Graybeal and Hartmann 1

ULTRA-HIGH PERFORMANCE CONCRETE MATERIAL PROPERTIES

B.A. Graybeal, P.E. Professional Service Industries, Inc.

6300 Georgetown Pike
McLean, VA 22101
(202) 493-3122
fax: (202) 493-3442
[email protected]

J.L. Hartmann, P.E. Federal Highway Administration

6300 Georgetown Pike
McLean, VA 22101
(202) 493-3059
fax: (202) 493-3442
[email protected]

Length = 4500 words

A paper submitted for presentation at the 2003 Transportation Research Board Conference
November 15, 2002

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 2

ABSTRACT
For centuries, concrete producers and researchers have been striving to improve the material properties of
concrete. Recently, a new class of concretes have been developed which exhibit significantly improved
strength and durability characteristics. The Federal Highway Administration at its Turner-Fairbank
Highway Research Center Structures Laboratory is currently investigating UHPC and its possible uses in
the transportation industry. The ongoing research program has tested full-scale highway bridge girders
and is now completing a material property characterization study.
Testing to date indicates that this material can attain a compressive strength of 200 MPa (29 ksi)
and a flexural tensile strength of 10.3 MPa (1.5 ksi). Testing has also indicated that flexural fatigue of the
base concrete material is unlikely to be a factor in structural applications. However, fatigue of the fiber
reinforcement may occur after cracking of the concrete matrix. Other material property testing being
completed includes direct and indirect tensile strength, creep, shrinkage, freeze-thaw resistance, scaling
resistance, chloride penetration resistance, and abrasion resistance.

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 3

ULTRA-HIGH PERFORMANCE CONCRETE MATERIAL PROPERTIES

B.A. Graybeal and J.L. Hartmann

INTRODUCTION
For centuries, concrete producers and researchers have been striving to improve the material properties of
concrete. Recently, a new class of concretes have been developed which exhibit significantly improved
strength and durability characteristics. These concretes, currently classified as Ultra-High Performance
Concrete (UHPC), can simultaneously have compressive strengths exceeding 200 MPa (30 ksi), tensile
strengths of 7 to 14 MPA (1 to 2 ksi), and durability properties that significantly exceed current High
Performance Concrete (HPC) behavior [1-4].
The Federal Highway Administration (FHWA) at its Turner-Fairbank Highway Research Center
(TFHRC) Structures Laboratory is currently investigating UHPC and its possible uses in the
transportation industry. The ongoing research program began in early 2001 with initial material property
testing, followed by structural testing of prestressed highway bridge girders. As a result of the promising
characteristics displayed during the initial material and structural testing, an extensive suite of material
property characterization tests was initiated. Characterization testing began in the fall of 2002 and should
be completed by mid-2003.

BACKGROUND
The UHPC currently being studied at FHWA is a steel fiber reinforced concrete that was originally
developed by Bouygues SA. This concrete is being produced and marketed by Lafarge under the trade
name Ductal. As a class, UHPC’s gain many of their enhanced properties from the fine grading and
tightly packed nature of its constituent materials [5-8]. These concretes have optimized mix proportions
wherein the largest particle is a fine sand and the smallest particles are multiple orders of magnitude
smaller. This type of concrete is being produced and distributed in Europe by a number of manufacturers,
including Lafarge and Eiffage. The UHPC from Lafarge is currently the only UHPC widely available in
the United States.
Table 1 shows a typical mix proportion for UHPC. As is evident from the table, the water to
cementitious material ratio is very low. This low ratio allows the concrete to contain less free water and
thus fewer voids in the final concrete matrix.
One of the primary benefits of this class of concretes is that it can exhibit significant tensile
strength and toughness. Much of this property enhancement is imparted to the concrete by the addition of
short, discontinuous fibers during the mixing procedure. Either steel or organic fibers can be used to add
tensile strength and control cracking in the material; however, for structural purposes, steel fibers provide
much improved tensile performance. The steel fibers used in the concrete currently being studied at
TFHRC are un-deformed 12 mm (0.5 in.) long and 0.2 mm (0.008 in.) diameter wires. The volume of
fibers in the mix can vary depending on the material properties desired. The current work utilizes two
percent steel fibers by volume in the concrete matrix.
The casting and curing of this concrete includes some differences from the procedures employed
for normal concrete. This concrete exhibits high fluidity and is virtually self-placing. Due to the
presence of fibers, internal vibration is prohibited; however, slight external vibration is normally used to
eliminate some entrapped air. The initial set of this concrete is usually achieved in 24 to 48 hours
depending on the surrounding temperature conditions. Final curing of the material is achieved by
subjecting the concrete to a steam cure (at least 90°C (194°F) and 95 percent humidity) for 48 hours.
According to the manufacturer, this curing process establishes a stabilized UHPC that has achieved its
final mechanical and durability properties by the conclusion of the cure.

