9.37(-5)
1.31(-6)
1.28(-6)
1.22(-6)
1.l0(-6)
3.53(-7)
1.36(-5)
1.26(-8)
a-
'"''"'
aa-
Table H.2-UNSCATIERED FLUX DENSITY FROM A DISK SOURCE WITH A

COSINE ANGULAR DISTRIBUTION (ISOTROPIC FLUX DENSITY)t
z/1'o
0.1
0.2
0.5
1.0
.'
2rll"'(0) GIE))
p/1'o '" 1.5
p/1'o '" 2.0
p/1'o - 5.0
p/1'o :: 10.0
9.65(-2)
9.10(-2)
8.60(-2)
8.13(-2)
5.65(-2)
3.57(-2)
1.12(-2)
2.16(-3)
1.19(-4)
5.05(-8)
2.85(-13)
2.79(-2)
2.51(-2)
2.27(-2)
2.05(-2)
1.06(-2)
4.55(-3)
5.01(-4)
1.98(-5)
5,43(-8)
3.72(-15)
8.56(-3)
7.25(-3)
6.15(-3)
5.23(-3)
1.77(-3)
4.24(-4)
9.12(-6)
2.73(-8)
4.86(-13)
4.18(-4)
2.57(-4)
1.59(-4)
9.80(-5)
3.59(-6)
3.73(-8)
8.01(-14)
6.52(-23)
5.06(-5)
1.87(-5)
6.96(-6)
2.59(-6)
2.75(-9)
1.84(-13)
1.13(-25)
3.83(-1)
3.74(-1)
3.66(-1)
3.58(-1)
3.09(-1)
2.54(-1)
1.50(-1)
6.95(-2)
1.79(-2)
4.82(-4)
1.87(-6)
1.54(-1)
1.49(-1)
1.44(-1)
1.40(-1)
1.12(-1)
8.37(-2)
3.84(-2)
1.25(-2)
1.77(-3)
1.01(-5)
3.73(-9)
5.25(-2)
4.98(-2)
4.72(-2)
4.48(-2)
3.14(-2)
1.94(-2)
5.34(-3)
8.10(-4)
2.77(-5)
2.90(-9)
1.64(-15)
1.68(-2)
1.54(-2)
1.42(-2)
1.31{-2}
7,41(-3)
3,42(-3)
4.07(-4)
1.66(-5)
4.61(-8)
3.20(-15)
3.84(-1)
3.78(-1)
3.72(-1)
3.67(-1)
3.30(-1)
2.84(-1)
1.85(-1)
9.37(-2)
2.64(-2)
8.07(-4)
3.42(-6)
2.81(-1)
2.77(-1)
2.72(-1)
2.68(-1)
2.39{-I)
2.03(-1)
1.28{-1)
5.20(-2)
1.65(-2)
4.57(-4)
1.80(-6)
1.86(-1)
1.82{-I)
1.79(-1)
1.75(-1)
1.53(-1)
1.27{-I)
7.48(-2)
3.29(-2)
7.38(-3)
1.37(-4)
3.07(-7)
9.50(-2)
9.27(-2)
9.04(-2)
8.82(-2)
7,43(-2)
5.85(-2)
2.94(-2)
1.02(-2)
1.49(-3)
8.77(-6)
3.29(-9)
3.70(-2)
3.57(-2)
3.47(-2)
3.33(-2)
2.60(-2)
1.85(-2)
6.85(-3)
1.46(-3)
8.58(-5)
3.79(-8)
2.17(-13)
2.06(-3)
1.87(-3)
1.69(-3)
1.54(-3)
7.77(-4)
2.96(-4)
1.72(-5)
1.74(-7)
2.54(-11)
2.39(-22)
2.15(-1)
2,12(-1)
2.09(-1)
2.07(-1)
1.89(-1)
1.79(-1)
1.76(-1)
1.74(-1)
1.72(-1)
1.56(-1)
1.43{-1)
1.41(-1)
1.39(-1)
1.37{-1)
1.24(-1)
9.85(-2)
9.69(-2)
9.54(-2)
9.38(-2)
8.37(-2)
5.18(-2)
5.08(-2)
4.98(-2)
4.88(-2)
4.23(-2)
3.93(-3)
3.74(-3)
3.56(-3)
3.38{-3)
2.39(-3)
4.98(-4)
4.51(.-4)
4.08(-4)
3.69(-4)
1.84(-4)
p/1'o '" 0
p/1'o - 0.2
p/1'o '" 0.5
P/1'o - 0.8
p/1'o '" 1.0
9.01(-1)1
8.78(-1)
8.56(-1)
8.35(-1)
7.08(-1)
5.70(-1)
3.26(-1)
1.48(-1)
3.75(-2)
9.96(-4)
3.83(-6)
8.97(-1)
8.75(-1)
8.53(-1)
8.33(-1)
7.06(-1)
5.70(-1)
3.26(-1)
1.48(-1)
3.75(-2)
9.96(-4)
3.81(-6)
8.77(-1)
8.55(-1)
8.35(-1)
8.15(-1)
6.95(-1)
5.64(-1)
3.26(-1)
1.48(-1)
3.75(-2)
9.96(-4)
3.83(-6)
7.88(-1)
7.70{-1)
7.53(-1)
7.36(-1)
6.35(-1)
5.23(-1)
3.12(-1)
1.46(-1)
3.74(-2)
9.96(-4)
3.82(-6)
4.30(-1)
4.19(-1)
4.08(-1)
3.97(-1)
3.36(-1)
2.71(-1)
1.57(-1)
7.19(-2)
1.83(-2)
4.91(-4)
1.89(-6)
0.02
0.03
0.1
0.2
0.5
1.0
2.0
5.0
10,0
8.04(-1)
7.88(-1)
7.72(-1)
7.57(-1)
6.60(-1)
5,46(-1)
3.23(-1)
1.48(-1)
3.75(-2)
9.96(-4)
3.83(-6)
7.98(-1)
7.82(-1)
7.67(-1)
7.52(-1)
6.56(-1)
5,43(-1)
3.22(-1)
1.48(-1)
3.75(-2)
9.96(-4)
3.83(-6)
7.61(-1)
7,46(-1)
7.32(-1)
7.18(-1)
6.28(-1)
5.23(-1)
3.15(-1)
1.47(-1)
3.75(-2)
9.96(-4)
3.83(-6)
6.30(-1)
6.18(-1)
6.06(-1)
5.95(-1)
5.23(-1)
4.39(-1)
2.71(-1)
1.31(_1)
3.52(-2)
9.84(-4)
3.83(-6)
0
0,01
0.02
0.03
0,1
0.2
0.5
1.0
2.0
5.0
10.0
5.53(-1)
5,45(-1)
5.37(-1)
5.29(-1)
4.78(-1)
4,14(-1)
2.71(-1)
1.36(-1)
3.67(-2)
9.96(-4)
3.83(-6)
5,44(-1)
5.36(-1)
5.28(-1)
5.21(-1)
4.71(-1)
4.08(-1)
2.67(-1)
1.34(-1)
3.64(-2)
9.95(-4)
3.83(-6)
4.94(-1)
4.87{-1)
4.79(-1)
4.73(-1)
4.27(-1)
3.70(-1)
2,43(-1)
1.23(-1)
3.43(-2)
9.79(-4)
3.82(-6)
2.93(-1)
2.89(-1)
2.86(-1)
2.83(-1)
2.60(-1)
2.88(-1)
2.84(-1)
2.81(-1)
2.77(-1)
2.55(-1)
2.61(-1)
2.57(-1)
2.54(-1)
2.51(-1)
2.31(-1)
0
0,01
0.02
0.03
0.1
0.2
0.5
1.0
2.0
5.0
10.0
0
am
om
0.02
0.03
0,1
<.A
p!1'o '" 1.2
8.35(-4)
6.55(-4)
5.13(-4)
4.03(-4)
7.51(-5)
7.13(-6)
7.94(-9)
1.57(-13)
1.25(-22)
1.01(-4)
6.15(-5)
3.75(-5)
2.28(-5)
7.24(-7)
5,49(-9)
3.19(-15)
2.24(-25)
2.52(-4)
2.06(-4)
1.69(-4)
1.39(-4)
3,46(-5)
4.80(-6)
1.35(-8)
8.95(-13)
5.82(-21)
;>
."
."
OJ
Z
0
~
:t
0.2
0.5
1.0
2.0
5.0
10.0
2.0
5.0
10.0
3.46(-2)
1.46(-3)
6.79(-5)
4.42(-2)
2.04(-2)
4.68(-3)
7.93(-5)
1.45(-7)
3.35(-4)
3.01(-5)
2.78(-7)
4.16(-13)
1.84(-22)
3.48( -6)
2.57(-8)
1.63( -12)
8.52(-25)
7.93(-7)
1.91(-2)
7.27(-3)
1.16(-3)
7.24(-6)
2.78(-9)
7.47(-2)
7.39(-2)
6.28(-2)
4.50(-2)
4,44( -2)
4.38(-2)
4.32(-2)
3.92{-2)
3.41(-2)
6.60(-3)
6.43(-3)
6.26(-3)
6.09(-3)
9.52(-4)
9.05(--4)
1.59{-1)
2.04{-1}
1.43(-1)
1.66(_1)
1.14(-1)
1.37(-1)
9.20(-2)
8.84(-2)
2.76(-2)
9.13(-4)
3.77(-6)
7.86(-2)
2,44(-2)
8.17(-4)
3.54(-6)
6.16(-2)
1.85(-2)
6.02(-4)
2.65(-6)
4.83(-2)
1.39(-2)
4.14(--4)
1.68(-6)
1.04(-1)
9.91(-2)
9.81(-2)
9.01(-2)
8.91(-2)
8.27(-2)
8.17(-2)
9.70(-2)
8.81(-2)
8.08(-2)
7.30(-2)
8.71(-2)
8.06(-2)
7.22(-2)
5.18(-2)
2.98(-2)
9.96(-3)
3.84(-4)
7.99(-2)
7.38(-2)
6.59(-2)
4.69(-2)
2.66(-2)
8.66(-3)
3.14(-4)
7.22(-2)
6.65(-2)
1.86(-6)
1,42(-6)
1.06(-1)
1.04(-1)
1.03(-1)
0.02
1.03(-1)
1.02(-1)
1.02(-1)
1.01 (-1)
9.50(-2)
8.55(-2)
6.22(-2)
3.67(-2)
1.28(-2)
5.43(-4)
9.40(-2)
8.45(-2)
6.15(-2)
3.62(-2)
9.60(-2)
8.90(-2)
7.99(-2)
5.79(-2)
3.38(-2)
1.26(-2)
1.16(-2)
5.32(-4)
2.87(-6)
2.80(-6)
4.75(-4)
2.45(-6)
0
0.01
0.02
0.03
0.1
0.2
0.5
1.0
2.0
5.0
10.0
1.94( -2}
1.92(-2)
1.9P(-2)
1.88(-2)
1.75{-2)
1.59(-2)
1.17(-2)
7.07(-3)
2.58(-3)
1.25(-4)
8.00(-7)
1.94(-2)
1.92(-2)
1.90(-2)
1.88(-2)
1.75(-2)
1.58(-2)
1.17(-2)
7.05(-3)
2.57(-3)
1.24(-4)
7.93(-7)
1.91(-2)
1.90(-2)
'-87(-2)
1.84(-2)
1.85(-2)
1.82{-2)
1.88( -2)
1.86( -2)
1.73(-2)
1.56(-2)
1.15(-3)
6.94(-3)
2.52(-3)
1.20(-4)
7.57(-7)
1.84(-2)
1.82(-2)
1.69(-2)
1.53(-2)
1.12(-2)
6.75(-3)
2.43(-3)
1.14(-4)
6.94(-7)
0
0.01
0.02
0.03
0.1
0.2
0.5
1.0
2.0
5.0
10.0
4.96(-3)
4.91(-3)
4.86(-3)
4.82(-3)
4.49(-3)
4.06(-3)
3.01(-3)
1.82(-3)
6.68(-4)
3.30(-5)
2.20(-7)
4.96(-3)
4.91(-3)
4.86(-3)
4.81(-3)
4.49(-3)
4.06(-3)
3.00(-3)
1.82(-3)
6.68(-4)
3.30(-5)
2.19(-7)
4.94(-3)
4.89(-3)
4.85(-3)
4.80(-3)
4.47(-3)
4.04(-3)
2.99(-3)
1.81(-3)
6.64(-4)
3.27(-5)
2.16(-7)
4.92(-3)
4.87(-3)
4.82(-3)
4.77(-3)
4.45(-3)
4.02(-3)
2.97(-3)
1.80(-3)
6.58(-4)
3.22(-5)
2.11(-7)
0.03
0.1
0.2
0.5
1.0
2.0
5.0
10.0
7.12{-2)
2.27(-1)
om
1.08(-1)
7.05(-2)
3.55(-2)
9.51(-3)
2.41(-4)
2.31(~1)
1.62(-1)
9.03(-2)
2.82(-2)
9.29(-4)
3.80(-6)
tFrom Trubey.2
tRead: 9.01 x 10-1 etc.
Effectively zero.
6.20(-2)
6.12(-2)
6.05(-2)
5.54(-2)
4.90(-2)
3.38(-2)
1.83(-2)
5.43(-3)
1.55(-4)
8.51(-4)
3.89(-3)
1.77(-3)
4.80(-4)
8.18(-4)
5.74(-4)
3.46(-4)
7.61(-5)
6.16(-6)
2.97(-3)
3.67(-5)
4.22{-B)
5.95( -5)
2.07(-8)
5.16(-7)
1.17(-7)
L28(-13)
1.78(-14)
7.29(-25)
1.80(-2)
1.78(-2)
1.66(-2)
1.50(-2)
1.10(-2)
6.57(-3)
2.35(-3)
1.08(-4)
6.41(-7)
1.79(-2)
1.78(-2)
1.76(-2)
1.74(-2)
1.62(-2)
1.46(-2)
1.07(-2)
6.37(-3)
2.26(-3)
1.02(-4)
5.82(-7)
1.72(-2)
1.70(-2)
1.68(-2)
1.67(-2)
1.55(-2)
1.39(-2)
1.02(-2)
6.02(-3)
2.11(-3)
9.08(-5)
4.88(-7)
1.57(-2)
1.56(-2)
1.54(-2)
1.52(-2)
1.41(-2)
1.27(-2)
9.16(-3)
5.34(-3)
1.81(-3)
7.16(-5)
3.36(-7)
7.10(-3)
7.00(-3)
6.90(-3)
6.80(-3)
6.16(-3)
5.35(-3)
3.51(-3)
1.74(-3)
4.28(-4)
6.57(-6)
6.77(-9)
1.80(-3)
1.76(-3)
1.72(-3)
1.68(-3)
1.44(-3)
1.15(-3)
5.91(-4)
1.95(-4)
2.12(-5)
2.88(-8)
5.46(-13)
4.89(-3)
4.84(-3)
4.79(-3)
4.74(-3)
4.42(-3)
4.00(-3)
2.95(-3)
1.79(-3)
6.52(-4)
3.18(-5)
2.06(-7)
4.86(-3)
4.81(-3)
4.76(-3)
4.71(-3)
4.39(-3)
3.97(-3)
2.93(-3)
1.77(-3)
6.45(-4)
3.12(-5)
2.01(-7)
4.80(-3)
4.75(-3)
4.71(-3)
4.66(-3)
4.34(-3)
3.92(-3)
2.89(-3)
1.74(-3)
6.33(-4)
3.03(-5)
0.91(-7)
4.68(-3)
4.64(-3)
4.59(-3)
4.54(-3)
4.23(-3)
3.82(-3)
2.81(-3)
1.69(-3)
6.07(-4)
2.83(-5)
1.72(-7)
3.57(-3)
3.53(-3)
3.49(-3)
3.45(-3)
3.19(-3)
2.85(-3)
2.04(-3)
1.17(-3)
3.81(-4)
1.34(-5)
5.06(-8)
1.77(-3)
1.74(-3)
1.72(-3)
1.70(-3)
1.54(-3)
1.33(-3)
8.73(-4)
4.31(-4)
1.05(-4)
1.54(-6)
1.37(-9)
5.91{-2)
4.16(-2)
2.32(-2)
7.33(-3)
2,45(-4}
1.01(-6)
2.26(-2)
1.14(-2)
5.07(-3)
'"'"'"Z
tl
S<
;I;
a-
...,'-"
658
APPENDIXH
and, for a cosine current source (isotropic flux), it is given by
(H.2)
where <1>0 is the isotropic flux at the source plane,
The variables for the rectangular source (see Tables H,3 and HA) are
r(llt,a,b) = dose rate on perpendicular to corner

H, W = height and width of rectangle (em)
a,b = dimensions of rectangle in units of detector distance, a = H/z, b =
W/z and S,G(E) and Ilt are as previously defined.
For an isotropic source the unscattered flux attenuation function for the
dose rate at a corner position is given by
rb r
471r(llt,a,b) =
Jo
SG(E)
Jo
_~, dx dy "'~[E ( )
2
r2
Ilt
~ E dlltJJ + a2 )]
(fora""b>l)
(H.3)
and, for the cosine current source (isotropic flux), it is given by
2r(llt,a,b) = ~
<1>0 G(E)
271
fb fa e-~';} dx,dy
0
"'~[E2(llt)
r r
- (1/)1 +a 2 ) E2(/lt~)]
(for
a"" b> 1)
(HA)
REFERENCES
1. J. H. Hubbell, R. L. Bach, and R. J. Herbold, Radiation Field from a Circular Disk
2. D. K. Trubey, Oak Ridge National Laboratory, unpublished calculations.
3. J. H. Hubbell, R. L. Bach, and J. c. Lamkin, Radiation Field from a Rectangular
APPENDIXH
659
Table H.3-UNSCAITERED FLUX AT A CORNER POSITION

FROM A RECTANGULAR PLANE SOURCE WITH AN
ISOTROPIC ANGULAR DI~TRIBUTIONt
41Tf/[S G(E)J
IU
b = 0.1
a = 0.1
b = 0.2
a = 0.1
b = 0.2
a= 0.2
b = 0.5
a = 0.1
b = 0.5
a = 0.2
b= 0.5
a= 0.5
0
0.01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
9.93(-3):1:
9.83(-3)
9.74(-3)
9.44(-3)
8.99(-3)
8.13(-3)
6.02(-3)
3.64(-3)
1.34(-3)
6.60(-5)
4.37(-7)
1.97(-2)
1.95(-2)
1.93(-2)
1.87(-2)
1.78(-2)
1.61(-2)
1.19(-2)
7.18(-3)
2.62(-3)
1.28(-4 )
8.26(-7)
3.90(-2)
3.86(-2)
3.82(-2)
3.70(-2)
3.52(-2)
3.18(-2)
2.35(-2)
1.42(-2)
5.14(-3)
2.47(-4)
1.54(-6)
4.62(-2)
4.57(-2)
4.53(-2)
4.39(-2)
4.17(-2)
3.75(-2)
2.75(-2)
1.64(-2)
5.79(-3)
2.60(-4)
1.50(-6)
9.16(-2)
9.07(-2)
8.97(-2)
8.69(-2)
8.25(-2)
7.43(-2)
5.44(-2)
3.23(-2)
1.14(-2)
5.04(-4)
2.83(-6)
2.16(-1)
2.14(-1)
2.11(-1)
2.05(-1)
1.94(-1)
1.74(-1)
1.26(-1)
7.38(-2)
2.53(-2)
1.04(-3)
5.15(--6)
1-'1
b = 1.0
a = 0.1
b = 1.0
a = 0.2
b = 1.0
a= 0.5
b = 1.0
a = 1.0
b = 2.0
a = 0.1
b = 2.0
a= 0.2
0
0.01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
7.83(-2)
7.75(-2)
7.66(-2)
7.40(-2)
7.00(-2)
6.26(-2)
4.47(-2)
2.56(-2)
8.49(-3)
3.85(-4)
1.67(-6)
1.55(-1)
1.54(-1)
1.52(-1)
1.47(-1)
1.39(-1)
1.24(-1)
8.85(-2)
5.06(-2)
1.67(-2)
7.55(-4)
3.16(-6)
3.69(-1)
3.65(-1)
3.60(-1)
3.48(-1)
3.29(-1)
2.93(-1)
2.07(-1)
1.17(-1)
3.74(-2)
1.77(-3)
5.77(-6)
6.40(-1)
6.32(-1)
6.24(-1)
6.01(-1)
5.65(-1)
4.99(-1)
3.45(-1)
1.87(-1 )
5.61(-2)
7.99(-3)
6.49(-6)
1.10(-1)
1.09(-1)
1.08(-1)
1.04(-1)
9.70(-2)
8.53(-2)
5.82(-2)
3.14(-2)
9.57(-3)
3.37(-4)
1.67(-6)
2.19(-1)
2.17(-1)
2.14(-1)
2.06(-1)
1.93(-1)
1.69(-1)
1.15(-1)
6.20(-2)
1.88(-2)
6.54(-4)
3.16(-6)
1-'1
b = 2.0
a = 0.5
b = 2.0
a = 1.0
b = 2.0
a = 2.0
b = 5.0
a = 0.1
b= 5.0
a = 0.2
b = 5.0
a = 0.5
0
0.01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
5.25(-1)
5.18(-1)
5.11(-1)
4.91(-1)
4.60(-1)
4.03(-1)
2_72(-1)
1.44(-1)
4.24(-2)
1.32(-3)
5.79(-6)
9.31(-1)
9.18(-1)
9.05(-1)
8.67(-1)
8.07(-1)
7.01(-1)
4.62(-1)
2.35(-1)
6.43(-2)
1.75(-3)
6.51(-6)
1.41(0)
1.38(0)
1.36(0)
1.30(0)
1.20(0)
1.02(0)
6.38(-1)
3.02(-1)
7,46(-2)
1.80(-3)
6.52(-6)
1.37(-1)
1.35(-1)
1.33(-1)
1.26(-1 )
1.16(-1)
9.93(-2)
6.38(-2)
3.27(-2)
9.69(-3)
3.37(-4)
1.67(-6)
2.73(-1)
2.68(-1)
2.64(-1)
2.51(-1)
2.31(-1)
1.97(-1)
1.27(-1)
6.46(-2)
1.90(-2)
6.54(-4)
3.16(-6)
6.57(-1)
6.46(-1)
6.35(-1)
6.03(-1)
5.55(-1)
4.72(-1)
3.00(-1)
1.50(-1)
4.28(-2)
1.35(-3)
5.79(-6)
660
APPENDIXH
Table H.3-(Continued)
41Tr/[S G(E)]
b = 5.0
b = 10.0
b = 5.0
b = 10.0
b = 10.0
a= 0.2
a = 0.5
1.47(-1)
1.44(-1)
1.41(-1)
1.33(-1)
1.21 (-1)
1.02(-1)
6.42(-2)
3.27(-2)
9.69(-3)
3.37(-4)
1.67(-6)
2.92(-1 )
2.86(-1)
2.81(-1)
2.65(-1 )
2.41(-1)
2.02(-1)
1.27(-1)
6.47(-2)
1.90(-2)
6.54(-4)
3.16(-6)
7.06(-1)
6.92(-1)
6.77(-1)
6.38(-1)
5.79(-1)
4.84(-1)
3.01(-1)
1.50(-1)
4.28(-2)
1.35(-3)
5.79(-6)
III
a = 1.0
b = 5.0
a= 2.0
0
0,01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
1.19(0)
1.17(0)
1.15(0)
1.08(0)
9.92(-1)
8.35(-1)
5.14(-1)
2.47(-1)
6.51(-2)
1.75(-3)
6.51(-6)
1.88(0)
1.84(0)
1.80(0)
1.69(0)
1.53(0)
1.26(0)
7.27(-1)
3.20(-1)
7.57(-2)
1.80(-3)
6.53(-6)
2.73(0)
2.65(0)
2.58(0)
2.39(0)
2.10(0)
1.64(0)
8.53(-1)
3.44(-1)
7.69(-2)
1.80(-3)
6.54(-6)
b = 10.0
b = 10.0
= 5.0
= 0.1
b = 10.0
b = 20.0
a = 5.0
b = 10.0
a = 10.0
b = 20.0
a= 2.0
1.28(0)
1.26(0)
1.23(0)
1.15(0)
1.04(0)
8.60(-1)
5.18(-1)
2.47(-1)
6.51(-2)
1.75(-3)
6.51(-6)
2.07(0)
2.02(0)
1.97(0)
1.83(0)
1.63(0)
1.31(0)
7.34(-1)
3.21(-1)
7.56(-2)
1.80(-3)
6.53(-6)
3.15(0)
3.04(0)
2.94(0)
2.67(0)
2.30(0)
1.74(0)
8.65(-1)
3.44(-1)
7.69(-2)
1.80(-3)
6.53(-6)
3.80(0)
3.64(0)
3.49(0)
3.10(0)
2.58(0)
1.86(0)
8.78(-1)
3.45(-1)
7.68(-2)
1.80(-3)
6.53(-6)
1.52(-1)
1.48(-1)
1.45(-1)
1.36(-1 )
1.22(-1)
1.02(-1)
6.39(-2)
3.27(-2)
9.69(-3)
3.37(-4)
1.67(-6)
3.02(-1)
2.95(-1)
2.88(-1)
2.70(-1)
2.43(-1)
2.03(-1)
1.27(-1)
6.47(-2)
1.90(-2)
6.54(-4)
3.16(-6)
b = 20.0
a = 0.5
b = 20.0
a = 1.0
b = 20.0
a= 2.0
b = 20.0
III
a = 5.0
b = 20.0
a = 10.0
b = 20.0
a = 20.0
0
0,01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
7.31(-1)
7.13(-1)
6.96(-1)
6.50(-1)
5.86(-1)
4.86(-1)
3.01(-1)
1.50(-1)
4.28(-2)
1.35(-3)
5.79(-6)
1.33(0)
1.30(0)
1.27(0)
1.18(0)
1.05(0)
8.63(-1)
5.18(-1)
2.47(-1)
6.51(-2)
1.75(-3)
6.51(-6)
2.17(0)
2.10(0)
2.04(0)
1.88(0)
1.65(0)
1.31 (0)
7.34(0)
3.21 (0)
7.56(-2)
1.80(-3)
6.53(-6)
3.38(0)
3.25(0)
3.12(0)
2.79(0)
2.36(0)
1.75(0)
8.63(-1)
3.44(-1)
7.69(-2)
1.80(-3)
6.53(-6)
4.22(0)
4.01 (0)
3.81 (0)
3.30(0)
2.68(0)
1.89(0)
8.79(-1)
3.45(-1)
7.68(-2)
1.80(-3)
6.53(-6)
4.88(0)
4.56(0)
4.28(0)
3.59(0)
2.81 (0)
1.92(0)
8.82(-1)
3.45(-1)
7.68(-2)
1.80(-3)
6.53(-6)
III
a = 1.0
0
0,01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
tFrom Hubbell el al.'
:fRead: 9.93
a = 0.1
10-3 etc.
a = 0.2
APPENDIX H
661
Table H.4-UNSCATTERED FLUX AT A CORNER POSITION

