U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000


Skip to content
Facebook iconYouTube iconTwitter iconFlickr iconLinkedInInstagram

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

Report
This report is an archived publication and may contain dated technical, contact, and link information

Federal Highway Administration > Publications > Research > Structures > Critical Literature Review of High-Performance Corrosion Reinforcements in Concrete Bridge Applications

Publication Number: FHWA-HRT-4-093
Date: July 2004

Previous | Table of Contents | Next

Chapter 3. EXPERIMENTAL METHODS AND FINDINGS

Background

Experimental techniques that historically have been employed to investigate corrosion phenomena associated with reinforcing steel in concrete (both plain carbon and corrosion resistant), have involved the following categories:

  1. Laboratory studies in synthetic aqueous solutions.
  2. Laboratory studies with cementiteous embedments.
  3. Test yard and field exposure of concrete specimens.
  4. Actual structures.

Typically, selecting one method over another recognizes a tradeoff between time and cost versus realistic simulation of actual service with each of these three factors increasing from 1-4. Also at issue is whether a specific method provides design relevant data (time-to-corrosion (ti, see figure 2) and Cl- threshold, for example) or simply information that facilitates ranking of materials. Thus, experiments in category 1 tend to focus on determining the interrelationship between Øcrit, pH, and Cl- concentration (see figure 7) using aqueous solutions that simulate pore water and employing either potentiodynamic anodic polarization or potentiostatic polarization to a relatively positive potential (or both). Methods in categories 2 and 3 use either mortar or concrete-coated specimens, often with relatively high w/c's, low cover, and admixed Cl- and measurement of corrosion potential or corrosion rate by polarization resistance or macrocell current determinations (or both). Time-to-corrosion can also be determined. Category 4 does not represent a specific experimental technique but rather a demonstration or prototype structure. Studies in each of these four areas are reviewed and discussed below.

Laboratory Studies in Synthetic Aqueous Solutions

McDonald et al. performed a comprehensive screening program on a wide variety of CRRs, including stainless steels, as a prelude to testing yard-type exposures.[33] This involved successive 1.75 hours wet-4.25 hours dry exposure in NaCl solutions of pH 7 and 13 of bars that had been bent about a mandrel of diameter 4 times that of the bars. The pH 7 solution simulated preconstruction atmospheric exposure or conditions within a concrete crack; the pH 13 solution simulated concrete pore water. Table 3 lists polarization resistance (PR) and corrosion rate (CR) results for the pH 7 case, and figure 8 shows the corrosion rate data graphically. These indicate that corrosion rate for the stainless steels varied by as much as an order of magnitude, with most data being in the range 0.1-1.0 mille-Amps per meter squared, depending on the specific alloy, or from 2-3 orders of magnitude less than for plain carbon steel.

Table 3. Polarization resistance and corrosion rate data for CRR candidates in accelerated screening tests at pH 7.
   

pH 7 + 3 w/o NaCl

Reinforcement Type

Bar Cond***

28 Days

56 Days

90 Days

PR, Ohm·m2

CR, mA/m2*

PR, Ohm·m2

CR, mA/m2*

PR, Ohm·m2

CR,·mA/m2*

Plain Carbon (Black)

As-Recorded

0.05

520

0.14

186

0.05

520

Hole

0.04

650

0.05

520

0.04

650

Type 304

As-Recorded

100.40

0.26

52.42

0.50

90.96

0.29

Hole

84.87

0.31

133.50

0.19

121.47

0.21

Type 304U**

As-Recorded

12.57

2.07

11.88

2.19

-

-

Hole

6.03

4.31

9.70

2.68

9.17

2.84

Type 304N

As-Recorded

43.10

0.60

59.10

0.44

76.6

0.34

Hole

52.70

0.49

67.63

0.38

124.88

0.21

Type 304 Clad

As-Recorded

93.41

0.28

81.21

0.32

186.77

0.14

Hole

10.60

2.45

12.33

2.11

12.73

2.04

Abrade

54.38

0.48

49.57

0.52

90.69

0.29

Nitronic 33TM

As-Recorded

38.53

0.67

57.11

0.46

97.63

0.27

Hole

289.30

0.09

89.73

0.29

442.4

0.06

Type 316

As-Recorded

48.67

0.53

21.57

1.21

29.78

0.87

Hole

56.36

0.46

135.20

0.19

130.5

0.20

Type 317LN

As-Recorded

107.90

0.24

131.20

0.20

179.1

0.15

Hole

37.20

0.70

93.88

0.28

168.34

0.15

Type XM19

As-Recorded

133.90

0.19

115.40

0.23

350

0.07

Hole

125.70

0.21

248.20

0.10

466.3

0.06

* Calculated as CR = 0.026*1000/PR.
** European source.
*** Six mm diameter hole (through cladding in the case of clad bars).

Similarly, table 4 and figure 9 show comparable data for a 0.30N KOH+0.05N NaOH solution (pH Approximately equal 13). Salt concentration as NaCl for each of 3 successive 56 day test periods was 3, 9, and 15 w/o, respectively. These yielded Cl--OH- ratios of 1.5, 4.5, and 7.5, respectively. The more corrosion-resistant materials were not tested in the higher pH solution. In all cases, corrosion rates were lower at pH 13 than at pH 7 by an amount that approached 1 order of magnitude.

