U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590

Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations

This report is an archived publication and may contain dated technical, contact, and link information
Publication Number: FHWA-HRT-06-103
Date: August 2006

Material Property Characterization of Ultra-High Performance Concrete



The UHPC used in this study is a patented product of a major worldwide concrete manufacturer. The product is a reactive powder concrete that is marketed under the name Ductal. This product has a number of different material compositions depending on the particular application. A typical composition is provided in table 1.

The constituent material proportions were determined, in part, based on an optimization of the granular mixture. This method allows for a finely graded and highly homogeneous concrete matrix. Fine sand, generally between 150 and 600 micrometers (µm), is dimensionally the largest granular material. The next largest particle is cement with an average diameter of approximately 15 µm. Of similar size is the crushed quartz with an average diameter of 10 µm. The smallest particle, the silica fume, has a diameter small enough to fill the interstitial voids between the cement and the crushed quartz particles.

Dimensionally, the largest constituent in the mix is the steel fibers. In this study, the fibers in the mix had a diameter of 0.2 millimeters (mm) (0.008 inch) and a length of 12.7 mm (0.5 inch). Given the relative sizes of the sand and the fibers, the steel fibers are able to reinforce the concrete matrix on the micro level. A further discussion of the properties of the steel fibers is provided in section 2.3.

Table 1. Typical UHPC composition.
Material Amount (kg/m3 (lb/yd3)) Percent by Weight
Portland Cement 712 (1,200) 28.5
Fine Sand1020 (1,720) 40.8
Silica Fume 231 (390) 9.3
Ground Quartz211 (355) 8.4
Superplasticizer30.7 (51.8) 1.2
Accelerator30.0 (50.5) 1.2
Steel Fibers156 (263) 6.2
Water109 (184) 4.4

1 kg/m3 = 1.686 lb/yd3

RDM = relative dynamic modulus (see p. 134)


As previously discussed, the UHPC used in this study is a proprietary product. The manufacturer has performed significant material property testing and has reported typical characteristics. Table 2 provides some of the material properties relevant to using this material in bridge applications. In general, these properties have not been verified and are provided here solely for completeness.

Table 2. Manufacturer-supplied material characteristics.
Material Characteristic Range
Compressive Strength (MPa) 180–225
Modulus of Elasticity (GPa) 55–58.5
Flexural Strength (MPa) 40–50
Chloride Ion Diffusion (m2/s) 1.9 x 10-14
Carbonation Penetration Depth (mm) < 0.5
Freeze-Thaw Resistance (RDM) 100%
Salt-Scaling Resistance (kg/m2) <0.012
Entrapped Air Content 2–4%
Post-Cure Shrinkage (microstrain) 0
Creep Coefficient 0.2–0.5
Density (kg/m3) 2,440–2,550

1 MPa = 145 psi

1 GPa = 145 ksi

1 m2/s = 1,550 inches2/s

1 kg/m2 = 0.205 lb/ft2

1 kg/m3 = 1.69 lb/yd3


The steel fibers used in this test program were straight steel wire fibers manufactured by Bekaert Corporation. The fibers have a nominal diameter of 0.2 mm (0.008 inch) and a nominal length of 12.7 mm (0.5 inch). The chemical composition of the fibers is shown in table 3. A thin brass coating is applied to the fibers during the drawing process; therefore, virgin fibers may be goldcolored. This coating disappears during the mixing process and is no longer clearly visible during the casting of the UHPC.

The intended function of these fibers within UHPC requires that the fibers have a very high tensile strength. The manufacturer’s specified minimum tensile strength is 2,600 MPa (377 ksi), and tension tests are performed as a means of quality control on the fiber production. The stress-strain behavior as recorded during one of these quality control tests is presented in figure 1. The results from three quality control tests were used to determine an average yield strength of 3,160 MPa (458 ksi) as calculated by the 0.2 percent offset method. The average modulus of elasticity was 205 gigapascals (GPa) (29,800 ksi), and the average ultimate strength was 3,270 MPa (474 ksi). These results clearly show that these high-strength steel wires have little reserve strength or ductility capacity beyond yield.

Table 3. Chemical composition of steel fibers.
Element 0 Composition (percent)
Carbon 0.69–0.76
Silicon 0.15–0.30
Manganese 0.40–0.60
Phosphorus ≤0.025
Sulfur ≤ 0.025
Chromium ≤ 0.08
Aluminum ≤ 0.003

Figure 1. Graph. Sample tensile stress-strain response for steel fiber reinforcement.

This graph shows the stress-strain response for one individual steel fiber. The fiber exhibits linear elastic behavior through approximately 2,500 megapascals before beginning to soften slightly. Yield for this fiber, as defined by the 0.2 percent offset method, occurs at 3,150 megapascals. The fiber ruptures at a stress of 3,250 megapascals (strain of approximately 0.018). The modulus of elasticity for this fiber is 210 gigapascals.