FROM A RECTANGULAR PLANE SOURCE WITH A
COSINE ANGULAR DISTRIBUTION (ISOTROPIC FLUXlt
2fj[(0) G(E)J
/.It
om
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
/.It
om
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
/.It
om
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
b
a
=0.1
=0.1
1.58(-3)+
1.56(-3)
1.55(-3)
1.53(-3)
1.43(-3)
1.29(-3)
9.54(-4)
5.78(-4)
2.12(-4)
1.04(-5)
6.92(-8)
b
a
=1.0
=0.1
1.12(-2)
1.11 (-2)
1.08(-2)
1.08(-2)
1.00(-2)
8.98(-3)
6.44(-3)
3.71(-3)
1.24(-3)
4.91(-5)
2.54(-7)
b
a
=2.0
=0.5
6.55(-2)
6.47(-2)
6.39(":'2)
6.30(-2)
5.77(-2)
5.08(-2)
3.49(-2)
1.89(-2)
5.78(-3)
1.95(-4)
8.63(-7)
b =0.2
a =0.1
b =0.2
a =0.2
b =0.5
a =0.1
b =0.5
a =0.2
b =0.5
a= 0.5
3.11(-3)
3.08(-3)
3.04(-3)
3.01(-3)
2.81 (-3)
2.54(-3)
1.88(-3)
1.13(-3)
4.14(-4)
2.01(-5)
1.30(-7)
6.12(-3)
6.00(-3)
5.94(-3)
5.53(-3)
5.00(-3)
3.69(-3)
2.22(-3)
8.08(-4)
3.87(-5)
2.45(-7)
7.09(-3)
7.01(-3)
6.94(-3)
6.87(-3)
6.39(-3)
5.76(-3)
4.22(-3)
2.51(-3)
8.90(-4)
4.00(-5)
2.31(-7)
1.40(-2)
1.38(-2)
1.37(-2)
1.36(-2)
1.26(-2)
1.14(-2)
8.30(-3)
4.93(-3)
1.74(-3)
7.70(-5)
4.36(-7)
3.21 (--2)
3.17(--2)
3.14(-2)
3.1 O(--2)
2.88(-2)
2.59(--2)
1.88(-2)
1.10(-2)
3.77(-3)
1.54(-4)
7.79(--7)
b =1.0
a =0.2
b =1.0
a =0.5
b =1.0
a =1.0
b =2.0
a =0.1
b =2.0
a =0.2
2.21(-2)
2.19(-2)
2.17(-2)
2.14(-2)
1.98(-2)
1.77(-2)
1.27(-2)
7.29(-3)
2.43(-3)
9.47(-5)
4.79(-7)
5.12(-2)
5.06(-2)
5.01(-2)
4.95(-2)
4.57(-2)
4.07(-2)
2.89(-2)
1.64(-2)
5.30(-3)
1.91(-4)
8.61(-7)
8.33(-2)
8.23(-2)
8.13(-2)
8.03(-2)
7.38(-2)
6.53(-2)
4.54(-2)
2.49(-2)
7.57(-3)
2.38(-4)
9.54(-7)
1.42(-2)
1.40(--2)
1.38(-2)
1.37(-2)
1.25(-2)
1.11(-2)
7.71(-3)
4.26(-3)
1.35(-3)
5.00(-5)
2.55(-7)
2.81(-2)
2.77(-2)
2.74(-2)
2.70(-2)
2.48(--2)
2.19(-2)
1.52(--2)
8.38(-3)
2.64(-3)
9.65(-5)
4.80(--7)
b =5.0
a =0.1
b =5.0
a= 0.2
b =5.0
a =0.5
1.56(-2)
1.53(-2)
1.51(-2)
1.49(-2)
1.35(-2)
1.18(-2)
8.02(-3)
4.33(-3)
1.35(-3)
5.00(-5)
2.55(-7)
3.08(-2)
3.04(-2)
3.00(-2)
2.95(-2)
2.68(-2)
2.34(-2)
1.58(-2)
8.53(-3)
2.65(-3)
9.65(-5)
4.80(-7)
7.23(-2)
7.12(-2)
7.02(--2)
6.92(-2)
6.26(-2)
5.45(-2)
3.64(--2)
1.93(-2)
5.81(-3)
1.95(-4)
8.63(--7)
b =2.0
a =1.0
1.09(-1)
1.08(-1)
1.06(-I)
1.05(-1)
9.52(-2)
8.32(-2)
5.58(-2)
2.92(-2)
8.34(-3)
2.44(-4)
9.56(-7)
6.06(~3)
b
a
=2.0
=2.0
1.48(-1)
1.45(-1)
1.43(-1)
1.41(-1)
1.27(-1)
1.09(-1)
7.05(-2)
3.49(-2)
9.26(-3)
2.49(-4)
9.58(-7)
662
APPENDIX H
Table H.4-(Continued)
2r/[(0) G(E)]
b =5.0
=5.0
b =10.0
a =0.1
b =10.0
a= 0.2
b =10.0
a =0.5
1.70(-1)
1.67(-1)
1.64(-1)
1.62(-1)
1.43(-1)
1.21(-1)
7.51(-2)
3.59(-2)
9.32(-3)
2.49(-4)
9.58(-7)
2.06(-1)
2.01(-1)
1.97(-1)
1.93(-1)
1.68(-1)
1.38(-1)
8.10(-2)
3.71(-2)
9.38(-3)
2.49(-4)
9.58(-7)
1.58(-2)
1.56(-2)
1.58(-2)
1.51(-2)
1.37(-2)
1.19(-2)
8.02(-3)
4.33(-3)
1.35(-3)
5.00(-5)
2.55(-7)
3.13(-2)
3.08(-2)
3.04(-2)
2.99(-2)
2.70(-2)
2.35(-2)
1.58(-2)
8.53(-3)
2.65(-3)
9.65(-5)
4.80(-7)
7.34(-2)
7.23(-2)
7.12(-2)
7.01(-2)
6.32(-2)
5.48(-2)
3.65(-2)
1.93(-2)
5.81(-3)
1.95(-4)
8.63(-7)
b =10.0
a =1.0
b =10.0
a = 2.0
b =10.0
a= 5.0
a =10.0
b =10.0
b =20.0
a =0.1
b =20.0
a =0.2
1.24(-1)
1.22(-1)
1.20(-1)
1.18(-1)
1.06(-1)
9.08(-2)
5.88(-2)
2.99(-2)
8.38(-3)
2.44(-4)
9.56(-7)
1.75(-1)
1.71(-1)
1.68(-1)
1.65(-1)
1.46(-1)
1.22(-1)
7.53(-2)
3.59(-2)
9.32(-3)
2.49(-4)
9.58(-7)
2.15(-1)
2.10(-1)
2.05(-1)
2.01(-1)
1.72(-1)
1.40(-1)
8.13(-2)
3.71(-2)
9.38(-3)
2.49(-4)
9.58(-7)
2.28(-1)
2.22(-1)
2.16(-1)
2.11(-1)
1.78(-1)
1.43(-1)
8.17(-2)
3.71(-2)
9.38(-3)
2.49(-4)
9.58(-7)
1.58(-2)
1.56(-2)
1.54(-2)
1.51(-2)
1.37(-2)
1.19(-2)
8.02(-3)
4.33(-3)
1.35(-3)
5.00(-5)
2.55(-7)
3.14(-2)
3.09(-2)
3.04(-2)
3.00(-2)
2.71(-2)
2.35(-2)
1.58(-2)
8.53(-3)
2.65(-3)
9.65(-5)
4.81(-7)
p.t
b =20.0
a =0.5
b =20.0
a =1.0
b =20.0
a =2.0
b =20.0
a =5.0
a =10.0
a =20.0
0
0,01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
7.37(-2)
7.26(-2)
7.14(-2)
7.03(-2)
6.33(-2)
5.48(-2)
3.65(-2)
1.93(-2)
5.81(-3)
1.95(-4)
8.63(-7)
1.25(-1)
1.23(-1)
1.21(-1)
1.19(-1)
1.06(-1)
9.08(-2)
5.88(-2)
2.99(-2)
8.38(-3)
2.44(-4)
9.56(-7)
1.76(-1)
1.72(-1)
1.69(-1)
1.66(-1)
1.46(-1)
1.22(-1)
7.53(-2)
3.59(-2)
9.32(-3)
2.49(-4)
9.58(-7)
2.18(-1)
2.12(-1)
2.07(-1)
2.02(-1)
1.73(-1)
1.40(-1)
8.13(-2)
3.71(-2)
9.38(-3)
2.49(-4)
9.58(-7)
2.32(-1)
2.26(-1)
2.20(-1)
2.14(-1)
1.79(-1)
1.43(-1)
8.17(-2)
3.71(-2)
9.38(-3)
2.49(-4)
9.58(-7)
2.39(-1)
2.31(-1)
2.24(-1)
2.18(-1)
1.80(-1)
1.44(-1)
8.17(-2)
3.71(-2)
9.38(-3)
2.49(-4)
9.58(-7)
p.t
b =5.0
a =1.0
0
0,01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
1.22(-1)
1.20(-1)
1.18(-1)
1.16(-1)
1.05(-1)
9.02(-2)
5.87(-2)
2.99(-2)
8.38(-3)
2.44(-4)
9.56(-7)
p.t
0
0,01
0.02
0.05
0.1
0.2
0.5
1.0
2.0
5.0
10.0
tFtom Trubey.'
b =5.0
=2.0
:lRead: 1.58
10-3 , etc.
b =20.0
b = 20.0
Random-Number Generators
Appendix I
The use of Monte Carlo methods in any calculation depends on a copious
supply of random numbers. Many early calculations relied heavily on a
table t of a million random digits generated electronically. Whether telephone directories were ever actually tried is not known; however, their use
was frequently postulated (too much bias for some digits). The advent of the
digital computer made a mathematical algorithm highly desirable since such
a function has the property of generating a reproducible sequence provided
one starts with the same initial numbers and provided the same computer
system is involved. The reproducible sequence is valuable in checking
problem results to ascertain that the nonrandom aspects of a program are
running properly. The necessity of limiting reproducibility to a given
computer system stems from the dependence of the generator on the
number of bits per word and on round-off methods. Both these characteristics vary widely with computer type. Computer manufacturers typically
supply a random-number-generator routine especially suited to their machines. A characteristic of many of these routines is that they generate the
same sequence for the same initial random number. It is thus advantageous if
the user is allowed to supply his own initial number so that he can repeat a
sequence or generate a new sequence at his option.
Number sequences produced by a mathematical relation are clearly
deterministic, not random, and are thus properly called pseudorandom. To
avoid semantic difficulties as well as logical contradictions, we acknowledge
that the number sequences we seek actually must have special properties that
further distinguish them from a random set. In the subsequent discussion,
unless otherwise stated, we use random to mean adequate for Monte Carlo
application. A classic paper by Hull and Dobell' serves as an excellent
introduction to the problem of devising random-number generators.
tRand Corporation, A Million Random Digits with 100,000 Normal Devidtes, Free Press of
Glencoe, New York, N. Y.,1955.
663
APPENDIX I
664
I.1
PROPERTIES OF RANDOM-NUMBER GENERATORS (RNG's)
The RNG with which we are concerned is a computer subroutine that

generates one random number each time it is called. It thus generates a
sequence of N terms
and it is the sequence of these N terms that must display the properties of
randomness. The two basic properties, although conceptually simple, are far
from trivial.
1. Equidistribution: The terms in the sequence should be uniformly
distributed over the unit interval. Their frequency distribution should be
characterized by a horizontal plot as shown in Fig. 1.1.
en
::;
a:
w
I-
u.
a:
w
'"
::;
:::>
o
VALUE OF TERM
Fig. I.l-Frequency distribution.
Further, large enough sub-sequences

'j, 'j+l ' ... , rj+k
imbedded in the full sequence should also be equidistributed. For example,

if a random-number generator can generate a sequence of a million terms,
any sub-sequence of 10,000 terms should also display equidistribution.
2. Independence: It is not enough that the sequence be equidistributed.
The sequence 0, 0.1, 0.2, ..., 0.9, 1.0,0,0.1,0.2, ..., 0.9, 1.0,0,0.1, ... is
equidistributed over [0, 1 J but certainly is not random. Thus each term of
APPENDIX I
665
the sequence should be independent of any other; i.e., each number in [0, 1]
should have equal probability of being generated each time the RNG is
called. Intuitively, patterns in the sequence should not be discernible.
Verifying that an RNG produces sequences satisfying both of these
properties is an unsolved problem. Many necessary conditions can be
defined, but the key theorems, those involving sufficiency, are yet to be
stated and proved.
Testing a generated sequence for equidistribution is simple, but being
assured that the next sequence to be generated will also be equidistributed is
a far greater challenge. In fact, some workers recommend that the RNG
sequence generated in a Monte Carlo calculation be analyzed during the
calculation so that if bad luck produced a pathologicalt sequence, the results
could be discarded. This is not a common inclusion in current Monte Carlo
codes, however.
Complete testing for independence is impossible, but techmques have
been developed which look for the appearance of independence. They
usually involve attempts to find a relation (called serial correlation) between
the nth term and the immediately preceding (e.g., 10) terms for all values of
n. Testing methods are discussed in the following sections.
I.2 RECURSION EQUATIONS FOR RNG's
A recursion relation is a function that relates the nth term of a sequence

to previous terms. Although some RNG's have used algorithms that used
more than the immediately preceding term to generate the nth term, the
common methods compute rn as a function only of rn-I and preselected
constants. Since a large quantity (on the order of 10,000 to 100,000) of
random numbers are used in a typical Monte Carlo calculation, speed of
generation is important. Thus the recursion relations used are designed to
minimize computing time while retaining sufficient randomness to be of
practical use.
Almost all RNG's used in shielding codes are designed for fixed-wardlength machines. The theory and' the tests must be modified to account for a
binary or a decimal machine as well as for the bit (or digit) length of the
computer word. For illustrative purposes we will assume a 36-bit binary
machine in this discussion.
tDefined later in this section.
666
APPENDIX!
There are only 236 numbers representable in a 36-bit word computer

(actually 2 3 s if one bit is reserved for a sign). An RNG in such a machine
can thus only generate at most 2 36 rational numbers. Rational numbers
requiring more than 36 bits (integers and fractions) as well as the
uncountably infinite number of irrational numbers cannot be generated.
Thus, although not critical (2 36 fractions would seem to cover the unit
interval quite well), the theoretical properties of equidistribution and
independence are irrevocably compromised in the computer approximation.
Most RNG's in common use generate sequences of positive random
integers, Xl' x2,"" Xi. The desired random fractions (0';;ri<1) are
obtained by the relation
ri = x;/M
where the integers Xi are uniformly distributed over r0, M]. In a binary
computer, if k bits are allotted to represent the magnitude of the number
(i.e., not including the sign bit, if any), M = 2k and the Xi values can range
from 0 to 2k - 1.
Many recursion schemes have been used to generate sequences of
pseudorandom integers. Recent work has shown that some of the commonly
used generators are inadequate.
Generators that have received the most attention are of the congruential
type. Three are discussed in the following paragraphs, the first for historical
reasons and the last two because of their widespread use.
1. Middle-digit square. The first RNG put into common use was that of
von Neumann 2 and Metropolis. 3 In this method an arbitrary first term with
a length of k digits is chosen and squared to obtain a 2k-digit number. The
middle k digits of this intermediate result are extracted to obtain the next
random number. The sequence is generated by repeating this process.
This generator appeared adequate and was used for many years.
However, when examined more carefully by newer testing methods, it
proved to be inadequate and is not in widespread use today.
2. Mixed congruential. This generator requires two constants as well as
an initial term of the sequence to be chosen. Its recursion relation is of the
formt
(for i = 1, 2, ...)
tThe expression A =: B (mod M) is read: A is congruent to B modulo M and means A is the
remainder of RIM.
APPENDIX [
667
in which A and c are the two constants and M = 2k (e.g., 236 ), where k is the
bit length of the computer word being used.
3. Multiplicative congruential. This is a special case of the mixed
generator in which c = O. It thus has the advantage of greater speed since an
addition step is eliminated. Its recursion relation is
Xi
== Axi _1(mod M)
(fori= 1, 2, ...)
These latter two generators have been extensively studied to determine if

one is clearly better than the other. Also, the effect of the constants A and c
and the starting value X I on the properties of randomness have been
investigated. The complexity of the problem is illustrated by the fact that
HuIl and DobeIl 4 recommended the mixed method as the best choice in
1964; this was foIlowed by the Fourier analysis by Coveyou and MacPherson,s which resulted in the opposite recommendations in 1967.
The constants A and c are usually fixed when the subroutine is written.
The first term, Xl' is provided as an input variable. Each has an effect on the
properties of the RNG. For example, if either A or XI is even, every term in a
multiplicath(e sequence will be even, which will halve the period of the
generator. The period (or cycle length), P, is the number of terms that will
be generated before the sequence repeats itself. In a binary machine with an
effective word length of k bits (sign bit not included), P";;' Zk.
In binary computers A is often chosen to be of the form
A = zP + b (p and b are small integers)

since the multiplication Axi-l can then be accomplished by the fast
shift-and-add operations in the computer. However, for small choices of p
(e.g., p";;' 8) where the speed advantage is greatest, the resulting generator
was found to be generally inadequate.
Coveyou and MacPherson s examined 32 generators of the mixed and
multiplicative type: 7 for 36-bit binary machines; 8 for 48-bit binary
machines; 7 for 3Z-bit binary machines; and 10 for decimal machines. They
varied values of A and c and used a Fourier analysis technique to examine the
distribution properties of one-to-ten term sub-sequences. Their condusions
are of interest:
1. There is at present no method of generation of pseudo-uniform sequences
better than the simple multiplicative congruence method with a carefully chosen
multiplier (A).
APPENDIX I
668
2. The multiplier A should not be close to a simple rational multiple of M; if

it is, the basic congruence shows that appreciable serial correlation will result.
3. The multiplier Ashould not be close to a simple rational multiple of the
square root of M; even though this choice may produce very small correlation
between adjacent pairs (adjacent terms in the sequence), serious difficulties
result in the triplet distribution.
4. The multiplier A should not be chosen with a small number of ones in the
binary representation (e.g., 2 7 ) to facilitate shift-and-add techniques.
5. Above all, the multiplier A must be adequately large (e.g., SI5).
6. The choice of multiplier must be made most carefully for 32-bit word
multiplication,
for (36-bit machines) the choice still needs some care. For
(48-bit) and
64-bit multiplication ..., a suitable choice should be easy.
These conclusions are reinforced by the work of Van Gelder,6 who

advises that the bit pattern for the chosen X should not contain long strings
of zeros. He further states: " ... there is no hope for multipliers below 10 3 ,
and it is best to stay above 10 4 For modulus 2 35 , the best range (for X)
seems to be about 2 1 Z to 2 13 , with at least one digit to break up the string
of zeros."
It is desirable to maximize the period P of the generator. Barnett 7 gives
necessary.and sufficient conditions on X and XI to maximize the period of a
multiplicative generator on a binary machine. They are, where xi'" XXi_I
(modM),
XI ~
(mod 8)
1 < X ~ 1 (mod 2)
i.e., X must be greater than 1 and odd. The maximum period obtainable on a
computer with a word length of k bits is 2k - z numbers where M ~ 2k . The
period P ~ (1/4)M, although the maximum value attainable by Xi, is still
M - 1. For mixed generators the period can be made to equal M. However,
this advantage is offset by the reduced computing speed of the mixed
method.
I.3 TESTING RNG's

Several classes of tests are applied to RNG's. Although not inclusive, the
following methods are illustrative of those in use:
669
APPENDIX I
1.3.1 Global Tests

Global tests are the simple tests that can be applied to large
sub-sequences of the generator. They make use of the following types of
relations:
lim
1
n - k+
L
"
fj
(k=1,2" .. )
="2
i=k
lim
"
2 _
n-k+1L
r
I
-~
(k= 1, 2, ...)
i=k
lim
(k = 1, 2, ...)
"~-
These three limits state that the average of any (n - k + 1) termt of a

random sub-sequence in the unit interval should approximate Y2 ; the average
of the squares, Y3 ; and the variance, Y!2' Observe that these tests all are
necessary conditions for equidistributions. They tell nothing about independence.
However, Coveyou 8 suggests one such global test for pair-wise correlation:
,
I1m
"~- n -
1
k+1
L
"
fjfH!
="4 + 12A
(i=1,2, ...)
j=k
1.3.2 Equidistribution
The simplest test of equidistribution is to partition the unit interval into
small subintervals of equal width (e.g., 0.001) and count the number of
terms of any sub-sequence that fall in each subinterval. The frequency
function thus obtained should approximate a constant, namely, Nil, where
N is the number of terms in the sub-sequence and I is the number of
subintervals chosen.
More sophisticated methods of analysis are available. Coveyou and
MacPherson' start with the necessary and sufficient condition for equidistribution
tNote that the choice of k (k = 1, 2, ... ) determines the selection of any sub-sequence within the
period of the generator.
670
APPENDIX I
'k
lim - 1:: e 21T1

n j=1
n-4
oo
Tj
= 0
for all k = 1, 2, ..., and develop a Fourier analysis that generates figures of
merit for the uniformity of distribution of n-tuples of consecutive terms in
the generated sequence. Their analysis covered n-tuples from 1 to 10 terms
in length.
1.3.3 Independence
Independence is the hardest property to analyze. No necessary conditions have been established. In fact, since each term is deterministically
computed from its predecessor, the terms are clearly dependent. One can
only hope that no obvious pattern emerges and that relatively short
sub-sequences are equidistributed. A standard method of analysis involve.s
the use of n-tuples; i.e., sub-sequences of n terms. We will illustrate the
technique by use of an example.
Consider an RNG that generates decimal digits (e.g., 2, 8, 7, 3, 3, ...). If
independence holds, each pair of all possible consecutive terms, e.g., (8,7),
should be uniformly distributed within the sequence. When a sequence has
been generated, each such pair (Xi, Xi+l) for i = 1, 2, ..., can be tabulated in
the matrix shown in Fig. 1.2. Each box contains the number of times the pair
appeared in the sequence. Hopefully, each pair occurs with the same
frequency as any other pair.
When this type of test is applied to RNG's generating fractions in the
unit interval, the partition of the unit interval creating the counting bins can
be as small as desired. Also, the dimensionality of the matrix can be
extended to handle any number of consecutive terms (n-tuples). We should
point out that a special type of equidistribution property is being used, that
of equidistribution of n-tuples, to examine apparent independence. In fact,
the Fourier technique of Coveyou discussed is exactly that. When the n-tuple
is one term, equidistribution over (0, 1) in the sense we have used is tested.
For n-tuples of two or more terms, apparent independence is the basic
consideration. Most RNG's are now tested for at least pair and triplet
correlation. In his earlier paper, Coveyou' suggests that one should test for
correlations of n-tuples up to 10 to 20 terms in length.
All the preceding tests admit to standard statistical analysis. A set of
sequences can be generated by a given RNG and each test can be applied.
The probability that the results represent random sequences can then be
estimated by chi-square tests and computation of correlation coefficients.
APPENDIX I
671
:;- 6
a:
;;:
.. 5
u.
::;;
ffi
l-
t;
a:
u:
o
o
SECONO TERM OF PAIR {x;
+ l'
Fig. I.2-Distribution of pairs.
The reader is referred to Refs. 1 and 4 for elementary discussion of such

statistical tests.
1.4 PATHOLOGICAL NUMBERS
In an infinite sequence of random numbers, all the weird sub-sequences
that one could imagine should be in evidence. Somewhere the sub-sequence
containing 100 consecutive values of 0.1 should appear, long monotonically
increasing sub-sequences should be included, etc. The sequence is called
pathological when it presents a pattern of this character. Thus, in examining
one sequence generated by an RNG, the fact that a clear pattern emerges
does not imply necessarily that the RNG is bad. After all, any conceivable
sub-sequence has a finite probability. However, in the finite sequences used
APPENDIX I
672
in a Monte Carlo calculation, the appearance of a pathological sub-sequence

could spell disaster. Any RNG that exhibits such behavior in any except the
rarest of instances should be avoided. In fact, it probably is good that
pseudorandom-number generators of the preceding type are sufficiently
nonrandom to preclude many of these situations. (The proper choice of A
and x. can guarantee that every number in the period appears only once
before the period is exhausted, for example.)
1.5 USEFUL RNG's

The major purpose of this discussion has been to introduce the concepts
and hazards inherent in RNG's. Many examples can be found where lengthy
and costly Monte Carlo calculations were invalidated because not enough
attention was paid to the quality of the RNG. The user of an RNG
subroutine should require sufficient evidence of its quality.
With this warning and the material presented here, one should be
prepared to examine some cookbook RNG codes that have recently
appeared. Marsaglia and Bray' propose using one- or two-line FORTRAN
statements in the body of a program rather than exit to a subroutine. They
use the multiplicative recursion relation and give FORTRAN statements for
the IBM-7094 and 360 and the UNIVAC 1108 computers. They also discuss
the use of two RNG's together, one to generate an ever-changing table of
pseudorandom numbers and the other to pick a number from the table at
(pseudo) random. Some flaws in their coding were corrected by Grosenbaugh.' 0 Finally, Seraphin" suggests that it is not so bad after all to exit to
a subroutine if it is coded in efficient assembly language, and he proposes a
multiplicative subroutine for the 32-bit IBM-360.
One point in using the preceding techniques may not be so obvious. The
earlier discussion of recursion relations assumed fixed-point (i.e., integer)
representations. In fact, most Monte Carlo codes are programmed and
implemented with floating-point numbers. Not only are the bits available for
integer representation different from the computer word length but also
multiplication is handled in a variety of ways. Thus care must be taken when
the two-step computation
Xi
== Axi-, (mod M)
= Xi
' M
APPENDIX I
673
is coded. The results must be properly floated and normalized for use in the
Monte Carlo program.
The last references9-11 recommending specific RNG's contain from zero
to three short lines discussing the tests to which they have been submitted. It
seems that a spirit of caveat emptor exists. Perhaps this is the just due of
Monre Carlo practitioners because the father of this subject, John von
Neumann, admonished that he who uses arithmetic means to generate
random numbers is, of course, in a state of sin. 2
REFERENCES
1. T. E. Hull and A. R. Dobell, SIAM (Soc. Ind. Appl. Math.) Rev., 4: 230-254 Vuly
1962).
2. J. von Neumann, Various Techniques Used in Connection With Random Digits,
National Bureau of Standards, Applied Mathematics Series 12, Superintendent of
Documents, U. S. Government Printing Office, Washington D. C., 1951.
3. N. Metropolis, Phase Shifts Middle Squares Wave Equation, in Symposium on Monte
Carlo Methods, H. A. Meyer (Ed.), John Wiley & Sons, Inc., New York, 1956.
4. T. E. Hull and A. R. Dobell, Mixed Congruential Random Number Generators for
Binary Machines,]. Assn. Camp. Mach., 11: 31-40 Vanuary 1964).
5. R. R. Coveyall and R. D. MacPherson, Fourier Analysis of Uniform Random Number
Generators,]. Assn. Camp. Mach., 14: 100 Vanuary 1967).
6. A. Van Gelder, Some New Results in Pseudo-Random Number Generation,]. Asm.
Camp. Mach., 14: 785 (October 1967).
7. V. D. Barnett, The Behavior of Pseudo-Random Sequences Generated on Computers
by the Multiplicative Congruential Method, Math. Comput., 16: 63 (1962).
8. R. R. Coveyou, Serial Correlation in the Generation of Pseudo-Random Numbers,].
Assn. Camp. Mach., 7: 72 Vanuary 1960).
9. G. Marsaglia and T. A. Bray, Oneline Random Number Generators and Their Use in
Combinations, Commun. ACM (Assn. Comput. Mach.), 11: 757 (November 1968).
10. L. R. Grosenbaugh, More on Fortran Random Number Generators, Commun. ACM
(Assn. Comput. Mach.), 12: 639 (November 1969).
11. D. S. Seraphin, A Fast Random Number Generator for IBM 360, Commun. ACM
(Assn. Comput. Mach.), 12: 695 (December 1969).
Demonstration Monte Carlo

Program
Appendix J
A slab penetration calculation is described in Sec. 5.9 as a sample Monte
Carlo problem to demonstrate some of the techniques presented in this
chapter. This appendix contains the input-data requirements, the logic flow
chart of the calculation, the FORTRAN listing of the program, the
description of a test problem for program check-out, and a specimen listing
of the input and output for the test problem.
We emphasize that the random-number-generator subroutine called
RAND is not a random-number generator but a table of random numbers
with a stepping procedure to make it possible to run the test problem and to
obtain the same results given here to ensure that the program is running
properly. Application of the program to realistic calculations will require a
bona fide random-number generator. Similarly, the subroutine XSEC is a
simplified cross-section interpolatien routine. Unlike RAND, however, it
may be used for realistic calculations for energy ranges over which the cross
sections vary in an approximately liuear fashion.
A number of explanatory comments (C statements) have been included
in the program listing which can be eliminated in subsequent copies and
thereby reduce the key-punch labor considerably.
For each computer system the card reader is referenced by one particular
digit and the ptinter, by another. The user of this program should change KX
to equal his computer's card-reader reference if other than 1 (line 29 in the
listing), and KY to equal his computer's printer reference if other than 3
(line 30 in the listing). If the available FORTRAN compiler does not accept
DATA statements, the user should delete the ninth and tenth lines in
subroutine RAND and insert 23 assignment statements (each to a card) for
IR; that is, IR = 22, IR = 16, IR = 20, etc., after the line 100 KNT = 23.
674
675
APPENDIX]
INPUTDATA REQUIREMENTS
Variable
NENEG
NANG
NHIST
NDIVT
IB
EO
TT
COA
AW
ANG(J)
FTD(I)
E(I)
Definition
Format
SIS
4EIO.3
6EIO.3
6EIO.0
6EI0.0
Number of print energy groups

Number of print angle interval (for each
surface)
Number of histories to be run
Number of slab increments for
storing abso;ptions
Random-number-generator base
Source energy (MeV)
Limit
,,7
,,6
,,20
Odd
Slab thickness (em)

Small angle below which theta is assumed
to be zero; usually 10-5
Atomie weight (amu)

Kth bound of the print angle
intervals; ascending order in
degrees
Flux density~to-dose conversion
factor for the Ith print
energy group
Energy bound of the Ith print
energy interval; ascending order in
NANG + 1
values
,,90
NENEG
values
NENEG+ 1
values
MeV
NCSE
110
EE(I)
6EIO.3
TOT(I)
6EIO.3
Number of energies at which cross

sections are input
Ith value of the energy at which
cross sections are input; ascending
order in MeV
Total macroscopic cross section at
,,40
NCSE values
NCSE values
EE(I) (em-I)
SCA(I)
6EIO.3
Scattering macroscopic cross section
at EE(I)(em- 1 )
NCSE values
676
APPENDIX]
FLOW CHART FOR SAMPLE MONTE CARLO PROGRAMt
B
o:=pcosq,
~=psin
STORE
REFLECTED
NEUTRON
E=E o
z=o
H=H+l
cos 8
>
=..;;.
= 1
STORE
ABSORBED
NEUTRON
= 0(2" - I)
-In n
1=-- l;,(E)
JJ.
'_ [1 +(A+l)'
2Ap + A
=z + )'L
E-E
STORE
TRANSMITIED
NEUTRON
= 211 - 1
tShaded numbers 1 to 5 refer to the outline of the program logic in Sec. 5.9.
1
]
677
APPENDIX]
FLOW CHART (Continued)
d=~
d=~
E=E'
cos Os '"
1 +AJ,L
;.
"';1 + 2AJ,L +A:Z
<
.--_ _1 - 111-
(X' '"
ki>---...,
bC'Ya - bd(j + aQ
0/ '" be
iJ' = bd
'Y' '" a'Y
t/>s '" 1J(2n - 1)
'Y''''-b~+ay
(X '" Q'
~ =~'
'Y '" 'Y'
a "'cosO s
b '" sin as '"
c"'cost/>s
(Chart continues on next page.)
678
APPENDIX]
FLOW CHART (Continued)
~5~
= coa 8j + cos 8j+1

'Yav.}
2.0
(aUj,8 j < 90)
~
= cos Bj+t - cos 8j
'Yav.}
2.0
(aUj, 8j > 90)
F(i,i) = N(M: i , tJ.8 0

Hmu'Yav.J
(aU ii)
~
D(i,j) "" F(i,j)*K(E av .j )
(all ii)
'J = N(M:,,!J.8
j)
I,}
H
N('
max
(aU ii)
N ('k)=N(M:"tJ.t,)
a I,
H
max
(aU i k)
\pm~THE/
RESULTS
1
STOP
APPENDIX]
679
FORTRAN LISTING
C
A MONTE CARLO OfMONSlRAllCN PROCEDURE
C
C
C
CALCULATES THE TRANSMITTEO AND REFLECTED CURRENT, FlUENCE,

AND DOSe AND NU~RER OF A8S0PPTIONS FOR A SEMI-INFtNUTE StAB
WITH A COSINE CURRFNT LOCATED ON A SLAB SURFACE.
DIMENSION XNIT(144)
DIMENSION FTOfSJ,FTC8,b),OTCB,b),FR(8,61,DRf8,b),COAfB),COB(8)
COMMON Nl,NNl,NOT,IOP,EAS,ANG(6),BNGlbl,Tlf21I,E(8I,T(8,61,R(8,6)
1 ,A{8,20t,l,G
"
EQUIVALENCE IFTI11,XNITIIJ),CDTfl),XNIT(49'.,IORfll,XNIT(91))
00 500 I = 1,144
500 XN1TI!) = O.
c
c
c
c
c
NH .. 0
READ INPUT PARAMETERS

NFNFG
NANG
NHIST
NOIVT
c
c
c
c
c
c
c
c
18
fa
TT
COA
AW
FTO
NO. PRINT ENERGY GROUPS

Ne. PRINT ANGLE INTERVALS
Ne. HISTORIES
NO. SLAB ABSORPTION INTERVALS

RANDOM NUMPER BASE (000)
SOURCE ENERGY
SLAB THICKNESS
(UT-OFF ANGLE
ATOMIC WEIGHT
FLUX-TO-DOSE CONVERSION FACTORS
KX '" I
KY '" 3
RFAo(KX,2QINENEG, NANG, NHIS1, NOIV1, IB
REAO(KX,191 EO, TT, COA, AW
C
C
C
READ THE PRINT ANGLE INTERVAL BCUNOS LESS THAN 90 DEG. AND
GENERATE THOSF GRF.ATER THAN 90 DEC.
NNl '" NANG+l

READ (KX,l9IfANG(KI, K=I,NN1)
00 24 M"" I, NtH
24 BNG(MI= IBv.a-ANGCM)
READ (KX,19. (FTO(lt,I=I,NENEGI
C
READ THE PRINT ENERGY BCUNCS.
NI = NENEG+l
READ (KX,191 (E(JI, Jzl,Nll
20 FORM6.HIOI5)
19 FnRMAT(6EI0.31
CONVERT PRINT
6.NGlES FROM DfGRFES
.0114533
COA*QfD
N"'l,NNl
ANG(NI= ANG(N1*QEO
25 8NG(NI '" 8NG(NI.QEO
Q!:D ""
CC6. '"
DO 25
ZERO DATA STORAGE ARRAYS

DO 23 Izl,NENEG
00 21 J"'I,NANG
TtI,JI = 0.0
21 RtI,JI 0.0
,00 23 K 1,NOIVT
23 A(I,KI zO.O
TC
RADIANS.
680
APPENDIX]
FORTRAN LISTING (Continued)
CAtCUtATF THE SLAB INCREMEN1 BCUNDS.
C
NOT. NOJ VT+l
TTl
1T/NOIYl
'II:
TlI lJ& 0.0

DO 22 t-2,NOl
22 TIft' TlfL-I' + TTl

C
8EGIN HISTORY GENERATION
C
C
C
C
SELECT A SOURCE POLAR OIRECTION FROM A COSINE DISTRIBUTION BETWEEN

ZERO AND ONE AND SELECT A SOURCE AZIMUTHAL DIRECTION ANGLE BETWEEN
C
C
MtNUS-PI AND PLUS-PI. THEN CAtCUL.aTe OJRECTIOH COSINES.

