Multiple Corrosion-Protection Systems for Reinforced Concrete Bridge Components

CHAPTER 4. EVALUATION

In this chapter, the experimental results presented in chapter 3 are compared and used to develop estimates of the life expectancy and cost effectiveness of the systems.

The comparative performance of the systems is measured based on the chloride content required to initiate corrosion, the average corrosion rate based on corrosion losses after corrosion initiation, and, in the case of coated bars, disbondment of the coating from the surface of the reinforcement.

Estimates of life expectancy and cost effectiveness are determined using a typical monolithic bridge deck with a thickness of 216 mm (8.5 inches) and concrete cover of 76 mm (3 inches) over the top layer of reinforcing steel. Life expectancy depends on the time to first repair. Cost effectiveness is based on the sum of the cost of a new bridge deck and the cost of repairs over a 75-year service life.

The results show that fusion-bonded epoxy coatings significantly improve not only corrosion resistance but also life expectancy and cost effectiveness of reinforcing steel and that additional protection provides only limited additional improvement. Some corrosion inhibitors and the multiple zinc-epoxy coating extend the time to first repair, but differences in the costs over a 75‑year design life are relatively small for coated bars.

COMPARATIVE PERFORMANCE

The systems under study are compared based on the chloride content required to initiate corrosion, also known as the critical chloride corrosion threshold (CCCT); the corrosion rate after initiation; and disbondment or separation of the coating from the surface of the reinforcement as it is subjected to a corrosive environment in concrete. The first two criteria can be used to provide an estimate of the life expectancy and, in turn, the cost effectiveness of structures such as bridge decks. Disbondment represents a noteworthy aspect of system performance, although its impact on life expectancy and cost effectiveness remains unclear.

Critical Chloride Corrosion Threshold (CCCT)

Table 31 lists the CCCT for each corrosion-protection system for bars cast in concrete with a w/c ratio of 0.45. The table also lists the number of samples and standard deviations. The standard deviations account for the effect of combining groups of data.

Table 31. Chloride content at corrosion initiation (kg/m³), taken as the CCCT.

Steel Designationa	Southern Exposure	Initiation Beam	Total Number of Samples	Weighted Average	Standard Deviation
Bars without coatings
Controls
Conv.	0.77	0.94	156	0.90	0.479
Conv.2	0.88	1.00	118	0.98	0.822
Conv. Average			274	0.94	0.673
Corrosion inhibitors in concrete
Conv.(DCI)	-	1.49	120	1.49	0.604
Conv.2(DCI)	3.72	-	18	3.72	1.69
Conv.(DCI) Average			138	1.78	1.02
Conv.(RH)	-	1.23	120	1.23	0.839
Conv.2(RH)	2.16	-	18	2.16	0.708
Conv.(RH) Average			138	1.35	0.769
Conv.(HY)	-	0.37	120	0.37	0.267
Conv.2(HY)	1.19	0.51	118	0.61	0.317
Conv.(HY) Average			238	0.49	0.425
Bars with coatings
Control and e poxies with increased adhesion
ECR	4.31	-	28	4.31	1.14
ECR(Chromate)	5.69	-	11	5.69	2.88
ECR(DuPont)	4.97	-	16	4.97	3.33
ECR(Valspar)	6.08	-	11	6.08	2.90
ECR Average			66	5.00	3.17
Conventional epoxy and epoxies with increased adhesion plus calcium nitrite in concrete
ECR(DCI)	5.80	-	12	5.80	1.24
ECR(Chromate)-DCI	2.77	-	6	2.77	1.87
ECR(DuPont)-DCI	5.73	-	11	5.73	1.58
ECR(Valspar)-DCI	5.35	-	4	5.35	0.179
ECR(DCI) Average			33	5.17	2.71
Corrosion inhibitors other than calcium nitrite in concrete
ECR(RH)	4.10	-	17	4.10	1.45
ECR(HY)	1.08	-	10	1.08	0.673
ECR(primer/Ca(NO2)2)	7.11	-	10	7.11	4.04
Bars with multiple coatings
MC	2.00	-	16	2.00	1.79

1 kg/m³ = 1.69 lb/yd³
- No specimens.
^a See table 1 for abbreviation definitions.

A w/c ratio of 0.35 was used for five systems: conventional steel, conventional ECR, and conventional ECR cast in concrete with each of the three corrosion inhibitors (calcium nitrite, Rheocrete^®, and Hycrete™), but CCCT values for these systems are not used for this comparison. As shown in table 13, no data are available for conventional ECR cast in concrete with a w/c ratio of 0.35. For each of the systems with a corrosion inhibitor in concrete with a w/c ratio of 0.35, six or fewer samples were taken in total. The corrosion threshold for conventional steel cast in concrete with a w/c ratio of 0.35 is about twice that of conventional steel cast in concrete with a w/c ratio of 0.45, while the corrosion threshold is lower at a w/c ratio of 0.35 than at a w/c ratio of 0.45 for the systems with ECR and a corrosion inhibitor. For these reasons, the effect of w/c ratio on corrosion threshold is considered to be uncertain based on the current test results and will not be addressed further.

In table 31, the values for conventional steel without and with the corrosion inhibitors calcium nitrite, Rheocrete^®, and Hycrete™ include values from both southern exposure specimens and initiation beams. The values for coated bars include chloride contents only from southern exposure specimens. The values for bars with 4 and 10 holes through the coating are combined and weighted based on the number of samples collected. As observed in chapter 3, the CCCTs for bars with coatings are consistently several times the CCCTs for bars without coatings. The higher thresholds of the coated bars, all of which have penetrations in the coating, are in all likelihood due to the lack of uniformity of the chloride content in the concrete and the low probability that a region of locally high chloride content will coincide with the point on a bar where the coating is penetrated.

Looking at the individual CCCT values, conventional reinforcement Conv. has an average CCCT of 0.90 kg/m³ (1.52 lb/yd³) while conventional reinforcement Conv.2 has an average CCCT of 0.98 kg/m³ (1.65 lb/yd₃). The two values are combined to obtain a weighted average CCCT for conventional reinforcement of 0.94 kg/m³ (1.58 lb/yd³). Similarly, the values for Conv. and Conv.2 reinforcement in concrete containing the corrosion inhibitors calcium nitrite, Rheocrete^®, and Hycrete™ are combined to obtain CCCT values of 1.78, 1.35, and 0.49 kg/m³ (3.00, 2.28, and 0.83 lb/yd³), respectively. The reason for the low CCCT for Hycrete™ is discussed in chapter 3.

Conventional ECR exhibits an average CCCT of 4.31 kg/m³ (7.26 lb/yd³) while the systems with ECR and increased adhesion have CCCT values ranging from 4.97 to 6.08 kg/m³ (8.38 to 10.3 lb/yd³). The use of an increased-adhesion epoxy should not affect the chloride threshold of the system, and data for the high-adhesion bars are treated as representing the same population as conventional ECR; therefore, the CCCT values for conventional ECR and ECR with increased adhesion are averaged to produce a single CCCT of 5.00 kg/m³ (8.43 lb/yd³).