As an alternative approach, Hurley and Scully performed potentiostatic exposures at potentials as high as +200 mVSCE on Types 304 and 316 stainless steel reinforcements in a saturated Ca(OH)2 solution with incremental Cl- as NaCl additions.[36] The threshold concentration, cth, represented as [Cl-]/[OH-] (molar basis), was related to a current density increase. Figure 10 shows their results as a plot of the [Cl-]/[OH-] molar threshold ratio as a function of potential.

Figure 8. Graphical representation of accelerated screening test data at PH 7. Bar chart. Corrosion rates of black bar and Type 304, 304 Clad (6 millimeters hole through cladding), 316, and 317LN, Nitronic 33, and XM19 stainless steels in PH 7 solution after 28, 56, and 90 days with chloride concentration 3, 9, and 15 weight percent, respectively. Corrosion rate for the bare steel is in the range 10 raised to the second power to 10 raised to the third power milli-Amperes per meter squared and for the stainless steels 0.1-1.0 milli-Amperes per meter squared.
Figure 8. Graphical representation of accelerated screening test data at pH 7.

 

Table 4. Polarization resistance and corrosion rate data for CRR candidates in accelerated screening tests at pH 13.
   

pH 13 + NaCl

Reinforcement Type

Bar Cond.**

First 56 Days

Second 56 Days

Third 56 Days

PR,·Ohm·m2

CR, mA/m2*

PR, Ohm·m2

CR, mA/m2*

PR, Ohm·m2

CR,·mA/m2*

Plain Carbon (Black)

As-Recorded

1.58

16

0.31

84

0.25

104

Hole

0.77

34

0.37

70

0.33

79

Type 304

As-Recorded

60.60

0.43

147.90

0.18

-

-

Hole

40.31

0.65

121.70

0.21

-

-

Type 304 Clad

As-Recorded

68.91

0.38

32.67

0.80

15.26

1.70

Hole

231.20

0.11

11.75

2.21

-

-

Abrade

68.33

0.38

-

-

67.8

0.38

Nitronic 33

As-Recorded

108.90

0.24

194.10

0.13

86.12

0.30

Hole

-

-

54.60

0.48

222.4

0.12

Type 316

As-Recorded

66.37

0.39

96.10

0.27

-

-

Hole

95.60

0.27

116.90

0.22

-

-

* Calculated as CR = 0.026*1000/PR.
** Six mm diameter hole (through cladding in the case of clad bars).

These authors also reported corrosion potential for various reinforcements based on tests in simulated pore water solutions and in Cl- contaminated concrete, as shown in figure 10, and reasoned that the +200 mVSCE control potential (figure 11) was an upper limit that should not be exceeded in service. As such, the thresholds at +200 mVSCE were projected as being conservative.

Comparable results were obtained by Bertolini et al. based upon both potentiodynamic polarization scans and potentiostatic tests at +200 mVSCE in solutions that simulated both alkaline and carbonated pore water.[37] Figure 12 reports their results as a listing of critical Cl- concentration for the different alloys that were investigated. A correspondence between cth and PRE/PREN is apparent, with the exception of the low carbon Type 304 and 316 stainless steels, for which relatively low thresholds are apparent. The reason for this is unclear.

Figure 9. Graphical representation of accelerated screening test data at PH 13. Bar chart. Corrosion rates of black bar and Type 304, 304 Clad (6 millimeters hole through cladding), 316, and Nitronic 33, stainless steels in PH 13 solution after successive 56-day periods with chloride concentrations of 3, 9, and 15 weight percent, respectively. Corrosion rate for the bare steel is in the range 10 to 10 raised to the second milli-Amperes per meter squared and for the stainless steels 0.1 to 1.0 milli-Amperes per meter squared.
Figure 9. Graphical representation of accelerated screening test data at pH 13.

 

Figure 10. Most positive corrosion potential for different alloys in solution and in concrete. Bar chart. Bar chart listing maximum corrosion potential for various reinforcement types in saturated CA(OH) subscript 2 solution with 0.1 molar NA CL and sand and glass beads and in chloride contaminated concrete. Values for Type 316 stainless steels (solid and clad) are in the range negative 0.07 to negative 0.23 volts (SCE and for black steel negative 0.49 to negative 0.59 volts (SCE).
Figure 10. Most positive corrosion potential for different alloys in solution and in concrete.

 

Figure 11. Threshold CL negative to OH negative ratio as a function of potential for stainless and carbon steels. Graph. Plot of CL negative to OH negative molar threshold ratio as a function of potential for black steel and clad and solid Type 316 stainless steel. The threshold ratio for Type 316 stainless steel exceeded the solubility limit of chlorides (CL negative to OH negative ratio is approximately equal to 100) for potentials more negative than 0 volts (SCE) but decreased almost an order of magnitude at positive 0.20 volts (SCE). The molar threshold ratio for black steel transitioned from about one at potentials more negative than negative 0.20 volts (SCE) to about 0.1 volts (SCE) at potentials positive to 0 volts (SCE).
Figure 11. Threshold Cl--OH- ratio as a function of potential for stainless and carbon steels.

 

Figure 12. Threshold CL negative concentration for different reinforcement types in aqueous solutions of different PH. Graph. Listing of threshold chloride concentrations for black steel and Types 410, 304, 304L, 316, 23CR4NI, and SMO stainless steels in aqueous solutions of PH 7.5, 9.0, 12.6, and 13.9. The threshold increased with increasing PH and, in general, increased in direct proportion to the pitting resistance equivalent number.
Figure 12. Threshold Cl- concentration for different reinforcement types in aqueous solutions of different pH.