CALL RANDe la,RI)
CTHE SQRTIRII
G= (THE
CAll RANDlla,AII
PHI = 3.141593*(2.0*Rl-1.0'
STHfA SQRTII.O-C-G.
RHO .. STHEA
AlFA RHO*COS(PHI.
P .. RHO.StNIPHI.
c
c
EA$ :: EO
z-
0.0
NH
= NH.l
s = 0.0
c
C
IF ALL HISTORIES ARE COMPLETE, BRANCH TO THE PRINT AREA.
IFCNHIST-NHI
c
c
c
c
c
c
~,
2, 2
GENERATE THE NEUTRON PAlH DISTANCE AND CALCULATE THE NEW 1

THEN DETER~INE IF THE NEUTRON HAS BEEN
TPAtlfSMITTEO, REFLECTED, AASDRBED, OR SCATTERED.
COOROINUE.
2 (ALL RAND(le,RII
CAlL XSfC(TCS,SCS,fASI
Pl = -AleG(RII/Tes
z
I+C.PL
IFIlT-II 1, 1, 8
C TRANSMITTED - STORE NEUTRON,
1 lOP
=I
THE~
START NEW HISTORY
CAL'l STRA
GO TO I
8 IF HI 9,9,3
C REFLECTED
q
- STORE NEUTRON, THEN START NFW HISTORY
lOP'" 2
CALL SHU
GO TO 1
3 CALL RAND(le,RII
IF(RI-SCS/TCSJ 10, 10, 11
C ABSORBED
- STORE NEUTRON, THEN START NEW HISTORY
11 I CP = 3
CAL L
STRA
GO TO 1
C SCATTERED
C
- CONTINUE CURRFNT HISTORY
APPENDIX]

C
C
GENERATE THE SCATTERED ANGLE (ISCTRCPIC IN eM COORDINATES),

CALCULATE THE ENERGY AFTER SCATTFR, AND THE NEW OI~ECTtON
C
C
COSINES.
10 S
:0:
S+I.0
CALL RANOfIe,RI)
(SA'" 2.0 IU-I.O
El ., EAS (O.5*ll.O+{(AW-l.O Z/IAW+1.0 Z)
1 +ll.O-(IAW-l.O' 2/fAW+l.01 Zt'.CSA))
EAS =El
~
CTHES
Il.O+AW*CSAI/SQPTCl.O+AW*AW+2.0*AW*CSA)
114 ., CTHES
BB '"' SORTII.O-AA.AA)
CAll RANOlle,Rl1
PHIS
:0:
3.141SQ3 -12.0-RI-I.O)
C '" COSIPHIS)
IFIPHIS. 12. 12, 13
12 0
-SORTCI.a-CtC)
eo
:0:
TO 14
13 0
SQRTll.O-CtC)
14 IF((1.O-A8SCG))-COA'tS,
15 AlFAP '" 68*C
BP '" 68*0
15,
16
GP ., AA*C
GO TO 11
16 AlFAP '" 186*C*C*AlFA - eB*O*BI/SQPT(l.O-G*GI + AA*AlFA

RP '" IBR*e_e*e + RR*O*ALFAI/SQRTfl.O-G*r.l + AA*e
GP :0: -B~*C.SQPT(}.O-G*G) + .A*G
11 Alf"A '" AlFAP
A '" RP
G '" GP
C
C
GENERATE A NEW PATH lFNGTH.
c..
GO TO 2
c
c
c
c
c
PRINT AREA
CALCULATE THE FLUX AND COSE FROM THE REFLECTED AND
TRANSMITTED CURRENT. NORMALIZE All RESULTS TO ONE
INCIDENT NEUTRON.
Jel,NANG
COACJI CCOSCANGIJII + COSCANGCJ+lJJt I 2.0
32 COB IJI =CCOSCBNGCJ+lll -COSC8NGCJIII
12.0
00 30 I ... } ,NENFG
00 30 Je},NANG
FTII,JJ '" TII,JJ I CNHIST.COACJII
DTCI,JI = FTCI,JI FTOCII
TCI,JJ '" TlI,JI I NHIST
FRCI,JI '" RII,JI I CNHIST COBIJJI
DRII,JI FRCI,JI FTDCIJ
30 RCI,JI '" ~(I,JI I NHIST
00 31 l-l,NENEG
DO 31 K = I,NOIYl
31 ACI,KI '" All,KJ I NHIST
5 00 32
PRINT THE RESULTS.
WRITE (KY,ItOJ
ItO FORMAT(lHl,'PARTICLES TRANSMITTED THROUGH THE SHIELD'.
WRITE IKl,ltlJ
Itl FORMATIIHO,'FlUXIENERGY,ANGlEI',/J
WR ITE CKY, 100 I
681
682
APPENDIX]

WRITE (K'(,1-101 IANGC"''' M=Z,NNlI
00 42 l-l.NENEG
~2
WRITE(3,601 ECI+ll,(FTCI,JI, J-t,NANGI
4~
FORMAT(lHO,'DOSEfENERGY,ANGLEJ',/1
WRITE
(I<Y,43)
WRITE (I<Y,100.
WRITE (KY,1101 (ANGfMI. M=Z,NNlJ
DO 44 l-l,NENEG
44 WRllE (3,601 Elt+ll. (OTCI,JI, J-l,NANGI

WRITE II<,Y,45.
45 FORMA T( HIO, CURRENT (ENERGY, ANGL E)' , "
WRITE IKY,100,
WRITE (KY,1101 fANGeMI, M-Z,NNI)

00 46 rcl,NENEG
46 WRITE (I<,Y,601 EII+II, (Tlf,JI, J-l,NANGI
WRITFIKY,481
48 FOR~ATIIHl,'PARTtClES REFLECTED BY THE SHIELD')
WRITE
WRITF
(I<.V,411
(I<,Y,100.
WRllE (I<.Y,1101 (8NGIMI,
on
4q
~.2,NNII
I -I ,NFNEG
WRITE II<.Y,601 FII+l), fFRlt,JI,J=l,NANGI
4q
WRITF fKY,4~1
"RITE (KY,lOOI
WRITE
00 50
50 WRITE
WRITE
WRITF
WRITF
(KY,II01 IBNGeMI, "=Z,NNl)

t-l,NENfG
(KY,bO) ECI+l" fOPfl,J', J-I,NANGI
CKY,451
CKY,lOO'
fKY,1101 CBNGCM" ~-2,NNll
0051 l-l,NENEG
WRITE CKY,bOI Efl+ll, CRCI,JI, J-I,NANGI
WRITE fKY,53'
FORMATCIHI,'PARTIClES ABSORBEO BY THE SHIELD"
W1l:ITF (KY,551
FORMATCIHO,'NO. OF PARTICLESCENERGy,THICKNESS.',I.
WRITE CKY,llO'
WRITE CKY,IIO' nUll, taz,NDT)
00 54 1"I,NENEG
WRITE CKY,bO' Efl+ll, CACI,kl, k-I,NDIVTI
FORMATC6EIZ.51
FORMATf34X,'ANGLffRAOIANSI'1
FORMATCbX, 'ENERGY', 5E1Z.5'
FORMATC34X, 'THICKNESS(C~I' I
51
53
5~
54
bO
100
liD
IZO
C
CALL EXIT
fNO
C
C
C
C
C
C
SU8ROUTINE RANOfll,XI
THf RANDOM NUM8ERS ARE SELECTED FROM THE TABLE, IR, AND DIVIDED RY
24 IN ORDER TO GENERATE 23 RANDOM NUM8ERS BETWEE~ 0.04 AND n.97
RAND STEPS THROUGH T~E TABLE 2~ TJ~ES WITH EACH OF THE 22 STEP
LENGTHS, SO EACH OF THE NUM8ERS APPEAR 22 TIMES IN THE SfQUENCF.
BF.FORE IT REPFATS, THUS PROVIOING A TOTAL SEQUENCE LENGTH OF
22-23 OR 506.
DIMENSION IR(23.
DATA lR/22,16,20,2,14,6,19,18,11,8,10,1,1,3,21,9,17,4,5,12,15,
-
13,231
IF ( KK.I' 100,101,100
100 "NT - 23
K - 12
IPT - 1
KK .. -1
101 IF (KNT - 23'
1,102,102
CHANGf STEP SIZE THROUG~ THE TABLE
C.
683
APPENDIX]

102 II<. '" TRIK)
K
K + 1
IF IK -
103 K
:II
104 KNl
22) 10"',104,103
1
.. 0
I KNT = KNT
INITIALIZE GENERATOR AT 11
IF (II - 13) 106,106,105
1 C5
106
101
108
'1 '" C
IF fill 105,10e,107
IPT :10 It
11 = 0
SELECT THE RANDOM
NU~BFR
X = IRfIPT)
X .. X 0.0416661
STEP DOWN TO NEXT
RANOO~
NUMBER
'PT '" 1 PT + II<.
IF (IP1 - 23) 110,1IO,lOQ

109 IPT :: IPl -
23
110 PETURN
END
SUBROUTINE STRl
C
C
C
THIS SU8ROUTINE STORES A NEUTRON WHENEVeR IT IS TRANSMITTED,

REFLECTED, OR ABSORBED.
I
C
COMMON Nt, NNI. NOT. lOP, EASt AHGrb., BNGC61,Tlf211, EISI.

118,6', RIB,6), AI8,201. It G
CALCULATE THE ENERGY GROUP SUBSCRIPT.
00 6 ]:II2,Nl
IF IEAS-EIII) 7,
1,
6 CONTINUE
r ..
J-l
GO TO C4, 5, lOt, lOP

TRANSMITTED - CALCULATE
4 00 2 J'"'2,NNI
IF fCCSIANGfJII-GI
~,
A~GLE
3,
SU8SCRI~T
AND STORE.
2 CONTI HUE
3 J '* J-l
Tfl,JI - Tfl,JI +1.0
GO TO 100
C REFLECTED
- CALCULATE ANGLE SUBSCRIPT AND STORE.
5 on 12 J-2,NNI
IFCG-COSC8NGfJIII 13, 13, 12
12 CONTI ~UE
13J=J-l
RCI,JI - RfI,JI +1.0
GO TO 100
C ABSORBED
- CALtULATE SLAB INCREMENT- SU8SCRIPT AND STOU.
10 00 15 K-2,NOT
IF fl-TlfKII 14, 14, 15
15 CONTINUE
14 K - K-1
Afl,KI 4CI,KI + 1.0
100 RETURN
END
SU8ROuTINE XSECC1CS,SCS,EASI
C
C
C
C
C
THTS SUBROUTINE RE_DS TDT_L _NO SCATTERED CROSS SECTION

DATA THE FIRS1 liME 11 IS CALLED. EACH TIME CALLEO. IT
DETERMINES TCS AND SCS _T ENERGY EAS BY LINEAR INTERPOL_TION
OVER THE UNPUT DATA.
684
APPENDIX]
c
c
C
C
DIMENSION fEf40), 1011401. SCAf401

FQUIV.lfNCECKTFST,KX)
READ INPUT DATA
IF (KTESl 10 KTESl
l' 10, 40, 10
REAOIKX,201 NCSf
REAOIKX,30IfEEII), 1- 1,NCSEI
REAOIKX,301
(TCTIII, I ~ 1,NCSFI
READCKX,301 (SCAli), I = l,NCSFI
20 FORMAT( I 10)
30 FOR~ATf6EI0.31
C
C
C
LOOK-UP CROSS SECTIONS

40 IFIEAS-EEfllJ
50 1=2
50,60. 70
GO TO 100
60 I I
GO 10 ZOO
I . ?,NCSF
10 00 80
IFffAS-EEfIJI 100, 200, 80
80 CONTI NUE
I NCSE
100 FACT'
Tes:o
SCS =
'EAS-FElI-11I1 'FFIII - fEll-III

TOTO-II + CTOTflJ-TOTtI-lH FACT
SCA(J-lJ + (SCAI I) - SCA(J-tll FACT
GO TO 300
200 TCS = TOT/II
SCS = SCAIII
300 RETURN
FNO
TEST PROBLEM
A test case has been run with tbe computer procedure for 3-MeV neutrons
incident on a 2-cm-thick semi-infinite slab of silicon. The reflected and
transmitted neutrons are sorted into six energy intervals and five angle
intervals, and the absorbed neutrons are sorted into the same six energy
intervals and into ten slab increments. The cross sections j that were input are
not realistic, but they make the neutron mean-free-path distance approximately 3 to 4 em. One-hundred histories were run; therefore one should not
expect the answers to be nonzero except for a few storage bins. Both the
input and output data are shown following the program listing. Summing the
output neutron current will show that 19 neutrons were reflected, 42 were
transmitted, and the 39 remainder were absorbed.
APPENDIX I
685
TEST PROBLEM
Input Data
100
3.00+00
0.00+00
2.00-09
0.00+00
3.00.00
5 8237
2.00+00
1.50+01
3.20-09
5.00-01
1.00-05
3.00+01
3.90-09
1.00+00
2.80+01
".50+01
It. 10-09
1.50+00
0.25
0.50
1.75
0.80
0.42
0.20
2.00
0.70
0.75
2.25
0.62
0.28
0,.28
6.00+01
It. 20-09
2.00+00
9.00+01
1t.42-09
2.50+00
12
0.00
1.50
0.98
0.1t5
0.10
0.29
1.00
2.50
1.25
0.55
3.00
0.1t9
O.ltl
0.41
0.40
0.37
0.23
0.25
0.28
0.28
0.31
0.26
0.27
Output Data
PARTICLES TRANSMITTED THROUGH THE SHIElO

FlUX(ENE~GY.ANGlE'
ANGlEIRAOIANSI
ENERGY 0.26180E 00 0.52360E 00 0.78540E 00
0.50000e 00 0.0
0.0
0.0
0.10000E 01 0.0
0.0
0.0
O.15QOOE 01 0.0
0.0
0.0
0.20000E 01 0.0
0.0
0.0
0.25000E 01 0.0
0.12713E-Ol
0.0
0.30000E 01 0.30520E-Ol 0.13101E 00 0.17799E 00
0.10"'12E 01 0.15708E 01
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.40000e-Ol
0.0
0.16569E 00 O.'OOOOE-Ol
DOSE.ENERGY,ANGlE.
ANGtEeRADUNS)
ENERGY 0.26180E aD O.52360E 00 0.18540E 00

0.50000E 00 0.0
0.0
0.0
0.10000E 01 0.0
0.0
0.0
0.15000E 01 0.0
0.0
0.0
0.20000E 01 0.0
0.0
0.0
0.25000E 01 0.0
O.53397E-I0
0.0
0.30000E 01 0.13490E-09 0.51905E-09 0.78671E-09
O.10lt12E 01 O.15708E 01
0.0
0.0
0.0
0.0
0.0
O.73233E-09
0.0
0.0
0.0
0.0
O.16800E-OQ
O.17680E-09
CURRENT{ENERGV.ANGLE)
ANGlE{RADIANS)
ENERGY
0.50000E 00
0.10000E 01
0.15000E 01
0.20000E 01
O.Z5000E 01
0.30000E 01
D. 2b 18DE 00 O.52360E 00 O.78540f 00 O.10412E-' 01 0.15708E 01
0.0
0.0
0.0
0.0
0.0
0.30000E-Ol
0.0
0.0
0.0
0.0
0.0
O.12000E 00
0.0
0.0
0.0
0.0
D.laoaDE-DI
O.14000E 00
0.0
0.0
0.0
0.0
0.0
O.IGOaOE 00
0.0
0.0
0.0
0.0
a.IDaaaE-a1
0.10000E-Ol
(Test problem continues on next page.)
686
APPENDIX]
TEST PROBLEM (Continued)

Output Data (continued)
PARTICLES REFLECTED B'l THE SHIELD
FLUXIENERGY,ANGLEI
ANGLEIRADIANSI
ENERGY 0.28798 01 0.26180E 01 0.23S62E 01 0.209"E 01 0.15708E 01
O.SOOOOE
O.IOOOOE
0.15000E
0.20000E
0.2S000E
0.30000E
00
01
01
01
01
01
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
O.D
O.~OO~OE 00 0.12S8SE 00 0.0

O.~OOOO-Ol
0.40040E 00 0.6292SE 00 0.67S98E 00 O.~ODOOE-OI
DOSE(ENERGY,ANGlE)
ENERGY
O.SOOOOE 00
O.IOOOOE 01
O.ISOOOE 01
0.20000E 01
0.2S000E 01
0.30000E 01
ANGLEIRADIANSI
0.28798E 01 0.Z6UOE 01 0.23S62F 01
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
o .16817E-08 0.S2857E-09
0.0
0.0
0.n698E-08 0.Z7813E-08
0.209~~E
0.0
0.0
~.O
01 0.15708E 01
0.0
0.0
0.0
0.0
0.0
0.0
O.100E-09
0.29878E-08 O.nUDE-I.
CURRE NT I ENERGY. ANGt E)
ANGl E I RAD IAN 5 I

0.209~~E 01
0.0
0.0
0.0
0.0
0.0
0.20000E-01 O.IOOOOE-Ol 0.0
0.20000E-01 O. SOOOO-Ol O. lO000E-01
ENERGY O.28798E 01 0.26180E 01 0.23S62E 01
0.500 DOE
0.10000E
O.ISOOOE
0.20000E
0.2S000E
0.30000E
00
01
01
01
01
01
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.,0
0.0
0.0
0.0
0.0
0.0
.o.15701E 01
0.0
It.o
0.0
0.0
O.lOOIDE-II
0.10000-01
PAR71CLES AISORBED BY THE SHIELD
NO. OF PARTIClESCENERGY,THICKNESSI
ENERGY
O.SOOOOE 00
0.10000E 01
0.15000E 01
0.2oo00E 01
0.25000E 01
0.30000E 01
THICKNESSIC"1
0.40000E 00 0.80000E 00 0.12000E 01 0.\60DOE 01 0.20000E 01
0.0
0.0
0.0
0.0
0.10000E-Ol 0.0
0.0
0.0
0.0
0.0
0.0
0.11000E 00 0.60000E-Ol 0.1I000E 00
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.60000E-Ol
0.0
0.0
0.0
0.0
0.0
O.<IODOO-OI
Moments-Method Results
for Fission-Neutron
Penetration in Be~ C~ CH~
CH2~ H~ and H20t
Appendix K
Source strengths are 1 neutron/sec for the point isotropic fission cases and 1
neutron cm- 2 sec- 1 for carbon, a plane isotropic fission source.
Table K.I-DiFFERENTIAL NUMBER SPECTRA AND DOSE FOR BERYLLIUM, A POINT ISOTROPIC
FISSION SOURCE
t,
E,MeV
10
20
0.330
0.383
0.445
3.15 (-1)1
3.27 (-1)
3.3. (-1)
3.44(-1)
3.51 (-1)
2.73
2.1
3.30
2.7.
2.77
(-1)
(-1)
(-1)
(-I)
(-1)
0.601
3.53 (-1)
0.734
1.10
1.34
3.5. (-1)
(-1)
(-I)
(-1)
(-1)
(-1)
1.2.
1.63
3.37 (-1)
3.1' (-1)
2.80 (-1)
I.
3.12
2.2
2.5.
2.25
1.28 (-1)
2.00
2.44
2.69
2.98
3.64
2.3. (-1)
1..0 (-1)
1..5 (-1)
1.40(-1)
.36 (-2)
1..2 (-1)
1.'2(-1)
.6' (-2)
7.45 (-2)
. 10 (-2)
1.07
6.72
2.70
2.7'
2.'
(-1)
(-2)
(-2)
(-2)
(-2)
4.44
5.43
6.63
8.10
9.89
5.5. (-2)
2. 5 (-2)
1.22 (-2)
'.12 (-3)
1.05 (-3)
4.42(-2)
2.2. (-2)
.20 (-3)
3.12 (-3)
.53 (-')
2.23
LI.
'.02
1.5.
'.55
(-2)
(-2)
(-3)
(-3)
(-.)
12.1
14.75
18.0
1.87 (-.)
2.17 (-5)
1.'7 (-6)
1.'2 (-')
1.51 (-5)
9.1. (-7)
7.12 (-5)
9.43 (-.)
'.17 (-7)
1.01 (-2)
7.5 (-3)
glcm'
30
.0
.0
120
Number Spectra. 411'r' No (r,E), neutrons 1-1 MeV-I
0.492
0.572
1.03 (-1)
1.2. (-1)
1.02 (-1)
1.06 (-1)
3.57
3.5
.
3.
'.21
(-2)
(-2)
(-2)
(-2)
(-2)
1.82
2.21
2.
2. 5
3.23
(-3)
(-3)
(-3)
(-3)
(-3)
1.52 (-.)
I. (-4)
2.52 (-.)
2.3. (-4)
2.7 (-')
l.35
I.
2.27
2.15
2.
(-5)
(-5)
(-5)
(-5)
(-5)
7.'6 (-2)
l.33(-1)
(-1)
1.40 (-1)
3.13
5.7.
.2
7.7.
(-2)
(-2)
(-2)
(-2)
(-3)
(-3)
(-3)
(-3)
(-2)
2.'2(-4)
. (-4)
7.30 (-4)
. (-4)
1.03 (-3)
2.19
O.
6 .5
.
(-5)
(-5)
(-5)
(-5)
7.41 (-2)
2.5.
5.12
7.4.
.6.
1.00
.0. (-2)
3.1. (-2)
LlO (-2)
(-2)
1.50 (-2)
7.2
2.51
.1.
1.2.
1.78
(-3)
(-3)
(-.)
(-3)
7.
1.83
8.13
1.21
1.81
.1
1.5'
.31
1.23
I.
(-5)
(-5)
(-.)
(-5)
(-5)
1.22
6.64
2.29
.93
2.72
1.62
.5
3.51
1.50
5.14
(-3)
(-.)
(-')
(-4)
(-5)
2.05
1..0
5.0
2.97
LI.
(-5)
(-5)
(-.)
(-.)
(-6)
1.11 (-1)
1.2.
(-2)
(-3)
(-3)
(-4)
(-.)
'.13 (-5)
5.93 (-.)
2.32(-7)
(~3)
7. 5 (-.)
1.19 (-6)
3.75 (-.)
(-4)
(-.)
(-5)
(-4)
(-4)
1.83 (-4)
1.18
'.50
2.16
.07
(-4)
(-5)
(-5)
(-.)
9.44 (-5)
I..'
1.22 (-.)
2.15 (-7)
5.63 (-9)
(-7)
3.61 (-.)
.34 (-10)
2.3. (-5)
2.3. (-.)
Dose,4'1fr' D(r), D in mrads/hr
3.69 (-3)
1.91 (-3)
2.33 (-')
tRead: 3.15 X IO-,ete.
t All data in this appendix have been taken from H. Goldstein, Fundamental Aspects of Reactor
Shielding, Addison--Wesley Publishing Company, Inc., Reading, Mass., 1959.
687
'"
00
00
Table K.2-DIFFERENTIAL NUMBER FLUX FOR CARBON. A PLANE ISOTROPIC FISSION SOURCE
No(x,E), neutron cm-2 sec- 1MeV- 1
x,g/cm 2
E,MeV
10
20
18.02
12.08
8.10
5.43
3.64
0.6453
0.8331
0.1512
0.1580
0.4676
X .10-
2.44
1.63
1.1
0.73
0.49
0.1748
0.3407
0.5311
0.7516
0.8507
0.33
0.221
0.148
0.099
0.066
0.7365
0.9966
0.1442
0.2113
0.3120
0.012
0.3 X 10-'
0.1401
0.4819
10-4
10-2
10- 1
Hr l
0.3134
0.4080
0.6391
0.8383
0.2194
10
10
10
10
10
0.8833
0.1819
0.2930
0.4234
0.4848
10
10
10 1
10 1
10 1
0.4175
0.5694
0.8384
0.1252
0.1888
X
X
X
X
X
X
X
X
X
X
X
X
X
10
10'
0.9070
0.3630
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
30
10-
10-'
10- 3
10-2
10- 1
0.1585
0.2076
0.2856
0.4576
0.1063
10- 1
10
10
10
10
0.4592
0.9959
0.1653
0.2436
0.2820
10
10
10
10 1
10 1
0.2419
0.3322
0.4966
0.7552
0.1159
10'
10'
0.5927
0.2740
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
90
60
10-
10-'
10- 3
10- 2
10- 1
0.2995
0.3942
0.4442
0.9661
0.1708
10- 1
10- 1
10
10
10
0.8427
0.2031
0.3596
0.5518
0.6543
10
10
10
10
10 1
0.5585
0.7783
0.1197
0.1885
0.3005
0.1765
0.1191
10
10'
X
X
X
X
X
X
X
X
X
X
X
X
10- 7
10-'
10-4
10-3
10-2
0.1020
0.1331
0.1445
0.3227
0.5213
120
10-7
10-'
10-4
10-3
10- 3
0.4588
0.6025
0.5984
0.1444
0.2262
10-8
10-
10-'
10- 3
10- 3
10-2
10- 1
10- 1
10- 1
10- 1
0.2502 X
0.6070 X
0.1087 X
0.1682 X
0.2008 X
10-2
10-2
10-1
10- 1
10- 1
0.1010 X
0.2370 X
0.4183 X
0.6427 X
0.7652 X
10-2
10-2
10-2
10-2
10- 2
10- 1
10- 1
10
10
10
0.1722 X
0.2406 X
0.3713 X
0.5881 X
0.9466 X
10- 1
10- 1
10- 1
10- 1
10- 1
0.6598 X
0.9193 X
0.1408 X
0.2214 X
0.3542 X
10- 2
10-2
10- 1
10- 1
10-'
10 1
10'
0.5892
0.5277
10
10'
0.2206
0.2385
10
10'
X
X
X
X
X
X
X
X
X
X
X
X
)-
""
."
"'Z
tl
X
i"
Table K.3-D1FFERENTIAL NUMBER SPECTRA FOR CH. A POINT ISOTROPIC FISSION SOURCE
41Tr2No (r,E), neutron sec-I MeV- 1
r, g/cm 2
E,MeV
10
20
30
60
90
120
18,02
12.08
8.10
5.43
0.9312
0.1100
0.1906
0.1397
X 10- 6
X 10- 3
X 10-'
X liJ '
0.4418
0.4664
0.6603
0.5142
X liJ6
X 10-'
X 10- 3
X 10-'
0.2511
0.2421
0.2843
0.2194
X 10- 6
X 10-'
X 10- 3
X 10-'
0.4272
0.3167
0.2005
0.1401
X 10- 7
X liJ'
X 10-'
X 10- 3
0.6668
0.3760
0.1392
0.8371
X 10-8
X 10- 6
X 10-'
X 10-'
0.1009
0.4431
0.1099
0.5378
X 10-8
X 10- 7
X liJ6
X liJ 6
3.64
2.44
1.63
1.10
0.3850
0.9867
0.1481
0.2007
X 10- 1
X liJ'
X 10
X 10
0.1099
0.2806
0.4117
0.5422
X liJ'
X 10-'
X 10-'
X liJ'
0.3872
0.8970
0.1250
0.1587
X 1iJ2
X 10-'
X 10-'
X liJ'
0.1620
0.2824
0.3424
0.4109
X 10- 3
X liJ 3
X 10- 3
X 10- 3
0.8243
0.1276
0.1476
0.1753
X 10-'
X 10-'
X 10-'
X 10-'
0.5170
0.7847
0.9034
0.1073
X 10-6
X liJ6
X 10- 6
X liJ'
0.734
0.492
0.33
0.2560 X 10
0.2981 X 10
0.3711 X 10
0.6802 X liJ '