Among the specimens with ECR and inhibitors, ECR in concrete with calcium nitrite (combining the results for conventional epoxy and the epoxies with increased adhesion) had a CCCT of 5.17 kg/m³ (8.69 lb/yd³), just above the CCCT for ECR alone. ECR in concrete with Rheocrete^®, however, exhibited a CCCT of 4.10 kg/m³ (6.93 lb/yd³), which is lower than the CCCT for ECR in concrete with no inhibitor. ECR in concrete with Hycrete™ has a CCCT of 1.08 kg/m³ (1.82 lb/yd³), well below that of ECR in concrete with no inhibitor (5.00 kg/m³ (8.45 lb/yd³)). Relatively speaking, this is similar to conventional reinforcement in concrete containing Hycrete™, which exhibits a CCCT of 0.49 kg/m³ (0.83 lb/yd³) compared to a CCCT of 0.94 kg/m₃ (1.59 lb/yd³) for conventional steel in concrete with no inhibitors. With a CCCT of 7.11 kg/m³ (12.0 lb/yd³), ECR(primer/Ca(NO₂)₂) has the highest corrosion threshold of any of the systems tested. The high CCCT value for ECR(primer/Ca(NO₂)₂) may be due to the fact that the primer places the corrosion inhibitor directly in contact with the steel surface.

In the southern exposure test, MC reinforcement exhibits a CCCT of 2.00 kg/m³ (3.38 lb/yd³), which is lower than that of ECR. A study examining the CCCT of galvanized reinforcement, however, found that galvanized steel has an average critical chloride corrosion threshold of 1.52 kg/m³ (2.57 lb/yd³), about 50 percent higher than the value for conventional reinforcement, which suggests that a higher value, such as that used for conventional ECR, would be more appropriate for MC bars.⁽⁵⁰⁾

Because the CCCT values given in table 31 are, in some cases, based on a small number of samples, Student's t-test can be used to determine if the differences are the statistically significant.

Student's t-test compares the means and variances of two data sets to determine the probability that any differences in the mean values could have arisen by chance; that is, that differences in the mean values 1 and 2 are due to natural variability, not differences in the systems. For example, = 0.05 indicates a 5 percent chance that the test will incorrectly identify (or a 95 percent chance of correctly identifying) a statistically significant difference in sample means when, in fact, there is no difference. For this analysis, a two-tailed test is performed, meaning that there is a probability of /2 that 1 is greater than 2 and /2 that 1 is less than 2 when, in fact, 1 and 2 are equal. An value of 0.20 is used as the threshold for statistical significance. If is greater than 0.20, the systems are considered to be performing in a similar manner. It is worthwhile to note that a threshold of 0.20 is higher than often selected and values of 0.02, 0.05, and 0.10 are more common. An value of 0.20 is used here to restrict the number of cases in which systems are treated as similar.

The results of the Student's t-test are summarized in table 32 and table 33 for bare and coated bars, respectively. The tables show that all of the differences for bare bars are statistically significant with < 0.001. For the coated bars, all of the differences are statistically significant with < 0.162, with two exceptions: the differences in chloride threshold between conventional ECR and ECR(DCI) (5.00 and 5.17 kg/m³ (8.45 and 8.69 lb/yd³)) and between ECR(DCI) and ECR (RH) (5.17 and 4.10 kg/m³ (8.69 and 6.93 lb/yd³)). Furthermore, the value of for the difference between the corrosion thresholds for ECR and ECR(RH) is fairly high at 0.162. Because concrete with Rheocrete^® raises the CCCT of conventional reinforcement relative to conventional reinforcement in concrete without an inhibitor and because there is no reason to expect that Rheocrete® would have a negative effect when used with ECR, it would seem appropriate to consider ECR cast in concrete with Rheocrete^® as having a CCCT no lower than that of conventional ECR alone. As mentioned earlier in this section, a CCCT of 2.00 kg/m³ (3.38 lb/yd³) for MC reinforcement appears to be low because galvanized bars have a higher CCCT than conventional bars. Thus, even though the differences between the CCCT for MC reinforcement and other coated bars are statistically significant, it would also seem appropriate to apply the CCCT of conventional ECR to the MC reinforcement.

Table 32. Student's t-test results ( values) for CCCT for bars without coatings

Steel Designationa	CCCT, kg/m3	Conv.	Conv.(RH)	Conv.(DCI)	Conv.(HY)
CCCT (kg/m3)		0.94	1.78	1.35	0.49
Conv.	0.94	1	2.89E-28	9.55E-11	1.40E-12
Conv.(DCI)	1.78	2.89E-28	1	6.06E-05	2.12E-39
Conv.(RH)	1.35	9.55E-11	6.06E-05	1	5.16E-25
Conv.(HY)	0.49	1.40E-12	2.12E-39	5.16E-25	1

1 kg/m³ = 1.69 lb/yd³
Bold indicates statistical significance.
^a See table 1 for abbreviation definitions.

Table 33. Student's t-test results ( values) for CCCT for bars with

Steel Designationa	CCCT, kg/m3	ECR	ECR(DCI)	ECR(RH)	ECR(HY)	ECR (primer/ Ca(NO2)2)	MC
CCCT (kg/m3)		5.00	5.17	4.10	1.08	7.11	2.00
ECR	5.00	1	0.470	0.162	0.001	0.029	0.044
ECR(DCI)	5.17	0.470	1	0.344	4.26E-04	0.009	0.055
ECR(RH)	4.10	0.162	0.344	1	6.59E-05	0.003	0.107
ECR(HY)	1.08	0.001	4.26E-04	6.59E-05	1	2.71E-04	0.089
ECR(primer/Ca(NO2)2)	7.11	0.029	0.009	0.003	2.71E-04	1	0.006
MC	2.00	0.044	0.055	0.107	0.089	0.006	1.000

1 kg/m³ = 1.69 lb/yd³
Bold indicates statistical significance.
^a See table 1 for abbreviation definitions.

Corrosion Rate

The systems can be compared based on corrosion rate for both bench-scale and field test specimens. As with the chloride threshold comparisons, emphasis is placed on results for concretes with a w/c ratio of 0.45.

The average corrosion rates based on corrosion losses after corrosion initiation for the bench-scale specimens are presented in chapter 3 in table 15 and table 20 for the southern exposure and cracked beam specimens and table 28 and table 29 for field test specimens without and with simulated cracks over the reinforcing steel. The tables include the number of specimens or bars for which the data is obtained as well as the standard deviation, providing a summary for corrosion rate that is equivalent to that presented for chloride content at corrosion initiation in table 31.

To aid in comparisons dealing with conventional reinforcement without and with corrosion inhibitors in the concrete, a modification is made to the data. Table 34 lists bench-scale macrocell corrosion rates for Conv. reinforcement without inhibitors and Conv.2 reinforcement without and with inhibitors. The bench-scale macrocell corrosion rates for Conv.2 reinforcement were 10.1 and 16.3 m/year (0.398 and 0.642 mil/year) in uncracked and cracked concrete (southern exposure and cracked beam specimens), respectively, about twice the respective a rates for Conv. reinforcement (5.69 and 7.00 m/year (0.224 and 0.276 mil/year)). All inhibitors reduced the corrosion rate of Conv.2 reinforcement compared to Conv.2 reinforcement with no inhibitor, but in cracked concrete, Conv.2(RH) and Conv.2(DCI) show corrosion rates greater than those observed for Conv. reinforcement. It is assumed that if the inhibitors had been used in conjunction with the earlier tests on Conv. reinforcement, the resulting corrosion rates would have been less than the corrosion rates measured for Conv. reinforcement with no inhibitor. Since all other protection systems (uncoated and coated) used the same heat of steel as Conv. reinforcement and because the mat-to-mat resistance values for the Conv.2 southern exposure and cracked beam specimens are lower than for the corresponding Conv. specimens (which would lead to higher corrosion rates), the only way to achieve a fair comparison between systems is to reduce the corrosion rates for the systems with Conv.2 reinforcement with inhibitors by the ratio of the Conv.2 corrosion rate to the Conv. corrosion rate.^(44,58) The designation Conv.* is used when referring to the modified corrosion rate data for conventional reinforcement in concrete with corrosion inhibitors. The modified corrosion rates for these systems are presented in table 34. Based on the modified corrosion rates, the estimated rates for Conv.*(RH) are 1.64 and 5.13 m/year (0.0646 and 0.202 mil/year) in uncracked and cracked concrete, respectively. Likewise, the estimated rates for Conv.*(DCI) are 3.77 and 6.24 m/year (0.148 and 0.246 mil/year) and are 0.706 and 1.80 m/year (0.0278 and 0.0709 mil/year) for Conv.*(HY).