Laboratory Studies with Cementitious Embedments

Sorensen et al. performed potentiostatic exposures at 0 to +200 mVSCE using mortar prisms with Type 304 and 316 reinforcement and 0-8 percent admixed Cl- exposed in saturated Ca(OH)2 solution with 1 M NaCl.[38] Figure 13 shows the results for Type 304 stainless, both welded and unwelded, as a plot of anodic current density versus admixed Cl- concentration. The passive current density was in the range 10-1-10-2 mA/m2 for the unwelded stainless, or about 1 order of magnitude less than in aqueous solutions (see figures 8 and 9). The greater susceptibility of welded reinforcement to corrosion was attributed to air voids being entrapped at the irregular weld profile and to surface oxides. Enhanced surface roughness may also have been a factor.[39] Figure 14 presents data for these same specimens in a time-to-corrosion format.

Figure 13. Current density as a function of admixed chloride concentration for mortar specimens potentiostatically polarized at 0 millivolts (SCE). Graph. Plot showing that the current density to anodically polarize black steel in mortar specimens to 0 millivolts (SCE) increased from 0.5 to 33 milli-Amperes per meter squared as admixed chloride concentration was increased from 0 to 0.35 weight percent cement. Current density for companion Type 304 welded reinforcement was in the range 0.04 to 0.25 milli-Amperes per meter squared until admixed chlorides were 3.5 to 4.9 weight percent, beyond which it increased sharply. For nonwelded Type 304, current density remained in this same range to a chloride concentration greater than 5 weight percent.
Figure 13. Current density as a function of admixed chloride concentration for mortar specimens potentiostatically polarized at 0 mVSCE.

Reduced performance for welded versus unwelded stainless steel was confirmed by experiments by Nürnberger et al., who performed preliminary potentiodynamic and subsequent, longer term potentiostatic tests upon both carbonated and uncarbonated mortar-coated austenitic and ferritic stainless steel specimens (welded and unwelded) as a function of admixed Cl- concentration.39 Figure 15 shows pitting potential data for their unwelded bars, including results for carbon steel specimens.

Figure 14. Time-to-corrosion as a function of admixed CL negative concentration. Graph. Plot of time-to-corrosion as a function of admixed chloride concentration in mortar specimens reinforced with black steel and welded and nonwelded Type 304 and 316 stainless steel. The data generally exhibit a directly proportional relationship with black steel in the range 50 to 110 days time to corrosion for specimens with 0.1 to 0.4 percent chlorides by mortar weight and with the unwelded Type 304 stainless steel extending to about 300 days at almost 0.8 percent chlorides. A number of the intermediate chloride concentration data were runouts.
Figure 14. Time-to-corrosion as a function of admixed Cl- concentration.

A = Austenitic
F = Ferritic
CS = Carbon Steel

Figure 15. Critical pitting potential as a function of admixed CL negative concentration for stainless and carbon steel specimens in both carbonated (C) and uncarbonated (UC) mortar. The number in the caption indicates the PREN for each alloy. Graph. Plot of the critical pitting potential of an austenitic and ferritic stainless steel (PREN 23.6 and 17, respectively) and black steel in both carbonated and uncarbonated mortar as a function of chloride concentration. For the black steel, the critical potential was about positive 500 millivolts (SCE) in uncarbonated and negative 450 millivolts (SCE) in carbonated, chloride-free mortar, but the value decreased to about negative 500 millivolts (SCE) for both materials with addition of 1 percent chloride by cement weight and changed little with further chloride additions. For the ferritic stainless, the decrease was more gradual and moderated as chlorides increased with the critical pitting potential being about 0 and negative 250 millivolts (SCE) for the uncarbonated and carbonated specimens, respectively, at five percent chlorides. The potential decrease was even more gradual for the austenitic stainless and reached values of approximately positive 400 and positive 150 millivolts (SCE) at five percent chlorides.
Figure 15. Critical pitting potential as a function of admixed Cl- concentration for stainless and carbon steel specimens in both carbonated (C) and uncarbonated (UC) mortar. The number in the caption indicates the PREN for each alloy.

Likewise, figure 16 compares these results with those for the welded condition. The trend is one where welding caused a relatively large reduction in Øcrit for the austenitic material but did not significantly impair performance for the ferritic or carbon steel specimens. This distinction presumably was a consequence of pitting susceptibility being relatively high in the ferritic and carbon steel, irrespective of whether or not these were welded.

Figure 16. Comparison of C subscript TH for welded and unwelded stainless and carbon steel specimens in carbonated and uncarbonated mortar. The number in the caption is the PREN. Where no carbon steel data are indicated, C subscript TH was zero. Graph. Plot showing that the critical chloride concentration for the austenitic steel in figure 15 decreased from about 5 to 1 weight percent chlorides (cement basis) upon welding. The thresholds were slightly lower in carbonated compared to uncarbonated mortar. The threshold for the unwelded ferritic steel was less than 1 percent in even the uncarbonated mortar and was less for both the welded and carbonated conditions. Likewise, the threshold was less than 0.5 percent for both welded and unwelded black steel in uncarbonated mortar and zero in both conditions for carbonated mortar.
Figure 16. Comparison of cth for welded and unwelded stainless and carbon steel specimens in carbonated and uncarbonated mortar. The number in the caption is the PREN. Where no carbon steel data are indicated, cth was zero.

