796 J. SPACECRAFT VOL. 9, NO.

11

Stress Corrosion Cracking and Hydrogen Embrittlement of

High-Strength Fasteners
JAMES K. STANLEY*
Aerospace Corporation, El Segundo, Calif.

Unexpected, brittle failures of high strength fasteners on aerospace vehicles have been caused by stress
corrosion cracking (SCC) and by hydrogen stress cracking (HSC). Confusion exists as to the nature of each
phenomenon. The poorly understood failure mechanisms are difficult to differentiate, especially in the field.
There is a growing acceptance of the term SCC to cover failures by both mechanisms. Data are given to
characterize the classes. For low-alloy carbon steels, heat treated to yield strengths below approximately
160 ksi, stress corrosion is not a problem, nor is hydrogen embrittlement (delayed cracking) very common.
Above this stress difficulties can occur. The high-strength precipitation hardening stainless steels have varying
degrees of resistance to stress corrosion cracking and hydrogen embrittlement, depending upon strength level
and heat treating procedures that influence the microstructure. Use of plane strain fracture toughness KIC
and the stress corrosion threshold of KISCC offers promise of selecting optimum bolting for a specific environ-
ment. The attractiveness of plane strain fracture toughness analysis is that it does not differentiate between
failure mechanisms; failure can be either SCC or HSC.

Introduction There are other types of fastener failures, e.g., overtorquing

and stress-rupture, but failures as a result of stress-corrosion

T HE design engineer, alone or in consultation with metal-

lurgical personnel, may or may not be fully cognizant of
fastener systems that could result in brittle failures as a result
and/or hydrogen embrittlement are most insidious. Because of
a lack of a standardized test method for SCC and for HSC,
statistical analyses of data are not feasible. The American
of stress corrosion cracking (SCC) or hydrogen stress cracking Society for Testing and Materials is actively working in both
(HSC).f The failure of high-strength fasteners in aerospace areas to develop standardized tests for both phenomena, t
equipment as a result of SCC and/or HSC occurs infrequently SCC and HSC, as two fracture mechanisms responsible
and happens unexpectedly. There is a growing acceptance for delayed failures, have caused many serious and unexpected
of the term SCC to cover both mechanisms. failures in high-strength fasteners. Failures have occurred
The incidence of fastener failure has been decreasing with in applications at stresses that were considered safe from
time because engineers have learned to avoid materials that stress (below the Y.S.) analyses even using generous factors of
have given difficulties in the past. Often, too, a difficulty safety. These occurrences have lead designers to use materi-
brings together a roomful of experts who generally come up als far from their true capability either by using less than
with a successful solution. Solutions are not always as optimum strengths of a high-strength steel or using steels heat
simple as a material's substitution, but may be complex, treated to the maximum strength, but using the steels at very
involving material changes, use of coatings, redesign, or low-strength levels, say 25 % of the yield strength.
reduction of stress, sometimes singly but often changing many As the strength level increases above about 160 ksi, both
factors simultaneously.1 the sensitivity to brittle fracture and the susceptibility to SCC
In aerospace hardware, fatigue of the fasteners has not been and HSC increases. Although steels with strengths in excess
a problem mainly because acceptance testing time and flight of 300 ksi are available, designers are reluctant to push for
times have been too short (less than 1 hr) to develop failures this strength level and have settled on levels of 200-260 ksi.
with the stress loadings. However, on the space shuttle that Some designers who have had the unfortunate experience of
aims for 100 flights fatigue of fasteners would become a either SCC or HSC have even backed off to strengths of
problem to consider. As higher stressed parts are used more 160-180 ksi.
attention must be given to steel cleanliness (inclusions act as Shotpeening, plating, and painting of low-alloy, high-
crack nuclei), to radiused threads (to reduce notch effect), strength martensitic fasteners as a means of preventing de-
and to larger head shank fillets (also to reduce notch effect). layed failures at ambient temperatures has been largely
High-strength fasteners should have threads and head shank unsatisfactory. Current interest lies in use of HSC or SCC
fillets rolled on after heat treatment and shanks should be resistant stainless types and superalloys (nickel-base and
ground and polished. Rolling threads after heat treatment cobalt-base).
has two benefits. It builds up a residual stress to counteract
in part a portion of the applied tensile load and it ensures
unbroken grain flow lines through the critical area created by Failures due to Stress Corrosion Cracking or
the notch effect of the threads. These considerations have Hydrogen Stress Cracking on the Titan III
been helpful in increasing fatigue life; they have also been Family of Vehicles
helpful in improving resistance to SCC.

Presented as Paper 72-385 at the AIAA/ASME/SAE 13th The Air Force in their Titan III program has had difficul-
Structures, Structural Dynamics, and Materials Conference, San ties with high-strength fasteners on the boosters over the past
Antonio, Texas. April 10-12, 1972; submitted April 20, 1972; six or seven years. Table 1 lists some of the fastener steels
revision received June 26, 1972.
Index category: Launch Vehicle and Missile Fabrication. I Standard test procedures are being developed in the following
* Manager, Applied Metallurgy. Member AIAA. areas: G.01.06 SCC and corrosion fatigue; G01.06.01 smooth test
t The term, hydrogen stress cracking, is also synonymous with specimens; G01.06.02 environments and materials; G01.06.04
hydrogen embrittlement, hydrogen cracking, and hydrogen in- precracked growth; E24.04 subcritical crack growth; and F7.01
duced delayed cracking. hydrogen embrittlement.
NOVEMBER 1972 CRACKING AND EMBRITTLEMENT OF HIGH-STRENGTH FASTENERS 797