Table 34. Equivalent corrosion rates for conventional reinforcement with inhibitors (m/year).

Steel Designation	Concretea	Macrocell Corrosion Rate	Ratio of Conv.2 to Conv. Rate	Modified Corrosion Rate (Conv.*)b
Corrosion rate (total area)
Conv.	U	5.69	NA	NA
	C	7.00	NA	NA
Conv.2	U	10.1	1.77	5.69
	C	16.3	2.32	7.00
Conv.2(RH)	U	2.91	1.77	1.64
	C	11.90	2.32	5.13
Conv.2(DCI)	U	6.67	1.77	3.77
	C	14.50	2.32	6.24
Conv.2(HY)	U	1.25	1.77	0.706
	C	4.17	2.32	1.80

1 m = 0.0394 mil
NA = Not applicable
^a U = Uncracked concrete (southern exposure), C = cracked concrete (cracked beam).
^b Estimated corrosion rate in conjunction with Conv. reinforcement.

To determine the statistical significance of the differences in corrosion rates between corrosion-protection systems, the two-tailed Student's t-test is again used with > 0.20 indicating that the observed differences are not statistically significant and that systems can be considered to perform in a similar manner. Bare bars, for which corrosion rates are presented based on total area, are examined separately from coated bars, for which corrosion rates are based on exposed area. The systems are compared first based on bench-scale results and then based on the field test results.

Table 35 and table 36 list the values of based on comparisons of corrosion rates for corrosion-protection systems with bare bars in southern exposure and cracked beam specimens, respectively. To prepare the tables, the corrosion rates and standard deviations for conventional reinforcement with inhibitors are scaled by the ratio of Conv.2 corrosion rate to Conv. corrosion rate prior to analysis, as shown in table 34. The comparisons for the southern exposure tests show that, at 5.69 m/year (0.224 mil/year), the corrosion rate is highest for Conv., followed by Conv.*(DCI), Conv.*(RH), and Conv.*(HY), with values of 3.77, 1.64, and 0.706 m/year (0.148, 0.0646, and 0.0278 mil/year), respectively. All differences are statistically significant with < 0.025. The comparisons for the cracked beam tests, however, show that while the order of the corrosion rates is the same as for the southern exposure specimens with values of 7.00, 6.24, 5.15 and 1.80 m/year (0.276, 0.246, 0.203, and 0.0709 mil/year) for Conv., Conv.*(DCI), Conv.*(RH), and Conv.*(HY), respectively, only the rate for Conv.*(HY) differs with statistical significance from the other values.

Table 35. Student's t-test results ( values) for average corrosion rates based on corrosion losses after corrosion initiation based on total area for bars

Steel Designationa	Corrosion Rate, m/year	Conv.	Conv.* (RH)	Conv.* (DCI)	Conv.* (HY)
Corrosion Rate (m/year)		5.69	1.64	3.77	0.706
Conv.	5.69	1	9.4E-06	0.005	1.5E-06
Conv.*(RH)	1.64	9.4E-06	1	0.012	0.025
Conv.*(DCI)	3.77	0.005	0.012	1	0.004
Conv.*(HY)	0.706	1.5E-06	0.025	0.004	1

1 m = 0.0394 mil
Bold indicates statistical significance.
^a See table 1 for abbreviation definitions.

Table 36. Student's t-test results ( values) for average corrosion rates based on corrosion losses after corrosion initiation based on total area for bars without coatings from cracked beam

Steel Designationa	Corrosion Rate, m/year	Conv.	Conv.* (RH)	Conv.* (DCI)	Conv.* (HY)
Corrosion Rate (m/year)		7.00	5.15	6.24	1.80
Conv.	7.00	1	0.22	0.31	0.004
Conv.*(RH)	5.15	0.22	1	0.28	0.003
Conv.*(DCI)	6.24	0.31	0.28	1	0.004
Conv.*(HY)	1.80	0.004	0.003	0.004	1

1 m = 0.0394 mil
Bold indicates statistical significance.
^a See table 1 for abbreviation definitions.

Table 37 and table 38 list the values of for comparisons of corrosion-protection systems with coated bars in the southern exposure and cracked beam tests, respectively. Since only one southern exposure specimen with Hycrete™ initiated corrosion, a statistical analysis cannot
be performed for that system.

Table 37. Student's t-test results ( values) for average corrosion rates based on corrosion losses after corrosion initiation based on exposed area for bars with coatings from southern exposure tests.

Steel Designationa	Corrosion Rate (/year)	ECR	ECR (DCI)	ECR (RH)	ECR (HY)	ECR (primer/ Ca(NO2)2)	ECR (Chromate)	ECR (DuPont)	ECR (Valspar)	MC
Corrosion Rate (m/year)		10.43	7.81	8.63	0.674	12.61	12.86	12.37	16.93	31.63
ECR	10.43	1	0.500	0.274	-	0.649	0.544	0.628	0.313	0.071
ECR(DCI)	7.81	0.500	1	0.291	-	0.468	0.344	0.391	0.324	0.150
ECR(RH)	8.63	0.274	0.291	1	-	0.613	0.569	0.529	0.977	0.471
ECR(HY)	0.674	-	-	-	-	-	-	-	-	-
ECR(primer/Ca(NO2)2)	12.61	0.621	0.468	0.613	-	1	0.964	0.966	0.577	0.140
ECR(Chromate)	12.86	0.544	0.344	0.569	-	0.964	1	0.919	0.576	0.142
ECR(DuPont)	12.37	0.628	0.391	0.529	-	0.966	0.919	1	0.532	0.132
ECR(Valspar)	16.93	0.313	0.324	0.977	-	0.577	0.576	0.532	1	0.220
MC	31.63	0.071	0.150	0.471	-	0.140	0.142	0.132	0.220	1

1 m = 0.0394 mil
Bold indicates statistical significance.
^a See table 1 for abbreviation definitions.

Table 38. Student's t-test results ( values) for average corrosion rates based on corrosion losses after corrosion initiation based on exposed area for bars with coatings from cracked beam

Steel Designationa	Corrosion Rate (/year)	ECR	ECR (DCI)	ECR (RH)	ECR (HY)	ECR (primer/ Ca(NO 2)2)	ECR (Chromate)	ECR (DuPont)	ECR (Valspar)	MC
Corrosion Rate (m/year)		8.07	11.2	17.0	10.8	8.73	20.4	25.9	16.4	68.6
ECR	8.07	1	0.364	0.014	0.506	0.826	0.064	0.005	0.158	0.005
ECR(DCI)	11.2	0.364	1	0.145	0.938	0.497	0.250	0.046	0.465	0.027
ECR(RH)	17.0	0.014	0.145	1	0.211	0.026	0.720	0.243	0.893	0.067
ECR(HY)	10.8	0.506	0.938	0.211	1	0.645	0.292	0.072	0.494	0.041
ECR(primer/Ca(NO2)2)	8.73	0.826	0.497	0.026	0.645	1	0.140	0.019	0.270	0.021
ECR(Chromate)	20.4	0.064	0.250	0.720	0.292	0.140	1	0.554	0.680	0.064
ECR(DuPont)	25.9	0.005	0.046	0.243	0.072	0.019	0.554	1	0.290	0.097
ECR(Valspar)	16.4	0.158	0.465	0.893	0.494	0.270	0.680	0.290	1	0.045
MC	68.6	0.005	0.027	0.067	0.041	0.021	0.064	0.097	0.045	1

1 m = 0.0394 mil
Bold indicates statistical significance.
^a See table 1 for abbreviation definitions.