Figure 17 attempts to interrelate cth data as reported by Sorensen et al. on a w/o cement basis for mortar specimens to [Cl-]/[OH-] assuming a 2:1 sand-to-cement ratio, 10 percent pore water, and OH- activity coefficient 0.7.[42] Also provided is the range of these two parameters (cth and [Cl-]/[OH-]) that has been reported historically for carbon steel in aqueous solutions or mortars (or both). The fact that the Sorensen et al. data include multiple values for the same material reflects the fact that different constant potentials were employed and cth decreased as potential was made more positive. The data suggest that cth for the stainless steel averaged about one order of magnitude greater than for carbon steel on a w/o cement basis, and from 1-2 orders of magnitude greater in terms of [Cl-]/[OH-].

Figure 17. Attempted cross correlation of CL negative threshold on a cement weight percent and on a CL negative to OH negative basis. Graph. This figure plots chloride threshold as reported by Sorensen et al. for mortar specimens on a cement weight basis to CL negative to OH negative assuming a 2.1 sand-to-cement ratio, 10 percent pore water, and OH negative activity coefficient 0.7. The data conform to a power law relationship.
Figure 17. Attempted cross correlation of Cl- threshold on a cement w/o and on a [Cl-]/[OH-] basis.

Test Yard-Type Exposures

Exposures in this category tend to span the gap between aqueous solution laboratory test cell experiments and actual field service structures. The accelerating features typically include relatively high w/c, low concrete cover, and frequent application of a concentrated salt solution under successive wetting and drying conditions.

Treadaway[40] and Treadaway et al.[41]] exposed concrete prisms reinforced with carbon steel, Types 405 and 430 (ferritic) stainless steels, and Types 304, 315, and 316 stainless steels outdoors in the United Kingdom for 10 years. Two mix designs were employed, one with an 8:1 aggregate:cement ratio, cement content 220 kg/m3, and w/c 0.75, and the second with parameters of 6:1, 290 kg/m3, and 0.60, respectively. Concrete cover was either 10 or 20 mm. Admixed Cl- concentration was as high as 3.2 percent (cement weight basis).

Figure 18 plots bar weight loss on an annual percentage basis versus admixed Cl- concentration for specimens with bars at both depths. This indicates that weight loss was highest for carbon steel and lowest for the Type 316 stainless steels. The other austenitics performed comparably. In general, corrosion rate increased with increasing Cl-, the effect being greatest for the carbon steel and minimal for the austenitics. In addition, corrosion intensity was greater for the lower cover steel and increased in proportion to the admixed Cl- concentration above 0.96 w/o. No explanation is apparent as to why the carbon steel exhibited higher corrosion rate with 0 compared to 0.32 and 0.96 w/o admixed Cl-. The ferritics performed satisfactorily at the lower Cl- concentrations, but suffered severe pitting at the higher concentrations.

Figure 18. Weight loss of different reinforcements during a 10-year United Kingdom exposure. Bar chart. Plot showing annual weight loss for 10-year exposure of concrete specimens with admixed chlorides for carbon steel and Types 430 and 316 stainless steels. The highest corrosion rate occurred for the carbon steel reinforcement followed by the 430 and, lastly, 316. In all cases, the rate was higher for specimens with 10 millimeters concrete cover than for ones with 20 millimeters cover. Also, corrosion rate for all reinforcement types was relatively low for admixed chloride concentrations of 0.96 weight percent concrete but increased in proportion to concentration above this.
Figure 18. Weight loss of different reinforcements during a 10-year United Kingdom exposure.

Corrosion resistance of a ferritic stainless steels was also investigated by Callaghan and Hearn who exposed 12 w/o Cr reinforcement with 12 and 25 mm cover in relatively poor quality concrete prisms (no admixed Cl-) to a severe marine environment for 4.5 years.[42] Severe pitting occurred on reinforcement in companion black bar specimens, whereas attack on the stainless steels was minimal. The authors reasoned that the good performance of the ferritic stainless steels in their exposures, compared to that of Treadaway et al., resulted from chlorides not being admixed such that a more protective passive film formed.[41] It was concluded that such stainless steels may be the best choice for moderately aggressive environments.

McDonald et al. performed a series of exposures using double matted concrete slab specimens with a range of reinforcement types that included Types 304 and 316 stainless steels.[43] The average cement content and w/c were 370 kg/m3 and 0.47, respectively, and clear cover was 25 mm. The exposures involved 4 days wetting with 15 w/o NaCl at 16-27 ºC and 3 days drying at 38 ºC for 12 weeks, followed by 12 weeks of continuous ponding. The cycle then was repeated.

Average macrocell current density (anodic) on straight black bar control specimens during the 96-week tests, which was determined as the voltage drop across a 10 O resistor between the top and bottom bars, was 27.6 mA/m2, or 0.03 mm/year.[1] All specimens were cracked after 48 weeks. For specimens with straight and bent Type 304 SS reinforcement in both mats and with both cracked and uncracked concrete, macrocell current density was 0.016-0.039 mA/m2, or 0.0002-0.0005 mm/year. However, for sound slabs with black bottom bars, this current density was 8.8 mA/m2 (0.01 mm/year), and minor rust staining was apparent on the concrete surface. For cracked slabs with black bottom bars, current density was even higher at 20.8 mA/m2 (0.02 mm/year). No surface staining was noted on this last specimen type, nor for any Type 304 stainless steel reinforcements.