Table 1 Fastener failures on the Titan III family of vehicles

Material Hardness, UTS and/or Application Probable failure mode

heat treatment
15-7 Mo RC46 Shear tie bolt between solid Stress corrosion cracking
core and liquid core
15-5 PH 180-190 ksi Solid rocket motor frangible Hydrogen stress cracking
bolt from galvanic coupling
17-4 PH H950 Fuel and oxidizer valve Hydrogen stress cracking
assemblies from galvanic coupling
440C RC51 Actuator adjustable bolt Stress corrosion cracking
H-ll Rc 52, 260 ksi Launch pad baseplate stems Hydrogen stress cracking
from galvanic coupling
431 180 ksi Marman clamp on hot gas Stress corrosion cracking
cooler Hydrogen stress cracking
Unitemp 212 180 ksi Solid rocket motor (?)
17-4 PH RC46; 1 hr 900°F Pressure valve Stress corrosion cracking
431 180 ksi Solid rocket motor Hydrogen stress cracking

and mode of failure. Failures were all in a marine atmos- The SCC occurs in specific enivronments and with environ-
phere. The precipitation hardening steels have all been mentally sensitive metals. Failures generally occur from
slowly replaced by the cold worked type of A286. The inadequate knowledge of environmental conditions. In most
440C and H-ll were continued in service but their heat cases, there is negligible loss of metal by general corrosion,
treatments were modified and/or protection by organic and at times the corrosion is imperceptible to the eye. Stress
coatings became a requirement. Type 212 was eliminated corrosion cracking requires highly anodic areas and a local-
but Type 431 was continued in service with organic coatings; ized pH, such as may exist in oxide film cracks, pits, crevices
long range solutions involved substitution of A286. and cold worked areas.
These failures occurred during the early years of T-III Hydrogen stress cracking is a phenomenon which occurs
development despite a program of stress corrosion control because of hydrogen penetration into the lattice in the pres-
having been evolved over this period. Tensile stresses (pre- ence of a tensile stress. It is generally agreed that corrosion
loads) on the fasteners are now minimized to 40% of yield, plays no direct role in this mechanism. However, corrosion
where possible, and materials are sought and heat treated often plays an indirect role as being the source of hydrogen.
where possible to an ultimate tensile strength of 160 ksi. The In classical HSC, the hydrogen goes into solid solution
importance of stress level, environment, and metallurgical being introduced by electrolytic charging, pickling, heat treat-
structure of the metal in SCC and/or HSC is well recognized ment, and by corrosion reactions. Hydrogen then causes
by the Program Office. Contact with dissimilar metals is delayed failure under static load in high-strength alloys and
recognized as the most likely source of hydrogen by corrosion the embrittling effect increases with increase in severity of
and is avoided or protected against. Chemical conversion notch, i.e., stress concentration.
coatings and anodizing on aluminum often retards such The amount of hydrogen which will cause HSC is exceed-
corrosion when high-strength fasteners are used in conjunc- ingly small and is of the order of 4 or 5 ppm. Damage has
tion with aluminum. been reported with hydrogen contents of even less than 1 ppm.
NASA has had a few failures by corrosion in 4330, 4340, The delayed failure characteristics of 4340 steel, § exposed
AM 355, and 17-7 PH.2 NASA plans to initiate studies of to distilled water and failed presumably by HSC, are shown
service influence on fracture behavior, i.e., use of fracture in Fig. 1. Similar failure characteristics are caused by SCC.
mechanics concepts. Note in Fig. 1 the short time-to-failure of the 4340 as the
ultimate tensile strength (UTS) of the steel increases and the
threshold stresses below which no failure occurs.3
Definition and Differentiation of Stress Corrosion
Cracking and Hydrogen Stress Cracking

Despite the effort expended, theories proposed, and data

collected, neither fracture mode is well understood. Some
of the difficulty arises because the two mechanisms are so
much alike. It is only in the laboratory, by electrochemical
means, that the two mechanisms can be uniquely differenti-
ated. In the field, it is difficult, if not impossible, to identify ^-TEMPERED AT 600°F TO A
which cracking phenomenon was responsible for the failure. 240ksi STRENGTH LEVEL
However, despite many similarities, the basic mechanisms are
different. Stress corrosion theory is not sufficiently ad-
vanced to predict failure times; this is also true of HSC. A ^-TEMPERED AT 400°F TO A
285ksi STRENGTH LEVEL
unifying mechanism of SCC is sorely needed.
Stress corrosion cracking is the phenomenon in which a 10 I0 2 I0 3
crack nucleates in a corrosive environment of a susceptible TIME TO FAILURE, min

metal while the metal is stressed in tension; the crack then

propagates by stress-induced corrosion of the advancing Fig. 1 Delayed failure characteristics of 4340 steel at various
crack tip. Cracking may occur intergranularly or trans- strength levels when exposed to distilled water.
granularly depending on the metal and its heat treatment.
Failure occurs with little or no plastic deformation; hence
the fractures are termed brittle failures. The stresses are § Nominal compositions of all alloys mentioned in this report
generally below the yield stress. are given in appendix.
798 J. K. STANLEY J. SPACECRAFT