0.7879 X 10-1
0.9896 X liJ 1
0.1940 X 10- 1
0.2210 X 10-'
0.2718 X 10-1
0.4947 X liJ'
0.5642 X 10- 3
0.6958 X 10- 3
0.2102 X liJ'
0.2392 X 10-'
0.2929 X 10-'
0.1287 X 10-'
0.1461 X 10-'
0.1782 X 10-'
?;
."
"'
Z
0
:><
Table K.4-DIFFERENTIAL NUMBER SPECTRA FOR CH,.o. A POINT ISOTROPIC FISSION SOURCE
41l'r2 No (r,E), neutron sec- t Me"t
r, g/cm 2
E,MeV
10
20
30
60
90
120
18.02
12.08
8.10
5.43
0..7280
0.7751
0.1204
0.7724
X Ilf 6
X Ilf4
X Ilf'
X Ilf2
0.3087
0.2769
0.3352
0.2003
X Ilf 6
X 10-4
X Ilr'
X Ilf2
0.1490)( Ilf6
0.1140 X 10-4
0.1041 X 10-'
0.5458 X 10-'
0.1425
0.6554
0.2627
0.9009
X 10- 7
X 10- 6
X 10-'
X 10-'
0.1282
0.3680
0.8284
0.1971
X 10-8
X 10- 7
X 10- 7
X 10- 6
0.1114
0.2118
0.3221
0.5711
X Ilf9
X 10-8
X 10-8
X 10-8
3.64
2.44
1.63
1.10
0.1883
0.4193
0.6053
0.8130
X 10- 1
X Ilf'
X 10-'
X Ilf'
0.3739
0.7368
0.9916
0.1258
X Ilf 2
X Ilf2
X Ilf2
X Ilf l
0.8267
0.1424
0.1793
0.2207
0.9384
0.1294
0.1487
0.1779
X 10-'
X 10- 4
X 10-4
X 10-4
0.2049
0.2847
0.3300
0.3960
X Ilf 6
X 10- 6
X 10-6
X 10-6
0.5743
0.7728
0.8774
0.1045
X 10-8
X 10-8
X 10-8
X 10- 7
0.734
0.492
0.33
0.1039 X 10
0.1259 X 10
0.1612 X 10
0.1538 X Ilf l
0.1806 X 10- 1
0.2246 X 10- 1
X 10-'
X 10- 2
X 10- 2
X 10- 2
0.2665 X 10- 2
0.3109 X 10- 2
0.3831 X 10- 2
0.2138 X Ilf 4
0.2504 X 10-4
0.3105 X 10-4
0.4745 X 10- 6
0.5523 X 10-6
0.6769 X 10-6
0.1255 X 10- 7
0.1470 X 10- 7
0.1820 X 10- 7
':;
."
'"
tl
~
:><
Table K.5-DIFFERENTIAL NUMBER SPECTRA FOR H2 , A POINT ISOTROPIC FISSION SOURCE

4rr,2 No (r,E), neutron sec-I MeV- t
r, em
E,MeV
10
18.017
13.348
8.947
5.997
0.1047
0.43476
0.11671
0.87118
3.824
1.555
0.7344
0.33
20
X 10-'
X Hf4
X 10-2
X Hf 2
0.032608
0.10505
0.16057
0.22197
0.6581
0.2814
0.63120
0.39413
30
X 10-6
X Hf4
X Hf 3
X Hf 2
0.012542
0.035939
0.057007
0.082713
60
90
X 10- 6
X 10-4
X 10- 3
X 10- 2
0.1322
0.49575
0.56543
0.18519
X 10-6
X 10-'
X Hf4
X 10- 3
0.42248
0.13801
0.97823
0.21545
X Hf 7
X 10-'
X 10-'
X 10-4
0.4784 X 10-2
0.011680
0.019296
0.029512
0.33947
0.59143
0.81843
0.11020
X 10- 3
X 10- 3
X 10-3
X 10-2
0.32015
0.55322
0.69773
0.83534
X 10-4
X 10-4
X 10-4
X 10-4
0.4415
0.18227
0.34302
0.17905
'"OJ
Z
><
~
Table K.6-DIFFERENTIAL NUMBER SPECTRA FOR WATER, A POINT ISOTROPIC FISSION SOURCE
411'r 2 No(r,E), neutron sec- l MeV- 1
',em
E, MeV
10.9
6.0
2.7
1.1
0.33
10
0
4.64
1.91
1.65
3.37
3.15
X 10-4
X 10- 2
X 10- 1
X 10- 1
X 10- 1
2.02
7.54
5.86
1.20
2.37
X 10-4
X 10- 3
X 10-2
X 10- 1
X 10- 1
20
8.77
2.90
1.62
3.14
5.06
X 10-'
X 10- 3
X 10- 2
X 10- 2
X 10- 2
60
30
3.79
1.10
4.37
6.70
1.20
X 10-'
X 10- 3
X 10- 3
X 10-3
X 10- 2
3.01
5.79
1.07
1.47
3.26
X 10- 6
X 10-'
X 10-4
X 10-4
X 10-4
90'
2.35
3.07
4.14
5.62
1.01
X 10- 7
X 10-6
X 10- 6
X 10-6
X 10-'
120
1.82
1.68
2.16
2.99
5.37
X 10-8
X 10- 7
X 10- 7
X 10- 7
X 10- 7
..,
""
~
Gamma-Ray Differential
Energy Spectra for
Water and Lead
Appendix L
r,
1000
r,
'\
100
.......
r-...
'"
~o
N~
r-...
10
.........
...........
-...... f.....
.......... '-
...
20
15
10
7
4
."..
.....
1/
1"-' -
r---
0.1
0.03
0.1
0.2
0.3
0.4
0.5
E, MeV
Fig. L.t-Point isotropic O.5-MeV gamma-ray source in water.
692
693
APPENDIX L
1000
,\
'1\
, \.
100
,"
.......
'\.
-, '\. ~'--
N~ 10
."
'\.
, "'
f'\.
-r-
..........
1\
0.1
r--
'\.
"-
--
20
15
10
7
4
......
1"-
0.1
.........
"-
"- I'.....
--
'"
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
110 r '"' 1
1.0
E, MeV
Fig. L.2-Point isotropic l.O-MeV gamma-ray source in water.
APPENDlXL
694
100
20
~~
10
15
~
t...-
i,~
10
I--
~ 7
l,....-
r-...
I...-
I-- ~
1'\
\
\
"
o.1
Jlr=1
0
~ I-
3
E. MeV
Fig. L.3-Point isotropic 4.0-Me V source in water.
695
APPENDIX L
100
10
-- -- ---- - --
I.
"
I----
'~
.....
\ .....
II~\
,,
\
\
----
--
I--
10
7
I-I.--
f.-- I-2
!...
.......
II. , 1
f.--
15
I-"
0.1
20
3
E,MeV
Fig. LA-Point isotropic 6.0-Me V gamma-ray source in water.
696
APPENDIX L
15
10
10
'/
'/
11
WV
1
/h
~V...-V
11,-1=
0
./
...-
(//1 V
~~
'Iff.V
I
0.0 1
15
0.00 1
1
1
0.1
0.2
0.3
0.4
0.5
E, MeV
Fig. L.S-Point isotropic O.5-MeV gamma-ray source in lead.
697
APPENDIX L
100
~/
10
20
15
10
/
/
'" . /
V/ V
h 1//V
I,, ,
1/V
V
1:
'/
// /
'j
./'"
./
./
.---
.-
1o ' , - 1_ -
./
;'
i'"
'j
1/,/
II
.
0, 1
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
E, MeV
Fig. L.6-Point isotropic l.O-Me V source in lead.
1.0
698
APPENDIXL
100
10
I V
1'-
""-
... .....
/'
10
I.-./
.......
II
15
l"-
II
20
......
l...- V
L-'
/l0 r ' 1
I/
o. 1
Fig.
L.7-Poi~t
3
E.MeV
isotropic 4.0-MeV gamma-ray source in lead.
APPEND/XL
699
100
"\.
f
,/
10
......
~ \
.....
r-...
.......
'-.
...........
1/
I
t\.
......
..........
.............
r- '--
'" .......
/lor:: 20
............
15
...... r--
10
7
,
"I
0.1
l..- l--
0.0 1
E. MeV
Fig. L.S-Point isotropic 8.0MeV gamma-ray source in lead.
Graphs for Neutron

Attenuation Calculations
Appendix M
This appendix consists of graphs of neutron-penetration results obtained
from Monre Carlo calculations by Clark, Betz, and Brown I and by Allen and
Futterer. 2 Table M.l gives the composition of materials used in the
calculations; Figs. M.l to M.lO show the dose (in ergs/g per incident
neutron/cm 2 ) due to monoenergetic beams of neutrons normally incident on
ordinary concrete, both on a slab andon a semi-infinite medium (half space).
Figures M.ll to M.l5 show the neutron-dose transmission factor as a
function of polyethylene thickness for monoenergetic neu trans incident at
various angles. When the thickness is adjusted according to the key included
at the top of each figure, these latter curves apply also to water, to concrete,
and to Nevada Test Site soil, both dry and water-saturated.
The use of these curves is explained in Sec. 6.4.3.
REFERENCES
1. F. H. Clark, N. A. Betz, and J. Brown, Monte Carlo Calculations of the Penetration of
Normally Incident Neutron Beams Through Concrete, USAEC Report ORNL-3926,
Oak Ridge National Laboratory, January 1967; F. A. R. Schmidt, Revised Neutron
Ketma Values at the Boundary of a Semi-Infinite Concrete Medium, USAEC Report
ORNL-TM-2284, Oak Ridge National Laboratory, 1968.

2. F.
J. Allen and A. T. Futterer, Neutron Transmission Data, Nucleonics, 21(8): 120
(1963).
700
APPENDIXM
701
Table M.1-COMPOSITIONS OF MATERIALS USED FOR

NEUTRONTRANSMISSION CALCULATlONSt
Material
Borated polyethylene
(8 wt.% B4 C):1:
Densitl'
glcm
0.97
Composition
Element
10 21 atoms/em 3
76.80
39.20
0.658
2.67
H
C
lOB
liB
Water
1.00
H
0
66.90
33.45
Concrete
2.26
H
0
13.75
45.87
1.743
20.15
Al
Si
Nevada Test Site
1.15
soil (dry)
H
0
8.553
22.68
2.014
9.533
Al
Si
Nevada Test Site
soil (100% saturated)
1.25
H
0
16.87
27.00
1.976
8.963
Al
Si
tFrom F. J. Allen and A. T. Futterer, Nucleonics, 21(8): 120 (1963).
*Several calculations made for pure polyethylene slabs (p 0.925 g/cm

up to 6 in. thick yielded approximately the same neutron transmission results
=
as those for the borated polyethylene.
APPENDIXM
702
CONCRETE BETWEEN SOURCE AND DETECTOR, em
60
BO
,::, ,
100
120
_'_~~ .c'c.
1---"'--- -
140
160
=i= . c . . ,
lBO
,,==:= 10-
-=t=-- += '
~
Fig. M.l-Transmission of 14-MeV normally incident neutrons in ordinary concrete.

(From Clark et at' )
APPENDIXM
703
CONCRETE 8ETWEEN SOURCE AND DETECTOR, em
60
80
100
120
10~
--
140
160
180
1~7
-.;:
~,
10-7
10-8
I~
"'.1 !-
"-'
10- 11
Slab results
Half space results
I
1
",
,
10- 12
10- 13
10-1 2
W
~
00
M
100
lW
Fig. M.2-Transmission of 12-MeV normally incident neutrons in ordinary concrete.

(From Clark et al. I)
706
APPENDIXM
CONCRETE BETWEEN SOURCE AND DETECTOR. em

60
BO
100
120
140
160
180
10-6
1
g
~
==
.~.. o:::~Ff=c-.'== 10-7

_~-i=C
---_.
.\----
"_.. , .... ,_
. . __
:.:c..
...
~"~~!~~~~_~~~~M~~.C~=..-=
==: _,,__....:1:::=.-.-_
10- ~.-"l-"-.-
t--f--_._- ......
",...,I'-...........-...
-j~~ , -
---+--
10...
N
10-9
=== c= i===:"===
10-8
<i
:;
-"c
g
,
~~~~~~~~"~""'~~-~~~"-~'~"'~~~~10.1O 1
10- 1-------.
""~-
.....:. f--
1----1- .--
iii
--1--- . . f -
",-
:;
..
-,,-,
"
"...
iii
~ ~~~4~~~~~~~~~
~~~~~~~
~~
~=::::::.-:::t=='=-I--~. ==-~~
~_..-:'
10
~ 10gJ
--
--1---
11
10-
r==
c=== F
Slab results
~ - - Half-space results
10- 12 L_---l
20
11
10..
,,"-
~!~~~~~~
-t--------
'--+-'-+.
12
10-
_~
..L
I __L_---l_ _-.L_ _...J 10-13
40
60
BO
100
120

(From Clark et al. I )
APPENDIXM
707
CONCRETE BElWEEN SOURCE AND DETECTOR, em
100
1~6
1~
.____.0_ .
1~
_.
lW
lW
"
10..7
10-8
",
.....'.
",
"
+.....",
",
<i.
~ 10- 10
II:
"
:::>
i=
1~7
..... '.
'\:
'\ ~"
10- 1
......
2=
Slab results
Half-space results
"
100
10- 14
1~
Fig. M.6~Transmission of 4-MeV normally incident neutrons in ordinary concrete.

(From Clark et al.' )
708
APPENDIXM
100
1~
1m
1~
1M
200
1~
1o~
Slab results
,0-
==
-
Half space results
"""
\.'
10- 10 N
~g
----i~,
,\,
10- 11
"
<i.
,,'
\:,
1\'!----
::;; 10-10
a:
w
10- 12 :!:
a:
w
"
"~ 10-1
.~
\:\
"w
"
0- 13 ~
;:
\\
\'.
10- 14
10- 13
20
40
60
10- 15
80
100
120

(From Clark et ai,' )
709
APPENDIXM

BD
'DO
120
140
160
'80
200
,~
O~
10-7
"'"
t'-..'
",
10-9
.-,,,
'\.
"....
'\.
'"
<t'
".
'\. '.
y"
0- 12
::;
a:
~
OJ
!!l
;::
"
===
Slab results
- Half-space results
10- 12
"', 1
"
'00
0 ..14
lW

(From Clark et al. 1 )
APPENDlXM
710
10- 11
'\.\
\\
1\\:--
1\\
10- 13 N
'\
10- 14 ::::::
- \\
",'
10- 15 :E
a:
w
><
w
::J
~\
gJ
10-16 ~
~,
10- 12
10- 17
Slab results
f- - 10-13
Half-space results
\\
W
10- 18
100
1W
CONCRETE BETWEEN SOURCE ANO DETECTOR, em
Fig. M.9-Transmission of 1.3~MeV normally incident neutrons in ordinary concrete.

(From Clark et aI, I)
APPENDIXM
711
,0-<
10-"
.--~_.-
10-7
~,
\\
r-.
0- 12
0- 13 N
~ ~
0- 14
!
i
.!'
~\
. ,,c
--".
.\- 1'-.-.C
f-.
\\
--
10- 17
10- 13
- - -
Slab results
Half space results
~
\\
~
10- 18
00
'00
1~
Fig. M.10-Transmission of O.7-MeV normally incident neutrons in ordinary concrete.

(From Clark et al. ')
712
APPENDIXM
2.5 em of polyethylene = 3.1 em of water

= 4.6 em of concrete
= 6.1 em of Nevada Test
Site soil (100% saturated) = 8.3 em of Nevada Test
--
~.
."
_L
1/
"......
't.
~
~ 10- 1
~
u.
1-..
""
"'-- 1'>:<'-
\"
""l:..
i'il..
~n
1." Ih
""- It>...
-_.._.
"'-
_.-
'"0\.... '" "-~ ""

,,
~::;
"-
""-
<n
-._--- - -
..:
a:
~~---
\""
I-
"
'll. ,
~ 10. 2
Site soil (Area 71
--
---'
--
-,,-'
~~.
_.
--I--.
-
0. concrete or NTS
45. concrete or NTS
- : 7~:: :~:;:~:I~~eN:rS
water
D 45. polyethylene or water
f-
f--
I~
o 70. polyethylene or water

-
-_.......
--
","---
'"
~'
=-~
""-
10
15
20
25
30
THICKNESS OF POLYETHYLENE, em
Fig. M.II-Penetration of 5-MeV neutrons incident on several materials at various

angles. (Based on data from Ref. 2.)
713
APPENDIXM
,0"
2.5 em of polyethylene
u,..---
.~.
,,\,
::0 3.1 em of water

= 5.1 em of concrete
Site soil (100% saturated)

8.9 em of Nevada Test
Site soil (Area 7)
8'l
l.
~,
"'-.. ...... ~
.-
"\.
a:
o
t;
'-
'" 10-2
LL
"-
"
Ml
"-
'" "-
"
.......
0-..
i'zA
"- ' ... ,
...'a:"
"
.0
~ 1(13
I-
0-"-
'"
'"
' ...
_.
_..
0. concrete or NTS
45. concrete or NTS

10-4 ~ 70. concrete or NTS
~ 0
0".
polyethylene or water
a 45. polyethylene or water .
f---
70. polyethylene or water
10- 5
10
'5
20
25
30
THiCKNESS OF POLYETHYLENE. em
Fig. M.12-Penetration of 3-MeV neutrons incident on several materials at various

714
APPENDIXM
I~
10"
"""l.
Q'~
1
2.5 em of polyethylene
.. "'".... ,
~
3.0 em of water
'" 4.7 em of concrete

Site soil (Area 7)
"'~I
,-....-
a:
.. 0. 2
z 1
-......
--0>
'I..
'"
iii
a:
'" f\, "" ~
,.
.....
f\: "
f-
~ 10-3
o
o
.....
.."
'""
'\
,.,,
"-
to
1-0 45. polyethylene or water
1-0 70. polyethylene or water
I
5
0"\
0. concrete or NTS
45.
concrete or NTS
1=. 70. concrete or NTS
~.
10
15
20
25
0'"
'"
'\
30