That Type 304 SS top reinforcement exhibited corrosion when coupled to black bottom bars is surprising, because it requires potential of the black bars minus voltage drop between the two mats to be more positive than Øcrit of the stainless steel. This finding contrasts with the results of Hope, who performed dual compartment aqueous exposures of stainless steel-black steel couples and concluded that these two reinforcement types could be mixed, provided concrete surrounding the black bars remained Cl--free.[44] Even if chlorides are present here, the black bars should corrode and provide galvanic cathodic protection to the stainless steel.

In contrast, macrocell corrosion rate for slabs with Type 316 stainless steel reinforcement was in the range 0.039-0.070 mA/m2 (0.0005-0.0009 mm/year), irrespective of the presence of a concrete crack or a black bar bottom mat. No rust staining or concrete cracking was noted.

Clemeña and Virmani reported results for 0.5 w/c concrete slabs that were exposed outdoors in Virginia to successive 3 days wet-4 days dry cycles using a saturated NaCl solution for 700 days.[45] Reinforcement types, which were both bent and straight, included carbon steel, Types 304, 316LN, and 2205 stainless steels, and Type 316 stainless steel clad bars. Clear cover over the reinforcement was 25 mm. Some of the clad bars contained two intentionally drilled 3 mm diameter holes that extended to the carbon steel core. Slabs for which the top bars were stainless steel and bottom bars were carbon steel were included. Researchers came to the following conclusions:

  1. In all cases, potential became progressively more negative as exposure time increased, and after 700 days was approximately -550 mVCSE (-480 mVSCE) for the carbon steel and -300 mVCSE (-230 mVSCE) for the stainless steels. The carbon steel slabs exhibited corrosion and concrete cracking, whereas no damage was apparent for the stainless steel slabs.
  2. The mean macrocell current density for slabs with carbon steel top bars averaged 6.36 mA/m2 (0.007 mm/year), whereas for the stainless steel specimens, this current was either zero or negative. Polarization resistance determinations showed corrosion rates to average 24.1 mA/m2 (0.03 mm/year) for the carbon steel and 0.8 mA/m2 (0.0009 mm/year) for the stainless steel, with relatively little difference between alloys in the carbon steel category. In addition, the general trend was that the carbon steel's corrosion rate increased with time, but this parameter was relatively constant for the stainless steels.
  3. Corrosion rate for the stainless clad reinforcement was essentially the same as that for the solid stainless bars. Also, no detrimental consequence of exposed core material at the intentional cladding defects, as indicated by potential, macrocell current, or polarization resistance, was apparent.

Rasheeduzzafar et al. also conducted outdoor exposures in eastern Saudi Arabia of concrete specimens reinforced with black, galvanized, ECR, and stainless clad reinforcing bars.[46] The mix design consisted of 390 kg/m3 Type V cement, w/c = 0.45, and admixed Cl- concentrations as NaCl of 2.4, 4.8, and 19.2 kg/m3. Clear cover was 25 mm. No details regarding the clad alloy were provided. After 7 years, the clad-reinforced slabs exhibited no indications of corrosion or cracking, whereas cracks were present in at least some of the slabs with each of the other reinforcement types.

Flint and Cox performed a series of seawater immersion and tidal exposures on concrete specimens with partly embedded Type 316 SS for up to 12.5 years.[47] This study focused on the possibility of crevice corrosion at the steel-concrete interface. Such attack was negligible but with some corrosion occurring at the steel-concrete-seawater interface, as has been reported to be particularly significant in the case of carbon steel reinforcement.[48]

Cross-Procedural Experiments

Several authors have conducted experimental programs that involved various combinations of aqueous solution, mortar-coated, and concrete embedded exposures. Among these are the tests of Darwin et al., who employed a two-compartment galvanic cell with simulated pore solution and mortar chunks in both compartments and 1.6 molal NaCl in one and no Cl- in the other.[49] In one set of tests, bare or uncoated specimens were employed; in a second set of tests, the specimens were mortar coated. In a third set, both sound and cracked macrocell-type concrete slab specimens were subjected to the Southern Exposure protocol.[50] Results were reported in terms of corrosion rate as determined from macrocell current, rather than cth, with the stated reason being that the latter parameter is achieved quickly in the vicinity of concrete cracks, such that all reinforcement types are likely to corrode at these locations. This rationale may be appropriate for some, but not all, types of reinforcement.

Figure 19 shows typical results for mortar-coated, aqueous solution exposed specimens during the 15-week period. Generally, corrosion rates were approximately constant after about week 10. On this basis, the long-term corrosion rate of black steel was highest. The rate for MMFXTM was approximately a factor of 2 less, while that for 2201 (unpickled) was less an additional factor of 2. In contrast, corrosion rates for 2201P (pickled), 2205, and 2205P (pickled) were typically 0.1 mm/year (0.086 mA/m2) or less. Also, although there was no significant distinction between the corrosion rates for 2205 and 2205P, the rate for 2201 exceeded that for 2201P.

Figure 19. Corrosion rate of various mortar-coated reinforcement types in an aqueous macrocell test arrangement. Graph. Plot of corrosion rate versus time to 15 weeks for 6 different types of mortar-coated reinforcements in a two compartment, aqueous solution macrocell corrosion test. As described in the text, the reinforcements graphed are black steel, MMFX, 2101, 2101P, 2205, and 2205P.
Figure 19. Corrosion rate of various mortar-coated reinforcement types in an aqueous macrocell test arrangement.