The important parameters in these delayed failures are

strength levels, steel composition and metallurgical structure [TESTS CONDUCTED IN 31/2% NOCL I
(microstructure), tensile stresses, environment (i.e., tendency [LOADED TO 0.9 OF NOTCHED UTS
— ANODIC CONDITION
for corrosion or introduction of hydrogen) and time. Tem- (METAL DISSOLUTION)-
perature seems to increase the likelihood of stress corrosion '-4PH AGED
1hr900F(H-900)
cracking more so than HSC; the latter occurs around room
temperature. The tensile stresses can be applied stresses,
residual stresses (as in roll formed threads, heat treatment, — I7-4PH AGED
ihr 950F(H-950)
welding, straightening, cold rolling) or a sum of these stresses.
If residual stresses are present, they are very difficult to
measure or estimate so that one does not know of what : -1.2
magnitude they are in a structure. For either SCC or HSC, 1
-CATHODIC CONDITION
-1.6 HYDROGEN GENERATED
there exists a stress threshold limit below which the stress
will not cause fracture. 1
-2.0
I
In the laboratory, the electrochemical behavior of the metal
TIME TO FAILURE, min
offers perhaps the best arguments that SCC and HSC are
separate, distinct phenomenon. Delayed cracking that
Fig. 2 Curves illustrate the type of data obtained with potentio-
occurs under cathodic polarization (hydrogen generation) static tests.
can hardly be attributed to SCC. Conversely, delayed
fracturing in situations where anodic polarization is causing
dissolution (corrosion) of the metal can hardly be identified In environmental cracking failures at the launch ranges, it
as HSC. The polarization vs time-to-failure curves identify is not possible to separate unequivocally HSC from SCC.
the two mechanisms uniquely. Obviously, such methods When there is a question as to whether it is one or the other
have limited field use. There are many test procedures that or possibly a combination, the failures have been referred to
are used to evaluate susceptibility to SCC and HSC, but as being the result of 1) stress-corrosion cracking (which
only a few of the common ones are mentioned. would include SCC and HSC), or 2) environmental stress
A common SCC test to establish susceptibility of bar stock cracking. For engineering purposes, this may be sufficient.
involves loading the specimen to some high percentage of its Pinpointing the actual failure mechanism may be of academic
yield stress, or sometimes UTS, and exposing it to alternate interest only, at least to an engineer.
immersion in a 3.5% NaCl solution for 10 min followed by The appearance of the fracture surface is similar, if not
50 min of drying in forced air. The cycling is continued identical, for stress corrosion and hydrogen failures and,
until the specimen fails or the test is discontinued. The hence, the failure modes can't be distinguished from each
sodium chloride is usually dissolved in distilled water or is other. This situation leads some to believe that the mechan-
acidified (say to a pH of 1.5). Imposed stresses generally isms are identical, whereas to others it just means that frac-
vary from 75 % to 90 % of the fatigue cracked notched tensile tures are similar. Electron fractography with transmission
strength and from 75% to 90% of their 0.2% yield strengths electron microscope or the newer scanning electron micro-
when unnotched. It is quite common to check austenitic scope cannot uniquely identify HSC. Electron microscopy
steels for SCC by exposing them to 42% boiling aqueous is a very useful tool when used in conjunction with other
Mg C12 (154C, 309F) solution; ferritic stainless steels, by tools. A knowledge of circumstances leading to failure is a
contrast, are relatively immune to cracking in Mg C12. valuable adjunct to successful failure analyses. The com-
Until recently, SCC tests were conducted on smooth pilations of electron fractographs by Air Force Materials
specimens. They are helpful in selection of materials for Laboratory (AFML) are helpful in deciding what the nature
environments or in development of coatings. They could of the failure mode may be.5
not be used, however, for establishing safe design loads. The occurence of intergranular and transgranular cracking
Often there is large scatter in these data. The SCC tests on or both is the result of hydrogen, a susceptible microstructure,
both smooth and precracked specimens used in facture the specimen geometry, and static or dynamic tensile loading.
toughness studies provide the designer with tools for material Fidelle, Legrand, and Couderc6 reviewing 39 studies involving
selection on bases of service environments that can be simu- fractography of materials failed by HSC and SCC found 23
lated in the laboratory. The testing methods as mentioned instances of fully or predominantly intergranular cracks,
need standardization. 8 fully or predominantly transgranular, 8 mixed, i.e., both
For the HSC test to establish susceptibility, a notched types. Intergranular cracking appears to be the most fre-
specimen is loaded to 75 % of its notched yield strength and quent type in both HSC and SCC.
hydrogen is cathodically charged into the steel while under
load. Sometimes a notched specimen is plated with electro-
lytic cadmium and then loaded. A notched specimen pro- Materials for High-Strength Fasteners
vides more susceptibility to HSC than an unnotched one.
Cathodically charging a steel with hydrogen provides con- The aerospace industry has emphasized the systems con-
siderably more hydrogen than is necessary to produce a cept of fastener usage for reliability. Design has been
delayed failure. Loadings are similar to those used for SCC improved so that stronger, lighter and more sophisticated
tests. fastener systems are available. Lately there has been thought
High-strength bolts must also pass a stress-durability given to environmentally resistant fasteners.
(static fatigue) test. The bolts are stressed to 72% of the Steel composition is important, as is the metallurgical
minimum tensile load and are held for 96 hr. If embrittled structure. The time for initiation of the first crack in either
by hydrogen they will fail. HSC or SCC is a reflection of the microstructure resistance
The potentiostatic procedure is becoming more popular to hydrogen diffusion or pit corrosion. Much is known
in the study of both types of phenomena. Figure 2 shows empirically about what compositions and/or structures are
how 13-8PH steel can be made to fail by either phenomenon susceptible to HSC or SCC, but little is known from so-called
by changing the impressed potential.4 This type of test can first principles about the relation of microstructure to crack
probably be standardized to develop the relative susceptibility initiation and propagation.7 Most data are basically phenom-
of a given steel in various heat treated conditions to either SCC enological.
or HSC. The test is definitely a laboratory tool and would Although much ordinarily is not said about nuts and
not yield design data, only guide lines. washers, they too must be carefully selected for the fastener
NOVEMBER 1972 CRACKING AND EMBRITTLEMENT OF HIGH-STRENGTH FASTENERS 799

system. Careful selection means that the nuts are also tested
in a SCC test; the hut is torqued on a bolt and the assembly STAINLESS STEELS
STRENGTH RANGE
ULTIMATE TENSILE STRENGTH, ksi
alternately submerged in salt water. Washers are used to AND
MEDIUM h————HIGH———H ULTRA-HIGH
SUPERALLOYS i i 150 200 250 300
distribute the load. They too should be compatible corro-
sion-wise. Being loaded in compression they do not fail by
HSC or SCC.
Materials for aerospace fasteners are categorized in sevaral
ways: 1) by grouping strength levels, 2) by composition and
metallographic structure, and 3) by relative susceptibility
to SCC and/or HSC. In the future, the steels will be cate-
gorized by fracture toughness parameters; some such compil-
ations are already appearing.