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 4

UHPC AT FHWA
The FHWA has an ongoing research program focusing on UHPC in its Structures Lab at the Turner-
Fairbank Highway Research Center. Initially this program considered the structural behavior of
AASHTO Type II prestressed girders fabricated with UHPC and without passive reinforcement [9].
The Type II girders were prestressed with a full complement (24) of 12.7 mm (0.5 in.) low relaxation
prestressing strands. This testing was performed in collaboration with the Virginia Department of
Transportation (VDOT) who expressed an interest in implementing UHPC in a highway bridge.
The girder testing included one flexural test and three shear tests. The flexural test of this 0.91 m
(36 in.) deep girder had a span of 23.9 m (78.5 ft). The girder carried a peak applied moment of 4400 kN-
m (3225 k-ft) and sustained a deflection of over 0.46 m (18 in.) prior to failure. Three shear tests have
been completed. As these girders did not contain any mild reinforcing steel, all shear forces were carried
by the concrete and fiber matrix. The testing showed that an AASHTO Type II composed UHPC can
carry between 1690 and 2200 kN (380 and 500 kips) of shear force.

COMPRESSION TESTING
As part of the UHPC girder testing described above, compression testing of both cast cylinders and cores
was performed. This testing produced data on the compressive strength, modulus of elasticity, and
Poisson’s ratio of the UHPC. These tests were completed according to ASTM C39 and C469.
Table 2 provides the results from the compression testing. These tests were completed in a 4400
kN (1000 kip) Forney hydraulic testing machine. Strains were measured using strain gages bonded
longitudinally and transversely to the surface of the specimen. The average values obtained from these
tests indicate that this UHPC has a strength between 193 and 207 MPa (28.0 and 30.3 ksi) at a 99 percent
confidence level and an average strength of 200 MPa (29 ksi). Also the modulus of elasticity is 52.4 GPa
(7600 ksi), the strain at peak compressive strength is approximately 0.0043, and Poisson’s ratio is 0.18.

FLEXURE TESTING
Flexure testing has also been completed on UHPC component specimens. Two cast prisms and seven
prisms cut from the Type II girders have been tested in four-point bending according to a method similar
to that described in ASTM C1018. The primary difference is that the prism cross-sectional area and span
length has been modified from that recommended in the specification. These static tests are described
below. Additionally, some fatigue testing has been completed following similar procedures but utilizing
a cyclic loading. These tests are also described below.

Static Testing
Static flexure tests have been completed on nine prisms. All of these prisms had a nominal 51 mm by 51
mm (2 in. by 2 in.) cross-section, however the span length and shear span varied. Table 3 provides the
results from these tests. These results show that, in this configuration, UHPC has a moment capacity that
is in excess of the peak elastic moment capacity. Also note that the tensile cracking strain is in the
vicinity of 300 microstrain and the tensile cracking stress is approximately 12 to 15.5 MPa (1.75 to 2.25
ksi).