The comparison between rates for the southern exposure test shows that only MC reinforcement, corroding at nearly 3 times the rate of conventional ECR, has a corrosion rate that is significantly different from that of the other systems ( < 0.20). ECR(HY) is also assumed to have a statistically significant difference in corrosion rate compared to the other systems because the corrosion rate of ECR(HY), 0.674 m/year (0.0265 mil/year), is over an order of magnitude lower than any of the other systems.

The results for the cracked beam tests show no statistically significant differences between ECR, ECR(DCI), ECR(HY), and ECR(primer/Ca(NO₂)₂) ( > 0.20), although it is worth noting that the corrosion rate for conventional ECR, 8.07 m/year (0.318 mil/year), is the lowest rate obtained for any of the systems in the cracked beam tests and that both conventional ECR and ECR(primer/Ca(NO₂)₂) exhibited lower rates in the cracked beam tests than in the southern exposure tests, highlighting some of the variability inherent in corrosion tests. ECR(RH), ECR(Chromate), ECR(DuPont), and ECR(Valspar) show significant differences in performance relative to ECR, with corrosion rates greater than that for ECR. In practice, however, there is no reason to expect that a corrosion inhibitor or an epoxy with increased adhesion will increase the corrosion rate. The greater corrosion rates observed are likely due to variations in the test or in concrete quality. The corrosion rate of the MC bars, 68.6 m/year (2.70 mil/year), is significantly different from that of all of the other systems-over 2 times the next closest value, 25.9 m/year (1.02 mil/year) for ECR(DuPont) and over 8 times greater than that of conventional ECR (8.07 m/year (0.318 mil/year)).

The results of Student's t-test for the field test specimens are shown in table 39 and table 40 for uncracked and cracked concrete, respectively. The analysis does not include bare bar systems because the only bare bar system evaluated using the field test specimens was conventional steel. The comparisons between average macrocell corrosion rates based on losses after corrosion initiation in uncracked concrete show no statistically significant differences in rates for ECR, ECR(RH), ECR(DCI), ECR(primer/Ca(NO₂)₂), ECR(Chromate), and ECR(DuPont) ( > 0.20). Although the rates for ECR(Valspar) and MC are statistically different from other corrosion-protection systems, they are not statistically different from conventional ECR, and these differences are likely due to variations in concrete quality. In addition, the corrosion rate for MC reinforcement (6.31 m/year (0.248 mil/year)) is only 11 percent higher than that of conventional ECR (5.68 m/year (0.224 mil/year)), compared to 3 times higher in the southern exposure tests. At 2.89 m/year (0.114 mil/year), the corrosion rate for ECR(HY) is 68 percent of the next closest corrosion rate and is statistically different from the rates for ECR, ECR(RH), ECR(Valspar), and MC.

Table 39. Student's t-test results ( values) for average corrosion rates based on corrosion losses after corrosion initiation based on exposed area for

Steel Designationa	Corrosion Rate (/year)	ECR	ECR (DCI)	ECR (RH)	ECR (HY)	ECR (primer/ Ca(NO2)2)	ECR (Chromate)	ECR (DuPont)	ECR (Valspar)	MC
Corrosion Rate (m/year)		5.68	4.26	5.43	2.89	4.49	4.83	5.14	9.11	6.31
ECR	5.68	1	0.425	0.915	0.136	0.574	0.732	0.949	0.269	0.763
ECR(DCI)	4.26	0.425	1	0.446	0.256	0.871	0.726	0.563	0.020	0.176
ECR(RH)	5.43	0.915	0.446	1	0.120	0.606	0.778	0.989	0.166	0.636
ECR(HY)	2.89	0.136	0.256	0.120	1	0.233	0.209	0.295	0.010	0.033
ECR(primer/Ca(NO2)2)	4.49	0.574	0.871	0.606	0.233	1	0.856	0.708	0.067	0.301
ECR(Chromate)	4.83	0.732	0.726	0.778	0.209	0.856	1	0.836	0.145	0.463
ECR(DuPont)	5.14	0.949	0.563	0.989	0.295	0.708	0.836	1	0.305	0.750
ECR(Valspar)	9.11	0.269	0.020	0.166	0.010	0.067	0.145	0.305	1	0.305
MC	6.31	0.763	0.176	0.636	0.033	0.301	0.463	0.750	0.305	1

1 m = 0.0394 mil
Bold indicates statistical significance.
^a See table 1 for abbreviation definitions.

Table 40. Student's t-test results ( values) for average corrosion rates based on corrosion losses after corrosion initiation based on exposed area for bars with coatings from field test specimens with cracked concrete.

Steel Designationa	Corrosion Rate (μm/year)	ECR	ECR (DCI)	ECR (RH)	ECR (HY)	ECR (primer/ Ca(NO2)2)	ECR (Chromate)	ECR (DuPont)	ECR (Valspar)	MC
Corrosion Rate (m/year)		8.13	5.79	8.38	4.32	4.65	8.94	6.50	7.64	8.11
ECR	8.13	1	0.275	0.923	0.234	0.200	0.745	0.656	0.733	0.918
ECR(DCI)	5.79	0.275	1	0.072	0.448	0.475	0.126	0.501	0.415	0.258
ECR(RH)	8.38	0.923	0.072	1	0.024	0.004	0.554	0.416	0.577	0.947
ECR(HY)	4.32	0.234	0.448	0.024	1	0.862	0.144	0.269	0.241	0.181
ECR(primer/Ca(NO2)2)	4.65	0.200	0.475	0.004	0.862	1	0.112	0.216	0.190	0.137
ECR(Chromate)	8.94	0.745	0.126	0.554	0.144	0.112	1	0.424	0.486	0.643
ECR(DuPont)	6.50	0.656	0.501	0.416	0.269	0.216	0.424	1	0.893	0.675
ECR(Valspar)	7.64	0.733	0.415	0.577	0.241	0.190	0.486	0.893	1	0.774
MC	8.11	0.918	0.258	0.947	0.181	0.137	0.643	0.675	0.774	1

1 m = 0.0394 mil
Bold indicates statistical significance.
^a See table 1 for abbreviation definitions.

The analysis of the cracked field test specimens shows that the corrosion rates for ECR, ECR(DCI), ECR(RH), ECR (HY), ECR(Chromate), ECR(DuPont), ECR(Valspar), and MC are not significantly different ( > 0.20). The rates for two systems combining ECR with inhibitors in the concrete, ECR(DCI) and ECR (HY), and ECR(primer/Ca(NO₂)₂) are lower and significantly different from at least two other systems. The rate for ECR(DCI) is significantly different from the rates for ECR(RH) and ECR(Chromate); the rate for ECR(HY) is significantly different from the rates for ECR(RH) ECR(Chromate) and MC; and the rate for ECR(primer/Ca(NO₂)₂) is significantly different from the rates for ECR, ECR(RH), ECR(Chromate), ECR(Valspar), and MC. As observed for the uncracked field test specimens, the specimens containing MC bars had a corrosion rate (8.11 m/year (0.319 mil/year)) that was close to that of conventional ECR (8.13 m/year (0.320 mil/year)).

Disbondment

The disbondment results are summarized in figure 110 through figure 112 for the bench-scale specimens and in figure 159 through figure 161 for the field test specimens. In all cases, the top bars exhibited far more disbondment than the bottom bars, indicating that higher chloride content, higher moisture content, or a combination of the two results in increased disbondment. The southern exposure specimens cast with one of the corrosion inhibitors exhibited significantly less disbondment than those cast in concrete without an inhibitor. The same observation does not hold for the cracked beam specimens or for the field test specimens, with or without cracks. The tests uniformly show that when cast in the same quality of concrete, bars with high adhesion epoxies exhibit no reduction in disbondment compared to conventional ECR bars.