Classification by Strength Level

One arbitrary classification of threaded fasteners is by
ultimate strength range: low strength—up to 125 ksi; medium
strength—125-160; high strength—180-260; ultra high
strength—280 and up. Figure 3 gives the strength ranges
for some popular alloy steel fasteners. Figure 4 gives the Fig. 4 Classification of stainless steels and superalloys by strength
strength ranges for some stainless steels and superalloys. range. Dashed lines indicate strengths obtainable by coldwork and
aging.

Classification by Composition and Metallurgical

Stainless Steels (Austenitic and Ferritic)
Structure
The austenitic steels of the Type 300 series are much too
Another classification by type of alloy is useful in under- weakU to be considered for high strength fasteners. Only
standing the behavior of these materials to SCC or HSC; through cold work and stress relief can high strengths be
these are: martensitic, stainless steels (austenitic and ferritic), developed in these materials; half-hard material will have an
precipitation hardening stainless steels, (semiaustenitic and UTS of 150 ksi while full-hard will have one of 185 ksi.
martensitic), superalloys (nickel-base and cobalt-base). Even though the annealed materials have low strength levels,
Experience has shown that some types of high-strength the alloys can be fractured in environments containing chlor-
steels are quite susceptible to SCC and HSC while others are ides. Pitting on these steels can occur at emergent slip lines,
not. Some change in degree of susceptibility to SCC results as well as at inclusions.7 The annealed materials are very
from the objectionable precipitation of grain boundary resistant to HSC.
carbides and presence of secondary phases. Precipitation hardening austenitic nickel chromium are also
Steels used at or above 160 ksi UTS should meet AMS- available. Alloy A286 is one of the first and is now one of
2300A cleanliness levels. Steels should be vacuum or con- the most popular high strength stainless steels. If the alloy
sumable arc melted to minimize inclusions which can cause is cold worked (60%) and aged (1200°F), strengths of over
pitting. Mechanical properties sometimes can be higher 200 ksi are attainable. Alloy A286 has outstanding resistance
and more uniform, with greater fracture toughness if vacuum to SCC and HSC. In poteniostatic experiments it has not
melted. been possible to fracture the alloy under anodic or cathodic
condition. Field experience confirms the alloy high resist-
Martensitic Steels ance to these phenomena.4
The ferritic stainless steels, Type 400 (e.g., 410,416,420, 422,
For carbon and low-alloy steels heat treated to strength levels
and 431) are heat treatable to strengths of around 200 ksi.
of about 160 ksi or less, SCC is not a problem nor is HSC
Although these steels are martensitic in structure, they are not
common. The common martensitic steels are shown in
Fig. 3. generally considered in the martensitic class inasmuch as they
have relatively high chromium contents, i.e., about 13%.
The maraging steels can develop strengths over 300,000 psi
These types have good general corrosion resistance but they
but so far they have had limited application at this level but
are susceptible to SCC. This susceptibility can be removed
at lower strengths are usable alloys. These steels are cooled
by tempering at 1100°F or higher, but high strength is sacri-
from the austenitizing temperature forming a soft and weak
ficed.
martensite; aging at 850-950°F hardens and strengthens it.
Precipitation Hardening Stainless Steels
STRENGTH RANGE
ULTIMATE TENSILE STRENGTH, ksi These precipitation hardening stainless steels beside
STEELS MEDIUM I-———— HIGH —~"° "-[ULTRA-HIGH
|(
)0 150 200 250 3C imparting excellent general corrosion resistance to the alloy
. _ ._!. ........ | I
4130 i————————————————i can attain rather high strengths. The objective of obtaining
a high-strength, unplated, corrosion resistant alloy appeared
to be attainable with the introduction of the precipitation
8740 n
HY-TUF hardening stainless-steel grades. This objective was to
D6AC circumvent problems such as HSC and/or SCC that had been
experienced with plated and unplated martensitic, low- and
MARAGING-200 high-alloy steels in the 200 to 250 ksi range.
MARAGING-250 i——i These precipitation hardening steels have varying degrees
MARAGING-300 m=] of resistance to SCC and HSC depending, of course, on

Fig. 3 Classification of low-alloy steels (martensitic types) by H MIL-Handbook-5A assigns Types 301, 302, 303, 304, 316, 321,
strength range. The wide range of some of the steels is obtained by and 347 yield strengths of 30 ksi and ultimate tensile strengths of
tempering at different temperatures. 75 ksi.
800 J. K. STANLEY J. SPACECRAFT