715
APPENDlXM
2.5 em of polyethylene - 2.7 em of water
'-'e..
-'It..\ -::'i.
- 4.3 em of concrete
'" 5.4 em of Nevada Test
'~
~~~~~~~~I~~~~~~~~-~B~.~1Ee~m~O~f~N~ev~ad~a~T~est~='
Site soil (Area 71
10-
a:
"\. '..
::
'. ~
\1 ", 'b,;.,~, ~
~ 1O-2~~gg~;;s:.~~
u;
'"~
~
......
'\
'~.,
~~
' ...~
~'e,
~ ~~~!~~~~~~M~~~f!~~~~~'
~
,
,
~
10-
10-4
0, concrete or NTS
45. concrete or NTS

70, concrete or NTS
:: 0
10..
...
-.
"-
045, polyethylene or water
- 0 70,
1"-'
\.
10-5 L -_ _L-1_ _-l.-1_----.l_ _---'------"_.L"Z--"'''-',,'--.J.J
polyethylene or water
10
15
20
THICKNESS OF POLYETHYLENE. em
25
30

716
APPENDIXM
"'=:=
2.5 em of polyethylene - 2.7 em of water

5.3 em of concrete
-~ -.,,~
- 6.B em of Nevada
Test Site soil (100%
saturated)
- 9.5 em of Nevada Test ::::::::::
Site soil (Area 7)
_
~"
\.',
"
h.:',
"\..
~.
Dc, concrete or NTS

45. concrete or NTS
10-5 ~ 70. concrete or NTS
t-
F 0
-~, ,,~
\.
-, \~,
f- 0 45. polyethylene or water
o 70, polyethylene or water
10-6
10
I~ ~' ,.'\. "

15
20
25
30
Fig. M.1S-Penetration of O.S-MeV neutrons incident on several materials at various

Graphs of the
'f! Function
Appendix N
A useful function in calculating the shield penetration of the uncollided

secondary gamma-tay flux is the
function, defined in Sec. 6.5. The
functions (n > 0) can be used to calculate the scattered flux. Values of these
functions have been given by Schmidt' for a slab shield and by Claiborne'
for a semi-infinite shield. The functions are plotted in Figs. N.1 to N.6, all of
which are taken from a compilation of Schmidt.'
"'n
"'0
REFERENCES
1. Fritz A. R. Schmidt, The Attenuation Properties of Concrete for Shielding of Neutrons
of Energy Less than 15 MeV, USAEC Report ORNL-RSIC-26, Oak Ridge National
2. H. C. Claiborne, Analytical Solutions for Heat Generation Distributions in Regular
Geometries, in Engineering Compendium on Radiation Shielding, Article 7.3,
Springer-Verlag, New York, 1968.
717
718
APPENDIXN
10"
,~~
10-3
~ ~IO~
"-
~ ~l~ f'..."'"~
,"'-1
~ ~0.6'" ~""- ""!~ O.~ [\' 1"'-
"'" ~
W0
10-"
~.\\ 1\\I~
10-7
1~ [\\
10-"
"'" ~
10- 10
Fig. N.la-The function
12
16
20
.t
1\ r\
\1\\\ \\ \
10-9
"" "
"""
\ '\\ \
24
28
r\ I~
~
32
r\
36
40
"'0 (Ill,a) fot a ~ 1 fot a slab shield. (From Schmidt.')
APPENDlXN
719
10"
H; 1
10-2
~,-= 1.0
I~
~ ~~"
~ ~ ~<3
1.2
~ ~~
1)~
~~
10-"
10-7
2.011~
2.5
10-8
3.0~
"o
10
12
"'
14
~ I'\.
~ .1'\..
~~
16
18
20
'"
22
24
Fig. N.lb-The function >/10 (l1',ex) for ex;;. 1 for a slab shield. (From Schmidt.')
APPENDIXN
720
10"
~ .......
~
f\\
I~'\ t< :'\ t'--. i'-.....
I'-.....
~ ~.~ ~"'
~ ::~ f'\ ."""
r0:=0~
~
0.\
.......
0.5
.........
..: 10-5
.=,
""""
'""
~ l\ 1\
.",-
O.~
1~
O~~ ~'
M1\"" I'\,"" i'" "'"
1.0\
\
1~
10-9
10.. 10
12
16
"-
20
\\1\ \
24
28
"\
1\
32
""
!~
36
40
~t
Fig. N.2a-The function 1/1,(IlI,OI.) fot
01.';
1 fot a slab shield. (Ftom Schmidt.')
APPENDIXN
721
10"
10- 1
10-'
1ll-3
.
~.
10-4
10-5
10-6
10-7
10-6
10-9
10- 10
10
12
14
16
18
20
22
24
"t
Fig. N.2b-The function
1/1, (Ilt,a) fot a ..
1 fot a slab shield. (Ftom Schmidt.')
722
APPENDlXN
1.00,----,---r="9"'--.-.,--r-----,---r---.-----,
...
~
0.025
0.900 1--hlL:::::n4:-+--=::'+-=-+--+--l--+-+---I
oLl~~~~~~
o
1.0
2.0
3.0
4.0
5.0
6.0
mean free paths
7.0
8.0
9.0
10
J.l t,
Fig. N.3-The function l/Jo(JJt,O:) for a semi-infinite shield. (From H. C. Claiborne,

Engineering Compendium on Radiation Shielding, p. 447, Springer-Verlag, New York,
1968.)
APPENDIXN
723
1.00
0.900
0.800
0.700
..5
.;;
0.600
0.500
0.400
0.300
0.200
0.100
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10
p.t, mean free paths
Fig. N.4-The function "'1 (J,Lt,O:) for a semi-infinite shield. (From H. C. Claiborne,
1968.)
724
APPENDIXN
1.00 ,,-,,-T:::::::+==r--,r---'--I'
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10
IJ,t. mean free paths
Fig. N.5-The function !/J2 (IJt,a) for a semi-infinite shield. (From H. C. Claiborne,
Engineering Compendium on Radiation Shielding, p.448, Springer-Verlag, New York,
1968.)
APPENDIX N
1.00
<:><>~
0.900
II...---
0.800
0.600
~
'i
0.500
4'
0.400
0.300
0.200
/,
-.
r----J.'0"-....,
"'5,
/'
............ ,
...............
"""" ""
q~'",,- ~ i'...
"-
".. ~ "-
i'...
.,'0
.........
"
"-
..............
~ '"-....,
"'}".s- . . . . . ,. . . . . . ""
f::.'~:~."
0.s- . ~~~ ~~ ~
~q~~t-.........,~~ ...........................
,::r-/~:~~I'--...
"<
1.4
0.100
...............
i"-......
oq,s~
""0 '.s" "...."
............
>5.. . . .
1"--0,1
............
/'
/
--
---
...............
,l0
0.
- ---- -
0.025
0.05 _ _
I~ ~
...-
/; ~ ---
0.700
725
Kl:--...t"1'--.
3.,2.0J~;~;:~{:::1:::l;:::
~ ~ ~~
t-==::
I-
o
o
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10
p.t, mean free paths
Fig. N.6-The function 1/13 (Ilt,ex) for a semi-infinite shield. (From H. C. Claiborne,
1968.)
Constants for Empirical

Expressions of Albedo Data
Appendix 0
Table O.I-CONSTANTS FOR THE EXPRESSIONt FITTING THE
MAERKER AND MUCKENTHALER DIFFERENTIAL DOSE ALBEDO
DATA FOR FAST NEUTRONS INCIDENT ON CONCRETE.
Value of constant for.:18o of
Constant
0.2 to 0.75
MeV
0.75 to 1.5
MeV
1.5 to 3
MeV
G.
G,
G,
G,
G.
G,
G.
G,
G,
6.585(-2)
5.048(-2)
3.710(-2)
1.544(-2)
7.837(-3)
0
0
0
0
7.045(-2)
4.393(-2)
7.088(-2)
1.898(-2)
2.408(-3)
-3.589(-3)
0
0
0
7.211(-2)
5.845(-2)
5.968(-2)
2.729(-2)
1.190(-2)
1.000(-3)
4.637(-3)
6.490(-3)
0
7.024(-2)
7.452(-2)
1.000(-1 )
5.591 (-2)
2.646(-2)
-6.908(-4)
-8.087(-4)
-1.459(-3)
-1.809(-3)
6.856(-2)
8.294(-2)
9.517(-2)
7.761(-2)
4.292(-2)
1.824(-2)
5.599(-3)
5.288(-3)
1.046(-2)
5.899(-2)
6.039(-2)
7.524(-2)
8.140(-2)
6.622(-2)
3.056(-2)
1.595(-2)
1.277(-2)
9.380(-3)
B.
B,
B,
B,
B.
6.27(-2)
1.50(-2)
5.3(-3)
0
0
9.00(-2)
8.5(-3)
9.7(-3)
0
0
8.80(-2)
1.30(-2)
6.0(-3)
0
0
9.05(-2)
2.15(-2)
2.30(-2)
0
0
8.744(-2)
2.817(-2)
2.344(-2)
1.779(-2)
8.517(-3)
6.374(-2)
1.382(-2)
1.178(-2)
1.084(-2)
6.801(-3)
K,
1.0
1.0
1.1
0.9
1.1
1.06
a.,
a.,
a,.
a"
a"
a,.
a21
a"
0.36
1.29
0
0.06
-3.06
0
-0.20
1.68
0
0.51
0.32
1.00
-0.04
-2.46
0
0.05
0.95
0
0.56
0.18
1.32
-0.14
-2.76
0
0.05
1.14
0
0.60
0.15
0.48
-0.61
-1.08
0
0.32
0.30
0
0.43
2.02
-0.38
0.05
-9.13
5.93
0.04
5.97
-4.39
...
3 to 4
MeV
4 to 6
MeV
6 to 8
MeV
0.35
0.95
0
0.10
-2.28
1.11
0
0
0
fEquation 7.3-1.
*From R. E. Maerker and F. J. Muckenthaler. Calculation and Measurement of the
Fast-Neutron Differential Dose Albedo for Concrete, Nucl. Sci. Eng., 22: 455 (1965).
Reach 6.585 X 1lf'. etc.
726
APPENDIX 0
727
Table O.2-COMPOSITIONS OF SOIL, IRON, AND POLYETHYLENE

USED IN MONTE CARLO CALCULATIONSt
Composition (in units of 1021 atoms/em 3 )
Nevada Test Site soil
Element
Dry
50%
100%
saturated
saturated
Iron
0.658
2.67
76.8
39.2
Boron-tO
Boron-It
Hydrogen
8.553
9.80
22.68
2.014
9.533
23.30
1.830
8.680
16.87
Carbon
Oxygen
Aluminum
Silicon
Iron
8% borated
polyethylene
27.00
1.976
8.963
84.9
Density,
g/cm
1.15
1.12
1.25
7.88
0.97
tFrom F. J. Allen, A. Futterer, and W. Wright, Dependence of Neutron

Albedos Upon Hydrogen Content of a Shield, Report BRL-1224, Ballistic
Research Laboratories, October 1963.
Table 0.3-VALUES OF THE CONSTANT k(E oit FOR THE EXPRESSION FITTING THE
D1FFERENTIAL-DOSE-ALBEDO DATA OF ALLEN et al.:j: FOR FAST NEUTRONS
INCIDENT ON VARIOUS MATERIALS
k(E o ) for incident energies of

Material
Concrete
Dry NTS soil
0.1 MeV
0.25 MeV
0.5 MeV
1.0 MeV
2.0 MeV
3.0 MeV
5.0 MeV
14.0 MeV
Fission
0.0948
0.0967
0.1027
0.0895
0.1062
0.1002
0.1323
0.1272
0.1164
0.1103
0.1030
0.0979
0.0834
0.0784
0.0552
0.0535
0.1110
0.1050
0.0868
0.0957
0.0952
0.1209
0.1074
0.0926
0.0746
0.0533
0.1015
50% saturated
NTS soil
100% saturated
NTS soil
Iron
0.0778
0.1750
0.0818
0.1752
0.0839
0.1801
0.1054
0.1182
0.0891
0.1477
0.0791
0.1508
0.0644
0.1158
0.0463
0.0802
0.0868
0.1366
tEquation 7.3-3.
+F. ]. Allen, A. Futterer, and W. Wright. Dependence of Neutron Albedos Upon Hydrogen Content of a Shield,
Report BRL-1224, Ballistic Research Laboratories, October 1963.
From R. L. French and M. B. Wells, An Angular Dependent Albedo for Fast-Neutron Reflection Calculations,
USAEC Report RRA-M31, Radiation Research Associates, Inc., November 1963.
>
""to
Z
'='
Table O.4-CONSTANTS FOR THE EXPRESSIONS! FITTING THE COLEMAN ET AL..; DIFFERENTIAL
AND TOTAL ALBEDO DATA FOR INTERMEDIATE-ENERGY NEUTRONS INCIDENT ON REINFORCED CONCRETE
Values of constants for l:iE o of
Constant
<I
<,
~I
/3,
71
7,
a
b
c
SI
S,
55.1 '0200
keY
15.2 to 55.1
keY
4.2 to 15.2
keY
1.15 to 4.2
keY
0.32 to 1.15
keY
87 to 320
eV
24 to 87
eV
6.6 to 24
eV
1.8 to 6.6
eV
0.5 to 1.8
eV
0.190
-0.020
0.020
0.300
0.11
0.91
0.20
0.56
0
0.880
-0.208
0.190
-0.025
0.025
0.295
0.11
0.91
0.225
0.69
0
0.865
-0.177
0.216
-0.047
-0.004
0.307
0.12
0.91
0.24
0.70
0
0.875
-0.200
0.210
-0.046
-0.005
0.310
0.12
0.91
0.24
0.70
0
0.875
-0.232
0.208
-0.042
-0.005
0.305
0.12
0.91
0.24
0.70
0
0.860
-0.205
0.210
-0.061
-0.003
0.296
0.125
0.865
0.28
0.72
0
0.845
-0.210
0.205
-0.068
-0.003
0.283
0.13
0.845
0.30
0.73
0
0.830
-0.228
0.202
-0.075
-0.002
0.270
0.13
0.82
0.32
0.74
0
0.815
-0.230
0.172
-0.059
0.021
0.218
0.105
0.65
0.40
0.77
0
0.817
-0.244
0.105
-0.036
0.115
0.125
0.080
0.48
0.255
-0.072
0.765
0.792
-0.232
t Equations 7.3-6 and 7.J.7.

:tW. A. Coleman, R. E. Maerker, F.]. Muckenthaler, and P. N. Stevens, Calculation of Doubly Differential Current Albedos for Epicadmium
Neutrons Incident on Concrete and Comparison of the Reflected Subcadmium Component with Experiment, Nucl. Sci. Eng., 27: 411 (1967).
>
."
."
"'
Z
0
X
0
730
APPENDIX 0
Table O.S-VALUES OF COEFFICIENTS AND FUNCTIONS FOR THE
SEMIEMPIRICAL FORMULA OF MOCKELt (Eq. 7.3-13)
Sourcet
Norm.
"
1.37(1 _ p)o ...
1.37(1 _p)o ...
Isotropic
Cosine
1.37(1 _ p)o ...

1.37(1 _ p)o .
1.33(1 _ p)o.,."
1.37(1 _ p)o ...
t A.
Mockel, Reflection and Transmission
f(p)
g(p)
0.2775
0.640
0.3882
(I _ p)o ....
(I _ p)o .
(I _ p)o .
0.067p
0
O.OSp,33
by a Strongly Absorbing Slab, NucL Sci. Eng., 22:
339 (1965).
tNorm., normal incidence; isotropic, isotropic current; and cosine, cosine current.
Table O.6-EXPRESSIONS FOR DIFFERENTIAL AND TOTAL TI-lERMALNEUTRON ALBEDOS DUE TO INCIDENT 0.5-eV TO 200-keV NEUTRONSt
/lEo
'" (/lEo,/Lo ,/L).t thermal neutrons

steradian- t source neutron-I
A, (/lEo,/Lo).
thermal neutrons!
source neutron
55.1 - 200 ke V
15.2 - 55.1 keY
4.2 - 15.2 keY
1.15 - 4.2 keY
0.32 -1.15 keY
87-320eV
24-87eV
6.6 - 24eV
1.8 - 6.6 eV
0.5 - 1.8 eV
p(1 + 1.73p)(0.0043 + 0.0058po)

p(1 + l.73p)(0.0052 + 0.0059po)
p(1 + l.73p)(0.0062 + O.0071po)
p(1 + l.73p)(O.0077 + 0.0073po)
p(1 + 1.73p)(0.0090 + 0.0099po)
p(1 + 1.73p)(0.011 + O.OI2po)
pI (0.0185 + 0.0150po) + p(0.0177 + 0.0235po) I
pI (0.0332 + 0.0085po) + p(0.0220 + 0.0268j.to) I
p[ (0.0595 + p(0.0290 + 0.0305po)]
p[ (0.124 - 0.035po) + p(0.020 + 0.053po) I
0.029 + 0.039po
0.035 + 0.040po
0.042 + 0.048j.to
0.052 + 0.049Po
0.061 + 0.067po
0.074 + 0.081po
0.095 + 0.096po
0.150 + 0.084po
0.248 + 0.064po
0.431 + O.OOlpo
tFrom W. A. Coleman, R. E. Maerker, F. J. Muckenthaler, and P. N. Stevens, NucL

Sci. Eng., 27: 419 (1967).
~P.o = cos 8 o ;fJ = cos O.
APPENDIX 0
731
Table O.7-EXPRESSI0NS FOR DIFFERENTIAL AND TOTAL THERMALNEUTRON ALBEDOS DUE TO INCIDENT 200-keV TO 9.57-MeV NEUTRONSt
0':2(aE O,#J.O,tl),:j: thermal neutrons

2.64 - 9.57 MeV
0.750 - 2_64 MeV
200 - 750 keY
A 2 (AE o ,ll0),
thermal neutrons/source
steradian- l source neutron t
neutron
11(1 + 1.7311)(0.0024 + 0.0040110)

11(1 + 1.7311)(0.0028 + 0.0044110)
11(1 + 1.7311)(0.0036 + 0.0049110)
0.016 + 0.027110
0.019 + 0.030110
0.024 + 0.033110
tFrom W. A. Coleman, R. E. Maerker, F.
J. Muckenthaler, and P. N. Stevens,
Nucl. Sci. Eng., 27: 422 (1967).

+#lo = cos 00 ; 1J = cos 8.
Table 0.8-VALUES OF PARAMETERS FOR CHILTONHUDDLESTON GAMMA-RAY DIFFERENTIAL DOSE

ALBEDO FORMULA (Eq. 7.4-1)t
Eo,
MeV
Water
0.2
0.662
1.00
2.50
6.13
-0.0187 0.0027
0.0309 0.0047
0.0470 0.0053
0.0995 0.0068
0.1861 0.0107
0.1327 0.0054
0.0253 0.0034
0.0151 0.0025
0.0058 0.0010
0.0035 0.0005
Concrete
0.2
0.662
1.00
2.50
6.13
0.0023 0.0033
0.0347 0.0050
0.0503 0.0056
0.0999 0.0078
0.1717 0.0103
0.0737 0.0065
0.0197 0.0035
0.0118 0.0025
0.0051 0.0011
0.0048 0.0005
Iron
0.2
0.662
1.00
2.50
6.13
0.0272 0.0033
0.0430 0.0045
0.0555 0.0049
0.1009 0.0073
0.1447 0.0101
-0.0100 0.0062
0.0063 0.0030
0.0045 0.0021
0.0044 0.0010
0.0077 0.0006
Lead
0.2
0.662
1.00
2.50
6.13
0.0044 0.0002
0.0308 0.0015
0.0452 0.0013
0.0882 0.0014
0.1126 0.0048
-0.0050 0.0004
-0.0100 0.0007
-0.0083 0.0004
0.0001 0.0002
0.0063 0.0003
Material
tFrom A. B. Chilton, C. M. Davisson, and L. A. Beach,

Transactions of the American Nuclear Society, 8: 656 (1965).
732
APPENDIX 0
Table O.9-CHILTON'S MODIFIED ALBEOO

FOR CONCRETE!
Parameter
137CS (0.662 MeV)
60 CO (1.25 MeV)
0.0455
0.0161
1.512
-0.606
-0.641
0.645
-0.157
0.0710
0.0114
1.555
-0.629
-0.605
0.539
-0.168
C
A,
A,
A,
A,
As
tFrom A. B. Chilton, Nucl. Sci. Eng., 27: 482 (1967).
_
C 10 26 K e (E o .8 s )+c'
:l:o:(Oo,O,//) -F(Oo.O,t/J) 1 +cosOo sed} (1 +2E o versOs)~
where
F(Oo ,8,(/J) = A I + A 2 vers 2 00 + A3 vers 2 0
+ A 4 vers 2 0 0 vers 2 8
+A s versOo versO vers
vers == 1 - cos <p.
Table O.IO-VALUES OF THE CONSTANT b FITTING THE EXPRESSION

OF HAGGMARK et al.! (Eq. 7.4-2) FOR GAMMARAY DIFFERENTIAL
DOSE ALBEDOS FOR ALUMINUM, IRON, AND CONCRETE
Material
Aluminum
Source
60
Concrete
0.50
0.75
1.00
0.0070
0.0090
0.0095
0.50
0,75
1.00
0.0132
0.0167
0.0194
0.50
0.75
1.00
0.0060
0.0065
0.0065
137Cs
0.50
0.75
1.00
0.0091
0.0120
0.0130
60
0.50
0.75
1.00
0.0075
0.0090
0.0100
0.50
0.75
1.00
0.0133
0.0165
0.0184
CO
137
Iron
cos 0 0
60
CS
CO
CO
137Cs
tFrom L. G. Haggrnark, T. H. Jones, N. E. Scofield, and W.], Gurney,

Differential Dose-Rate Measurements of Backscattered Gamma Rays from
Concrete, Aluminum and Steel, Nucl. Sci. Eng., 23: 138 (1965).
Table O.I1-CURVEFITTED EXPRESSIONS FOR CAPTURE GAMMARAY DIFFERENTIAL

AND TOTAL ALBEDOS ARISING FROM THE SLOWING DOWN AND CAPTURE
OF INCIDENT EPICADMIUMt
!li3 0
55.1
15.2
4.2
1.15
0.32
87
24
6.6
1.8
200 keY
55.1 keY
15.2 keY
4.2 keY
1.15 keY
320 keY
87 eV
24 eV
6.6 eV
0.5 - 1.8 eV
(fads hr- I
"'D2(n.1 )(!li3 0,/10,/1),+

steradian- I )/(incident unit neutron current)
J(f7 /1(0.43 + 2.17/1- 1.67/1' )(0.39 + 0.51/10)(1.10)

10- 7 /1(0.39 + 1.78/1-1.39/1')(0.50 + 0.68/10)(1.11)
10-7 /1(0.70 + 2.53/1- 2.07/1' )(0.37 + 0.46/10)(1.12 - 0.01/10)
10- 7 /1(0.68 + 2.59/1- 2.08/1' )(0.40 + 0.54/10)(1.12 - 0.01/10)
J(f7 /1(0.66 + 3.18/1- 2.60/1' )(0.46 + 0.54/10)(1.14 - 0.02/10)
10- 7 /1(0.89 + 3.52/1 - 3.09/1' )(0.41 + 0.59/10 )(1.18 - 0.06/10)
10- 7 /1(1.14 + 3.98/1 - 3.77/1' )(0.38 + 0.62/10 )(1.30 - 0.15/10)
J(f7 /1(1.40 + 2.73/1- 2.50/1' )(0.45 + 0.65/10}(1.11 - 0.04/10)
J(f7 /1[ 1.34 + 0.78/10 + /1(0.15 + 2.30/10) - /1 (0.60 + 2.11/10)]
X (1.09 - 0.03/10)
10- 7 /1[2.12 + 0.72/10 + /1(-1.46 + 2.92/10) + /-l' (0.40- 2.68/-l 0 )]
X (1.07 - 0.02/-l 0 )
tFrom W. A. Coleman, R. E. Maerker, F.

+/-lo = cos 0 0 ; /1 = cos O.
J. Muckenthaler, and P. N.
AD 2(n.1 )(!li30,/10),
(rads/hr )/(incident
unit neutron current)
J(f7 (1.40 + 1.83/-l 0 )

J(f7 (1.52 + 2.07/-l 0 )
10- 7 (1.74 + 2.15/10)
10- 7 (1.93 + 2.57/-l 0 )
1(f7 (2.43 + 2.79/-l 0 )
10- 7 (2.56 + 3.42/-l 0 )
1(f7 (2.95 + 4.20/-l 0 )
1(f7 (3.08 + 4.18/-l0)
10- 7 (3.89 + 4.14/-l0)
1(f7 (4.52 + 4.33/-l 0 )
Stevens, Nucl. Sci. Eng., 27: 420 (1967).
to
"o><
Radiation Penetration of
Cylindrical Ducts t
Appendix P
Table P.I-UNCOLLIDED FLUX DENSITY AT EXIT OF CYLINDRICAL
DUCT DUE TO RADIAnON ARRIVING THROUGH DUCT WALLS
(ISOTROPIC SOURCE)
m
0.001
0.002
0.005
0.01
0.02
0.05
0.1
0.2
0.5
0.75
1.0
Flux density per unit surface source intensity for shield thickness of
JlZ =0.1 t
JlZ =0.2
JlZ - 0.5
1.823
1.823
1.824
1.825
1.828
1.834
1.842
1.851
1.825
1.765
1.689
1.223
1.223
1.224
1.226
1.229
1.239
1.254
1.272
1.277
1.240
1.185
0.5602
0.5607
0.5620
0.5643
0.5688
0.5820
0.6024
0.6360
0.6841
0.6804
0.6560
IlZ
=1.0
0.2198
0.2207
0.2215
0.2235
0.2278
0.2406
0.2613
0.2982
0.3659
0.3819
0.3767
JlZ = 2.0
0.04912
0.04935
0.05002
0.05119
0.05363
0.06171
0.07621
0.1049
0.1687
0.1923
0.1980
IlZ
= 5.0
1.166 x Hf'
1.184 x Hf'
1.242 x Hf'
1.349 x Hf'
1.614 x Hf'
2.977 x Hf'
7.042 x Hf'
1.795 x 1(f2
5.223 x 1(f2
6.990 x 10-2
7.753 x 1(f2
fNumber of relaxation lengths.
+All data in this appendix have been taken from D. K. Trubey> A Calculation of Radiation
Penetration of Cylindrical Duct Walls, USAEC Report ORNL-CF-63Z-64, Oak Ridge National
Laboratory, Feb. 28, 1963.
734
APPENDIXP
735
Table P.2-UNCOLLIDED FLUX DENSITY AT EXIT OF CYLINDRICAL

DUCT DUE TO RADIATION ARRIVING THROUGH DUCT WALLS
(COSINE SOURCE)
Flux density per unit surface source intensity foc shield thickness of
(i)
,..z = O.It
,..z = 0.2
IlZ = 0.5
IlZ = 1.0
IlZ = 2.0
,..z = 5.0
0.001
0.002
0.005
0.01
0.02
0.05
0.1
0.2
0.5
0.75
1.0
1.445
1.446
1.446
1.447
1.450
1.455
1.459
1.453
1.339
1.190
1.036
1.149
1.149
1.150
1.152
1.156
1.166
1.179
1.188
1.116
0.9949
0.8643
0.6539
0.6545
0.6564
0.6595
0.6657
0.6836
0.7095
0.7462
0.7510
0.6827
0.5953
0.2976
0.2983
0.3002
0.3034
0.3100
0.3298
0.3610
0.4120
0.4714
0.4465
0.3956
0.07544
0.07582
0.07698
0.07896
0.08314
0.09703
0.1217
0.1682
0.2494
0.2553
0.2340
0.002026
0.002059
0.002166
0.002365
0.002862
0.005458
0.01323
0.03343
0.08728
0.1036
0.1014
tNumber of relaxation lengths.
736
APPENDIX P
Table P.3-RATIO OF LlNE-OF-SIGHT RADIATION TO TOTAL

UNCOLLIDED FLUX DENSITY IN DUCT
Fraction of total flux density for shield thickness of
(i-)
pZ = 0.1 t
J.lZ = 0.2
J.lZ = 0.5
pZ= 1.0
pZ= 2.0
J.lZ = 5.0
<I(f"
<1(f"
<1(f3
0.001
0.004
0.020
0.062
0.16
0.40
0.53
0.64
<1(f3
0.002
0.010 .
0.038
0.11
0.29
0.42
0.52
0.68
0.76
0.82
<10-"
<10-"
<1(f3
0.001
0.005
0.025
0.076
0.19
0.45
0.61
0.71
<1(f3
0.002
0.011
0.040
0.12
0.27
0.43
0.54
0.71
0.79
0.85
Isotropic InfinitePlane Source
0.001
0.002
0.005
0.01
0.02
0.05
0.1
0.2
0.5
0.75
1.0
<10-"
<1(f5
<1(f5
<10-"
<1(f3
<1(f3
0.003
0.011
0.059
0.11
0.17
<Hf"
<1(f5
<10-"
<10-"
<1(f3
0.001
0.004
O.ot5
0.083
0.15
0.23
<10-"
<1(f5
<10-"
<10-"
1(f3
0.002
0.008
0.031
0.14
0.25
0.34
<1(f5
<1(f5
<1(f"
1(f3
1(f3
0.005
0.019
0.062
0.23
0.37
0.48
Cosine Infinite~Plane Source
0.001
0.002
0.005
O.ot
0.02
0.05
0.1
0.2
0.5
0.75
1.0
<1(f"
<1(f5
<10-"
<10-"
<1(f3
0.002
0.007
0.026
0.14
0.25
0.36
<1(f"
<10- 5
<10-"
<1(f"
<1(f3
0.002
0.008
0.031
0.16
0.28
0.40
tNumber of relaxation lengths.
<1(f5
<1(f5
<10-"
<1(f3
<1(f3
0.004
0.014
0.050
0.22
0.37
0.50
<1(f5
<1(f"
<10-"
<1(f3
0.002
0.008
0.027
0.12
0.31
0.48
0.60
Solutions to Exercises
CHAPTER 2
2.1
Let N\t) he the number of atoms that have disintegrated at time t, where N(O):; No
(a)
P, (t)
(h)
=.{; p(t') dt' =f;"l.Noe>..t' dt'

= No(l- e>..I)
t'd'
- I Op(t)
t_N o (1-e -At )_
.>..1
f()
It-_,,-l-e
Iop(t)dt
f 2 (t)
No
I '- p(')'d't _ Noe >..I _.>..1

- I;; p(t') dt' e
r;;;;- -
(d)
2.2 (a) See accompanying sketch.
p_~
..L_.!.R~--I
Solution 2.2-Solid-angle calculation.
737
738
dn
dA
= n dA = cos ~ dA
1r rl
R
(R' +p')'1>
-,
fa
2rrp dp, r = R + p~ , cos () =-
= Jo
fa
dn
= Jo
21TRp dp
(R' + p' )'h. =
21TR
]a
(R'+P')\l o
(h) As R -+ 0 or a-+=,n-+ 21T (half space).

2.3 (a)
AdA
I'
'- .---~--/I-- __
f/J
I
2:.-,..':::o+'/
----J
Solution 2.3-Spherical-shell source.
dA=R'sinOdOdf/J
The differential source = Sa R 2 sin 8 dO dl/J particles cm-2 sec- 1
1
4," Sa R 2 sin 8 dO dep particles em- 2
sec- 1 steradian-,.t
The solid angle at dA sub tended by a sphere of unit cross-sectional area centered
at D is l/R' steradian (see figure). Thus, for flux density, (R),
d<l>(R) = (:1T Sa R' sin 0 dO df/J)
(R) =
L .I:'
2
<P=O
8=0
(;2)
1 Sa sin 0 dO df/J
-4
1T
739
(b) From symmetry the net current through the plane P is zero.
2.4 (a) From symmetry the flux density is (1/2)5" particles cm- 2 sec-I,
(b)
----r--...
.................j
...
Solution 2.4-Hemispheric shell source.

The differential source strength is given by
1
ds = 4 rr Sa R 2 sin 8 dO df/> particles em-2 sec-J steradian- J
The solid angle subtended at dA from a unit area about point D in plane P (see
sketch) is cos O/R 2 Thus the differential current through Pis given by
J(R) =
2~
=o
f.~/2 I
0 0
0
-4 Sa sin 0 cos 0 dO
1r
, 2 O]2~
1 Sa partle
' Ies em-2 sec-1
="21 Sa [1"2 sm
0 = "4
2,5
d'treete d downwar d
(ai
0,
~~r=t====i~_n
Solution 2.5-Disk and point sources.
740
Let lIl 1 (D 1 ) be the flux-density contribution from the disk and tP z (D z ) be the
flux contribution from the point source. Using the parameter in the sketch,
where r2 = p2 +
ni and r dr = pdp.
(D2+R2)~
1>,(D 1
)=i'
!Sa(!)dr
2
r=D,
Thus
(b) Let J, (D,) be the cutrent through p from the disk and ],(D 2 ) be the cutrent
through p from the point source.
[I](D;+R')%
ZSaDI -;: D
in direction n
J '2 (D)
2
5 (cos
8) -In d-trectlOn-n
4rr
D~
The net current J(D) at the point D is
J(D) = J(D,) - J2 (D 2 )
-_I'2 Sa [ 1 -
D']
--nr-
1 (Scos8)_
_
_
(Vi + R 2)%: - 411'
In dIrectIOn n
741
2.6
----dS
Surface source
Solution 2.6-Plane array of point source.
Consider in the accompanying sketch the solid angle drl about n passing through
the area dA, where dA is the projected area on the unit sphere about the
differential surface source area dS. Slc 2 = average surface source (particles cm- z
sec- t ) and (1/417)(5Ic 2 ) dS = particles sec- 1 steradian- t from dS into dn. about
11. But dS=dAlcosO. Thus (1/4rr)(S/c 2 )(dA/cos 0) particles sec steradian- I
pass through dA at point P, and the flux density at point P is
dA) (dAI)
IS
( 477" c 2 cos ()
= 47l' c 2 cos 8 particles em
-2
sec
_I
.-1
steradIan
2.7 (a) Referring to the sketch for part (a) and using polar coordinates about P,
Solution 2.7(a)-Nonabsorbing slab source G(O).
742
dV= R' sin 8 d8 dq, dR

dn = sin 8 d8 dq,
4~ Sv dV= particles sec- t
;2
G{8) dn =
steradian-I into dfl about n
steradians subtended at dv by a unit area at P perpendicular to n
X/cosO.' 1
-4 >vR' sin 8 d8 dq, dR
ReO"
1
= -4 Sv
1r
G(O) = 4
1f
cos 8
cos
1
-;;2
particles cm-' sec-I into dn
sin 8 d8 dq,
8 SvX particles em-2 sec- t steradian- 1
(b) Referring to the sketch for part (b),
Solution 2.7(b)-1(8) from slab source.
<1>(8) dS = G(8) dA
dA=cos8dS
<1>(8) dS = G(8) cos 8 dS
<1>(8) = G(8) cos 8
Thus
743
(c)
<1>. =
In
<1>(0) dn =
20 LO/2 I
-4 SvX sin 0 dO dt/!
<1>=0
8=0
"
=-S X
2 v
(d) The equivalent isotropic surface source emits particles in all directions (into as
well as out of the slab). Thus <1>. = SvX.
(e) Referring to the sketch for part (e),
n
dA
t/!--""I
Solution 2.7(e)-H(0) function.
H(O)dO=
20
G(O)dn= L20 4
<1>=0
'If
I 0 SvX sin 0 dO dt/!
cos
I
= 2 Sv X tan 0 dO
= -1 SvX [ In cos 0
l8.
83
(COS
1 SvX In --.0)
= -2
1
COsu2
The values for each interval are shown in the table.
744
SOLUTION 2.7(f)
Interval
0" to 10
40 to 50"
75 to 85
80 to 90
1
0.715
0.259
0.174
N(E) =
2.8 (a)
Cos 02
Cos 81
0.985
0.643
0.0872
0
Cos 8 1
Cos 02
In Cos
Cos 02
1.016
1.19
2.97
0.0159
0.174
1.09
0.77E~e-O.776E
d: N(E) = 0.;7
F~e-O.776B _ 0.77(0.776)E~e-O.776E
Setting
dN =0
dE
I
~=
2E
A"
0.776E" = 0
2~0.776E=0
E= 0.645 MeV
if = J;;' N(E) E dE = J;;' 0. 77E 'la e-0.776E dE
J;;' N(E) dE
J;;' 0.77E"'e-0 .776E dE
(b)
Letx
= 0.776E;E = x/0.776
I
dE=-- dx
0.776
if
Jo~(0f7-6)\x(~)dx _ (0.776f'laf(5/2)
-r(_x_)11
o 0.776 e
-x
(_I_)dx - (0.776) lif(3/2)

0.776
= (0.776ft (3/~!~~{2) = (0.776ft (3/2) = 1.93 MeV

__
VB -
11
(0.5251.93X 10-18 ) -_ 1.91 X 10
em/sec
745
(c)
N(E) = 0.77E~e-0.776E
mv 2 dE
E=-'-=mv
2 'dr
N(v)
=N(E) dE = 0.77 (!!!.)~ ve-O .77 6 mv' /' mv

dr
= 0 77
m*'
2
---r,-v2e-O.776 mv
n
'/
r
r~ 3
(0.776 mv')
_JoN(v)vdvJOvexp2
dv
v=
=
r~
d
r~ ,
(0.776 mv') d
JO N(v) v
JO v exp 2
v
_ 0.77 6 mv
Let y -
_
2
' ' v - ( 0.776m
)~ y ~
2 )~_~
dv = ( 0.776m y dy
ro (
r~ (
v=
Jo
)*' Y e
*,.y [
]~ y -~ dy
2~
2(0.776m)
2 ) 'Y[
1
]~ _~
0.776m ye
2(0.776m) y dy
Ji;' ye'Y dy
0.776m Ji;' y~e'Y dy
_(_2 )~...!QL _(_2 )~ [_I ]~

r(1/2) -_(_2)~(~)
Vii
0.776m
r(3/2) -
- -=
mv'
E
- =m
- [4(
v
0.776m
(1/2)
2
)]--=1.64MeV
4
0.776m" 0.776"
0.776m
746
CHAPTER 3
3 .1 (a) Partial densities:
(0.00125) = 0.000976 g/em3 of air
Po = 0.2099 (0.00143) = 0.000300 g/em 3 of air
PAr =0.0096 (0.00178) = 0.0000171 g/em 3 of air
PN
=0.7805
(h) Weight percents:
wN -- 0.000976
0.00129 X
0 %5%
1 0 - 75. 0
0.0003
Wo = 0.00129 X 100% = 23.2%
wAr -- 0.0000171
0.00129
(c) Atomic densities: (Avogadro's number
0602 X 10"
NN = '
14
% - 3%
100 - 1. 0
IS
6.02 X 1023 )