Fatigue Testing
Fatigue testing using the same four-point loading configuration as used for the static testing has also been
completed. These fatigue tests focused on two different types of behavior. Specifically, both the fatigue
strength of the cracked concrete matrix where the steel fibers carry the majority of the tensile forces and
the fatigue strength of the uncracked concrete matrix were investigated. The objective of the testing was
to determine if either the steel fibers or the uncracked concrete matrix experience any fatigue related
degradation when subject to repetitive loadings.
First, two tests designed to produce steel fiber fatigue were completed. In these tests, the prism
was loaded statically until it reached its cracking load. Subsequent to cracking, the prism was unloaded
and then a cyclic flexural fatigue loading was applied. The load level for the fatigue cycling ranged from

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 5

approximately 10 percent to 60 percent of the cracking load. Fatigue related degradation and eventually
failure was observed in both of these specimens. The initial crack grew larger and extended further into
the specimen as cycling proceeded and some of the highest stressed fibers began to pull out. Eventually,
audible evidence of fiber breakage was heard. One of the specimens failed after 9,950 cycles while the
other failed after 129,700 cycles. Figure 1 shows an electron microscopy photograph of both an
undamaged fiber as well as a fiber that ruptured under fatigue loading.
The remaining fatigue tests focused on the fatigue behavior of the UHPC matrix prior to cracking.
A four-point loading setup was used to provide a region of uniform flexural tension on one face of the
prism. The prism was subjected to cyclic loading with a minimum flexural tensile load of 0.7 MPa (0.1
ksi) and a tensile stress range of 5.2 MPa to 13.8 MPa (0.75 ksi to 2.0 ksi). Figure 2 shows the results of
these tests. These preliminary results indicate that the UHPC matrix cracking stress under cyclic loading
is toward the lower end of the observed static flexural tensile cracking stress range. This may be due to
slight degradation of the matrix due to the cycling; however, the small sample size precludes a definitive
conclusion.

BATCHING OF UHPC
The UHPC batching process differs significantly from that employed in the production of typical or high-
performance concrete. Although the actual procedures will vary slightly depending on the individual
mixer used, in general, the mixing process requires more time and imparted energy than is required for
other concretes.
The batching that has been completed at TFHRC has been performed in a two cubic foot (0.057
cubic meter) pan mixer. The batching procedure includes a dry mixing of the pre-blended elements prior
to the addition of the water and the first half of the superplasticizer to the mixer. The remaining
superplasticizer in introduced to the mix once the water and first half of superplasticizer have been
thoroughly dispersed. Mixing continues until the UHPC reaches its break point. The break point occurs
when the water and superplasticizer have been sufficiently dispersed throughout the mix to allow the
UHPC to begin to flow like a homogeneous, viscous fluid. Subsequent to the break point, the steel fibers
are introduced and mixing continues until the fibers are also well dispersed throughout the mix. In total,
the mixing of one batch of UHPC in this pan mixer required approximately 25 minutes. Higher energy
mixers should require less time.

ONGOING RESEARCH
In mid-2002 the second phase of FHWA’s UHPC research program was initiated. This phase is focusing
on characterizing a diverse range of UHPC material properties. In particular, the research concentrated
on classifying this concrete according to the FHWA’s definition of a High Performance Concrete (HPC).
The FHWA definition of HPC focuses on strength and durability aspects of the concrete [10].
The strength criteria include compressive strength, modulus of elasticity, shrinkage, and creep. The
durability criteria include freeze-thaw resistance, scaling resistance, abrasion resistance, and chloride
permeability.
The material testing outlined in this study is intended to answer many questions regarding the
material behavior of UHPC. As both testing and practical use of UHPC are in their infancy, a great deal
of initial research is required to determine how UHPC can best be utilized.
Compression testing of cylinders and cubes is a widely performed and easily reproduced method
used to determine the properties of concrete. Within this program, tests of various size cylinders and
cubes will be performed. These tests will be completed in accordance with ASTM C39 and ASTM C109,
respectively. Due to the high compressive strength of UHPC, standard test specimen sizes and
preparation techniques must be modified. In particular, given that UHPC has a compressive strength of
approximately 30 ksi (207 MPa), compression testing of four inch diameter cylinders requires a test
machine capacity approaching 400 kips (1780 kN). For this reason, test cylinder diameters will range
from two to four inches. Additionally, the high strength of this concrete precludes the use of standard end
capping methods. For this reason, the ends of cylinders must be ground. In an attempt to circumvent the