Bars in cracked concrete exhibited far more disbondment than those in uncracked specimens in both the bench-scale and field tests. Bottom bars fared better in uncracked concrete than in cracked concrete, presumably because of the higher chloride content lower in the specimens in the presence of a crack.

Bars in southern exposure specimens with a w/c ratio of 0.35 consistently exhibited less disbondment than bars in southern exposure specimens with a w/c ratio of 0.45, in some cases exhibiting no disbondment (see figure 112). The results were mixed for bars in cracked beam specimens. As shown in figure 112, the ECR(primer/Ca(NO₂)₂) and ECR(DCI) bars cast in concrete with w/c of 0.35 exhibited no disbondment, while ECR, ECR(RH), and ECR(HY) exhibited nearly the same disbondment as those in the specimens with the higher w/c ratio. MC bars exhibited somewhat less disbondment than for the other systems. This trend is especially true for the bottom bars in concrete with cracks (see figure 111 and figure 160). Finally, as shown in figure 153, top bars that were electrically isolated (bars not connected across a 10-ohm resistor to a bottom bar in the same specimen) exhibited far less disbondment than those that were connected, indicating the negative impact of macrocell corrosion and the importance of maintaining electrical isolation between coated bars in reinforced concrete structures.

LIFE EXPECTANCY

The life expectancy of a bridge deck (the structure used to compare life expectancy and cost effectiveness in this study) is based on an estimate of the time to first repair combined with the time between repairs to establish the present cost of a deck over a 75-year design life. In many cases, the time to first repair is based on experience. For example, the South Dakota Department of Transportation (SDDOT) has estimated that the time to first repair for bridge decks containing conventional steel is 10 years under harsh environmental conditions and 25 years in arid conditions.⁽²³⁾ The latter matches the time to first repair estimated by the KDOT.^(23,36) In 2001, the time to first repair for bridge decks containing ECR was estimated to be 35 and 40 years by KDOT and SDDOT, respectively. The estimate for decks with ECR was based on the fact that, as of the 2001, no bridge decks containing ECR had required repair due to corrosion-induced damage since its first use in the late 1970s.⁽³⁶⁾ Other estimates of time to first repair use models that are based on the time required for chloride to diffuse through uncracked concrete.^(59-62) Models of this type usually include a preselected time for the corrosion products to cause the concrete to crack following corrosion initiation. Diffusion-based models have two key drawbacks: (1) they do not account for the role played by cracks in the concrete in allowing rapid penetration of chlorides to the level of the reinforcing steel and (2) a preselected time for the corrosion products to cause the concrete to crack is not based on actual corrosion rates. Because surveys of bridge decks with ages ranging from several months to over 20 years demonstrate that reinforced concrete bridge decks exhibit significant cracking parallel to and directly above the reinforcing bars, estimates of time to first repair are based principally on corrosion in the presence of cracks. (See references 3, 43, and 63-66.)

The procedures used in the current analysis are based on field and laboratory evidence addressing corrosion initiation and propagation in cracked concrete combined with experience with deck repair. Using this approach, the time to first repair depends on: (1) the time required for the chloride content of the concrete to reach the critical chloride initiation threshold for the system, (2) the time required after initiation for corrosion products to cause cracking and spalling of the concrete cover, and (3) the time between first cracking and the time that the repair is made.

Time to Corrosion Initiation

The time to corrosion initiation is estimated based on chloride contents measured at crack locations on bridge decks in Kansas and the CCCT (water-soluble chloride content) for each corrosion-protection system, as previously discussed.^(3,43,67)

The chloride contents in bridge decks are based on two studies.^(43,67) In those studies, chloride samples were obtained in bridge decks using a vacuum drill. The samples were obtained in increments of 19 mm (0.75 inches) to a depth of 95 mm (3.8 inches) both at and away from cracks in bridge decks primarily in northeast Kansas. Figure 162 shows the relationship between the average chloride content at crack locations interpolated to a depth of 76.2 mm (3 inches) versus age for bridges with an average annual daily traffic (AADT) greater than 7,500 (high traffic bridges).⁽³⁾ The decks in the survey were cast monolithically and with high-density conventional and silica fume overlays. Figure 162 demonstrates that the chloride content at cracks is independent of the type of deck.

1 kg/m³ = 1.69 lb/yd³

Figure 162. Graph. Chloride content taken at cracks interpolated at a depth of 76.2 mm (3 inches) versus age for bridges with an AADT greater than 7,500.

Based on the data shown in figure 162, the chloride content C (in kg/m³) can be expressed as a function of age at the time of sampling T (in months) by the trendline shown in the equation in figure 163.

Figure 163. Equation. Chloride content trendline.

Using the equation in figure 163, the average time to reach a specific critical chloride threshold T_c can be expressed as a function of the critical chloride threshold C_c, as shown in figure 164.

Figure 164. Equation. Average time to critical chloride threshold.

Table 31 lists the chloride contents at corrosion initiation (from table 13 and table 17), taken as the CCCT, for each corrosion-protection system in this study cast in concrete with a w/c ratio of 0.45, a realistic value for bridge decks. The values are combined and weighted based on the number of samples collected.

The CCCT values listed in table 31, with some modifications, are used in conjunction with figure 164 to determine the time to corrosion initiation for each corrosion-protection system. The values, expressed in years, are listed in table 41.

Table 41. Estimated time to corrosion initiation for corrosion-protection systems in a bridge deck with 76.2-mm (3-inch) cover on top reinforcing steel.

Steel Designationa	Chloride Threshold, kg/m3b	Age at Corrosion Initiation in Bridge Decks, years
Bars without coatings
Control
Conv.	1.58	2.2
Corrosion inhibitors in concrete
Conv.(DCI)	3.00	6.0
Conv.(RH)	2.28	4.1
Conv.(HY)	0.83	0.2 (1.0)c
Bars with coatings
Control
ECR	8.29	20.0
Epoxies with increased adhesion
ECR(Chromate)	8.29	20.0
ECR(DuPont)	8.29	20.0
ECR(Valspar)	8.29	20.0
Corrosion inhibitors in concrete
ECR(DCI)	8.29	20.0
ECR(RH)	8.29	20.0
ECR(HY)	1.82	2.8
ECR(primer/Ca(NO2)2)	12.0	29.7
Bars with multiple coatings
MC	4.92	20.0

1 kg/m³ = 1.69 lb/yd^3
a See table 1 for abbreviation definitions.
^b See text for explanation of differences of values from those in table 31.
^c Rounded up from 0.2 years.

As shown in the table, the system combining uncoated conventional bars with concrete containing Hycrete™ has the lowest calculated time to initiation of corrosion, 0.2 years. Because salt is not applied to bridge decks until the first winter, this value is rounded up to 1 year. Conventional reinforcement with no inhibitor initiates corrosion after 2.2 years. Rheocrete^® and calcium nitrite extend the initiation time of conventional reinforcement to 4.1 and 6.1 years, respectively.