Figure 20 shows the average corrosion rate for the different reinforcement types during 48 weeks of Southern Exposure. These data are qualitatively consistent with those from the mortar-coated tests, although steady state appears to have been reached for the more active metals in the case of Southern Exposure; but it is unclear that this was always so for the mortar coated specimens. Corrosion rate for the stainless steels was consistently 2-3 orders of magnitude less than for the actively corroding reinforcements. Also, alloys that performed well apparently remained passive, irrespective of test condition.

Figure 20. Corrosion rate of various reinforcement types in concrete slabs undergoing Southern Exposure testing. Graph. Plot of corrosion rate to 48 weeks for 5 various reinforcement types undergoing Southern Exposure testing. The reinforcements graphed are black steel, MMFX, 2101P, 2205, and 2205P. Corrosion rate for carbon steel after 48 weeks approached 10 micro-meters per year and was about 2 micro-meters per year for MMFX. Corrosion rates for 2201P, 2205, and 2205P were scattered in the range 10 to the negative third to 10 to the negative one micro-meter per year.
Figure 20. Corrosion rate of various reinforcement types in concrete slabs undergoing Southern Exposure testing.

As noted above, there was a difference in performance for 2201 SS but not 2205 SS in the aqueous exposures, depending on surface treatment (as-received versus pickled). Unpickled specimens were not included in the Southern Exposure concrete specimens, so no such comparison is possible here. This distinction between specimens with the two preparation methods apparently resulted because surface penetrating nonmetallic inclusions or iron/steel particles from handling equipment became embedded and corroded (or both). This leaves surface irregularities that, for metals with relatively low passivation tendency in a sufficiently high Cl- concentration electrolyte, can result in pit initiation according to the mechanism described above.

Figure 21 compares corrosion rates for the different environments investigated by Darwin et al. and indicates that these increased generally in proportion to the anticipated conductivity of the electrolyte (highest corrosion rates in the aqueous solution and lowest in sound concrete).[53] Comparisons here should be made with caution, however, because bars with little or no tendency for passivation (black steel) were apparently active upon initial exposure, whereas bars with an intermediate passivation tendency apparently either activated slowly, or active corrosion spread progressively with time. On this basis, the low resistivity electrolyte test conditions with high initial Cl- failed to provide due credit for a corrosion initiation period, which may be extensive. As such, the aqueous exposures provide information on which different reinforcements can be ranked, but do not facilitate life-prediction modeling. This rationale disregards the presence of mortar or concrete cracks, however. If (1) chlorides accumulate rapidly at the base of these cracks, (2) the resulting corrosion initiates early in the structure's life cycle, and (3) this corrosion controls service life, then the data can be viewed as applicable.

Figure 21. Comparison of corrosion rates in the different environments (multiple listings of same alloy represent results for duplicate specimens). Bar chart. Listing of corrosion rate for black steel, MMFX, and Types 304 (clad), 2201, and 2205 (pickled and as-received). Rates for bare bars exceed those of mortar-coated bars and of bars in concrete subjected to Southern Exposure. Corrosion rate for cracked concrete Southern Exposure slabs exceeded those for uncracked ones.
Type of Steel
Figure 21. Comparison of corrosion rates in the different environments (multiple listings of same alloy represent results for duplicate specimens).

Specific Reinforcement Alloys

General

Data for several relatively unique corrosion-resistant alloys (in addition to the standard ferritic, austenitic, and duplex SSs) were represented in some of the assessments discussed above. These include Nitronic 33, MMFX, 3Cr12TM, and clad SS. Each of these is discussed in greater detail below.

Nitronic 33

Limited data for Nitronic 33, which conforms to American Society for Testing and Materials (ASTM) Specification A580, Grade XM-29, were reported above in conjunction with tables 3 and 4 and figure 8. In addition, Jenkins reported the results of laboratory aqueous exposures and field evaluations.[51] This alloy is of particular interest to the U.S. Navy because of its nonmagnetic properties in addition to relatively good corrosion resistance and high strength. Nominally, the composition of Nitronic 33 is 18 percent Cr, 12 percent Mn, and 3.5 percent Ni. The material tested by Jenkins had a yield strength of 800 MPa. Potentiodynamic polarization scans in aqueous solutions indicated a full range of passivity for pH = 12.1, 11.6, and 11.2 and Cl- concentrations of 0-6,000 parts per million (ppm). At pH 10.0, full passivity was exhibited at a Cl- concentration of 2,000 ppm, but at 6,000 ppm, Øcrit was +100 mV (reference electrode not stated). The program also involved evaluating driven piles that were placed at the Port of Tacoma, WA. No Nitronic 33 corrosion activity was detected, although age of the pilings was only 17 months when measurements were taken.

MMFX

Within the past several years, a proprietary alloy, initially designated as MMFX (subsequently MMFX-I and then MMFX-II), has been marketed as a CRR alternative for concrete bridge deck service. Composition is nominally that of low carbon steel but with 9-10 w/o Cr. As such, it does not meet the classification of stainless steel, because a minimum of 12 w/o Cr is required for this designation. Nonetheless, such a Cr amount may contribute to enhanced corrosion resistance compared to carbon steel. Mechanical strength of MMFX exceeds that of conventional reinforcement. For example, Darwin et al. determined the average 0.2 percent offset yield strength, tensile strength, and elongation for 5 specimens in each of 3 heats as 910 MPa, 1139 MPa, and 7.1 percent, respectively.[52]