1 MPa = 145 psi

1 GPa = 145,000 psi


2.4.1 Fiber Orientation Effect on Mechanical Properties

Stiel, Karihaloo, and Fehling have conducted a research program investigating the effect of fiber orientation on the mechanical properties of UHPC.(9) These researchers focused on a patented UHPC marketed under the name CARDIFRC®. This UHPC is composed of similar constituent materials and in similar proportions to the UHPC investigated in the present study. One primary difference is that CARDIFRC contains two lengths of steel fibers and a total fiber volumetric percentage of 6 percent.

This research program focused on the effect of UHPC flow direction during casting on the compressive and flexural tensile behaviors of the concrete. Fiber reinforcement tends to align with the direction of flow during casting. This research program investigated the tensile and compressive behaviors of UHPC when loaded parallel to and perpendicular to the direction of flow during casting. The compression tests were performed on 100-mm (4-inch) cubes. The three-point bending flexure tests were performed on 100-mm by 100-mm (4- by 4-inch) prisms with a 500-mm (20-inch) length.

The cube compression tests indicated that preferential fiber alignment has no significant effect on either the compressive strength or the modulus of elasticity of UHPC. However, the threepoint flexure tests showed that the peak equivalent flexural strength of the UHPC prisms was decreased by a factor of more than three when the fibers were preferentially aligned perpendicular to the principal flexural tensile forces. This preferential fiber alignment was clearly apparent on failure surfaces of the prisms. These prisms also did not exhibit the traditional postcracking toughness behaviors normally associated with UHPC and frequently exhibited an abrupt load decrease immediately after first cracking. All of these findings point to the importance of understanding the structural loadings that will be carried by a UHPC member and following correct placement techniques when casting UHPC members.

2.4.2 Permeability of Cracked Concrete

Rapoport et al. investigated the permeability of steel fiber-reinforced concrete as compared to normal concrete.(10) The research focused on creating small cracks in 0.5 percent and 1.0 percent steel fiber-reinforced concrete, then determining the permeability of the concrete. The two primary findings of interest from this study are as follows. First, this study confirmed the findings of other researchers that cracks less than 0.1 mm (0.004 inch) wide have little impact on the permeability of normal concrete.(11) Second, this study confirmed that steel fiber reinforcement reduces the total permeability of a strained section of concrete by changing the cracking mechanism from a few large width cracks to many small width cracks. As would be expected, the concrete with the higher volume percentage of fiber reinforcement displayed more distributed cracking and had a lower permeability.

2.4.3 Creep and Shrinkage of UHPC

Recall the very low post-steam treatment creep and shrinkage values presented in table 2. Lafarge, the manufacturer and distributor of the UHPC discussed in this report, has performed significant research focusing on the creep and shrinkage behaviors of this concrete. Some results of this research were presented in Acker wherein the microstructural behaviors leading to creep and shrinkage of UHPC, HPC, and normal concrete are discussed.(12) Additional discussion with further experimental results is presented in Acker.(13)

Acker argues that creep and shrinkage are closely related behaviors that cannot generally be uncoupled and studied separately. He indicates that shrinkage is primarily caused by selfdesiccation of the concrete binder resulting in the irreversible collapse of calcium-silicate-hydrate (CSH) sheets. As UHPC contains a very low water-to-cementitious materials ratio, this concrete completely self-desiccates between casting and the conclusion of steam treatment. Thus, UHPC exhibits no post-treatment shrinkage.

In regard to creep, Acker restates previous research indicating that the CSH phase is the only constituent in UHPC that exhibits creep. Also, he points out that concrete creep tends to be much more pronounced when it occurs as the concrete is desiccating. Thus, the collapsed CSH microstructure and the lack of internal water both work to reduce UHPC creep.

2.4.4 Abrasion Resistance of HSC via ASTM C944

Horszczaruk studied the abrasion resistance of high-strength fiber-reinforced concrete using the ASTM C944 standard procedure.(14,15) This is the same procedure that was followed in the abrasion tests discussed in section 3.13.4 of the present report. Horszczaruk’s study focused on 83 to 100 MPa (12 ksi to 14.5 ksi) compressive strength concretes containing basalt aggregates (2.5 to 12.7 mm (0.1 to 0.5 inch) diameter) and natural river sands (less than 2.5 mm (0.1 inch) diameter). The testing followed ASTM C944, except that the duration of test was increased from 2 to 40 minutes to allow for differentiation between concretes.

The relevant results from this study include the following. The linear best-fit approximation of the concrete mass loss per 2-minute abrading cycle ranged from 0.14 to 0.78 grams (0.005 to 0.027 ounce). Of the 10 concretes tested, six of them ranged from 0.14 to 0.25 grams (0.005 to 0.009 ounce). Horszczaruk also indicates that the rate of mass loss was relatively consistent throughout the abrading, with no clear increased abrasion resistance during the abrading of the smooth exterior face of the concrete.

Previous    Table of Contents    Next


Federal Highway Administration | 1200 New Jersey Avenue, SE | Washington, DC 20590 | 202-366-4000
Turner-Fairbank Highway Research Center | 6300 Georgetown Pike | McLean, VA | 22101