Based on the statistical analysis of chloride threshold values presented earlier in this chapter for systems with coated bars, the differences in CCCT value for ECR, the three types of ECR with increased adhesion, and ECR with calcium nitrite or Rheocrete^® added to the concrete are not significant. Therefore, the values are averaged (weighted based on the number of samples) to obtain a single CCCT value, 4.92 kg/m³ (8.29 lb/yd³). Because galvanized reinforcement has a higher corrosion threshold than conventional reinforcing steel, the CCCT value obtained in the southern exposure tests for MC bars, 2.00 kg/m³ (3.37 lb/yd³), is considered unrealistic.⁽⁵⁰⁾ Thus, the threshold used for most of the other coated bars, 4.92 kg/m³ (8.29 lb/yd³), is used for the MC reinforcement, as well. For this value of CCCT, the time to corrosion initiation is 20.0 years. The use of the calcium nitrite primer under the epoxy significantly increases the time to 29.7 years, the highest of any of the systems in this study. ECR with Hycrete™ initiates corrosion after 2.8 years.

Time to Cracking After Corrosion Initiation

The time required to generate enough corrosion products to crack concrete is a function of the total corrosion rate and the corrosion loss required to cause cracking. The latter is a function of the bar size and the concrete cover. (See references 44 and 68-71.) Total corrosion rates will be covered first. To do this, equivalent field test macrocell corrosion rates must be calculated for conventional reinforcement cast in concrete with corrosion inhibitors because corrosion rates in field test specimens serve as the basis for calculating the time to cracking and because field tests were not performed on conventional reinforcement cast in concrete with corrosion inhibitors.

Equivalent Field Test Corrosion Rates for Conventional Reinforcement Cast in Concrete with Corrosion Inhibitors

As shown in chapter 3, corrosion rates and losses were consistently lower in the field test specimens than in the bench-scale specimens. Because the results for the field test specimens serve as the basis for the life-cycle and cost-effectiveness calculations and because field test specimens were not used for conventional reinforcement cast in concrete with corrosion inhibitors, an estimate of corrosion rates in field test specimens is needed for these systems. This is done using the ratio of the field test to bench-scale macrocell corrosion rates for conventional reinforcement in concrete without and with cracks above the reinforcement. Figure 165 and figure 166 compare the average macrocell corrosion rates after corrosion initiation based on total area for the field test and bench-scale specimens containing conventional reinforcement in uncracked and cracked concrete, respectively. The figures also show the range in the rates for each type of specimen. As shown, the average macrocell corrosion rates in the field test specimens equal 15.5 and 13.4 percent of the rates in the bench-scale specimens in uncracked and cracked concrete, respectively. These values are used to convert the corrosion rates shown in table 34 to equivalent macrocell corrosion rates in field tests shown in table 42.

1 m = 0.0394 mil

Figure 165. Graph. Comparison between average macrocell corrosion rates after corrosion initiation based on total area for bench-scale and field test specimens with conventional reinforcement in uncracked concrete.

1 m = 0.0394 mil

Figure 166. Graph. Comparison between average macrocell corrosion rates after corrosion initiation based on total area for bench-scale and field test specimens with conventional reinforcement in cracked concrete.

Table 42. Equivalent field test specimen macrocell corrosion rates (m/year)

Steel Designationa	Concretea	Benchscale Corrosion Rate	Equivalent FTS Corrosion Rate b
FTS(Conv.)	U	NA	0.882
	C	NA	0.939
Conv.*(RH)	U	1.64	0.255
	C	6.34	0.939
Conv.*(DCI)	U	3.77	0.584
	C	6.34	0.939
Conv.*(HY)	U	0.706	0.109
	C	1.80	0.241

1 m = 0.0394 mil
FTS = Field test specimen
NA = Not applicable
^a See table 1 for abbreviation definitions. U = uncracked concrete, C = cracked concrete.
^b Estimated using a ratio of FTS to bench-scale rate of 0.155 in uncracked concrete and 0.134 in cracked concrete.

The corrosion rates for bare bar systems assume the entire area of steel is corroding; however, autopsy results from field test specimens indicated corrosion occurs in localized regions on the bars (see figure 150 and figure 151). As discussed in chapter 3, a visual inspection of conventional reinforcement from field test specimens at the end of testing indicated that corrosion covers only about one-third the total area of bars in uncracked concrete and about 40 percent of the total area of bars in cracked concrete. Because corrosion only occurs on limited regions of the bar, the corrosion rates for bare bars in uncracked and cracked concrete are multiplied by 3 and 2.5, respectively, to obtain a macrocell corrosion rate based on effective area for these systems. The results are listed in table 43. As before, conventional reinforcement in concrete with no inhibitor has the greatest corrosion rates based on effective area in both uncracked and cracked concrete, 2.65 and 2.35 m/year (0.104 and 0.0925 mil/year), respectively. Conv.*(HY) has the lowest corrosion rates based on effective area in uncracked and cracked concrete, 0.328 and 0.602 m/year (0.0129 and 0.0237 mil/year), respectively. These adjustments appear to be appropriate not only based on the observed area undergoing corrosion but also based on the time to cracking observed for the Conv. field test specimens, as discussed in chapter 3.

Table 43. Equivalent field test specimen macrocell corrosion rates (m/year) for bare bar corrosion-protection systems based on effective area.

Steel Designationa	Concretea	FTS Corrosion Rate (Total Area)b	FTS Corrosion Rate (Effective Area)c
Conv.	U	0.882	2.65
	C	0.939	2.35
Conv.*(RH)	U	0.255	0.765
	C	0.939	2.35
Conv.*(DCI)	U	0.584	1.75
	C	0.939	2.35
Conv.*(HY)	U	0.109	0.328
	C	0.241	0.602

1 m = 0.0394 mil
FTS = Field test specimen
^a See table 1 for abbreviation definitions. U = uncracked concrete, C = cracked concrete.
^b See table 42.
^c Estimated using a ratio of 3 in uncracked concrete and 2.5 in cracked concrete.

Total Corrosion Rates

The field test specimens provide realistic models of bridge decks in terms of both configuration and exposure. Because the measurements are based on macrocell corrosion, which represents only a portion of the total corrosion loss, a comparison of the linear polarization resistance and macrocell corrosion results for the southern exposure and cracked beam specimens is used in conjunction with macrocell readings from the field tests to estimate the total corrosion rates in the field tests, and thus, in bridge decks.

To estimate the relationships between the macrocell and total corrosion rates for the different systems and degrees of exposure evaluated in this study, the average corrosion rates based on losses after corrosion initiation for southern exposure (table 15) and cracked beam (table 20) specimens are compared with the average corrosion rates based on total corrosion losses after corrosion initiation as measured using linear polarization resistance (table 22). Because of differences in exposure and corrosion mechanisms, separate comparisons are made for southern exposure and cracked beam specimens for uncoated bars, epoxy-coated bars, and MC bars.

The relationships between total and macrocell corrosion rates are shown in figure 167 and figure 168 for the bench-scale specimens containing uncoated bars, in figure 169 and figure 170 for specimens containing epoxy-coated bars, and in figure 171 and figure 172 for specimens containing MC bars. In each pair of figures, the first represents the results for the southern exposure specimens and the second represents the results for the cracked beam specimens. The figures show trend lines originating at the origin, which give the ratio of total to macrocell corrosion rate. As shown in figure 167 and figure 168 for conventional reinforcement, the total corrosion rates average 1.79 and 3.49 times macrocell corrosion rates for southern exposure and cracked beam specimens, respectively. The multiples are 3.15 and 12.36 for epoxy-coated bars and 4.90 and 5.82 for the MC bars.

1 m = 0.0394 mil

Figure 167. Graph. Total versus macrocell corrosion rate after corrosion initiation for southern exposure specimens with conventional reinforcement.

1 m = 0.0394 mil

Figure 168. Graph. Total versus macrocell corrosion rate after corrosion initiation for cracked beam specimens with conventional reinforcement.

1 m = 0.0394 mil

Figure 169. Graph. Total versus macrocell corrosion rate after corrosion initiation for southern exposure specimens with ECR.