Corrosion resistance of this steel is ascribed to a special thermomechanical treatment that yields a microstructure comprised of packet martensite and nanosheets of untransformed austenite. This is stated to reduce or eliminate microcells that otherwise cause corrosion. However, microcells invariably exist irrespective of microstructure. Also, it is unclear why microcells should necessarily be a factor in bridge deck service where corrosion, once initiated, is controlled predominantly by macrocells. It can be reasoned that the MMFX microstructure may have a reduced exchange current density for the oxygen or hydrogen reduction reactions (or for both), in which case reduced corrosion rate could result. Corrosion data for MMFX, as reported by Darwin et al., were indicated above in conjunction with figures 20 and 21, where a reduced rate compared to carbon steel by a factor of 2-3 is apparent.[53] In addition, Lopez conducted salt fog exposures of MMFX specimens to 1,000 hours using 5 w/o NaCl and a repetitive 1-hour fog (ambient temperature)-1-hour dry (35 ºC) cycle; the results are shown in figure 22.[53] While the data scatter is relatively large, corrosion rate of the MMFX was, on average, 44 percent less than for the carbon steel. It should be interesting to determine via parallel experiments if corrosion performance of MMFX in concrete service affords any advantage compared to 12 w/o Cr ferritic SSs (see below).

Figure 22. Corrosion rate of MMFX and carbon steel under cyclic salt fog exposure. Graph. Weight loss data for a pair of MMFX and carbon steel specimens under cyclic salt fog chamber exposure to 1,600 hours. The data are scattered but generally show corrosion rate of MMFX to be about one-half that of black steel.
Figure 22. Corrosion rate of MMFX and carbon steel under cyclic salt fog exposure.

Also of concern is the relatively high strength of MMFX compared to carbon steel reinforcement. Based on concrete beam tests, Ansley concluded that designers using MMFX reinforcement should consider detailing and the lack of a distinct yield point for this material.[54] Ansley also indicated that lap splices and hook elements that are adequate for Grade 60 reinforcement may be insufficient with MMFX.

Proprietary Ferritic Stainless Steels

Ferritic stainless steels, designated as 3Cr12 and CRS 100, with approximately 12 w/o Cr are being produced as reinforcement. The composition of these is essentially the same; the 3Cr12 conforms to European Standard EN 10088 Grade 1.4003. These are relatively low-cost alternatives to Types 304 and 316 and may prove suitable choices for moderate, although probably not severe, chloride exposures.

Clad Stainless Steel

Bars of this type are potentially available from two suppliers. Both employ Type 316 stainless steel as the cladding and innovative manufacturing technologies. One company packs a stainless tube with carbon steel scrap and hot rolls this into reinforcing bars. This production method allows any cladding alloy to be employed instead of 316 SS, with cost varying accordingly. The other company plasma-coats steel billets and then rolls these. The plasma method lacks the flexibility of the first method, because the technology is relatively insensitive to clad material cost.

Several of the evaluation programs noted above included clad SS reinforcement.[39, 48] As such, clad bars potentially afford the corrosion resistance of the SS cladding but at a reduced cost compared to solid SS bars. Concerns include corrosion at cut ends where the carbon steel core is exposed and corrosion at cladding defects. On one hand, the relatively small anode-to-cathode area ratio that is likely to accompany clad defects or exposed bar ends should facilitate such an attack. On the other hand, stainless steels serve as a relatively poor catalytic surface for the cathode reaction, as revealed by the experiments of Sorensen et al., who performed potentiodynamic polarization scans and determined that the current density at -700 mVSCE was 9 ±2 mA/m2 for Type 304 SS, 31 ±11 mA/m2 for Type 316, and 143 ±68 mA/m2 for carbon steel.42

Cui and Sagüés[55,56] and Clemeña and Virmani[49] both exposed concrete specimens containing SS clad reinforcement, where holes had been drilled intentionally through the cladding to the carbon steel core. In the Cui and Sagüés experiments, the concrete was admixed with either 5 or 8 w/o Cl- (cement weight basis) and exposed indoors under ambient laboratory conditions or at 40 ºC and 100 percent relative humidity. Clemeña's and Virmani's experiments were wet-dry cyclic ponded with a saturated NaCl solution, as noted above. Cui and Sagüés measured corrosion activity at the clad defect sites, although the magnitude of this tended to moderate with time in some cases. This finding that corrosion moderated with time was attributed to progressively developing corrosion products.

Companion numerical modeling results were in general agreement with the experimental findings, and predicted that corrosion rate should increase with increasing defect size and decrease with increasing concrete resistivity. For a 1 mm diameter clad break and 30 kO·cm concrete, corrosion rate was projected as approximately 14 mA/m2 (0.18 mm/year). Clemeña and Virmani, however, reported no electrochemical indications of corrosion on their clad bars with defects after 700 days' exposure.[49] This could have resulted because their specimens contained no admixed chlorides, such that time was provided before chlorides arrived at the steel depth for the passive film to stabilize. Clearly, more experimental information is needed before definitive conclusions can be reached regarding the importance of exposed carbon steel core material.

Actual Structures

While there has been only limited utilization of CRR in concrete construction to date, a small inventory of such structures does exist, nonetheless. These can provide information that can be used to develop a database of experience. A summary of these structures is provided below.

Progresso Pier

Probably the best example of stainless steel reinforced concrete structures in aggressive, Cl- environments is the 1,752-m long pier at the Port of Progreso de Castro in Yucatán, Mexico. Constructed between 1939-1941, the pier is now more than 60 years old. Approximately 200,000 kg of 30 mm diameter, nondeformed Type 304 SS was employed. The pier consists of 146 12-m span hinged arches, each of which is supported at the ends by reinforced girders positioned on massive nonreinforced concrete piles. Figure 23 shows a general photograph of the pier.