strength level but also on the temperature of the aging treat- Classification by Relative Resistance to Stress
ment. Aging below 1000°F may cause these steels to be Corrosion Cracking
susceptible to SCC and HSC, but aging them above this
temperature the steels have excellent resistance to these A third classification involves rating SCC according to
cracking phenomena. relative susceptibility and can be used as a rough guide to
These precipitation-hardening types are of two basic material selection based on experience and some laboratory
classes: martensitic and semiaustenitic. The martensitic work. This classification comprises: 1) alloys and heat
types are 17-4 PH, 15-5 PH, and PH 13-8 Mo. In these treatments that can be used without restriction; 2) alloys and
alloys the martensitic structure forms on cooling from a heat treatments that can be used if used with caution; 3) alloys
solution treatment; subsequent aging between 900°F and and heat treatments that should not be used.
1150°F strengthens the martensite by precipitation hardening This information has been abstracted and expanded, on
and tempering to realize a gamut of strength values. Typical the basis of other aerospace findings, and is given in tabular
semiaustenitic steels are 17-7 PH, PH 15-7 Mo, PH 14-8 Mo form, Tables 2-4. The tables apply only to SCC in environ-
and AM 350. In these alloys, the composition has been ments of sodium chloride solutions, salt sprays, alternate
adjusted so that the austenite forms on solution treating and immersion (wetting and drying) and marine atmospheres,
is retained at room temperature. In this condition it is and are primarily for smooth specimens. Similar tables for
readily fabricated (cold worked). The hardening is obtained HSC are not available.
by reheating the austenite to 1400°F or 1750°F (called con- These ratings are not to be construed as particularly exact
ditioning), cooling, and finally aging at 950°F or 1050°F. because no attempt has been made to evaluate the effect of
If either of these steels is overaged, i.e., beyond highest stress, environment, metallographic structure and time. This
strength, both the fracture toughness and SCC resistance are type of tabular information represents the type of SCC data
improved. The fracture toughness and SCC resistance of that was available before the advent of fracture mechanics.
martenistic steels are significantly higher than the semiausten- No attempt, incidentally, has been made here to incorporate
itic types. data obtained by fracture mechanics and given elsewhere.
Other environments that are of concern to the aerospace
Superalloys (Nickel-Base and Cobalt-Base) materials engineer are the metal-propellant compatibilities.
Cracking with some of the steels has also been reported in
In general, as the nickel content increases in the austenitic phosphate, sulfate, nitrate, and sulfide solutions. None of
steels, the more they are resistant to SCC and to HSC. Some these are of concern here.
high-strength stainless fasteners are also made from super-
alloys (i.e., high-strength nickel-base and cobalt-base alloys).
The nickel-base Inco 718 superalloy, cold worked and precipi- Promise of Fracture Toughness Criteria
tation hardened will have strengths in excess of 200,000 psi.
The new and highly alloyed MP-35N exhibits the corrosion Improvement of methods of analyzing fracture in various
resistance of the best nickel-base alloys. It has a high environments through the use of fracture mechanics allows
resistance to SCC and to HSC in salt environments and the fracture process (whether by SCC or HSC) to be studied
marine atmospheres. on a macroscopic basis independent of the influence of

Table 2 Materials with a high resistance to stress corrosion cracking

Materials Type Heat treatment" Remarks

300 series stainless Austenitic Annealing Stressed material can crack in
types 303, 304, 316, chloride solutions. Annealed
321, 347 materials are not of high strength.
Cold worked materials can develop
high strength but they must be
stress relieved.
17-4 PH Martensitic HI 000 and above
17-7 PH Semiaustenitic CH900 Strength is developed by cold work
(60%) and aging (900°F)
PH 13-8 Mo Martensitic H1000 and above
15-5 PH Martensitic HI 000 and above
PH 15-7 Mo Semiaustenitic CH900 Strength is developed by cold work
(60%) and aging (900°F)
PH 14-8 Mo Semiaustenitic CH900 As for PH 15-7 Mo above
AM 350 Semiaustenitic SCT1000 and above
AM 355 Semiaustenitic SCT1000 and above
Custom-455 Semiaustenitic HI 000 and above
A 286 Austenitic Solution treated and aged
A 286 Austenitic Solution treated and aged High strength is developed by cold
(CW and Aged) work (60%) and aging (1200°F)
Inconel 718 Face centered cubic Solution treated and aged
Inconel X-750 Face centered cubic Solution treated and aged
Rene-41 Face centered cubic Solution treated and aged
MP35N Face centered cubic Solution treated and aged Solution annealed and cold worked
60% and aged
Waspaloy Face centered Solution treated and aged
Low Alloy Steels Martensitic Quenched and High resistance to SCC if tempered
4130,4140,4340, tempered to attain strength of 160 ksi or
8740 lower
Maraging Steel Martensitic Solution treated and aged High resistance if heat treated to
200 ksi or lower
a
For heat treatments refer to Aerospace Structural Metals Handbook, Metals Handbook ASM, or steel producer's literature.
NOVEMBER 1972 CRACKING AND EMBRITTLEMENT OF HIGH-STRENGTH FASTENERS 801

Table 3 Materials with a high resistance to stress corrosion cracking if used with caution

Materials Type Heat treatment Remarks

Low alloy steels Martensitic Quenched and tempered Good resistance to SCC if tempered
4130, 4140, 4340, to approximately 160-180 ksi
HY-TUF
8740, D6AC,
Maraging steel Martenstic Solution treated and aged All three grades ; Maraging-200,
-250 and -300
400 series stainless Martensitic Quenched and tempered Not susceptible if tempered at
410, 416, 422, 431 1100°F or higher
15-5 PH Martensitic H950-H1000
PH 13-8 Mo Martensitic H950-H1000
17-4 PH Martensitic H950-H1000
AM 355 Semiaustenitic SCT 950-H1000

Table 4 Materials with a low resistance to stress corrosion cracking

Materials Type Heat treatment Remarks

Low alloy steels Martensitic Quenched and tempered Very susceptible to SCC if tempered
4130, 4140,4340, to attain strengths of 1 80 ksi and
8740, D6AC, higher
HY-TUF
H-ll Martensitic Quenched and tempered
17-7 PH Semiaustenitic All heat treatments
except CH 900
PH 15-7 Mo Semiaustenitic All heat treatments
except CH 900
AM 355 Semiaustenitic Heat treatments below
SCT 900
400 Series stainless Martensitic Quenched and Very susceptible in the secondary
410, 416, 422, 431 tempered hardening range, 500°-1000°F