1 m = 0.0394 mil

Figure 170. Graph. Total versus macrocell corrosion rate after corrosion initiation for cracked beam specimens with ECR.

1 m = 0.0394 mil

Figure 171. Graph. Total versus macrocell corrosion rate after corrosion initiation for southern exposure specimens with MC reinforcement.

1 m = 0.0394 mil

Figure 172. Graph. Total versus macrocell corrosion rate after corrosion initiation for cracked beam specimens with MC reinforcement.

As demonstrated in chapter 3, figure 167 through figure 172 show that corrosion rates are significantly higher for cracked concrete (cracks directly above and parallel to the reinforcement) than for uncracked concrete. The figures also show that the total corrosion rate is a higher multiple of the macrocell corrosion rate for coated bars than for uncoated bars and that the multiple is consistently higher for cracked concrete than for uncracked concrete. The multiples are closest between uncracked and cracked concrete for the MC bars, in all likelihood due to the combined effects of the amphoteric nature (corrodes in alkaline as well an acidic conditions) of zinc and the galvanic protection provided by the zinc.

The multiples developed in figure 167 through figure 172 are applied to the field test macrocell corrosion rates, including the equivalent rates for conventional bars cast in concrete containing corrosion inhibitors, to develop estimates of total corrosion rates in bridge decks. The macrocell corrosion rates used for the coated bars are based on the statistical analyses presented earlier in the chapter. Because the differences in corrosion rate are not statistically significant, an average macrocell rate based on exposed area of 5.66 m/year (0.223 mil/year) is used for all but one system in uncracked concrete; the corrosion rate for that system, ECR(HY), (2.89 m/year (0.114 mil/year)) does exhibit a statistically significant difference. Based on the results of the analysis for bars in cracked concrete, an average macrocell corrosion rate of 7.95 m/year (0.313 mil/year) on an exposed area equal to that provided by a 3.2-mm (0.125-inch)-diameter penetration is used for conventional ECR, ECR with increased adhesion, ECR with corrosion inhibitor Rheocrete^® in the concrete, and MC bars. Rates of 5.79, 4.32, and 4.65 m/year (0.228, 0.170, and 0.183 mil/year) are used, respectively, for ECR in concrete with the corrosion inhibitors calcium nitrite and Hycrete™ and ECR with primer containing microencapsulated calcium nitrite. In the latter case, it is unlikely the calcium nitrite provided by the primer would be adequate for the full time to first repair. Thus, for purposes of the analysis, a corrosion rate equal to that used for conventional ECR is assumed beginning 10 years after corrosion initiation. The estimated total corrosion rates in bridge decks are shown in table 44.

Table 44. Estimated total corrosion rates in bridge decks (m/year) for corrosion-protection systems.

Steel Designationa	Concreteb	Macrocell Corrosion Ratec	Estimated Total Corrosion Rated
Corrosion rate (total area)
Conv.	U	2.65	4.74
	C	2.35	8.19
Conv.*(RH)	U	0.765	1.37
	C	2.35	8.19
Conv.*(DCI)	U	1.75	3.14
	C	2.35	8.19
Conv.*(HY)	U	0.328	0.588
	C	0.602	2.10
Corrosion rate (exposed area)
ECR	U	5.66	17.8
	C	7.95	98.3
ECR(Chromate)	U	5.66	17.8
	C	7.95	98.3
ECR(DuPont)	U	5.66	17.8
	C	7.95	98.3
ECR(Valspar)	U	5.66	17.8
	C	7.95	98.3
ECR(RH)	U	5.66	17.8
	C	7.95	98.3
ECR(DCI)	U	5.66	17.8
	C	5.79	71.6
ECR(HY)	U	2.89	9.1
	C	4.32	53.4
ECR(primer/Ca(NO2)2)	U	5.66	17.8
	C	4.65	57.5e
MC	U	5.66	27.7
	C	7.95	46.3f

1 m = 0.0394 mil
^a See table 1 for abbreviation definitions.
^b U = uncracked concrete, C = cracked concrete.
^c Macrocell corrosion rates for field test specimens.
^d Macrocell corrosion rates multiplied by a ratio of 1.79, 3.15, and 4.90 for bare, coated, and MC bars in uncracked concrete and by 3.49, 12.36, and 5.82 for bare, coated, and MC bars in cracked concrete.
^e Rate converts to value for ECR 10 years after corrosion initiation.
^f Rate converts to value for ECR after 50-m (1.9-mil) zinc layer is consumed.

For systems with bare bars in uncracked concrete, Conv. has the highest total corrosion rate, 4.74 m/year (0.187 mil/year), while Conv.*(HY) has the lowest estimated total corrosion rate, 0.588 m/year (0.0231 mil/year). For systems with bare bars in cracked concrete, Conv. has the greatest total corrosion rate, 8.19 m/year (0.322 mil/year), and Conv.*(HY) again has the lowest estimated total corrosion rate, 0.0827 mil/year (2.10 m/year).

For systems with coated bars in uncracked concrete, conventional ECR, ECR with increased adhesion, and ECR with Rheocrete^® are assigned total corrosion rates of 17.8 and 98.3 m/year (0.701 and 3.87 mil/year) in uncracked and cracked concrete, respectively. ECR with calcium nitrite is assigned total corrosion rates of 17.8 and 71.6 m/year (0.701 and 2.82 mil/year) in uncracked and cracked concrete, and ECR with primer containing microencapsulated calcium nitrite is assigned total corrosion rates of 17.8 and 57.5 m/year (0.701 and 2.26 mil/year) in uncracked and cracked concrete (the latter for the first 10 years, at which point it changes to 98.3 m/year (3.87 mil/year), the value for conventional ECR). MC has a higher corrosion rate than ECR in uncracked concrete, 27.7 m/year (1.09 mil/year), but a lower corrosion rate than ECR in cracked concrete, 46.2 m/year (1.82 mil/year); this is due to the difference in multipliers compared to bars without the zinc coating under the epoxy. This behavior will govern until the 50-m (2-mil) zinc layer is consumed, after which the bar is treated as corroding as conventional ECR. ECR with Hycrete™ has estimated total corrosion rates in both uncracked and cracked concrete of 9.1 and 53.4 m/year (0.36 and 2.10 mil/year), respectively.

Corrosion Loss to Cause Concrete Cracking

A number of experimental studies have been performed to determine the corrosion loss on steel reinforcing bars required to crack (or delaminate) concrete. (See references 44 and 67-71.) All included corrosion of bare bars, and two included bars on which only a portion of the bar surface could corrode.^(44,71) One of those efforts was conducted in concert with this study and included finite element analyses to supplement the experimental data.⁽⁴⁴⁾ Based on this work, which is summarized in appendix B, the corrosion loss in m or mil required to crack concrete can be expressed as shown in figure 173 and figure 174.

Figure 173. Equation. Corrosion loss to crack concrete in SI units.

Figure 174. Equation. Corrosion loss to crack concrete in English units.

Where:

x_crit = Corrosion loss at crack initiation, m or mil.

C = Cover, mm or inches.

D = Bar diameter, mm or inches.

L_f = Fractional length of bar corroding, L_corroding/L_bar.

A_f = Fractional area of bar corroding, A_corroding/A_bar.

For a conventional steel No. 16 (No. 5) bar with a concrete cover of 76.2 mm (3.0 inches), L_f= A_f = 1.0, and the value of x_crit is 56 m (2.2 mil). For a No. 16 (No. 5) epoxy-coated bar with a damage pattern equal to that used for the field test specimens (3-mm (0.125-inch)-diameter holes spaced at 0.124 m (4.9 inches) on each side of the bar), the fractional length of exposed bar L_f is 0.024, the fractional area of exposed bar A_f is 0.0023, and the value of x_crit is 2,430 m (95.7 mil). For the purposes of the analysis, the bars in bridge decks are assumed to have the same damage pattern as the bars in the field tests.