X 0.000976 = 4.20 X 10' 9 atoms/em 3 of air
0602 X 1024
No = '
16
X 0.000300 = 1.13 X 10'9 atoms/em 3 of air
N Ar -
0602 X 10 24
17
. 39.9
X 0.0000171 = 2.58 X 10 atoms/em 3 of air
(d) Electron density:

Pe = (4.2 X 10'9) 7 + (1.13 X 10'9) 8 + (2.58 X 10 17 ) 18
= 3.89 X 10 20 electrons/em 3
(e) Average Z
_
3.89 X 10'
- (4.2 X 10'9) + (1.13 X 10 '9 ) + (2.58 X 10 17 )
Z - ...,.--,.---:-:-=-"-=-:--::-'c..:....,,..,.-_,.-_......
3.89 X 10'
5.36 X 1019 - 7.26 electrons/atom
3.2 (a) Atomic density:

0602 X 1024
Molecules H,O = . 2 + 16
X 1 g/cm 3 = 3.34 X 10"
molecules/em 3
There are thus 3.34 X 10 22 atoms/em 3 of oxygen and 6.68 X 10 22 atoms/em 3

of hydrogen. The atomic density is then:
N(H,O)
= (3.34 X 10") + (6.68
X 10")
= 1.0 X 10"
atoms/em3
747
(h) Electron density:

Pe = (3.34
1022 ) 8 + (6.68
10 22 )
= 3.34
1023 electronslcm'
(c) Atomic number:
- 3.34 X 1023
Z = 1.0 X 1023 = 3.34 electrons/atom
3.3 (a) Discussion question.
= 3.45 + 0.041 = 3.49 hlatom
(h) Total cross section: at

Then
~ = 0.6021~ 10
3.49
10"'\ = 0.15 cm'/g
=0.15 cm'/g X 0.00125 g/cm' =1.88 X 10" em"
Ilt
Il t (N)
P
(e)
24
=0.602
24
14
10
(0.806 + 0.025) X 10.. 4
24
Il t (0) - 0.602 X 10
16
Ilt (A)
= 0.60~ X
24
10
9.9
=0.0357 em'/g
(0.921 + 0.032) X 10.. 4 = 0.0358 em'/g
=0.0338 em'/g
(2.07 + 0.17) X 10.. 4
Il t air = 0.0357 (0.755) + 0.0358 (0.232) + 0.0338 (0.013)

P
= 0.0357 em'/g
Il, (air)
= 0.0357 (0.00129) = 4.61

1
10" em"
10'
A = -=--= 2.17 X 10' m

Il 4.61
(d)
(e) At 0.1 MeV:

ah=
co
22
22
22
22
4.06 (3.34 X 10 ) + 0.493 (6.68 3X 10

3.34 X 1022 molecules/em
_ 3.94 (3.34 X 10 ) + 0.493 (6.68 X 10

3.34 X 1022
507 hi moecue
I
I
=.
4 95 hi I I
.
rno eell e
uincoh -
22
ape
= 0.071 (3.34 X 10
3.34 X 1022
app
=0
+ 0 = 0 071 hi
.
rno eeu e
At 3.0 MeV:
aInCO h-_ 0.921 (3.34
22
10 ) + 0.1151 (6.68
22
3.34 X 10
22
10
-1 16 hi
-.
I I
moecue
748
0coh :;:: 0incoh :;:: 1.16l-/molecule
_ 0.032 (3.34 X 10 22 ) + 0.00052 (6.68 X 10 22 ) = 0.033 b/moleeule

app 3.34 X 10 22
ape
=0
glem'
k t h'Ie k ness:;:: 0.04
3.4 DIS
2
I 3:;:: 0.01 48
.7 g em
em
Disk volume, V= 1T(2') X 0.0138 = 0.186 em 3

N A /=
(0.602 X 10'4)
27
2.7 = 6.02 X 10" atoms/em 3
The total scattering coefficient includes coherent events:

Il eo h = NA/aeoh = (6.02 X 10 22 ) (6.79 X 10'4) = 0.409 em-I
The scattering coefficient for events that involve energy loss does not include
coherent events,
Il;"eoh = N 1 /a;"eoh = (6.02 X 10
22
(6.41 X 10-24 ) = 0.386 em-I
Since 0.1 MeV is below the threshold for pair production, the photon absorption
coefficient is due only to photoelectric events:
Il pe = NA/a pe = (6.02 X 10 22 ) (0.78 X 10-24 ) = 0.047 em-I
An interactions are given by the sum

11, = Ileo h + IIp , = 0.409 + 0.047 = 0.456 em
-I
(a) The reaction rate (RR):;:: <Pllt V, where V is the total volume over which
interactions are to be considered.
<PIl,V= 10'0 (0.456) (0.186) = 8.48 X 10' photonslsee interaction
(b) Absotption rate = <Pllpc V = 10' 0 (0.047) (0.186) = 8.74 X 10 7 photonslseel absorbed.
(c) Scattering rate = <Plleoh V = 10 1 0 (0.409) (0.186) = 7.61 X 10' photonslsee

(d)
<I-out
:;:: 4>itl -
total degraded or absorbed
+ J.1pe )t
0
=10. 11-0.433(0.0148)J =0.994x 10'0 photonsem-' see- 1
"'" fJ>jn - <Pin (J.ljncoh
(e)
.
F ractIon =
reaction rate
.
.
total atoms In dIsk
_
8.48 X 10'
_
X -14
- (6.02 X 1022 ) (0.186) - 7.57 10
749
3.5
"'0
Figure for Solution 3.5.
pdp=rdr
I!.t = 0.0148 em
dV = 211P dp I!.t = 211(0.0148) r dr
2
coso=r
Pe
= (6.02 X 10 22
atoms/em 3 ) (13 electrons/atom) = 7.83 X 10 23 electrons/em
'!Jri ~ a(O)Pe = (7.83 X 1023 ) (10-26) [29.5 -78.3(~) + 56.6 (;\)]

d<l> (P)
sour~e steradia~s at P = (dJ.ls
steradIan
2../2
"',(P) =
r=2
10
1o
(7.83
X [29.5 _ 78.3(2)
r
_ (7 27
= (7.27
dn.
em
10 )
dV)-\r
+ 56.~(4)1
f2.J2
2
10-
<l>
T]
J,r [211(0.0148) rdr]
[29.5 78.3(2) 56.6(4)] d

- ----:,- + ---:;-- r
r
10 6 ) [29.5 In r+ 78.3(2) _ 56.6(2)] 2.,fi

r
= (7.27 X 10 ) [29.51nv'z + 78.3
,2
~Jz -1)
- 56.6 (0.25 - 0.50)]
= (7.27 X 10 6 ) 1.34 = 9.75 X 10 6 photons em-' see- 1
750
_ r: a(v) <1>(v) dv r: a(v) n(v) v dv
3.6
a=
r: <1>(v) dv
f': n(v) v dv
f': (Cv') Av'emv ' 12k]" v dv
r: Av e-mv2/2kT v dv
2
Cr: p2 e-mv'j2kT dv
=
r: v
e- mv 'j2kT dv
mv' dx:;::-vdv
m
kT
Letx=-vdv;-dx
v= (2kT)1I
- - x \I andv ,2kT
=--x
'
2kT'
kT'
m
'
m
m
r=
C), (2kT)\I
- - x \Ix
e _kT dx
Thena=
m_
= 2kT x kT dx
--xe ,m
m
Reference to a table of integrals shows that both numerator and denominator give
gamma functions:
a= C
2kT)\I
2kT
m
{mrr
_
a = C \8kT
Since a
.11
r(3/2) = C(2kT)
(2)
m
[1/2f(1/2)]
1
=~f;~)\1
\
)\1
C/V,
.!:.= (mrr)\I
v
.,,'
\8kT'
=8kT
m7f
E = mv' = ~(8kT)
2
2 mrr
E=4kT
rr
3.7
2(37 X 1010)
<Il = '
2
4rr(100)
(The factor 2 appears because

disintegration. )
5.88 X lOS photons em-2 sec- 1
60 Co
emits two photons, 1.33 and 1.17 MeV, per
751
With an average energy per disintegration of 1.25 MeV, the intensity is

1= E
= 1.25 (5.88 X 10') = 7.35 X 10'
MeV em-. ,ee- I
(aJ Air:
K= I(~k) = (7.35 X 10') 0.0268 = 1.97 X 10
MeV g-' sec-I
K ai , = (1.97 X 104 ) (1.6 X 10-6 ergs/MeV) = 0.0315 ergg- I see-'
(b) Water:
-'
K H,O -- 0 .03 15 (0.0300)
0.0268 -- 0.0353 erg g- 1
sec
(c) Iron:
0.0253)
K pe = O.03 15 ( 0.0268
-I
= 0.0296 ecg g
sec
-I
(d) Lead:
0.0350)
-,
Kpb = 0.0315 ( 0.0268 = 0.0411 erg g
sec
_I
(e) Tissue:
0.0290)
D = K = 0.0315 ( 0.0268 = 0.034 erg g-I sec-I
3600 rads/hr )
= 0.034 ( 00
-,
-1
= 1.23 rads/hr
1
ergs g sec
3.8
"II' - "11=1 - cos 0; cos 0 = 1 - "II' + "II

eos 2 0 = (1 -"II' + "11)2; eos 2 0 - 1 = _sin 2 0 = (1 - "II' + "11)2 - 1
_sin 2 0 = (1 - "11')2 + 2(1 - "11')("11) +"112 - 1
= 1- 2"11' +"11'2 +2"11- 2"11"11' +"112 _1
= 2("11 - "II') +"112 - 2"11'"11 + "11'2
= 2("11 - "II') t ("II - "II')'
(a) Thus 11(0) =
1~" (-t)' [~ + ~ + 2("11 -
"II') + ("II - "11')2 ]
The probability of scattering into solid angle dn:;; 2n sin (J d8 (see figures for
Solution 3.8) is given by 0(8) dn/ae , where U c = total Compton cross section.
Since
752
Figure for Solution 3.8.
"A'
=:
1 + A- cos 8, dA.' = sin 0 dO
do
dodA'dO
,.
0(0)= dn = dA' dO dn=o(}.,)smO 21TsinO
and
o(}.,') = 21TO(0)
so that
(b)
CHAPTER 4
4.1
(0-0) = cos 8, where 8 is the angle between the planes containing dA and dS. When
0< 1T/2, cos 0 > 0 and the leakage through dS along n is also outward (toward n)
through dS. When 1T/2 < 0 < 1T, the leakage along n through ds is opposite in
direction to 0, but then cos (J < O. Thus the sign of (noO) properly is positive when
the leakage through dS is outward and negative when it is inward.
4.2 We use the Compton wavelengtht }., = liE, which implies dE = _}.,-2 dA, and first
prove two lemmas:
Lemma 1: J(}.,) = 1(E) where J(}.,)
=}"(}") and J(E) =E(E).
Proof' Between two values of wavelength,

demands that
Aa > Ab.
tIf E = h"fA. is used, the results will differ only by the constant I1/).
conservation of particles
753
(The limits are reversed since A increases as E decreases.) Thus
The relation must hold for all Aa and Ab; the integrands must be equal, and
(X) = A' (E); thus A(X) = E (E)

Lemma 2: A' :!:,(A')p(A''''A,O''''O) (A') = :!:,(E)p(E'...E.O...O)(E')
Using conservation of particles scattering from E' (A') into the energy interval
(Eb,Ea) where Ea
rEa
> Eb, the following must hold (where E' = 1/A'):

,
JE=Eb :!:,(E ) p(E "'E,
O... 0 ) ...
~.
(E ) dE
fAb
~ ,\'
A=Aa ",,(,,)
, '\ O 0
p(X "'''.
... ) (A) a"
Again the integrands must be equal. Thus
Now, multiplying the Boltzmann equation (Eq. 4.2-8) by E, we obtain
V' OI(E) + :!:,(E) I(E)= fn , J;'=E :!:,(E') p(E...E,O...O) (E') E dE' dO' + E S(E)
Making the change of variable'.to i\ and using the two lemmas. we obtain
IdE'
X ~
cIA' cIA
d.,n' + E S(E)
Now. recalling A'(A') =I(A).dE/cIA' = - (A)-2. and formally defming S(A) =E S(E).
we obtain
v .0
I(A) + :!:,(A)I(A) =
fn' foA :!:,(A') p(X...A.O....O) I(X')~. cIA' dO' + S(X)
754
SOLUTIONS
m EXERCISES
4.3 The slab geometry is shown in the figure for solution 4.3.
.r+--------....~~~-----J'--x
I
---:
y
Solution 4.3, geometry.
We first note that because of the azimuthal symmetry 3'N3y

jn = cos 8. Thus
= 3<1>/3z = 0
and
3<1> .)
.3<1>
3<1>(r 0)
V'O (r,O) = O'V = 0 ( - I = 0\ - = cos 0
'
3x
3x
3x
We define ~,p(O'....O) = ~,(O,O'); the balance equation becomes
cos 0
3<1>~~0) + ~t (r,O) =
fo' ~,(O,O') (r,Sl') dfl' + S(r,O)
The differential flux density in the preceding equation has units of particles
cm-2 sec- t steradian-I. Let J.1.::::: cos 8,dJ.1 = -sin () dO, and
= -dll dl/J. The flux
density per unit J.1 is given by
an
2.
(x,ll) = ~=o (r,O) difJ = 2" (r,O)

(Note: <Il(r,S1) is not a function of l}J because of azimuthal symmetry.) Integrating
each term over I/J. we obtain
2
3<1>(x,ll) + ~t (x,ll) = ~=o
Il ~
f,"
, (x,O") dO
0,,,,,(0,0)
difJ + S(x,ll)
where fJ'1r S(r,n) dl/J : : : S(x,p.). Now integrate the inscattering integral over
sin 0' dO' difJ' = -dll' difJ'):
f,1r
,
,
/',
f,o2"f,2.
0
0 ~,(O,O) (x,O ) (sin 0 ) dO difJ difJ
f 2" f2. fl
=
~=o ~'=O
1// (dil' : : :
,... ) (x,O'"
l=-I ~,(O,..
) dll difJ difJ
755
Substituting this in the balance equation above completes the derivation. Note that
the inscattering integral cannot be integrated over I/J or 1/1' since. in general,
Ls(Q,n') can be a function of I/J and 1//. However, see the solution to problem 4.4
for a method of handling this integration.
4.4 The point source geometry is shown in sketch a for solution 4.4.
1/J,
Solution 4.4, sketch a.
Care must be taken in developing the angle dependencies. The polar angle (} between
rand n has an implicit dependence on the polar angle Or of the radius vector f.
Sketch b for solution 4.4 illustrates the relations:
/-'=-.,-----''-----..'--~-
--- -- - - - - -
Solution 4.4, sketch b.

where a is the unit vector along rand b is the unit vector associated with Or in
spherical coordinates. The sketch shows the change dO r and dO (both negative
changes) as Irl goes from, to ,+ dr. Also observe that'" = ("/2) - 0 and that dO, =
dO. Thus
ail>
1 ail>
ail>
1 ail> )
Vnil>("O,;n;E)=nVil>(r,O,;n;E)=n ( ;;;-a+;aO, b =Or- (n'a) +--(n'b)
roO,
756
But
(n'a) ; cos 0
(Ob)
= -
cos a =
-Sin(; - a: )= -sin 8
and
. 8) -a4>
V -Del> = cos 0 -a4> + -I ( - 8m
ar
ao,
. 0) -d4>
cos (J -a4> + -1 ( -sm
r
dO
ar
Let
J1 = cos (),
~~
Since sin 2
()
-sin
(J,
and apply the chain rule:
= 1 _ Jl2 ,
a4> 1 - Jl' a4>

vn 4>(rn
" E); JaTl - +r - -3fJ
The balance equation becomes
; IE' In' ~,(r,E') p(E'....E,n'....n) 4>(r,n',E') dn' dE + S(r,n,E)

As in problem 4.3, the differential flux density ep(r,n,E) is in units of particles
em-2 sec- 1 MeV-I steradian- t Now
2.
f~"O 4>(r,n,E) dJi; ;
21T 4>(r,n,E) '" 4>(r,Jl,E)
Integrate the balance equation over 1/J to obtain
Jl
a4>(r,Jl,E)
ar
1 - Jl' a4>(r,Jl,E) ~ ( ) ~(
aJl + .Jt r,E ' T,J.1,E)
r
where again
2.
1o
S(r,n,E) " S(r,IJ.,E)
757
Because of the azimuthal symmetry, the in scattering integral can be written
Li
E
'11
l
, ,E )
,l-2
<f>(r,J,l
Jl =-1
1/1'
'J,i
1r
' , ,
~s(r,E ) prE ->E,O ->0)
dl/J
, "
d>/J dJ,l dE
We now assume p(E'~E,n'--+n) is a function only of E', E, and 11o. where 110 ';
n'o = f(J,l' ,J,l). Define
~s[r,E'->E,J,lo(J,l',J,l)]
20 f20
=2rr f,,=0
"'=o~,(r,E') p(E'->E,O'->O) d>/J' d>/J
to obtain
J,l a<f>(r,J,l,E) + 1 - J,l' a<f>(r,J,l,E) + ~t(r,E) <f>(r,J,l,E)

ar
r
aJ,l
fl
",d
= f'=0 "=_I~S[r,E->E,J,lo(J,l,J,l)I<f>(r,J,l,E)dJ,l E +5(r,J,l,E)

~
4.5 Using the results of problems 4.2 and 4.3, we can immediately write the Boltzmann
equation as
J,l
aI(~;:,A) + ~t(A) I(x,J,l,A)
However, for photon scattering by the Compton process (see Sec. 3.2.1), the
following relation holds:
A - A' = 1 - cos 0 0 = I - 0'0

where 8 0 is the scattering angle. Thus
~ CA')
s
CA'->A O'->U =
p
and we can write (since cos 8 0
)
;:
{~s(A"O' '0); 1 -
cos 0.. = A - A'

O;I-cos Oo*A-A
(1' n)
The inscattering integral can now be integrated over l/J to obtain
758
Letting k =
/l
21f~s (~), we obtain
aI(X,/l,A)"
')
,
ax
+ "" I(x,/l," = S(x,/l, ,,) u (x)
r~
Jo
I n'
k 5(1 + A' - A - 0"0) I( 0' A') dO' ciA'

2rr
x,
The delta function in the source term forces the source to be zero everywhere
except in the y-z plane at x = o.
CHAPTER 5
5.1
Event space:
(a) The sel {1,2,3,4,5,6}.
(b) The sel of unordered pairs (21
members) {(1,1); (1,2); (1,3); (1,4);
(1,5); (1,6); (2,2); (2,3); (2,4); (2,5);
(2,6); (3,3); (3,4), (3,5); (3,6); (4,4);
(4,5); (4,6); (5,5); (5,6); (6,6)}.
(c) The sel of ordered pairs (36
members) {(1,1); (1,2); ... ; (1,6);
(2,1); (2,2); ... ; (2,6); (3,1); ... ;
(3,6); ... ; (6,6)).
(d) The continuous set of all energies
into which fission neutrons can be
Typical event:
A die landing 3.
A pair of dice showing (2,4)
(note that the dice are
indistinguishable).
A die landing 4 on the first

throw and 2 on the second
throw.
A neutron emitted from the
fission process at 1.0 MeV.
born.
(e) The set of all energies from E
(energy of the incident photon)
to E' (energy resulting from
0
180 scatter).
or
the set of all photon scattering
angles from 0 to 180,
or
the set of all electron scattering
angles from 0 to 90,
etc.
A photon leaving the collision

point with energy 1.0 MeV,
or
0
a photon scattering through 75 ,
or
an electron scattered at an
0
angle of 30 ,
etc.
5.2 For 5.1(a):

(a) The PDF is the point function having a value of Y6 at 1,2,3,4,5,6 and zero
elsewhere.
":t
I
I
, ,, ,
I
I
759
Solution to Exercise 5.2(a). Values of PDF.
(b) The CDF is the function
, r-i
2/3
t-tl
1/3
I
I
Solution to Exetcise 5.2(b). Values of CDF.
(c) A random-variable function is not necessary to define the PDF. At least an

ordering of the six events is necessary to plot- the CDF.
For 5.1(b):
(cl) Values of the PDF are listed in the accompanying matrix. It is zero elsewhere.
2/36
2/36
2/36
2/36
2/36
2/36
2/36
2/36
2/36
, /36
2/36
2/36
2/36
, /36
2/36
2/36
1/36
2/36
1/36
1/36
1/36
Solution to Exercise 5.2(d). Values of PDF.
760
(e) To obtain the CDF, we must order the possible outcomes. One way to accomplish
this is to assign integers to the outcomes. Let f(i,j) = k be the function defined in
the accompanying matrix.
11
15
18
20
10
14
17
19
13
16
12
21
fl/,jl
k,
1= 1.2 6
j
i, j + I, ...,6
Solution to Exercise 5.2(e). Values of CDF.
The CDF then becomes
..-....-.-
5/6
5/9
--
--
.-
.-
..-
---
..-
..-
5/18
.-
..-
.-
(f) For the PDF, no; for the CDF, yes.

For 5.1(c):
(g) The PDF has a value of Y36 for each ordered pair and is zero elsewhere.
761
(h) The CDP must be constructed as above. It will be a step function increasing
at each step.
%6
5.3 (a) Figure 2.1 is not a PDF. (It has values greater than 1). Figure 2.4 is a PDF. Its
values are in (0,1). and the integral under the curve is 1. (b) Figure 2.1 can be converted to a PDF by integrating the curve from 0 to infinity and dividing the plotted
values by the value of the integral.
5.4 See graph for solution 5.4.
1.0,--,---,,-----,---,<:='0.....,---,----,-----,----,
10
12
14
16
18
E, MeV
Solution to Exercise 5.4.
5.5 (aJ Jg C sin x dx = 1 implies C = 1/2. Thus f(x) = (sin xJ/2 and 0 < f(x) < 1.
(b) CDF = J~ (sin y)/2 dy = (1 - cos x)/2.
(cJ See graph for solution 5.5.
762
Solution to Exercise 5.5(c).
(d) With 0, 0.1,0.2, ... ,0.9, 1.0, the following choices of x are obtained:
I
o
"
e
"e
0.1
0.2
"l
"0 0'" 0~
e"
e"
e" "
e
e"
"!
0
0.3
0,4
":
"!
"
"
0.5
0.6
-"
'"0
"e
0.7
0.8
0.9
1.0
x (units of rr)
Solution to Exercise 5.5 (d).
(e) They are distributed as the PDF. (Notice the bunching toward x = "/2).
(f) Probably so, if enough random samples, n, were chosen with equal probability
from (0,1). (But no guarantees for a finite number of choices-you might have a
run of bad luck!)
(g) Choosing the samples equally distributed in (0,1) forced the values of x to be
distributed exactly as the PDF. Thus we forced more choices to be made in the
important region around rr/2 where the PDF has its maximum vaIue~'This was, in
fact, systematic sampling (See Sec. 5.3.4). The samples were chosen from the
biased distribution
r (x) = f(x)h(x)
where h(x) = 2/sin x and were adjusted by l/h(x).
763
5.6 (a) The graph is not a PDF. The values are not constrained within (0,1).
(b) The following table for 5.6(b) was the result of 300 trials:
SOLUTION TO EXERCISE 5.6(b)-AUTHOR'S RESULTS
rolls
100
200
300
Expected
distribution
Selected
Selected
Selected
No. of
rejections
12
20
33
20%
16%
18%
23
44
64
39%
36%
35%
24
58
86
41%
48%
47%
41
78
117
No. of
17%
33%
50%
That the distribution is not closer to the expected one is revealed in the smaller
table for the distribution of results of tossing the first die (to select the letter to
be tested):
SOLUTION TO EXERCISE 5.6(b)-ANALYSIS
OF AUTHORS
No. of choices
Letter
in 300 tolls
A
B
108
106
86
Perhaps it can be explained by the fact that the authors randomly selected the
two dice from their well-used set of "liar's dice." Of course, a better explanation
is that these results are well within the accuracy to be expected for 300 trials.
CHAPTER 6
6.1 Use Fig. 6.1, and interpolate to obtain curves at 85 and 95 em. Use the 120-cm curve
directly.
(a) The spectrum at 1 is given by the 95-cm curve X 2 X 10 10 The spectrum at 2 is
the sum of the 8S-cm and the 120-cm spectra X 1 X 10 10 . Observe that at
point 2 the contribution of source A is inconsequential.
764
SOLUTION TO EXERCISE 6.1(.)

E,t MeV
47Tr2 N o
No
r
4rrr2 No
No
2
r
41fT No
No
Spectrum at 1 +
Spectrum at 2+
85
em
= 95
em
= 120
em
r =
0.1
4(-5)
4.4(-10)
4(-6)
3.3(-11)
1(-6)
5.5(-12)
0.66
4.4
1.5(-5)
1.7(-10)
1.5(-6)
1.2(-11)
3.5(-7)
1.9(-12)
0.24
1.7
9(-6)
9.9(-11)
1.1(-6)
9.1(-12)
2.2(-7)
1.2(-12)
0.18
0.99
6(-6)
6.6(-11)
9(-7)
7.4(-12)
2.0(-7)
1.1(-12)
0.15
0.66
4(-6)
4.4(-11)
6(-7)
4.9(-12)
1.4(-7)
7.7(13)
0.Q98
0.44
5(-6)
5.5(-11)
8(-7)
6.6(-12)
1.6(-7)
8.8(-13)
0.13
0.55
E,t MeV
2
85
4rrr No
em
No
2
r = 95
41fT No
em
No
4rrr2 N o
r= 120
em
No
Spectrum at 1
Spectrum at 2+
r =
10
11
5(-6)
5.5(-11)
9(-7)
7.4(-12)
1.8(-7)
9.9(-13)
0.15
0.55
4(-6)
4.4(-11 )
8(-7)
6.6(-12)
1.5(-7)
8.3(-13)
0.13
0.44
3(-6)
3.3(-11)
6(-7)
4.9(-12)
2(-6)
2.2(-11)
3( -7)
2.5(-12)
1 (-6)
1.1(-11)
1.3(-7)
1.1(-12)
4(-7)
4.4(-12)
0.098
0.33
0.05
0.22
0.02
0.11
0.04
t4(-5) = 4 X 10-5 , ete.

tNeutron em ~ sec- l MeV- t
(b) Plot the energy-flux-density spectra (ENo ) and numerically integrate:
SOLUTION TO EXERCISE 6.1(b)

E,MeV
--- - - - - - -_.-
At 1,
EN o
At 2,
EN o
0.1
10
0.07
0.24
0.36
0.45
0.39
0.65
0.90
0.91
0.78
0.45
0.20
0.44
1.7
2.0
2.0
1.8
2.8
3.3
3.1
2.6
2.0
1.1
Using fj.E = 0.4 MeV for numerical integration:
f;OEN O (point 1) dE "" 5.3 MeV em"" see- l
11
0.44
765
4.0
2.0
1.0
~V
>
~ 0.6
./'...
0.4
('II
\
~
0.8
l,
$
.........
If
>
:;;
.;
w 0.2
o. 1
0.08
0.06
0.04
10
12
E, MeV
Solution to Exercise 6.1(b). Graphs of energy-flux density at points 1 and 2.
(c) Use Fig. 6.20:

At point 1
At point 2
ot>tn
= (10') (1.6 X 10.... ) + (10') (2.2 X 10-11 ) = 16 neutrons em"
see-'
766
6.2 We first observe that 1 mrad/hr = 2.78 X 10-5 ergs g-t sec- t
flux density incident (essentially normally) to the slab.
Next we compute the
Solution to Exercise 6.2, geometry.
7.63 em = radius of beam tube
d<l>;
10 12 21r,d,
41rp
cosO;, +152 ;p
152
cosO ; - P
<1>; f7.63 ( I012 J'dri52; 1.152.19 152(10 12 ) dp

2
p2
,=0
2p
P
p=152
; 5(152)10"
( 1)
-P
152 1
.
I S2
= 6.2 X 10 8 neutrons em"""2 sec-1 at point B
To compute the attenuation, we use the curves of Appendix M and the fission
spectrum given in Fig. 2.18 to construct Table A. The choices of E were dictated by
the curves of Appendix M. The spectrum NolAE was obtained from Fig. 2.18 multiplied by AB, and the sum over all energy intervals was verified as 1.00.
SOLlTFION TO EXERCISE 6.2---CONCRETE ATTENUATION DATA

B mill
10.... ,
MoV
l,MeV
6E,MeV
13-18
2.'
9-11
10
2.'
N a (E)I4E,t
neutron.,
MoV
N(E)4E',
neutro...
11-13
I'
5.'
12
7-'
8
2.'
S-7
3.5-5
2.5-3.5
1,6-2.5
2~
1.5
I.'
1.0-1.6
1.3
0-1.0
'.7
I.'
3(-5)*
2(-4)
8(--4)
4(-3)
2(-2)
8(-2)
1.5(-1)
2(_1)
3.1(-1)
3.1(-1)
1.5{-4)
4(-4)
t.6{~3)
8(-3}
4(-2)
1.2(-1)
1.5(-1)
l.8(-I)
1.9{-1)
3.1(-1)
8(-9)
8(-9)
1(-8)
9.5(-9)
9(-9)
3.2(_9)
4(-9)
6{-IO)
1.1(_1l)
1.7(_12)
7(-10)
6.5(-10)
9(-10)
8..5(-10)
6.6(-10)
1.6(-10)
t.5(_10)
9{-12)
6(-14)
1.3(-15)
4.3{_1l)
4.5(-11)
?O{-H)
6.5(-11)
".I{-tl)
8.5(-12)
6.5(-12)
1.3(-13)
2(-16)
1.2(-18)
D,'(li. 60 cm).
(erpg-IIeC-I)1
(neutron
cm"'llec-l)
D'(B, 90 cmJ,
(Irp I-I K<:"')/
(_troo
cm"'1 tee -)
D'(E,120 em).
(eJf;'IS-1 1CC-1 )/
neutron
Q111 tee-I)
tSwn OYer aU energies = 1.00.

P(-5)"3)(10"',
767
The three normalized dose spectra were obtained from Appendix M; we extrapolated
where necessary. The objective is to get a curve of total dose rate vs. slab thickness for
the whole spectrum. We now form the product N(E) till D'(E,x) = D(E,x),
using ell = 6.2(8) neutrons em'""2 sec -1, and construct Table B.
SOLUTION TO EXERCISE 6.2-CONCRETE ATTENUATION DATA
E,
N(E) till,
MeV
neutrons
D(E,60),
-I
-1
ergs g sec
D(E,90),
-I
-I
ergs g sec
D(E,120),
-I
-I
ergs g sec
14
12
10
8
6
4
3
2
1.3
0.7
Sum
1.5(-4)t
4(-4)
1.6(-3)
8(-3)
4(-2)
1.2(-1)
1.5(-1)
1.8(-1)
1.9(-1)
3.1(-1)
1.00
7.5(-4)
2.0(-3)
1.0(-2)
4.7(-2)
2.2(-1)
2.4(-1)
3.8(-1)
7.0(-2)
2.0(-3)
3.3(-4)
0.97 = D(60)
7.0(-5)
1.7(-4)
0.9(-3)
4.2(-3)
1.6(-2)
7.5(-3)
1.5(-2)
1.0(-3)
7.0(-6)
2.5(-7)
4.5(-2) = D(90)
3.8(-6)
1.6(--5)
0.7(-4)
3.2(-4)
1.0(-3)
6.5(--4)
6.0(-4)
1.5(--5)
2.5(-8)
2.2(--10)
2.7(---3) = D(120)
tI.5(-4) = 1.5 X 10'4.

Plotting D(60), D(90), and D(120) on semilog paper, we find, as expected (the
curves in Appendix M suggest it), that they fall on a straight line, and
D(x) "" 290 e-o 096X
Thus
-I
290
I
4 -0.096x = n 2.78 X 10'" -In 104 + n 10 -- 13.9
x"" 0.09
13.9 = 145 em
6.3 We use Fig.4.11 and the thickness computed in Problem 6.2 (x = 145 em).
Extrapolating the 152-cm curve of Fig. 4.11, we find that the thermal-flux density is
about 5 X 10'" per incident unit flux density. Thus
=:: 3.1 X 102 neutrons em-z sec-I
768
6.4 The source strength of the disk is 5000 X 3.7 X 10 J 0 X 2
3.7 X 10 14 photons/sec
If we assume that there is no self-shielding, the equivalent surface source is
3.7 X 10"
12 h
-Z-I
3.14 X 102 - 1.18 X 10 p otons em sec emitted isotropically
We plot Berger's coefficients vs. energy (Appendix F) using the 0 to 7 mean-free-path
data (75 em '" 4.7 mean free paths) to find C = 0.96 and D = 0.082.
1.4
1.3
1.2
Berger's coefficients
1.1
0.3
1.0
0.9
0.2
0.8
o
0.1
0.7
0.6
o
o
0.5
0.5
1.0
2.0
3.0
4.0
E. MeV
Solution to Exercise 6.4. Interpolation of Berger's coefficients for water.
==
rdr==pdp
75 em
R'" 10 em
Geometry of Solution to Exercise 6.4.
DR :;
rdr
lO S
r=O
769
)()
/laE
~ e-IJP 1 + CJJ.peDIlP
2p
where Sa = 1.18 X 10 12 photons em-'2 sec-I.

Note that the first parentheses enclose the buildup factor and that the coefficient of
the first term will be the dose rate without buildup.
Let x
= /lP; dP = dx//l; P = x//l, then
e-x
Sa/laE[f~b-dx+C i~b e-x(I-D) dx ]
DR=--
/ld = 0.064(75)
/ld(I-D)
~dx
~d
= 4.80; /lb = /l(d' + R')\I = 0.064(5625 + 100)\1 = 4.84
= 4.80(0.918) = 4.406; /l(d' + R')\I
DR = 1.18 X 10
(0.03)(1.25)
'
[1.45 X 10-
3 _
(1 - D)
= 4.84(0.918) = 4.443
1.37 X 10-3
!l.:~
+ 0.918
X(0.0122 _ 0.0117)] t
= 2.21 X 10 10 [8.0 X 10- 5 + 1.05(5 X 10-4 )]
= 2.21 X 10 10 (8 X 10-5
+ 5.2:' X 10-4)
(.)DR = 2.21 X 10 10 (6.05 X 10-4 ) = 1.34 X 10 7 MeV

(b) The buildup factor is 6.05/0.8 = 7.6.
tValues of E 1 arc from Appendix G.
g-I
sec- 1
770
6.5 Since the concrete shield is hydrogenous, we can use neutron removal theory. From
Table 4.3 the removal cross section for lead is oR = 3.5 barns/atom. Thus
_ 0.6 X 10'24 (11.3) 3 5 10-24

207
. X
= 0.115 em"
The added factor of fast-neutron removal is simply
The fast-neutron dose rate is reduced by a factor of 0.563.

6.6 Since the average prompt fission photon energy is in the MeV region, an
approximate, conservative answer can be obtained by using the minimum attenua
tion coefficient, which occurs at about 3.5 MeV for lead. From Fig. 3.8, a == 15
barns/atom. Thus
24
_0.6X10
(11.3) 5)-049
"
207
(1 - . em
Jl-
The gamma dose rate is reduced by a factor of 0.082. To obtain a more accurate
result, we would follow a procedure for the gamma-ray spectrum similar to the
treatment of the neutron spectrum in the solution to Problem 6.2.
Author Index
Bray, T. A., 672

Broder, D. L., 9