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 6

end grinding requirement, cube specimens are also included in the test matrix. Both two and four inch
cubes will be tested in uniaxial compression. The modulus of elasticity and Poisson’s ratio will also be
collected during cylinder compression testing in accordance with ASTM C469.
Unlike normal concrete, UHPC exhibits significant tensile strength both before and after cracking
of the concrete matrix. Normal relationships between concrete compressive and tensile strengths are not
valid for this material. For this reason, the ongoing test program includes two direct and two indirect
methods of measuring the UHPC’s tensile strength. The two widely used concrete tension tests, namely
the split cylinder test and the prismatic section flexural test, will both be conducted on a matrix of
samples. These tests will be conducted according to ASTM C496 and ASTM C1018, respectively. Two
direct tension tests will also be conducted. These include the mortar briquette test as described in
AASHTO Specification T132 and the notched cylinder test as recommended by RILEM TC 162-TDF
[11]. It is expected that there will be significant scatter in the results obtained from these four tests;
however, the results should provide an indication of the lower bound of the tensile strength of UHPC.
Use of UHPC in prestressed and post-tensioned applications requires that the material exhibit
dimensional stability over time. Determination of the material stability will be investigated through both
creep and shrinkage testing. The creep testing will be performed on cylinders according to ASTM C512
and the shrinkage testing will be performed on prisms according to ASTM C157.
As with any bridge construction material, durability plays an important role in both the safety and
economy of a structure. Bridge construction materials must be capable of maintaining structural integrity
in an aggressively deleterious environment over time with little or no maintenance in order for the
condition of the nation’s inventory to improve. Four durability tests will be completed on a body of
UHPC samples in this phase of the program. First, freeze-thaw testing will be completed according to
ASTM C666. Scaling tests will then be performed on UHPC slabs according to ASTM C672. Chloride
penetration testing will be performed both by the rapid method outlined in ASTM C1202 and by the
standard method outlined in AASHTO Specification T259. Finally, abrasion testing will be performed
according to the rotating-cutter method described in ASTM C944.
Additional testing on this UHPC is expected to include a number of non-destructive evaluation
techniques, including computed tomography, ultrasonics, and scanning electron microscopy.

PRACTICAL APPLICATIONS
The possible applications of UHPC are many and varied. In theory, UHPC can be used for any
application, either structural or architectural, for which concrete would normally be specified. However,
most typical concrete applications do not efficiently utilize UHPC’s materials properties. Although
projected UHPC life-cycle costs are favorable relative to typical bridge construction materials,
applications need to be developed that are also first-cost conscious. In a direct comparison, commercially
available UHPC currently costs approximately ten times normal concrete. Additionally, the batching and
placement requirements of UHPC preclude its use in cast-in-place applications. Therefore, until this
technology progresses, the current practical uses of UHPC will be limited to precast applications where
increased strength and durability are required.
In the transportation field, UHPC may be best suited for use in prestressed bridge superstructures.
The increased tensile and compressive strengths may allow for longer spans while simultaneously
decreasing girder cross-sectional dimensions. Another possible bridge application is in precast deck
panels, where the increased durability of UHPC could allow for longer-life bridge decks. Other possible
applications include bridge substructures in areas with high ductility demand (seismic zones), thin-walled
tunnel liners, and blast-resistant barriers and structures.

CONCLUSIONS
Ultra-High Performance Concrete exhibits properties that clearly distinguish it from normal concrete.
The FHWA’s ongoing research is intended to quantify the relevant characteristics of this material with an
aim toward its use in the highway industry. Testing to date indicates that this material can attain a
compressive strength 200 MPa (29 ksi) and a flexural tensile strength of 10.3 MPa (1.5 ksi). Testing has

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 7

also indicated that flexural fatigue of the uncracked concrete matrix is unlikely to be a factor in structural
applications, however fatigue of the fiber reinforcement may occur after cracking and thus needs
additional investigation. Further material testing is in progress and preliminary results are expected in
early 2003.

