U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
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-09-020
Date: April 2009
Corrosion Resistant Alloys for Reinforced Concrete
Two specimen types, SDS and MS, which were intended to represent a reinforced concrete bridge deck exposed to deicing salts, and two other specimen types, 3BTC and FC, which represented a marine substructure element, were exposed to chlorides. Reinforcements included stainless steels 316 (UNS-S31603), 304 (UNS-S30400), 2304 (UNS-S32304), 2101 (ASTM A955-98), and 3Cr12 (UNS-S41003); two types of 316 clad BB, AASHTO MP 13M/MP 13-04, and MMFX-2 (ASTM A1035); and BB (ASTM A615), the last being for baseline comparison purposes. Bars were cast into concrete specimens in the as-received condition, which was either as-rolled or pickled depending upon the source, after solvent cleaning. Specimen configurations included the following:
These were in addition to a standard specimen for which all bars were straight, and none of the above conditions were present (standard condition for MS specimens was with bars wire brushed). Three concrete mix designs, termed STD1 (high permeability), STD2 (intermediate permeability), and STD3 (permeability between that of STD1 and STD2) were employed. All specimens were tested outdoors where the SDS and FC were fully exposed and the MS and 3BTC were sheltered. A second set of MS specimens was tested under controlled temperature and relative humidity (25 oC and 50 percent).
Based on the study, several conclusions were reached. The reinforcements, other than BB, were classified into two groups as either improved performance or high performance where alloys in the former category initiated corrosion during the project time frame, and ones in the latter did not, at least in cases for specimens of the standard configuration (STD-all straight bars in the as-received condition without crevices and no simulated concrete cracks). Improved performers were 3Cr12, MMFX-2, and 2101 (BB-reinforced specimens were included in this grouping also for reference purposes). The other alloys were high performers. These alloys ranked according to time for corrosion to initiate as BB < 2101 < 3Cr12, MMFX-2.
No SDS specimens with 304 and clad or solid 316 bar, besides those with BB lower layer, initiated corrosion, and no sustained macrocell currents were detected during the exposure period. Test times for the 304 reinforced specimens were as long as 440 days and 1,726 days for the 316 reinforced specimens. The MS and 3BTC specimens with these reinforcements exhibited both anodic and cathodic macrocell current "spikes" to as high as 16 μ A for 316 and 0.8 μ A for 304. However, net mass loss associated with these was calculated as 0 μ A. In general, macrocell current activity was less for 304 than for 316, which is in contrast to the normally perceived better corrosion resistance of the latter alloy. Also, macrocell currents were less for specimens in the higher quality concrete (STD2) compared to the lower (STD1). Corrosion potentials remained relatively positive, and the macrocell current activity was not considered indicative of corrosion initiation.
For improved performance and BB reinforcements, Ti and CT were distributed over a range rather than being a discrete value. Chloride threshold for corrosion initiation of 3Cr12 and MMFX-2 reinforced SDS specimens was about four times greater than for BB specimens and slightly less than four times greater in the case of 2101 specimens. Weibull analysis of the [Cl-] data indicated that at 2 percent of bars being active, CT for BB was 1.0 kg/m3 or 0.36 wt percent cement and 4.0 kg/m3 or 1.44 wt percent cement for 3Cr12 and MMFX. On the other hand, there was little difference in Ti between each of the three improved performance reinforcements compared to BB for STD1 MS specimens. For STD2 MS specimens, however, Ti for MMFX-2 and 2101 was from 3.4 to more than 5.7 times greater than for BB (limited data precluded this determination for 3Cr12). For 3BTC specimens, this ratio for these same two alloys (MMFX-2 and 2101) in STD2 concrete was from 2.1 to 5.2. The results imply that the enhanced corrosion resistance that is derived from these reinforcements relative to BB increases with increasing concrete quality.