Brown, ]., 291, 700
Burrus, W. R., 490
Burton, B. S., ]r., 275, 280
Butkov, E., 463
Butler,]., 6, 190-192
Aalton, E., 566-570, 572-575

Abbott, L. S., 1, 8
Adamson, J., 560, 561-565
Alberg, M., 503-504, 506, 508
Albert, R. D., 178, 187-189, 286, 299
AIl,n, F,]" 272, 291, 319-320, 323-328,
393-394,700-701,727-728
Ambarzumian, V. A., 163
Anderson, D. C., 189-190
Aronson, R. 6, 465
Attix, F. H., 616
Auxier,]. A., 101, 109-110, 509, 616
Avery, A. F., 6, 190,198
Cain, V. R., 6, 218, 382, 395, 398

CanaH, U., 194
Capo, M, A" 7,171-172,424,635-639
Carlson, A. D., 495, 497, 500
Carlson, B. G., 7, 133
Cashwell, E., 6, 228, 239-240
Casper, A. W., 188, 276
Cerbone, R.]., 494-497
Certaine,].,6, 159
Chaltron, W. F., 527
Chandrasekhar, S., 133, 333-334
Chapman, G. T., 6, 179, 184
Chapman,]. M., 388-393, 400-401, 403
Chase, D. M., 371
Cheka,]. S., 509
Chilton, A. B., 7, 307, 327, 345-347, 381,
383-384,388-391,393-394,641,
731-732
Cierjacks, S., 496, 498
Claiborne, H. C., 8, 27, 63, 111, 362, 420,
519,717,722-725
Clack, F. H, S" 6-7, 232, 291, 456, 458,
641,700,702-711
Clark, 0., 333
Clifford, C. E. (Canada), 401, 405-406
Cliffo,d, C. E. (U,S,A,), 2, 4, 8, 362, 381,
490,493-495,498,514-515
Coleman, W. A., 331, 339-340, 353, 396,
729-731,733
Collier, W. D., 198
Collins, D. C., 378
Counney,], C" 57, 482, 485, 487-489,
501,503
Bach, R. L' 175-176,360,653

Baksys, I. G., 338
Barrett, M. J 'J 347
Bayer, C. E., 398
Beach, L. A., 346, 731
Beaudry, R. J., 527. 529-531, 535-536,
538,542-543,545-547,549,551-558
Beck, H' t 503-504, 508
Bekh",], A" 478-479, 595-597, 599
Bellman, R. 163
Bendall, D. E., 190-192, 194, 375, 377
Benenson, R. E., 375-376
Bennett, C. M., 598, 600-601
Benoit,]. \V" 432
Berger, M. J.t 6-7, 68, 79, 150, 171-172,
286,304,306,342-343,346-347,425,
437
Betz, N. A., 291, 700
Bjeldanes, M. N., 574
Blatt, J. M., 240
Blizard, E. P.,,6, 8, 55, 68, 92,180,291292,299,430,579,582-583,586,592,
594
Blosser, T. V., 580
Bokhari, M. S., 57, 482, 485, 487-489,
501,503-,04,515
Bowman, L. A., 423
t
771
772
AUTHOR INDEX
Coveyall, R. R., 6, 218, 667, 669-670

Cranberg, L., 54
Cranford, W" 465
Curtis, C. c., 198
Gritzner, M. L., 432

Grodstein, G. W., 5
Grosenbaugh, L. R., 672
Grotenhuis, M., 7,425
Grundl,]. A.,55
Davisson, C. M., 346, 731
Gumey, W.]" 8, 347, 350, 732
Diaz, J. A., 334

Dobell, A. R., 663, 667
Doggett,]., 6, 342-343
Doty, D. R., 394
Dresner, L., 8
Dudley, T. E., 13
Dudziak, D.]., 181, 184-185
Dunn, W. L., 184
Edlund, M. C" 333, 340
Edwards, W., 6
Emmett, M. B., 291
Enginol, T. B., 465
Engle, L. B., 52-53
Engle, W. W.,jr., 7, 425
Eriksson, B., 231
Everett, C., 6, 228, 239-240
Fano, U" 5-6, 149-150, 155
Fasano, A. N., 375-377
Fedorcnko, R. p" 464-465
Fenning, F. W" 3
Fermi, E., 332, 340
Fisher, E., 550
Fisher, p. C., 52-53
Flynn, W. A., 595
Foderaro, A., 68, 84, 91,291,426,428-
429
Forti, P., 496
Freestone, R. M., Jr., 490, 580
French, R. L., 7,298,324-325,329,354,
425,442,444,509,514,728
Friedman, S. T., 471, 480, 482
Frye, G., 54
Futterer, A. T., 272, 291, 319, 325-328,
700-701,727-728
Garrett, C. W., 11, 63, 119, 207, 261
Gelbard, E. M., 148,248
Gilley, L. W., 509
Glasstone, S., 333, 340, 595
Gocrtzel, G., 6, 218
Goldberg, M. D., 88
Goldstein, H., I, 5-~6, 8, 54-55, 58, 68, 70,
78,80-82,88,150,171-172,265-268,
422-426,627,687
Gousscv, M. G., 291
Grant,]. S., 401
Greenhaw, C. R., 425
Greenspan, H., 338
Haffner,]. W., 189,299

Hageman, L. A., 148
Haggmark, L. G., 347-348, 350, 732
Halliday, D. B., 3, 387-388
Halpern, 0., 333, 340
Hansen, K. F., 167
Harrison,]. R., 3
Hayward, E., 6
Haywood, F. F., 509-513
Heaton, H. T. II, 498
Held, K., 465
Henderson, B.]., 101-103
Henry, K. M., 490
Henry, R. L., 327-330
Herbold, R. J., 175, 653
Hjarnc, L. H., 196, 567
Holden, N. E., 13
HolI, R.]., 575-577

Holland, L. B., 278
Hoover, L.]., 426
Horton, C. c., 1,3,6,8,366,387-388
Howerton, R.j., 91
Hubbell,]. H.,6-7, 68, 79, 115, 172, 175-
177,360-361,653,655,660
Huddleston, C. M., 327, 345, 389, 391, 394
Hull, T. E., 663, 667
Hungerford, H. E., 451, 519-520, 527-
528,550
lise man, H., 194
Ingold, W. c., 391
Irving, D. C., 6
jaeger, R. G., 8, 607
jaeger, T., 8
jerri, A. ]., 400
johns, H. E., 505
johnson, W. P., 432
] oncs, A. R., ] 11
Jones,]. D., 6
Jones, T. D., ]09, 616
Jones, T. H., 347, 350, 732
] udd, A. M., 560

Kahn,H.,6,230,240
Kalaba, R., 163
Kalos, M. H., 6, 218, 244
Kam, F. B. F., 456, 458, 641
Keith, J. R., 432-433
Kimel, W. R., 8
AUTHOR INDEX
King, R. W", 54
Kirkbride,]., 49
Klahr, C. N., 430, 465
Kodras, F. D., 300
Komarovskii, A. N., 445
Kapp, H. J ,,148
Kopsch, D., 496
Kouts, H. J ,,411-413
Kovalev, E. E., 291
Krakar,].].,598
Krell,..t,566
Kropp, L., 496
Krumbein, A. D., 265
Kruse, W. G., 598
Kusler, I., 124
Lacy, P. S., 579
Lahti, G. P., 460, 462
Lamkin, J. C., 176, 360, 653
Lanning, W. D., 132
Lathrop, K. D., 133,425-426
Leachman, R. B., 56
Le Doux,]. C., 381, 383-384, 388-391,
393
Lee, C. E., 133, 146
Leimdorfer, M., 196, 231-232, 341-342,
349-351,567
Leonard, B. P., J r" 4
Lewellen, J. W" 5 79~580
Lewin,]., 8
Lietzke, M. P., 424, 635
Littler, D.]., 3
Loy, E. T., 509
Lueneburg, R., 333
Lundin, M. 1., 509
Lyday, R, O"J'" 400
Lynch, R. E., 432
MacDonald, J. E., 6
MacPherson, R. D., 667, 669
McLaughlin, ]. E., 503-504
McNair, N. I., 560
Maerker, R. E., 278, 319-320,324,330331, 337-340, 353~354, 362, 396-400,
490,726,729-731,733
Maienschein, F. c., 4,13,49-52
Mantey, R. F., 527
Marable,]. H., 426
Marcum,]. I., 438-439
Marsaglia, G., 672
Marshall,]. D., 378-380, 435, 506-507
Maskewitz, B. F., 8
Mason, E. A., 167
Mathews, D. R., 167
May, V. M., 88
Meem, L., 4
773
Mendelson, M. R., 13
Metropolis, N., 666
Miller,). E., 6, 494
Miller, R. A., 465
Mills, C. B., 549
Mingle,]. 0., 167
Mittelman, P., 6
Mizuta, H., 167
Mackel, A., 334-335, 730
Mooney, L. G., 327, 329-330, 419, 444,
471,509,514
Morgan, I. L., 607
Morgenau, H., 463
Morris, E. E., 347
Moses, E., 334
Moteff,]., 6
Muckenthaler, F. 1., 7, 278, 319--324,
330-331,337-340,353-354,396-400,
426,490,726,729-731,733
Mueller, G. 0., 425
Murphy, G. M., 463
Mynatt, F. R., 7, 132, 147, 426, 490
Nebe,]., 496
Nelms, A. T., 74, 76
Nereson, N., S4
Neuer,]. R., S5
Neufeld,]., 108-110, 398
Newell,)., 398
Newson, H., 2
Nichols,]. P., 456, 458-459, 461
Nicks, ~., 198
O'Brien, K., 185-186, 503 --504,508
O'Reilly, B. D., 265
Peelle, R. W., 4, 13, 49, 50
Penny, S. K., 8, 426
Perkins,]. F., 54, 610
Pertine, G., 198
Peterson, D. M., 282
Peterson, E. G., 194-195
Piercey, D. C., 375, 377
Plechaty, E. F., 91
Pomraning, G. C., 335, 340
Ponti, C., 198
Preusch, H., 194, 198
Prevost, R. ]., 327, 329-330
Price, B. T., 3, 8
Price, W.)., 184, 516
Profio, E. A., 494
Raby, T, M" 579
Rasa, D.)., 342-346, 348, 389
Rickover, H., 2
Ritts,].)., 101-102, 105,110,611
Rockwell, T. III, 8, 550
774
AUTHOR INDEX
Roesch, W. C., 616

Rosen, L. 54
Rothberg, D. M., 185-186
t
Sabian, G. E. 425
Sanders, F. W., 509
Sandlin, R' 566
Schaeffer, N. M., 1,4,354,419,519
Schmidt, F. A. R., 717-721
Schmucker,]. E., 181, 184-185
Schrack, R. A., 498
Schreiber, P. W., 300
Schubart, H. 198
Schwartz, R. B., 498
Scofield, N. E., 347, 350, 732
Scoles, J. F., 54
Selph, W. E., 11,63,119,207,261,279,
313,398
Seraphin, D. S., 672
Serduke,]. T., 275. 281
t
Shamberger, R' t 3
Sheffield, R. D., 465
Shelton, F. H' t 432-433
Shimizu, A., 167
Shure, K., 6,132, 185-186, 189-190
Simon, A., 381
Skarsgard, L. D., 505
Slotin, L., 2
Smith, C. V., 275, 281
Smith, D. C. G., 563
Smith, N. M., Jr., 413
Smith, W. R., 579, 584-585, 587, 590,
592
Snyder, W. 5.,101, 108-110, 398, 616
Solomito, E., 93. 101, 178, 611
Song, Y. T., 326. 388, 393-394
Spanier,}., 248
Spencer, L. V., 5-6,8,149-150,155,334
Spinney, K. T., 3, 8, 190, 192, 195, 201,
300,366
Stehn, J. R., 88
Steinberg, H., 465
Stephen, D. W., 575-577
Stevens, P. N., 6, 7, 27, 63, 101, 119, 301,
331,396,426,611,729-731,733
Stockton,]. R., 178
Storrs, C. L., 184
Straker, E. A., 6, 432-434, 438-441, 490
Suvorov, A. P., 9, 464-465
Taylor,].J., 7,171, 306

Terrell, C. W., 391-393, 402
Thompson, R.}., 595

Tipton, C. R., Jr., 451
Tobias, M. L., 424, 635
Todd. H. A., 515
Tonks, L., 409-410
Troubetzkoy, E. S., 89-91, 463, 465
Trubey, D. K., 6, 8, 111, 172, 175, 177,
179,247,291,300,302,304,371,423,
425,427,432,629-631,641,644-649,
657,662,734
Tsypin, S. G., 9
Turner, M. A., 579, 584-585, 588-590,
592
Unseld, H., 496
Verbinski, V. V., 57, 482, 485, 487-490,
498,501,503-504,515
Veselkin, A., 9
Vogel, R. E., 435
Vondy, D. R., 424, 635
von Neumann,J., 666,673
Waldman,J., 347
Walker, R. L., 425
Ward, D. R., 509
Watt, B. E., 54
Webster, W. M., 432, 434
Weinberg, A. M., 129, 131
Weisskopf, V. F., 240
Wells, M. B., 6, 324-325, 329, 336-337,
340-342,345-349,352-354,432,435,
437, 506-507, 728
Welton, T. A., 3, 178, 187-189, 286, 299
Western, G. T., 277
Whitesides, G. E., 57, 482, 485, 487-489,
501, 503-504
Wick, G. C., 133
Wigner, E. P., 129, 131, 533
Wilkins, J. E., Jr., 6, 58, ISO, 171-172,
422-423, 426
Willoughby, D. F., 495
Wing, G. M., 163
Worlton, J., 133
W,ight, W., 272-273, 319, 325-328, 727728
Yegorov, Yu., 9
Yost, K.J., 91, 93, 218
Zerby, C. D., 6, 432, 435
Zinn, W., 2
Zolle" L. K., 179, 182, 478-479
Subject Index
fast-neutron, Monte Carlo, 319-322

borated polyethylene, 324-327
concrete, 319-328, 726, 728
curve fit, 321-322, 325, 327,
Absorbed dose, 95-98, 101, 106, 112, 116,

cad. 97, 104
Absorbed dose maximum, 616
Absorbed dose rate, 26,97
Absorption coefficients, narrow-beam,
726-728
dependence on hydrogen content,
82-83
326,329
Absorption of neutrons, 84
Activation cross sections, 93-94
Activation neutrons, 18
Activation-product gamma rays. 16
Agesta reactor, 566~574
experimental verification,
322-323,328-330
iron, 324-327, 727-728
soils, 324-327, 727-728
steel with borated polyethylene
vs. thickness, 328-330
water, 324-327
gammaray, 341-350
aluminum, 731-732
concrete, 343-350, 731-732
contributions by positron
annihilation, 341-342
formulas for concrete, 345,
calculational model, 567-570

description of reactor and shield, 566--570
magnetite concrete in, 566
measurements, 570-574
primary shield, 568
streaming in shield. 571
AIM-6 diffusion code, 548-549. 559
Aircraft Shield Test Reactor (ASTR), 4
Air dose, 99
Air Force, early nuclear program, 2~5
Air transport, 430-444
air-over-ground calculations, 437-439
boundary effects, 431
comparison of calculation and measure
ment, 440-442
differential data, 435
effect of air density, 435, 437
fission-neutron, 432, 434-435, 437
geometry, 431
infinite air medium, 432-437, 439
interface effects, 439-444
neutron (14 MeV), 432-433
single-scattering, 432
348,731-732
rron, 342-343,348,731-732
lead, 731-732
for spherical walls, 349-351
vs. atomic number, 342-344
water, 342-343, 731-732
geometry for definitions. 315
intermediate-neutron, 331-332, 729
materials data available, 314
Mockel's coefficients, 730
neutrons incident at nonthermal
energies, 339-340
less than 200 keY, 339-340
up to 9.57 MeV, 340
Albedo, 228, 313-356
neutrons incident at thermal energy,
application to wall scattering,
332-339
355-356
analytic results, 333-336

fits to Monte Carlo results, 336-339,
assumption, 314
differential types, 315-317
effective, 314, 318
730-731
775
776
SUBJECT INDEX
secondary gamma-ray, 350-355

fits from Monte Carlo data, 352--354,
733
thermal-neutron, 319, 730-731
three types, 316
total types, 315, 317,-318
Albert-Welton point kernel, 178, 187-188,
286-287,299
Angular current, 29
Angular current density, 34--39
Angular distribution
monodirectional approximation of,
46-48
plane isotropic source, 4S
Angular energy-flux density. 30
Angular flux density, 29, 34
group, 31-32
ANISN code, 7
Annihilation radiation, 16
Anthracene scintillation dosimeter
(ASD),475
Argonne National Laboratory
(ANL),7
graphite pile, 1
Associated spherical-harmonic functions,
625
Associated spherical harmonics, 129
Atomics International (AI) Shield
Test Facility, 596
Atoms for Peace Conference, 6
Attenuation
exponential, 122
geometric vs. material, 120 --123
neutron (see specific transport method,
e.g., Moments method)
shield calculations, 261-309
See also Shield attenuation
calculations
slab, 121--123
spherical shell, 123
Attenuation coefficient, 66
See also Cross sections
linear, 83, 115
photon, 80-83
Attenuation functions for gamma
rays, 653
slab geometry, 653-662
cosine source, 656-657, 661-662
disk source, 653-658
isotropic source, 654-655, 659--660
rectangular source, 658-662
ATTOW removal-diffusion code, 198-199
Average value, 24
Azimuthal angle, 42
Barn book, 92,94

Barrier factor, 121
BEPO,3
Beryllium measurements, 478
Beryllium oxide measurements,
478-482
Bettis Atomic Power Laboratory, 2
Biasing, 216, 223--225
collision parameters, 234-236
effect on variance, 250
particle weight, 223-225, 228
path length, 229
plane isotropic source, 224-225
point isotropic source, 223-224
Binding energy of neutrons, 15, 91,
93,97
Biological damage, 94-112
Biological dose, 94-95
Boiling water reactor, 574-579
calculations, 578
measurements, 579
plant layout, 576
Boltzmann transport equation, 7, 119-120,
123-129,208,216-217
applied to discrete ordinates, 134-135
approaches to solving, 128-129
conservative form, 135
integral operator, 136
integro-differential form, 127-128
in moments method, 150-151
second form, 131
transformations, 128
Boral
effect on captures, 481
Borated graphite
in Dounreay reactor, 560-562
in Fermi reactor, 537-541
Borated materials, thermal absorption
in, 93
Boron, 449-450
in shield fabrication, 525
Boron carbide measurements, 478
Boron steel
in Fermi reactor, 540-541, 545
Boron trifluoride detector, 473
Bremsstrahlung, 17, 83
in dose calculations, 101
production of photoneutrons, 497, 501
BREN (Bare Reactor Experiment Nevada),
509
British program, 3-4, 6
Brookhaven National Laboratory (BNL),
3, 65
SUBJECT INDEX
data files, 492
Broomstick experiment, 490-492, 494
Buildup factor, effective, 424
Buildup factors for gamma rays
applications, 284-286
Berger form, 171-172, 286,304,307,
628,631,640
coefficients for air, 640-641
aluminum, 636
concretes, 640
iron, 637
lead, 639
lithium hydride, 640-641
sand,640-641
tin, 638
water, 635
wood,640-641
in ducts, 368, 407
energy absorption, 170
energy deposition, 170, 631-632
equivalent Z method, 423
exposure, 170, 626-628, 630
finite media, 422
infinite-plane source, 285
interpolation method, 424
linear form, 171-172,628
multiple layers, 423-424
narrow-beam geometry, 284
polynomial form, 171-172, 305, 635
quadratic form, 171-172, 628
in shield designs, 578, 589, 598
in shield heating, 421-425
Taylor form, 171-173, 306, 408
wide-beam geometry, 284-285
Bulk Shielding Reactor (BSR), 1,4,
482-485,591
C-17 code, 282
Cable-route openings (see Ducts)
Calcium borate in Fermi reactor, 539-543,
546
Capture gamma rays
energy distribution, 15
sources of, 14-15
spectra, 15
Carbon steel in Fermi reactor,
540-541,545
CarlsofJ's Sn (see Discrete-ordinates
Sn method)
Center-of-mass coordinates, 236
Charged particles
equilibrium, 99-101
heating, 429-430
sources of, 12
Chicago Pile-3, 2
777
Coherent scattering of photons,

80-81,115
Collided radiation, 119-120
Collision kernel, 21.7
Collision parameters, 234-242
biasing, 235-236
direction cosines, 241-242
gamma rays, 235, 240
neutrons, 235-240
Compton effect, 70-74, 97-98, 104, 341
energy deposition, 73
fractional energy loss, 73
in Monte Carlo, 240
Compton wavelength, 71--72,117
Conceptual design, 520-524
calculations for, 521--522
detailed considerations, 522-524
economic considerations in, 523-524
radiation streaming in, 523
Concrete, 1-3, 447--451
barytes, 3, 450---451
colemanite, 450-451
construction criteria, 534
in Dounreay reactor, 560-562
effect of water content, 456, 459, 461
effective atomic number, 425
ferrophosphorus, 450-451
heating criteria, 533
iron-portland, 3, 450-451
lumnite, 450""-451
magnetite, 544, 566
monoenergetic-neutron transport
data, 291
in Pathfinder reactor, 576
serpentine, 544
Convair, Fort Worth, 2-5
Conversion factors (see Neutron
conversion factors)
Coordinate systems, 617
center-of-mass, 86-87
laboratory, 86
transformations, 617 --619
Cosine distribution, 45-~48
Coupled Monte Carlo and discrete-ordinate
calculational links, 7
CRAM, 548-550, 556
Cross section, differential, 72, 237
energy absorption, 420
Cross sections, 63-67
See also Removal cross section
absorption, 64
activation, 93-94
effect of state of medium, 64
evaluated data, 64-65
778
SUBJECT INDEX
gamma-ray, 68-83
Compton energy absorption, 73
Compton scattering, 73, 81
pair production, 74-80
photoelectric effect, 68-70

macroscopic, 65-66
microsl;opic, 64-65
neutron, 53-94
effect of minima, 490-492, 503

elastic, 84--88
inelastic, 64, 89-91
nonelastic, 93
radiative capture, 92
total, 64
units of, 65
Crud, 575-578
Cumulative distribution function (CDF),
21,211-215,258
Cumulative scattering probability, 237
Current, 35
angular, 29
differential net, 35
net, 35
scalar, 36
Current density, 26, 34-39
angular, 34-39
examples, 37 -39
net, 60
Curvature coefficient in discrete ordinates,
139
Delbruck scattering, 80
Demonstration Monte Carlo program,
251-254, 674-686
flow chart, 676-678
FORTRAN listing, 679-684
input data requirements, 675
test problem, 684-686
Depleted uranium, 454-455, 457
Detectors
boron trifluoride, 473
concept, 26
gamma-ray (see Gamma~ray detectors)
neutron (see Neutron detectors)
thermoluminescent (TLD), 476
Diamond difference technique, 133,
144-145
breakdown, 145
Differential cross section, 72, 237
Differential distributions, 19-24
first-moment, 24, 60
function, 29, 42
histogram form, 22
nth-moment, 24
Differential energy*flux density, 31
Differential flux density, 30, 43

Differential net current, 35
Differential particle density, 28
Differential phase-space cell, 124, 134
Differential solid angle, 25
Diffusion theory, 160-162, 190
equation, 161
fust-flight correction, 190
group approach, 161--162
Dirac delta function, 621-622
Direction cosines, 241-242
Directional distributions, 42-48
cosine, 46-48
isotropic, 47
monodirectional approximation, 48
normalization, 46-47
quadrature schemes in, 146
source term in, 143
Discreteordinates Sri method, 7, 132-149,
494
advantages and disadvantages,
148-149
anisotropy of gamma-ray scattering,
426
calculation, removal term of, 140--142
curvature coefficient in, 139
derivation of finite-difference equation,
136-144
group angular flux in, 136
multigroup calculations for, 140
negative flux fix-up, 145
numerical solution, 144-148
ray effect, 148
spherical geometry, 134
Dispersion, 248-250
Distribution functions
cumulative, 21, 211-215, 258
sampling from, 209-215
Divergence theorem, 126, 620-621
Domain of function, 210, 214
Dose, 94-112
absorbed, 95-98, 101, 106, 112, 116
maximum, 108-111, 616
rate, 26, 97
air, 99
biological,94-95
first-collision, 99-103
free-air, 99
free-field, 99
maximum permissible (MPD), 112
multicollision,112
Dose buildup factor, 170
Dose current, 271-272
Dose distribution factor (OF), 108
SUBJECT INDEX
Dose equivalent, 105-108, 112
maximum, 108-111, 616
Dosimeters
anthracene scintillation (ASD), 475
glass (see Gamma-ray detectors)
Hurst fast-neutron, 472, 511
DOT, 7, 426
Dounreay Fast Reactor, 559-566
borated graphite in, 560-562
calculation model, 562

graphite in, 559
resiloD in, 559-560
rotating top shield, 559-566
streaming, 565-566
DTF code, 486-490, 494

DTK code, 486-490, 502-504, 548
Ducts, 356-404
bent, 378-404
cylindrical, cosine source, 735-736
isotropic source, 734, 736
line-of-sight component, 357-367
cosine" (J distribution, 358-359,
362
cylinder annulus, 366-367
cylinder ducts, 365-366
long thin ducts, 363
rectangular ducts, 360-363
rectangular slots, 363-365
types of plane sources, 358-359

relative magnitude of line-of-sight
and waIl-penetration components,
371-372
wall-penetration component, 367-375
comparison with experiment, 374-375
cylindrical ducts, 370-372
partially penetrating cylindrical
ducts, 372-374
ray analysis, 368-370
wall-scattered component, 375-404
albedo methods, 380-404
albedo Monte Carlo, 395-404
analog Monte Carlo, 376-380
coplanar vs. noncoplanar three-legged
ducts, 401
corner-lip correction, 383-384
double reflected component, 382
experimental studies, 400-404
fIrst-order reflection, 380
rectangular-duct formula, 391
rectangular ducts, 388-394
Simon-Clifford technique, 385-388
V-shape, 400-402
Z-shape, 400-402
Economic considerations in conceptual

design, 523-524
Effect of medium, 57 -58
Effective buildup factor, 424
Effective relaxation length, 288, 371
Elastic neutron scattering, 84-88
angular distribution, 87
anisotropy, 88
ENDFjB (evaluated nuclear data
filej, 65, 492, 494-495
Energy-absorption cross section, 420
Energy cutoff in Monte Carlo, 255
Energy deposition
defInition, 96-98
methods, 419-430
rate, 67
Energy distribution
of fIssion neutrons, 54-56
of particle-reaction neutrons, 18
Energy distributions, 48-50
effect of medium, 57-59
sampling from, 218-219
Energy fluence, 33
Energy imparted by ionizing radiation, 96-98
Energy removal rate, 67
Energy spectra, 20
See also Differential distributions;
Energy distributions
Engineering
construction follow-up, 525
fIeld drawings, 525
in shield design, 524-525
Enrico Fermi Atomic Power Plant,
519,527-559
biological dose criteria, 532
borated graphite in, 537-541
boron steel in, 540-541, 545
calcium borate in, 539-543, 546
calculational techniques, 548-552
carbon steel in, 540-541, 545
comparison of measurements and
calculations, 553-559
description of plant, 528-532
exotic materials, 548
graphite shield, 537-541
heat-removal system, 532
intermediate heat exchanger, 544
lead in, 541
magnetite concrete in, 544
operating floor shield, 544
primary neutron shield, 535-541
primary shield, 528
primary sodium-pipe shields, 547
radial biological shield, 544
779
780
SUBJECT INDEX
radiation-damage criteria, 532-533

radiation-heating criteria, 533
reactor shield, 534-547
remote maintenance, 533
rotating-plug shield, 539-540
safety standards, 534
secondary shield wall, 541-543
serpentine concrete in, 541
shield costs, 548--549
shield criteria, 532--534
shield systems for the auxiliary facilities,
547
shield tcst program, 527
sodium coolant, 529-559
sodium pump shield plugs, 544

stainless steel in, 540-541
steel shot in, 539,545
streaming, in shield, 550
streaming paths, 538-547
thermal shield, 537, 542-543
Entranceways (see Ducts)
Epithermal region, 85
Equation of continuity in diffusion theory,
160-162
Equidistribution, 664
of random-number generators, 664,
669-670
Euler's constant, 306
Evaluated nuclear data file (see ENDF/B)
Evaluation of integrals by Monte Carlo
method,216-218
Event, 209, 258
probability, 207--208
space, 210
Excitation energy, 239
in neutron capture, 91
Expected value, 212-213, 248-250
Experimental shielding, 471-516
importance of, 471
purposes of, 472
Experiments
accelerators, 493-502
air transport and air-ground
interface, 507 -514
60Co gamma-ray, 508---514
cross-section evaluation, 490-492
137CS gamma-ray source, 503-507
duct penetration, 514-516
fixed sources, 503-507
LlNAC, 498-503
neutron, (14 MeV), 507-514
neutron spectral and angular
distribution, 482-490
reactor source, 477--493
various environments, 516
Exponential attenuation, 122

Exponential function, 305
Exponential integral functions, 292,
642-652
graphs, 644-652
Exponential transform, 229, 232
Exposure, 103-105
Exposure buildup factors, 170, 626-628,
630
See also Buildup factors
Exposure rate, 26, 104
Fallout shielding, 8
Fast breeder reactor, 527-566
Fast-neutron energy region, 85
Fast reactor system, special shielding
problems of, 527
Federal Register, 109
Fermi age
in moments method, 149
theory, 189
Fermi Nuclear Plant (see Enrico Fermi
Atomic Power Plant)
Finite-difference cell, 132, 134
Finite-difference equation, 132
First-collision dose, 99-103
First-flight estimate, 244-247
First-last collision model, 442-444, 509
Fission chambers, 473
Fission gamma rays, 13
delayed, 13
energies, 13
intermediate-life, 13
prompt, 13
shordife, 13
spectra, 49-51
Fission neutrons, 17
See also Neutron spectra from
fission
air transport, 432, 434--435, 437
average energy, 17
delayed emissions, 17