ACKNOWLEDGEMENT
The research which is the subject of this paper was funded by the Federal Highway Administration with
contributions from Lafarge North America and Prestressed Services Inc. of Lexington, Kentucky. The
authors gratefully acknowledge this support. The publication of this article does not necessarily indicate
approval or endorsement of the Federal Highway Administration or the United States Government, the
findings, opinions, conclusions, or recommendations either inferred or specifically expressed herein.
This paper is intended as an academic discussion, not as engineering advice, and no reliance upon
this paper is permitted. Independent advice by the professional of record as to the application of the
concepts and opinions contained herein to any specific project should be sought.

REFERENCES

1. Bonneau, O., M. Lachemi, E. Dallaire, J. Dugat, and P.-C. Aïtcin, “Mechanical Properties and
Durability of Two Industrial Reactive Powder Concretes,” ACI Materials Journal, V. 94, No. 4, July-
August 1997, pp. 286-290.

2. Bonneau, O., C. Poulin, J. Dugat, P. Richard, and P.-C. Aïtcin, “Reactive Powder Concretes: From
Theory to Practice,” Concrete International, April 1996, pp. 47-49.

3. Dugat, J., N. Roux, and G. Bernier, “Mechanical Properties of Reactive Powder Concretes,”
Materials and Structures, V. 29, No. 188, 1996, pp. 233-240.

4. Roux, N., C. Andrade, and M. Sanjuan, “Experimental Study of Durability of Reactive Powder
Concretes,” Journal of Materials in Civil Engineering, V. 8, No. 1, Feb. 1996, pp. 1-6.

5. Bonneau, O., C. Vernet, M. Moranville, and P.-C. Aïtcin, “Characterization of the Granular Packing
and Percolation Threshold of Reactive Powder Concrete,” Cement and Concrete Research, V. 30, No.
12, 2000, pp. 1861-1867.

6. Cheyrezy, M., V. Maret, and L. Frouin, “Microstructural Analysis of RPC (Reactive Powder
Concrete),” Cement and Concrete Research, V. 25, No. 7, 1995, pp. 1491-1500.

7. De Larrard, F., and T. Sedran, “Optimization of Ultra-High-Performance Concrete by the Use of a

Packing Model,” Cement and Concrete Research, V. 24, No. 6, 1994, pp. 997-1009.

8. Richard, P., and M. Cheyrezy, “Composition of Reactive Powder Concretes,” Cement and Concrete
Research, V. 25, No. 7, 1995, pp. 1501-1511

9. Hartmann, J., and B. Graybeal, “Testing of Ultra-High-Performance Concrete Girders,” PCI Journal,
V. 47, No. 1, Jan.-Feb. 2002, pp. 148-149.

10. Goodspeed, C., S. Vanikar, and R. Cook, “High-Performance Concrete Defined for Highway
Structures,” Concrete International, V. 18, Feb. 1996, pp. 62-67.

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 8

11. RILEM, TC 162-TDF, “Test and Design Methods for Steel Fibre Reinforced Concrete –
Recommendations: Uni-axial Tension Test for Steel Fibre Reinforced Concrete,” Materials and
Structures, V. 34, Jan.-Feb. 2001, pp. 3-6.

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 9

LEGEND TO TABLES AND FIGURES

TABLE 1 Typical Composition of UHPC

TABLE 2 Compression Testing Results

TABLE 3 Prism Flexure Testing Results

FIGURE 1 Fatigue failure surface of a steel fiber: (a.) virgin fiber, and (b.) fatigued fiber.

FIGURE 2 Peak flexural tensile stress versus loading cycles for UHPC in flexural fatigue.