Time-to-corrosion for STD1 MS specimens was shorter than for SDS specimens, and macrocell currents for the former were an order of magnitude or more greater than for the latter. Apparently, the MS type specimens and exposure conditions provided a more severe testing of the reinforcements than the SDS specimens. The finding that Ti for improved performance bars was about the same as for BB in STD1 MS specimens but about four times greater in SDS ones may have resulted because of this.
The MS specimens exposed outdoors exhibited shorter Ti and greater macrocell current activity than identical ones tested at constant temperature and relative humidity. The former condition is thought to have fostered a higher level of sorptive moisture and Cl- transport, such that the corrosion threshold was reached in a shorter time.
The FC specimens, which were exposed in the tidal zone on the Intracoastal Waterway at Crescent Beach, FL, with improved performance and BB reinforcements, typically initiated corrosion within the first several days. This is thought to have resulted because of poor concrete quality and possible cracks that provided direct water access to the reinforcement. With one exception, the high performance reinforcements have remained passive. The exception was a 316 reinforced column that was damaged during installation.
A ranking of these CRR based on Cl- threshold failed to correlate with results from previously performed short-term potentiostatic tests in synthetic pore solution to which chlorides were incrementally added. This calls into question the usefulness of this and perhaps other accelerated test methods for evaluating corrosion resistant alloys for service as reinforcements in concrete.
In specimens with lower layer BB, corrosion of the BB was often extensive and to the point that delaminations occurred along the plane of these bars. If a corrosion-resistant steel upper bar layer is to be combined with a BB lower one, the concrete should be of sufficient quality and with limited or no cracks such that Cl- concentration at the lower bars remains below the black steel threshold for the design life of the structure.
Corrosion occurred at defects and unprotected embedded bar ends for stainless clad reinforcements. The likelihood that this attack or corrosion at clad defects will ultimately cause concrete cracking and spalling is uncertain, but it depends on corrosion morphology, geometry, and rate factors. Corrosion-induced concrete cracking and spalling probably result in time if the surface area of the exposed core carbon steel and extent of attack exceed a certain value. Further research is needed to define this threshold.
Life-cycle cost analyses for CRR should consider not only differences in CT and macrocell current, but also the possibility that concrete cover can be reduced. This, in turn, could lower superstructure and, hence, substructure size, weight, and initial cost, accordingly. Also, lower cover may reduce the number and width of concrete cracks, leading to less corrosion of top bars and lower maintenance costs in the long term.
An example analysis was performed that calculated Ti of a concrete structure reinforced first with BB and second, with an improved performance rebar. The calculation was based on an effective Cl- diffusion coefficient of 10-12 m2/s, concrete cover 63 mm, surface [Cl-] 18 kg/m3, and CT for the corrosion resistant alloy as four times greater than for BB. The analysis yielded Ti for BB as ranging from 17 years to 24 years and for the corrosion resistant alloy from 43 years to 86 years as the percentage of bars being active increased from 2 percent to 20 percent. Limitations of this analysis are that, first, the above input parameters for the calculation are mean values, whereas these are, in fact, distributed such that corrosion will initiate at some locations sooner than projected. Second, enhanced inward Cl- migration along any concrete cracks was not considered. Further, corrosion should initiate sooner at concrete corners because inward Cl- diffusion is from two directions here rather than just one. On the other hand, the above difference in CT for black compared to the improved performance bar was based on data acquired from highly permeable concrete. However, this difference is expected to be greater for concrete with a Cl- diffusion coefficient of 10-12 m2/s.
The CT and macrocell current data indicate that the intended service life of major reinforced concrete bridge structures (75-100 years) can confidently be achieved with the solid high performance reinforcements that were investigated. This may be the case also for the clad reinforcements, provided there is adequate control of surface defects and bar ends are protected. This same service life may also result with the improved performance bars, provided design and construction quality control are good, and concrete cracking is minimal but with a lesser degree of confidence and margin for error compared to the high performance reinforcement.