energy distribution of, 54-56
energy range, 17
number, 17
Fission plate, 492
Fission-product gamma rays, 13-14,
51--54
beta-ray energy, 54
disintegration rate, 54
energies, 14
function of reactor operating time, 53
half-lives, 14
long-lived,52
total gamma-ray energy release, 54
SUBJECT INDEX
Fitted-parameter data application, 277-283
Fluence,33
Fluorescent radiation, 68
Flux, 28
See also Flux density
removal, 194, 196
Flux density, 26, 28
angular, 29, 34
angular energy, 30
definition, 124
differential, 30, 43
differential energy, 31
examples, 37--38, 60
group, 31-32
incremental sphere concept, 31
special limiting process, 26, 33
thermalneutron, 298-301
total, 31, 32
total energy, 31-32
FORTRAN,251
Free-air dose, 99
Free-field dose, 99
Frequency functions, 210
Gamma-ray attenuation (see specific transport method, e.g., Moments method)
Gamma-ray attenuation functions (see
Attenuation functions for gamma rays)
Gamma-ray buildup factors (see Buildup

factors for gamma rays)
Gamma-ray cross sections (see Cross
sections, gamma-ray)
Gammaray detectors, 475
anthracene scintillation dosimeter,
475,591
film, 476
Geiger-Mueller, 571
Geiger-tube, 579
glass, 476
glass dosimeters, 571
ion chambers, 571
ionization chamber, 599
NaI(Tl) spectrometer, 475--503
calibration, 505-506
solid-state spectrometers, 475
Gamma-ray doses in phantoms, 111
Gamma-ray energy release vs. time
after fission, 608-610
Gamma-ray heating, 309, 420-429
comparison of Sn, Monte Carlo, and
buildup-factor methods, 426--427
differential energy spectra, 425
multigroup diffusion code, 425
Sn code, 425
781
Gamma-ray sources, 12-17

See also Activation-product gamma rays;
Capture gamma rays; Fission gamma
rays; Fission-product gamma rays;
Secondary gamma rays
inelastic neutron scattering, 605-607
relative importance, 12-13
Gamma rays
activationproduct, 16
attenuation functions for, 653
See also Attenuation functions for
gamma rays
buildup factors for (see Buildup factors
for gamma rays)
capture (see Capture gamma rays)
fission, 13
fission-product, 13-14, 51-54
inelastic-scattering, 15-16
kerma rate factors for, 102-103
reaction-product, 16
General Dynamics, 478
Aerosystems Test Facility, 596
General Electric Company,S
Graphite, 449
in Dounreay reactor, 559
in Fermi reactor, 537--541
specifications, 534
Green's functions (see Kernel technique)
Ground Test Reactor (GTR), 4
Group angular current, 35
Group angular flux density, 31--32
Group angular flux in discrete ordinates,
136
Group flux density, 31--32
Group net current, 35
Gulf General Atomic, 494, 497
Hanford,2
Hanford reactor, 1-2
Haydite, I
Health Physics Research Reactor
(HPRR),509
Heavy water, natural uranium, 566--574
Helium-3-filled detectors, 473
History generator, 209
routine for, 218
Hurst fast-neutron dosimeter, 472, 511
ICRP (see International Commission on
Radiological Protection)
ICRU (see International Commission on
Radiation Units and Measurements)
Importance functions, 217-218, 230
Independence, in random-number sequences,
664-665,670--671
782
SUBJECT INDEX
Inelastic neutron scattering, 84, 88-91

gamma rays from, 90-91
threshold, 88-89
typical cross section, 89
Inelastic-scattering gamma rays, 15-16
emission time, 15
energies, 16
Inscatter term
in discrete ordinates, 139-140, 142
in transport equation, 125, 127
Interaction loss term in transport equation,
127
Interaction probability, 29
See also Cross sections
Interaction rates, 63
Interactions of radiation, 63-117
Intermediate neutrons, energy of, 85
International Atomic Energy Agency
(IAEA),8
International Commission on Radiation
Units and Measurements (ICRU), 27, 33,
95,99,103-104,106,112
International Commission on Radiological
Protection (ICRP), 107-108
Invariant imbedding, 7,163-168
advantages and disadvantages, 167-168
application, 167
Ionization chambers
gamma-ray 475
neutron, 473
Ionizing radiation, energy imparted by,
96-98
Iron, 449
neutron cross section, 494
spectra, 492
Isotropic distributions, 43-45, 47, 60
Isotropic point sources, 44, 60-61
Kerma,99-103
Kerma rate, 26, 100
factors for gamma rays, 102-103
factors for neutrons, 101-102
Kernel technique, 168-188
gamma rays, 169-177
neutrons, 178--188
plane disk source, 172-17 5
rectangular source, 176-177
KFK data files, 492
Klein-Nishina formula, 72, 341
Knolls Atomic Power Laboratory
(KAPL), 2, 5
Kronecker delta function, 624
Laplacian operator, 620
Last-flight estimate, 244-247
Law of large numbers, 216
Lead, 448, 449

in Fermi reactor, 541
in Savannah reactor, 581, 587
Legendre polynomial, 360-362, 622-623
associated, 623-625
in discrete ordinates, 139-144
in moments method, 151-152, 156
in spherical harmonies, 132
Lethargy, 33
Lid Tank Shielding Facility, 3, 178~180,
492-493
LIDO reactor, 4
LlNAC (linear accelerator), 494
Line source, 40
Linear attenuation coefficient, 83, 115
Linear energy transfer (LET), 105
Lithium hydride, 454-456, 458, 460, 462
in SNAP-lOA shield, 596
L05 code, 378
Lockheed, 5
Los Alamos Scientific Laboratory
(LASL),6-7
MAC
difference from Spinney method,
195-196
removal-diffusion method, 194-196
Macroscopic cross sections (see Cross
sections, macroscopic)
Macroscopic transfer coefficient,
multigroup, 140
Magic nuclei, 91
Magnetite concrete
in Agesta reactor, 566
in Fermi reactor, 544
Manhattan Project, 1
Masonite, 2
Mass attenuation coefficients, 66, 83, 115
Massachusetts Institute of Technology, 2
Materials Testing Reactor (MTR), 4
Mathematical experiment, 20
Maximum absorbed dose, 108-111
Maximum dose equivalent, 108-111
Maximum permissible dose (MPD), 112
Maxwell-Boltzmann distribution of
thermal neutrons, 85
Mean value, 212-213
Mean-value theorem, 137-138
Measured data applications, 274-283
gamma-ray data, 276, 280~281
neutron data, 275-279
Microscopic cross sections (see Cross
sections, microscopic)
SUBJECT INDEX
Mock-up tests, 472, 478

MODRIC, 300
Moments method, 5-6,149-159
application, 265--269
definition of moments, 149
in design cases, 588, 598
energy degradation in scattering,
159
Fermi age in, 149
gamma-ray differential data, 268-269,
692-695
accuracy, 268
buildup factors, 269
finite media, 268
lead, 696-699
spectra, 269
water, 692--695
neutron differential energy spectra,
265-268,687-691
beryllium, 687
carbon, 588
CH,689
CH 2 ,690
finite media, 267-268
H 2 ,691
H 2 0,691
sample problem, 268
spectral hardening with depth, 266
surface source, 268
neutron transport, 158-159
polynomial expansion of flux densities,
155-157
sequence of moments calculation, 153
undetermined parameters to construct
the flux densities, 158
Monte Carlo codes, debugging of, 255
Monte Carlo data applications, 270-274
comparison with moments method, 270
complexity of, 270
neutron attenuation, in concrete,
702-716
in soil, polyethylene, and water,
712-716
principle of reciprocity, 270-271
sample results, 272-273
Mon'te Carlo methods, 6, 207-259
biasing, 216, 223-225, 250
collision parameters, 234-242
cross-section preparation, 235
demonstration program, 251-254,
674-686
evaluation of integrals by, 216-218
exponential transform, 229-232
first-flight estimate, 244-247
importance function, 217-218, 230
783
last-flight estimate, 244-247

neutron elastic scattering, 236-238
neutron inelastic scattering, 238--240
particle absorptions, 240--241
particle scoring, 242-247
path-length sampling, 225-- 234
probability distribution function,
210-215,258
programming suggestions, 254-257
random sampling, 209-215
rejection technique, 214-216, 258
splitting, 234
statistical estimates, 243-244
statistical variance, 247-251
transport equation, 208, .216-217
MORSE code, 6, 256
Most-probable value, 24
Multicollision dose, 112
Multigroup calculations for discrete
ordinates, 140
Multigroup cross section in Monte
Carlo, 256-257
NaI(Tl) gamma-ray spectrometer, 475-503
calibration, 505-506
Narrow-beam absorption coefficients,
82~83
National Bureau of Standards (NBS), 5 - 6

National Neutron Cross Section Center, 65
Nautilus, U.S.S., 5
Navy nuclear-powered submarine
program, 2-5
Negative flux fix-up in discrete
ordinates, 145
Net current, 35
Net current density, 60
Net leakage term in transport
equation, 125-127
Neutrinos, 12
Neutron attenuation (see specific
transport method, e.g., Moments method)
Neutron binding energies, 15, 91,93,97
Neutron conversion factors
fluence-tokerma, 611--613
reactions in neutron kerma calculations,
611-615
Neutron cross section (see Cross sections,
neutron)
Neutron detectors
active, 472-474
BF 3 counters, 473, 515, 554, 579, 591
emulsion, 475
fission counters, 473, 553. 563, 598
fission foils, 553
foils, 474, 563, 573
784
3 Hc-filled, 473
Hurst fast-neutron dosimeter, 472, 511
6LiF,486
long counter, 591
Ne-211 organic seintillators, 494, 497
Nc-213, 490--491, 501

Ne-908 lithium gas scintillator, 494
passive, 472--475
proportional counter, 474
proton-recoil counters, 554
solid-state, 474
sulfur pellets, 502
track-length, 474
Neutron doses in phantoms, 109-110
Neutron elastic scattering
average angle, 428 -429
average cnl:rgy loss in, 429
in Monte Carlo calculations, 236-238
Neutron emission, 18
Neutron heating, 427--429
Neutron inelastic scattering, 15 --16
in Monte Carlo, 238---240

Neutron kerma factors, 101--102
Neutron radiative capture, 91~93, 98
Neutron reactions, 83-94
contrasted with gamma-ray reactions,
83--84
nonelastic events, 90~-91
Neutron scattering
capture, 86-87
diffraction, 86
elastic, 8488
inelastic, 84, 88 91
potential,86
resonance, 86~-87
Neutron sources, 17-19
polonium--beryllium,18
Neutron spectra from fission, 54-57
average energy, 56
empirical expression, 54-55
23SU, 23JU,and U9pu, 56
Neutron spectrometer, 486--491
Neutrons
absorption of, 84
activation, 18
binding energy of, 91, 93,97
fission, 17
intermediate, energy of, 85
kerma rate factors for, 101-102
thermal, energy of, 85
Neutrons per fission, 56
Nevada Test Site (NTS), 509
NIOBE, 133, 300, 486-490, 502-504,

548
Nonelastic neutron cross section, 93
SUBJECT INDEX
Nonelastic reactions, 93
Normalization
of distribution (unction, 211 - 212
of functions, 21
of source strength, 46 -47
NRN removal-diffusion method,
196,-198,567
Nuclear Aerospace Research Facility,
2-5,478
Nuclear Data Sheets, 94
Nuclear Development Associates, Inc., 5 -6
Nuclear emulsions, 475
Nuclear-powered aircraft, 2-5
Nuclear-powered submarine, 2--5
Nuclear propulsion, 11
Numerical integration techniques, 289--290
Oak
Ridg~
National Laboratory (ORNL),
178,256,319,482,492,509,514,591
early shielding program. 2--9
X-I0reactor,1---2
Oak Ridge Tower Shielding Facility, 4
05R system, 6, 256, 492
06R,256
1OF code, 494-496
Optimization, 462-465
conditions for a minimum, 463
gradient nonlinear programming, 465
linear perturbation theory, 464--465
of shadow shields, 465
shield synthesis, 463-465
with discrete-ordinates transport, 465
Outside Test Tank (OIT), 477-479
P n approximation, 131
Pair production, 74-81, 104

Paraffinized wood, 1
Particle absorptions in Monte Carlo
methods, 240~-241
Particle death, 208
Particle density, 26-27
differential, 28
Particle history, Q
Particle-reaction neutrons, energy
distribution, 18
Particle scoring in Monte Carlo
methods, 242--247
Pathfinder Atomic Power Plant,
574-579
calculation, 578
measurements, 579
plant layout, 576
Path length
biasing, 229
multiregion geometry, 226-228
sampling of, 225-234
SUBJECT INDEX
single infmite region, 226--228

track length. 225--234
Phantom, cylindrical, 616
Phantoms, for dose calculations, 95,
109-111
gamma-ray doses in, 111
neutron doses in. 109-110
Phase space, 27, 30, 124
in diffusion theory, 161
invariant imbedding, 163-164
Photoelectric effect, 68--70, 80. 104
Photon att~nuation coefficient, 80-83
Photon interactions, 68-73
See also Compton effect: Pair production:
Photoelectric effect
relative importance, 78, 81-83
Photoneutrons,18
threshold photon energies, 18
Photons
coherent scattering of, 80-81, 115
Physical dose, 94-95
Plane source, 44
angular distribution, 45
isotropic, 44
Plutonium--beryllium neutron source,
18
Point isotropic source, 626
Point kernel
Albert-Welton, 178, 187-188, 286-287,
299
applications, 6-7, 40, 288-298, 550,
578,588
See also Kernel technique
first-collision approximation, 294-296
geometric transformations, 291-294
last-collision approximation for cylinder,
297-298
last-collision approximation for slab,
289,296-298
from Monte Carlo calculations, 290-291
shield heating, 421-425
Point source, 39
Polar angle, 42
Polonium-beryllium neutron source, 18
Polyethykne, 453-454, 458, 460
in Savannah reactor, 581, 587
Portholes (see Ducts)
Positron emission, 16
Pratt and Whitney,S
Pressurized-water reactor, 579-595
Probability density function (PDF),
210-215,258
Probability of an event, 207-209, 258
distribution, 210
Proof testing of shield, 526-527
785
Pseudorandom numbers, 215, 663

Purdue University, 519
QAD,178
Quadrature schemes in discrete ordinates,
146
Quality factor (QF), 106-108, 110
Rad,97,104
Radiation
annihilation, 16
collided,119-120
fluorescent, 68
interactions of, 63-117
responses to, 94-112
scattered, 119-120
uncollided, 119-123, 292
unscattered, 119-123
Radiation detectors
See also Gamma-ray detectors; Neutron
detectors
interpretations of output, 476-477
in a mixed field, 476
types of, 471-498
Radiation heating (see Shield heating)
Radiation-heating criteria, 533
Radiation intensity, 26-39
Radiation production, 11-19
See also Gamma-ray sources; Neutron
sources
Radiation Shielding Information
Center (RSIC), 8, 477, 482
Radiation streaming in conceptual
design, 523
Radiation surveys, 527
Radiation transport, 119-201
geometric vs. material attenuation,
120-123
slab geometry, 121-122
Rand Corporation, 6
Random number, 214-215, 226, 233-235,
237,258
Random-number generators, 215, 663-673
middle-digit square, 666
mixed congruential, 666-667
multiplicative congruential, 667-668
recursion equations, 665-668
tests of, 668-671
Random-number sequences, 664--665,
670-671
independence in, 664
Random numbers
pathological sequence of, 665, 671-672
Random sampling, 209---:-215
for cylindrical source, 220
786
SUBJECT INDEX
for disk source, 219

from energy distribution, 218-219
ini.tial direction of source particle,
222
cosine source, 222
isotropic source, 221
of spatial position, 219
for spherical shell source, 220
of track length, 225-234
Random variable, 210, 258
Random walk, 208~209
RASH E removal-diffusion method,
192-194,562,567
Ray tracing, 550, 596
Rayleigh scattering, 80
RBE (see Relati.ve biological effectiveness)
Reaction-product gamma rays, 16
Reaction rates, 66-67,83,115
Reactors (see specific reactors and reactor
types)
Receiver concept, 11, 26
Reflection, 46
Rejection technique, 214-216, 258
Relative biological effectiveness (RBE), 94,
105
dose, 105--108
Rem (roentgen equivalent man), 106
Removal cross section, 3, 178-188,368,
492
additive property, 181
dependence on geometry, 179
for hydrogen-deficient shields, 181~ 187
macroscopic, 179, 183
microscopic, 179-180, 186
minimum thickness of hydrogen, 180
relation to transport cross section, 181
Removal--diffusion methods, 6,188-201,
560-562,567
ATTOW,198---201
differences in current methods,
199-201
MAC, 194-196, 199-201
NRN, 196, 201
RASH E, 192-194, 199-201
SABINE, 198-201
Spinney method, 190-192
Removal flux, 194, 196
Removal term of discrete ordinates
calculation, 140-142
Removal theory, 548, 578, 588, 596
difference between slabs and
homogeneous mixtures, 287-288
kernel application, 286-288
RENUPAK, 178, 300
Resilon in Dounreay reactor, 559-560
Resonance-region energy, 85
Responses to radiation, 94-112
Roentgen, 104
Roentgen equivalent man (Rem), 106
Russian roulette technique, 234
Sn method (see Discrete-ordinates

Sn method)
SABINE removal--diffusion method,
198
Sample space, 258
Sampling, systematic, 250
Savannah N.S., 579-595
attenuation calculations, 587-591
lead in, 581-587
main shielding, 580
maintenance shields, 581-583
polyethylene in, 453-454, 458, 460
secondary gamma flux, 589
shield design criteria, 583-586
Scalar current, 36
Scattered radiation, 119-120
Secondary gamma rays, 84, 91
in air,S
in dose considerations, 99
graphs for semi-infinite shield, 722-725
graphs for slab shields, 718-721
kernel technique, 301-309
measurements, 481
semi-infinite shield, 308, 722-725
slab shield, 304, 718-721
sources of, 14-15
Serpentine concrete in Fermi reactor, 541
Shado~ shield in SNAP-lOA, 598
Shield attenuation calculations, 261-309
interpolation and curve fitting, 265
types of distributions, 264
use of equilibrium distributions, 265
use of parametric data, 264-283
use of unit source, 264
Shield design, 519-606
iterations, 519-527
principles of, 519--520
Shield heating, 3, 419-430
buildup factor, 421-425
by charged particles, 429-430
in conceptual design, 523
energy-deposition methods, 419-430
by gamma rays, 420-429
by neutrons, 427-429
point kernel, 421-425
ratio of neutron to gamma-ray
heating, 427-429
in SNAP-lOA, 598
SUBJECT INDEX
Shield-material experiments, 477-500
Shield materials, 443-460
See also specific substance
capture-gamma-ray dose, 456, 460
comparis'on of gamma-ray attenuation,
456-457
comparison of neutron. attenuation, 456,
458
for Lenin, 452
for mobile reactors, 450-462
for Savannah N .S., 452
selection of, 445-447,521-522
shipping and storage, 452-453
for stationary reactor, 447-450
testing, 526
thermal problems, 452
used in test reactors, 448
Ship propulsion, 579-595
attenuation calculation, 587-591
main shielding, 580
maintenance shields, 581-583
secondary gamma flux, 589
Sievert's integral, 292
Simon-Clifford technique, 578
method for cylindrical ducts, 385-388
Single-collision dose (see First-collision dose)
Slab attenuation, 121-~123
Slab-penetration measurements, 477 -483
SNApIOA
flight test, 595-602
radiation limits, 596
shadow shield in, 598
shield analysis, 595~598
shield heating, 598
test results, 595-602
Solid angle, 24, 60
differential, 25
Solid-state detectors, 474
Solid-state gamma-ray spectrometers, 475
Source analysis, 261-263
discrete-ordinate Sn' 263
fission gamma rays, 262
fission, product gamma rays, 262
list of components, 262
low-energy neutrons, 262
Monte Carlo solution, 263
reactor codes vs. shielding codes, 263
Source term
in discrete ordinates, 143
in transport equation, 127
Space Nuclear Auxiliary Power, 595
See also SNAP-lOA
Space power, 595-602
See also SNAP-lOA
787
Spatial distribution, 39-42

line, 39-40
point, 39-40
surface, 39-41
volume, 39, 42-43
Spectra in water
gamma rays, 506-507
neutrons, 498
Spectrum, 49
See also Energy disttibutions
Spherical-harmonic functions, 625
Spherical harmonics method, 7, 129-132
accuracy limitations, 131-132
truncation number, 131
Spherical shell attenuation, 123
Spinney method, 190-192, 300, 562,
567
multigroup diffusion equations, 191
of removal-diffusion calculations,
190-192
slowing-down length, 191
variations, 192--201
Spiral duct, 2
Splitting, 234
Stainless steel
See also Iron
in Fermi reactor, 540-541
Standard deviation, 213
Standard-man composition, 111,614
Statistical estimation, 217,243-244
Statistical variance, 247-251
average or mean, 248-250
central limit theorem, 247-248
dispersion, 248
expected value, 248-250
Statistical weight, 208
Stochastic processes, 209
Steel shot in Fermi reactor, 539, 545
Steradian, 25
Streaming, 3
See also Ducts
in Agesta shield, 571
in Dounreay shield, 565-566
in Fermi shield, 550
Surface source, 40-41
Survivorship function (distribution), 21,60
TDe, 548, 562

Thermal-neutron-tlux-densityestimation
methods, 298-301
multigroup diffusion, 299-300
Spinney method, 300
transfusion, 300-302
Thermal neutrons
energy of, 85
788
SUBJECT INDEX
u.
Maxwell-Boltzmann distribution of, 85

Thcrmoluminescent detectors (TLD), 476
Thomson nuclear scattering, 80
Thomson units (T.U.), 72
Time-of-flight measurement,
494~502
Titanium hydride, 454, 456

Total energy-flux density, 31--32
Total flux density, 31-32
Total nct current, 35
Tower Shielding Facility, 322, 340, 514
Track length, 29
See also Flux density
detector for, 474
per unit volume, 29, 30
Transfer kernel, 217
Transport 0q uation
Boltzmann (see Boltzmann transport equation)
jnscatter term in, 125, 127
interaction loss term in, 127
nct leakage term in, 125-127
source term in, 127
Transport kernel, 217
Tungsten, 453-454, 457,460,462
Uncollided radiation, 119-123, 292
United Kingdom early shielding
program, 3-4, 6
University of lllinois, 7
Unscattercd radiation, 119~-123
Uranium, depleted, 454,-455, 457
S. Army Ballistic Research Laboratories,
319
Value functions, 218
Variance, 248-249
of function, 213
statistical,247-251
Vector derivative, 620
Vector operators, 619-620
Vents (sec Ducts)
Voids, 404-414
See also Ducts
disk-shaped, 404-409
ray analysis, 409
small random, 412
spherical, in water, 411
Volterra integral equation, 152,154
Volume source, 41-42
Water
gamma-ray and neutron spectra in,
498,506-507
measurements in, 448, 482
transport in, 498-503
Wigner effect, 533
Westinghouse Bettis Laboratory,S
Windscale reactors, 3
X-I0 reactor, 2, 492
Zirconium hydride measurements, 478
r-------------NOTICE - - - - - - - - - - - ,
This book was prepared under the sponsorship of the United States Government.
Neither the United States nor the United States Atomic Energy Commission, nor
any of their employees, nor any of their contractors, subcontractors, or their
employees, makes any warranty, express or implied, or assumes any legal liability
or responsibility for the accuracy, completeness or usefulness of any information,
apparatus, product or process disclosed, or represents that its use would not
infringe privately owned rights.
<oU.S, GOVERNMENT PRINTING OFFiCE: 1981740-145/2740

Reactor Shielding For Nuclear Engineers

Uploaded by

Copyright:

Available Formats

Reactor Shielding For Nuclear Engineers

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Reactor Shielding For Nuclear Engineers

Uploaded by

Copyright:

Available Formats

Available as TID25951 for $28.

Springfield, Virginia 22161

shielding directed at an advanced undergraduate or graduate level. The reader

weapons shielding and in isotope source applicati<.Hls. The study of space

2.1.1 Gamma-Ray Sources

2.2 Basic Mathematical and Physical Concepts

2.3 Spatial and Directional Characteristics

2.4 Energy Distributions

2.4.1 Energy Distributions of Gamma-Ray Sources

3 INTERACTIONS OF RADIATION WITH MATTER

3.1.1 Microscopic Cross Section

3.2 Radiation Interactions .

3.3 Responses to Radiation

Transport Equation and PhaseSpace Geometry

4.5 Moments Method

4.7 Invariant Imbedding Method

4.9 Combination Removal-Diffusion Methods

5 MONTE CARLO METHODS FOR RADIATION TRANSPORT

Selection from an Energy Distribution

5.4 Path Length . . . . . . . . .

Neutron Elastic Scattering

6 SIllELD ATTENUATION CALCULATIONS

Moments-Method Differential Energy Spectra

6.4 Simplified Solutions

Applications of Gamma-Ray Buildup Factors

6.5 Application of Kernel Technique to Calculations of

7 ALBEDOS, DUCTS, AND VOIDS

7.3 Neutron Albedos

7.9 WallPenetration Component

7.10 Wall-Scattered Component

Analog Monte Carlo Calculations

7.11.1 Single Voids

8 SHIELD HEATING, AIR TRANSPORT, SHIELD MATERIALS,

8.2 Air Transport

8.2.1 Infinite Air Medium. .

8.3 Shield Materials

8.4 Shield Optimization

Considerations in Materials Selection

9.2 Shield-Material Measurements

9.3 Phenomenological Measurements

Air-Transport and Air-Ground Interface Experiments

10.2 Fast Breeder: Enrico Fermi

10.3 Fast Breeder: Dounreay Fast Reactor

Heavy Water, Natural Uranium: Agesta

10.5 Boiling Water: Pathfinder

10.6 Ship Propulsion: N.S. Savannah

Description of Ship, Reactor, and Main Shielding

10.7 Space Powe" SNAP-lOA Flight Test

10.7.1 Shield Analysis. .

GAMMA RAYS FROM INELASTIC NEUTRON SCATTERING

Appendix B: NEUTRON FLUENCETOKERMA CONVERSION

CYLINDRICAL TISSUE PHANTOM

COORDINATE SYSTEMS. VECTOR OPERATIONS,

COEFFICIENTS FOR GAMMA-RAY BUILDUP FACTORS

Appendix G: GRAPHS AND FORMULAS OF EXPONENTIAL

Appendix H: TABLES OF ATTENUATION FUNCTIONS

Appendix I: RANDOMNUMBER GENERATORS

1.3.1 Global Tests . .