specimen geometry effects and dependent only on stress level where degradation occurs there is a threshold stress below
and environment. The fracture toughness approach gives which no SCC or HSC occurs, see Fig. 5. Note how severely
for the first time a quantitative knowledge of the effects of a the salt environment reduces the stress intensity factor Kt
particular environment on a steel stressed below the yield (Ref. 9). Distilled water, as well as moisture, is seldom con-
stress. Such quantitative data will be required by the de- sidered an aggressive environment. Yet moisture can have
signer once he learns how to use it. The appeal of the method, a controlling influence on the fracture behavior of high-
aside from being quantitative, is that it reflects a metals strength steels.10
behavior in an environment which may lead to either SCC or In tough alloys failure occurs at longer time intervals.
HSC. More importantly, the method does not differentiate Crack propagation may be slower and the alloy may tolerate
between mechanisms leading to failure. a longer crack before fracturing. A small crack may cause
The Kiscc parameter, hence, indicates with good repro- SCC in a material of low toughness whereas a larger crack
ducibility, the stress-crack-size threshold below which sub- may be required to fail a tougher material. The tougher
critical cracks will not propagate to a critical size leading to material may require a longer time to fail because crack
catastrophic failure in a gaseous, liquid or complex environ- growth is slower in a tougher material and not because SCC
ments in a period of usually 500-1000 hr. Both KIC and KISCC or HSC is slower. Failure time then becomes a measure of
have units of ksi(in.)1/2. the rate of crack growth.
To overcome disadvantages with smooth specimens used By relating the environmental applied stress intensity
previously, where long times may be involved for crack factor Kiscc to the plane strain fracture toughness KJC the
nucleation, Brown and others have used specimens with pre- differences in toughness of alloys or their heat treatments
existing cracks. Such specimens eliminate crack initiation
periods when surface films break down and pitting starts,
removes possibility of drawing erroneous conclusions that
4340 STEEL
alloys are immune to SCC, and specimens with pre-existing HEAT TREATED TO 225 ksi UTS
cracks permit use of fracture mechanics concepts. Brown8 PRECRACKED SPECIMEN EXPOSED
TO 3 1/2% NoCl SOLUTION
introduced the concept of the threshold Kiscc. Very quickly
CANTILEVER BEAM LOADING
Brown's idea became popular and many investigations have
shown the value of this approach.
The use of plane strain** fracture toughness criteria, i.e.,
KIC and the KISCC, offer promise of selecting fastener materials
that are not susceptible to either SCC or HSC on the launch THRESHOLD
pad or in any other environment. The fracture mechanics
approach can show if a metal is affected by the stress-envir- , NO FAILURE
onment and to what degree. The analysis also shows that io2 f io^ i o 4 i o 5

TIME TO FAILURE, min

** Plane strain conditions refer to the square fracture produced
by SCC or HSC. Plane stress conditions would involve slant or Fig. 5 Typical behavior of a high strength 4340 in a corrosive envi-
shear fractures; these are not observed with SCC or HSC ordin- ronment. Note how severely the fracture toughness, KIC, is reduced
arily. in the salt solution.
802 J. K. STANLEY J. SPACECRAFT

Table 5 Fracture toughness rating of alloys and heat treatments on their resistance to salt water environments12

Ultimate Seacoast test Accelerated test

Material Heat treatment tensile strength,
ksi KISCC KIC KISCC KIC

Inconel 718fl 1950°F AC, 8 hr 189.6 106 0.87 130 0.98

1350° + FCto
1200°F for 24 hr
17-4 PHa H1150 151.6 93.9 0.77 110 0.89
AISI 304 Annealed 84.0 53.5 0.77 59.7 0.86
4340 800°F temp. 204.8 48.3 0.72 29.7 0.44
17-4 PH H900 202.4 38.5 0.69 40.3 0.72
H-ll (AM)" 1100°Ftemp. 282.6 39.5 0.62 23.2 0.24
410a 1125°F temp. 128.8 52.4 0.55 49.6 0.52
H-ll (AM) 1000°F temp. 300.3 16.7 0.52 8.6 0.27
18Ni(250)Mar. 900°F 269.5 55.6 0.50 72.9 0.65
H-ll (VM) 1000°F temp. 11.4 0.40 10.8 0.38
4340 475°F temp. 267.2 13.3 0.29 11.1 0.24
AM355 (FH) SCT 1000 33.1 0.28 50.3 0.42
AM355 SCT 1000 169.4 24.5 0.24 36.7 0.43
410 650°F temp. 197.0 22.0 0.24 23.8 0.26
AM355 . SCT 850 195.9 10.7 0.22 24.9 0.52
AM355 (FH) SCT 850 9.7 0.15 6.2 0.10
AISI 304 Sensitized 83.9 8.5 0.12 15.2 0.22
lOOhr. 1100°F
3
Plane strain conditions maintained only at low-stress intensities. Therefore, values are approximate. True plane strain KISCC could not be obtained.
b
Explanation of abbreviations: AM—air melt, VM—vacuum melt, FH—fully hardened.

can be normalized. The ratio KIScc/KIC serves as a normal- Recommendations

izing parameter to compare steels and their heat treatments.10
See Fig. 6. Recommendations take on a two-fold aspect. 1) What
The most widely used fracture mechanics specimen is the can the designer and/or metallurgist do about the SCC and
precracked cantilever-beam specimen. * * By dead-weight load- HSC problem ? 2) What can industry, profession, society or
ing this specimen in an environment and recording the time government agency do about test standardization, data
to failure, the plane-strain stress-intensity threshold KIScc accumulation, and dissemination of the information ?
can be calculated.
The type of data obtained by fracture mechanics analyses
What Can the Designer or Metallurgist Do about SCC and HSC?
is well-illustrated by the work of Freedman12 who obtained
KIC and KISCC data on ferrous and nickel alloys. Single edge There are some practical approaches to the problem of
notched and fatigue cracked specimens were tension loaded failed fasteners by SCC and HSC; soine of these are quite
in a salt solution for 1000 hr (accelerated test). Identical obvious.
specimens were tension loaded in racks exposed at the sea- The first obvious solution is to use alternate materials.
coast (Playa del Rey, Calif.). Times to failure at seacoast There are materials that have good resistance to SCC. On
varied from 49 hr to 7668 hr. Some tests were run 12,843 hr the basis of the fracture toughness criteria KISCc and the ratio
without failure. Kiscc/Kic tempered with experience, a selection could be
Freedman's data (Table 5) gives a rating of susceptibility made. Unfortunately fracture toughness criteria are only
of various alloys to accelerated (laboratory) and seacoast now being collected and all fastener materials have not been
testing. The ratio of KISCc/KIC for seawater ranks the alloys tested. However, there are some highly resistant materials
from Inconel 718 (most resistant) to sensitized Type 304 of the stainless steel (so-called corrosion resistant) type
(most susceptible). Unexplained differences are to be noted available, e.g., A286, A286 CW, Inconel 718 CW, etc.
in the ratio for the accelerated test. Variations between the The second obvious solution is to keep the aggressive
accelerated and seacoast tests can probably be explained by environment in the case of SCC, and a hydrogen source in
the unappreciated variability of the aggressive environments. the case of HSC, away from the steel. Although aerospace
engineers shy away from coatings, platings of cadmium or
aluminum may be helpful on low-alloy steel martensites.
Extreme care in electroplating and required subsequent
265 ksi UTS baking must be exercised at high-strength levels to prevent
HSC. In some aerospace applications, organic coatings,
paints, (Locktite, Lock-safe, greases and even baked-on solid
lubricants) can protect against the environments. If torquing
and retorquing are done on the bolts, these operations can
remove these types of coatings and reapplication of the
protection is required.
206 ksi UTS A less obvious solution to the problem would involve
designs that would eliminate or minimize factors that promote
SCC. One should, for instance, avoid crevices, deep re-
10 10',2 I0 4
cesses, sharp corners, notches of any kind, and dissimilar
TIME TO FAILURE, min metals unless one metal is insulated from the other in some
manner.
Fig. 6 Delayed fracture characteristics of 4340 steel plotted in a In any new design, where new alloys are to be tried and
normalized manner using the stress intensity ratio. new environments are to be experienced, it is highly recom-
NOVEMBER 1972 CRACKING AND EMBRITTLEMENT OF HIGH-STRENGTH FASTENERS 803