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 10

TABLE 1 Typical Composition of UHPC

Material Amount (lb/yd3) Percent by Weight

Portland Cement 1200 28.7
Fine Sand 1720 41.1
Silica Fume 390 9.3
Ground Quartz 355 8.5
Superplasticizer 22 0.5
Steel Fibers 270 6.4
Water 230 5.5
1 lb/yd3 = 0.593 kg/m3

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 11

TABLE 2 Compression Testing Results

Axial Strain
Peak at Peak Axial Transverse Poission's
Specimen Diameter Height Strength Strength Modulus Modulus Ratio
(in.) (in.) (ksi) (µε) (ksi) (ksi)
1744-13 3.0 5.0 27.7 - - - -
1744-11 3.0 5.0 28.0 - - - -
1744-12 3.0 5.1 28.4 4250 7340 44500 0.17
CA01 2.76 5.54 31.0 4400 7690 41340 0.19
CA02 2.76 5.58 28.9 4470 7490 43580 0.17
CA03 2.76 5.47 28.0 3980 7560 42590 0.18
CA04 2.76 5.55 27.8 3650 8360 - -
CA05 2.76 5.42 32.4 5890 6880 40630 0.17
CB01 3.98 7.98 28.8 4020 7830 41350 0.19
CB02 3.98 8.06 30.8 4540 7370 - -
CB03 3.99 8.04 28.5 4160 7480 - -
CB04 3.99 7.97 28.7 4070 7650 40930 0.19
CB05 3.98 7.91 28.5 4290 7620 - -
CB06 3.98 7.97 30.7 4610 7730 43830 0.18
Mean 29.2 4360 7580 42340 0.18
- indicates data was not obtained
1 ksi = 6.895 MPa
1 in. = 25.4 mm

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
Graybeal and Hartmann 12

TABLE 3 Prism Flexure Testing Results

Bottom Cracking
Flange Bottom Cracking Cracking Mmaximum
Shear Stress Load Cracking Flange Tensile Tensile Maximum ––––––––––
Specimen Width Depth Span Span Rate‡ Moment Strain Stress† Stress‡ Moment Mcracking
(in.) (in.) (in.) (in.) (ksi/sec) (k-in.) (µε) (ksi) (ksi) (k-in.)
31744-5 2.0 2.0 9 3 static 3.23 341 2.59 2.42 6.58 2.03
31744-6 2.0 2.0 6 2 static 2.78 300 2.28 2.08 6.39 2.30
PA01 2.18 1.94 12 4.5 0.6 2.48 - - 1.81 3.19 1.29
PA08 2.23 1.80 12 4.5 2.5 3.13 383 2.91 2.60 - -
PA10 1.95 2.06 12 4.5 0.25 2.57 264 2.01 1.86 - -
PA12 2.08 2.07 12 4.5 static 2.58 255 1.94 1.73 3.86 1.50
PA13 1.85 1.90 12 4.5 0.6 2.44 310 2.36 2.20 3.65 1.49
PA14 1.90 1.92 12 4.5 2.5 2.71 351 2.67 2.32 3.15 1.16
PA15 1.91 2.10 12 4.5 2.5 3.14 315 2.39 2.23 3.42 1.09
† Based on Hooke’s law, tensile strain reading, and modulus of elasticity of 7600 ksi.
‡ Based on load and cross-sectional dimensions.
- indicates data was not obtained
1 ksi = 6.895 MPa
1 k-in. = 113 N-m
1 in. = 25.4 mm

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 13

(a.)

(b.)

FIGURE 1 Fatigue failure surface of a steel fiber: (a.) virgin fiber, and (b.) fatigued fiber.

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.
 Graybeal and Hartmann 14

2.5
Peak Tensile Stress (ksi)

1.5

1
Crack Observed
0.5
Run-out

0
0 2 4 6 8
6
Cycles (x10 )

FIGURE 2 Peak flexural tensile stress versus loading cycles for UHPC in flexural fatigue.

TRB 2003 Annual Meeting CD-ROM Paper revised from original submittal.