Several inspections and evaluations of the pier condition have been performed during the past decade.[57, 58] These have indicated no visible signs of deterioration, despite the fact that no maintenance apparently has been performed. Figure 24 shows a photograph of two arches and piling and reveals the generally good condition of the structure. It has been determined that the concrete is of relatively poor quality with a w/c of 0.50-0.70, porosity 19-24 percent, and resistivity 0.6-2.5 kO·cm. Based on acid soluble analyses of samples from two cores, Cl- concentration at the bar depth (78 mm) was determined as 1.2 w/o concrete. This, coupled with the tropical marine environment, constitutes severe exposure conditions for the reinforcement. In the most recent inspection, no corrosion stains or corrosion-induced cracks were observed along the entire structure.[61] However, stress corrosion cracking (SCC) was stated to have affected exposed bent end hooks. Occurrence of SCC in Type 304 SS in near-neutral or alkaline Cl- environments at ambient temperatures is unexpected and warrants further investigation. Also, severe localized corrosion was noted at a location where the reinforcement had been exposed previously. Figure 25 provides a photograph of this damage.

Figure 23. Perspective view of the Progresso pier. Photo. Perspective photograph of the stainless steel reinforced concrete pier at Progresso.
(photograph courtesy of Dr. E.I. Moreno)
Figure 23. Perspective view of the Progresso pier.

Detroit, MI (Interstate (I)-696 over Lenox Road)

As a part of this bridge construction project, Type 304 SS reinforcement was placed in the eastbound lanes in 1984, and ECR was used in the westbound lanes. During an inspection in 1993, cores were removed from the eastbound lanes to assess the concrete and reinforcement. Three of the cores contained 16 mm diameter reinforcement. No concrete cracking that could be attributed to corrosion was noted, neither on the bridge generally nor in the cores. Minor corrosion staining in conjunction with a noncorrosion-related crack was apparent on one of the extracted bar sections. Acid soluble Cl- determinations at the reinforcement depth (75-165 mm) indicated values of 0.54, 0.26, and 0.22 w/o cement (0.078, 0.037, and 0.032 w/o concrete) for the three cores, which is near the threshold for black steel. The inspection and analysis results are described in greater detail elsewhere.[59]]

Figure 24. Photograph of hinged arches and piling. Photo. Photograph of a typical hinged arch and piling on the Progresso pier.
Figure 24. Photograph of hinged arches and piling.
Figure 25. Corrosion of an exposed reinforcing bar. Photo. Example of severe localized corrosion on an exposed stainless steel reinforcing bar on the Progresso pier.
Figure 25. Corrosion of an exposed reinforcing bar.

Trenton, NJ (I-295 over Arena Drive)

The northbound and southbound bridges for this project were constructed in 1983-1984 using ECR in the northbound bridge and Type 304 clad SS from a source in England in the southbound bridge. The design employed steel girders, stay-in-place metal decking, a reinforced concrete bridge deck, and a 25-37 mm latex-modified concrete overlay. During an inspection in 1993, minor delamination of the overlay was detected in some locations.[59] Four cores, two from delaminated and two from sound locations, containing a total of nine clad SS segments, were acquired. No corrosion was apparent on any of these, except beneath a plastic capped cut bar end where the carbon steel core was exposed. This attack was attributed to a low pH environment that developed here, such that passivity was not maintained. No adhesive had been used in conjunction with placing the cap. Acid soluble Cl- concentration at the steel depth (50-62 mm including overlay) was low and in the range 0.009-0.013 w/o concrete.

Ontario, Canada

In 1996, a 21-m long, single-span bridge was constructed as a demonstration project on Highway 407 over Mullet Creek in Ontario, Canada. The two-lane structure consists of a 235-mm thick concrete deck slab reinforced with 11,000 kg of Type 316LN SS in both mats at a design cover for the upper mat of 80 ±20 mm on prestressed concrete I-beams. Probes were installed for corrosion monitoring purposes. The SS was shipped in coils that subsequently were straightened using conventional carbon steel equipment. This resulted in iron embeddments that were visually unappealing but did not compromise performance. Inspections performed within the first year of construction revealed corrosion potentials for the reinforcement in the passive range (-0.26 VCSE with standard deviation 0.06). Minor shrinkage cracks were disclosed on parapet walls but with none on the deck.[60]

In 1998, the Ontario Ministry of Transportation (OMT) established a policy that the top mat of 400 series bridges (100,000 or more vehicles per day) and all barrier walls would be stainless steel. Requiring this for the top mat only was based on a research study that showed no occurrence of corrosion for black steel electrically connected to Types 316LN or 2205 stainless steels, as long as chlorides were not present at the black steel.[61]

In conjunction with the above policy, OMT constructed a two-lane, three-span, continuous 225-mm thick by 37.5-m long concrete deck on galvanized steel plate girders bridge on Highway 9 over the South Holland Canal in Ontario, Canada, in 1999 using Type 316L stainless steel clad bars as the top mat. Despite bar quality control and delivery schedule issues, researchers concluded that the use of this reinforcement type is viable.[62]


Footnote

[1] This review employs the units used by the original authors. The following conversions may be used for converting between the different units:
     1 mA/m2 = 0.1 µA/cm2 = 0.011 mm/year = 11.5 µm/year = 0.43 mils/year.

 

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