Table 6 Nominal compositions of all alloys mentioned in the text

Material Elements given in percent

C Cr Ni Co Fe Mo W Cb Ti Al Other
4130 0.3 0.95 bal 0.20
4140 0.4 0.95 bal 0.20
4340 0.4 0.8 1.90 bal 0.25
8740 0.4 0.5 0.55 bal 0.25
HY-TUF 0.25 0.30 1.80 bal 0.40 1.5 Si
D6AC 0.45 1.00 0.55 bal 1.00 0.07 V
HP 9-4 0.45 0.27 8.00 4.00 bal 0.25
Maraging-200 0.03 18.5 8.50 bal 3.25 0.20 0.10
Maraging-250 0.03 18.5 7.50 bal 4.80 0.40 0.10
Maraging-300 0.03 18.5 9.00 bal 4.80 0.60 0.10
Type 410 0.15 12.5 bal
Type 41 6 0.15 13.0 bal 0.60
Type 422 0.23 12.0 0.80 bal 1.00 0.25 V
Type 431 0.15 16.0 1.85
Type 440C 1.10 17.0 0.75 bal 0.55
H-ll 0.40 5.0 bal 1.30 0.5 V
17-4 PH 0.07 16.50 4.0 bal 0.3 0.4 Cu
17-7 PH 0.09 17.00 7.1 bal 1.10
PH 15-7 Mo 0.09 15.0 12.1 bal 2.50 1.10
PH 13-8 Mo 0.05 12.75 8.0 bal 2.25
PH 14-8 Mo 0.05 14.35 8.1 bal 2.50 1.10
15-5 PH 0.07 14.75 4.5 bal 0.3 0.35 Cu
AM 350 0.10 16.5 4.5 bal 2.9 0.1 N
AM 355 0.12 15.5 4.5 bal 2.9 0.1 N
Inconel-718 0.05 15.0 26.0 bal 1.25 2.15 0.2 0.3 V
Haynes-25 0.10 20.0 10.0 bal 15.0
(L605, WF-11)
Custom-455 0.05 11.75 8.5 bal 0.5 0.3 1.1 2.0 Cu
A-286 0.05 14.75 25.25 bal 1.3 2.15 0.15 0.3 V
MP35N 20.0 35.0 35.0 10.0
Hastelloy-X 0.10 22.0 bal 1.5 18.59.0 0.6
Rene-41 0.09 19.0 bal 11.0 10.0 3.1 1.5
Waspaloy 0.07 19.5 bal 13.5 2.0 4.3 3.0 1.4
Inconel-X750 0.04 15.5 bal 7.0 0.95 2.5 0.7
Type 303 0.15 18.0 9.0 bal
Type 304 0.08 18.5 9.5 bal
Type 309 0.20 23.0 13.5 bal
Type 316 0.08 17.0 12.0 bal 2.25
Type 321 0.08 18.0 11.0 bal 5xC
Type 347 0.08 18.0 11.0 bal 10 x C
Unitemp 212 0.08 16.0 25.0 bal 0.50 4.0 0.15

mended that fracture toughness tests be conducted to ascertain sufficiently advanced to be able to predict dangerous stress-
the possibility of some susceptibility to degradation, if not environment structure combinations that could lead to
complete failure. failure.
The high strength fasteners which occasionally fail in
What Can Industry, Profession, Society or Agency Do about SCC service can be divided into three categories: 1) by ultimate
and HSC ? tensile strength level, 2) by metallurgical type, and 3) by
relative susceptibility to salt solutions or marine atmospheres.
There is a need for standardization of fracture toughness Fracture toughness analysis offers the best hope of obtain-
tests using the precracked specimens. The American Society ing data on SCC or HSC. The analysis does not differentiate
for Testing and Materials already has a tentative specification between mechanisms. The use of plane-strain fracture
on testing. Without standardization, the data that are and toughness KIC and stress corrosion threshold KIScc offers
are becoming available have limited usefulness in material promise of selecting fastener materials that are not susceptible
selection or design. A standard test must be in existence to SCC or HSC failures. The KIC or KISCC parameters are
long enough so that sufficient data can be accumulated for becoming more meaningful to the designer than elongation
statistical study. The accumulating data should be compiled and reduction of area because these "measures of ductility"
into usable form by industry, the metallurgical profession, have little to do with the performance of a structure.
government agency or technical society. Table 6 gives the nominal compositions of all alloys
mentioned in the text.
Conclusions
References
The current situation on the use of fasteners in aerospace
applications has been appraised. Confusion exists as to the 1
physical nature of HSC and SCC. Examples of fastener Stanley, J. K., "Solutions to Some Stress Corrosion Cracking
failures on the Titan III family of vehicles illustrate the Problems in Aerospace Situations," Proceedings of the First Joint
Aerospace and Marine Corrosion Technology Seminar, Houston,
nature of the problem. Although there is a distinct difference Texas, 1969, pp. 36-43.
between SCC and HSC phenomena on the laboratory scale, 2
Lucas, W. R., "Report of the ADHOC Committee on Failure
it is often difficult if not impossible to differentiate between of High Strength Materials," Aug. 1971, NASA Marshall Space
the two mechanisms in the field. Corrosion theory is not Flight Center, Huntsville, Ala.
804 J. K. STANLEY J. SPACECRAFT

3 8
Hanna, G. L., Troiano, A. R., and Steigerwald, E. A., "A Brown, B. F., "A New Stress Corrosion Cracking Test for
Mechanism for Embrittlement of High Strength Steels by Aqueous High Strength Alloys," Materials Research and Standards, Vol. 16,
Environments," Transactions American Society for Metals, Vol. 57, March 1966, pp. 129-133.
9
1964, pp. 658-671. Brown, B. F., "The Application of Fracture Mechanics to
4
Dull, D. L. and Raymond, L., "A Method of Evaluating Stress-Corrosion Cracking," Metallurgical Review, Vol. 13, 1968,
Relative Susceptibility of Bolting Materials to Stress Corrosion pp. 171-183.
10
Cracking," presented at 1972 Westec Conference, March 1972; Mulherin, J. H., "Stress Corrosion Susceptibility of High
also Corrosion, to be published. Strength Steel in Relation to Fracture Toughness," Transactions of
5
Phillips, A., Kerlins, V., and Whiteson, B. V., "Electron Fracto- the ASME: Journal of Basic Engineering, Vol. 88, 1966, pp. 772-
graphic Handbook," AFML Rept. ML-TDR-64-416, Aug. 1968, 782.
11
Air Force Materials Lab., Wright-Patterson Air Force Base, Ohio. Tentative Method of Test for Plane-Strain Fracture Toughness
6
Fidelle, J. P., Legrand, J., and Couderc, C., "A Fractographic of Metallic Materials, E399-70T, Book of Standards, American
Study of Hydrogen Gas Embrittlement in Steels," Transactions of Society for Testing and Materials, Philadelphia, Pt. 31, July 1971,
the American Institute of Mining, Metallurgical Engineers, 1972, to pp. 911-1927.
12
be published. Freedman, A. J., "Development of an Accelerated Stress-
7
Tiner, N. A. and Filpin, C. B., "Microprocesses in Stress Corrosion Test for Ferrous and Nickel Alloys," Finary Summary
Corrosion of Martensitic Steels, Corrosion, Vol. 22, 1966, pp. 271- Rept., NOR 68-58, April 1968, Northrop Corp., Hawthorne,
279. Calif.

NOVEMBER 1972 J. SPACECRAFT VOL. 9, NO. 11

Application of an Improved Transpiration Cooling Concept

to Space Shuttle Type Vehicles
JOHN R. SCHUSTER* AND THOMAS G. LEEf
McDonnell Douglas Astronautics Company—West
Huntington Beach, Calif.

In order to reduce coolant requirements transpiration cooled structures should be operated at the maximum
possible surface temperature. At high-surface temperature a flow instability can occur, because of the pressure
drop characteristics of gases, leading to burnout of the porous structure. This instability can be avoided by
overlaying the sintered metallic structure with a ceramic coating of much higher permeability. By applying this
design concept to the leading-edge and interference heating regions of a space shuttle type vehicle substantial
reductions can be realized in coolant requirement. Since the thickness of the ceramic layer is inversely pro-
portional to the required coolant flow rate, the leading-edge coolant savings are offset by increasing structure
weight. For interference heating, however, the ceramic layer is thin and large net savings in weight can be obtained.

Nomenclature T = temperature, deg R

a = porous matrix viscous pressure drop constant, ft ~ 2
13 = porous matrix inertial pressure drop constant, ft" 1
Cp = specific heat, Btu/lbm-°R £ = effective emissivity of surface, dimensionless
D = diameter, ft y = porous matrix void fraction, dimensionless
gc = conversion factor, 32.17 lbm-ft/lbf-sec2 F = porous matrix tortuosity for heat conduction, dimension-
H = enthalpy, Btu/lbm less
A// = net enthalpy rise in coolant, Btu/lbm /x = coolant viscosity, Ibm/ft-sec
k = conductivity, Btu/ft-sec-°R p = coolant density, lbm/ft 3
L = length, ft a = Stefan-Boltzmann constant, 0.476 x 10-12Btu/ft2-sec-°R4
M = coolant molar weight, Ibm/lb-mole a = elemental surface area normal vector, ft 2
m = coolant flux, Ibm/ft2-sec 9 = included angle between a and freestream vector, rad
Nc = convective heat-transfer coefficient, Ibm/ft2-sec
P = pressure, lbf/ft 2
Qc = convective heat flux, Btu/ft2-sec Subscripts
R = gas constant, lbf-ft/lb-mole-°R
A =
value at point A
B =
value at point B
Submitted April 17, 1972; revision received July 17, 1972. This C =
value for coolant
work was supported by McDonnell Douglas Independent Research e =
effective value
and Development funds, and was performed under the direction / =
value for coolant liquid phase
of James E. Rogan. o =
value without transpiration cooling or value in coolant
Index categories: Heat Conduction; Boundary Layers and Con- storage reservoir
vective Heat Transfer—Laminar; Structural Composite Materials pore = value for pore within the porous matrix
(Including Coatings). r = recovery value
* Senior Engineer Scientist, Aerothermodynamics. s = value for freestream species evaluated at surface temper-
f Section Chief. Heat Transfer and Materials Performance. ature and pressure