Propagation Time

The time from corrosion initiation to initial delamination of the concrete cover due to the formation of stress-induced cracks caused by expansive corrosion products (propagation time) for each system is found by dividing the corrosion losses required to crack concrete, calculated above, by the estimated total corrosion rates listed in table 44. Because bridge decks inevitably develop cracks over and parallel to the reinforcement due to settlement of plastic concrete and shrinkage of the hardened concrete, the comparisons using the corrosion rates in cracked concrete should provide a more accurate representation of corrosion in bridge decks and are used for the balance of the analysis.

The estimated times to first cracking after corrosion initiation are presented in table 45. Based on the corrosion rate in cracked concrete, ECR with Hycrete™ has the longest propagation time, about 46 years, while the other coated bar systems have propagation times ranging from 25 to 34 years. Conventional reinforcement cast in concrete with Hycrete™ has a propagation time of about 27 years. The other systems involving conventional reinforcement cast in concrete without and with a corrosion inhibitor have propagation times of 6.8 years.

Table 45. Estimated times to formation of initial delamination cracks after corrosion initiation (propagation times) (years).

Steel Designationa	Cracked Concrete
Conv.	6.8
Conv.*(RH)	6.8
Conv.*(DCI)	6.8
Conv.*(HY)	26.6
ECR	24.8
ECR(Chromate)	24.8
ECR(DuPont)	24.8
ECR(Valspar)	24.8
ECR(RH)	24.8
ECR(DCI)	34.0
ECR(HY)	45.6
ECR(primer/Ca(NO2)2)	28.9
MC	25.5

^a See table 1 for abbreviation definitions.

Time to First Repair

The time to first repair for each corrosion- protection system is found by combining the time to corrosion initiation, the time to initial delamination cracking of the concrete after corrosion initiation, and the time between first cracking to the time when the deck is repaired. The latter period is based on the observation that a bridge deck is not fully repaired when the first crack forms. Rather, the bridge typically undergoes a series of short-term temporary repairs. To account for the period of temporary repairs, a 10-year delay between first cracking and repair is assumed based on the experience of the KDOT.

Table 46 compares expected times to first repair for all corrosion-protection systems based on the corrosion rate in cracked concrete. As shown in the table, conventional reinforcement has an expected time to first repair of 19 years, which is within the range of 10 to 25 years predicted by KDOT and SDDOT maintenance engineers.⁽²³⁾ Among systems with conventional reinforcement with inhibitors, conventional reinforcement used in conjunction with concrete containing Rheocrete^® and calcium nitrite have expected times to first repair of 21 and 23 years, slightly greater than those observed for conventional reinforcement without inhibitors. At 38 years, conventional reinforcement used in conjunction with concrete containing Hycrete™ has more than twice the age to first repair as conventional reinforcement with no inhibitor. ECR and ECR with increased adhesion have expected times to first repair of 55 years compared to the 35 to 40 years estimated by KDOT and SDDOT.⁽²³⁾ Most bridges containing ECR have not yet reached this age. Systems containing ECR used in concrete containing corrosion inhibitors have times to first repair ranging from 55 to 69 years, with ECR with calcium nitrite and ECR with primer containing microencapsulated calcium nitrite giving the longest times at 64 and 69 years, respectively. Systems with MC reinforcement have an expected time to first repair of 56 years. Thus, based in the assumption that cracked concrete will dominate corrosion behavior, all systems will require at least one repair during the assumed 75-year design life of the deck.

Table 46. Time to first repair based on corrosion rate in cracked concrete

Steel Designationa	Time to Initiationb	Time from Initiation to Crackingc	Time from Cracking to Repaird	Expected Time to First Repair
Conv.	2.2	6.8	10	19
Conv.*(RH)	4.0	6.8	10	21
Conv.*(DCI)	6.0	6.8	10	23
Conv.*(HY)	1.0	26.6	10	38
ECR	20.0	24.8	10	55
ECR(Chromate)	20.0	24.8	10	55
ECR(DuPont)	20.0	24.8	10	55
ECR(Valspar)	20.0	24.8	10	55
ECR(RH)	20.0	24.8	10	55
ECR(DCI)	20.0	34.0	10	64
ECR(HY)	2.8	45.6	10	58
ECR(primer/Ca(NO2)2)	29.7	28.9	10	69
MC	20.0	25.5	10	56

^a See table 1 for abbreviation definitions.
^b See table 41.
^c See table 45 for cracked concrete.

COST EFFECTIVENESS

A 75-year economic life is used to compare the costs associated with the various corrosion-protection systems for a typical bridge deck. A 46-m (150-ft)-long, 11-m (36-ft)-wide, 216-mm (8.5-inch)-thick bridge deck is used in the analysis. Costs include those for initial construction and repair over the 75-year period. With the exception of steel and admixture prices, costs are based on experience in Kansas and South Dakota for the years 2004 through 2008. Prices during this period are considered to be more indicative of the long term than prices between 2008 and 2011, which are representative of a depressed construction market. User costs are not included in the analysis.

Initial Cost

The material costs for reinforcement and inhibitors used in this analysis are provided by the material suppliers. For conventional reinforcement, the base cost is $0.77/kg ($0.35/lb). ECR and ECR with increased adhesion from DuPont™ and Valspar^® have a base cost of $0.99/kg ($0.45/lb). ECR with chromate pretreatment and ECR with the calcium nitrite primer have a base cost of $1.10/kg ($0.50/lb). MC reinforcement has a base cost of $1.65/kg ($0.75/lb). A placement cost of $1.14/kg ($0.52/lb) is used for all reinforcement. A steel reinforcement density of 163 kg/m³ (275 lb/yd₃) is used, based on the average quantity of steel used in 12 bridge decks constructed in Kansas between 2004 and 2007.⁽⁵⁸⁾ A 216-mm (8.5-inch)-thick bridge deck requires 35.2 kg/m² (64.9 lb/yd²) of steel based on the surface area of deck, as shown in figure 175.

Figure 175. Equation. Typical quantity of reinforcement.

Using the required reinforcement per unit surface area determined using the equation in figure 175, the reinforcement costs for each system are calculated, as shown in the equations in figure 176 through figure 179 and listed in table 47.

Figure 176. Equation. Reinforcement cost for conventional steel.

Figure 177. Equation. Reinforcement cost for ECR, ECR(Dupont), and ECR(Valspar).

Figure 178. Equation. Reinforcement cost for ECR(Chromate) and ECR(primer/Ca(NO₂)₂).

Figure 179. Equation. Reinforcement cost for MC reinforcement.

Table 47. Total in-place cost for reinforcement per unit area of bridge deck.

Steel Designationa	Reinforcement Cost		Reinforcement Used		Total Cost
	$/kg	$/lb	kg/m 2	lb/yd 2	$/m 2	$/yd 2
Conv.	1.91	0.87	35.2	64.9	67.23	56.35
ECR	2.13	0.97	35.2	64.9	74.98	62.84
ECR(Chromate)	2.24	1.02	35.2	64.9	78.85	66.08
ECR(DuPont)	2.13	0.97	35.2	64.9	74.98	62.84
ECR(Valspar)	2.13	0.97	35.2	64.9	74.98	62.84
ECR(primer/Ca(NO2)2)	2.24	1.02	35.2	64.9	78.85	66.08
MC	2.79	1.27	35.2	64.9	98.21	82.31

^a See table 1 for abbreviation definitions.

The base in-place cost of concrete with no inhibitors used in this study is $735.75/m³ ($562.51/yd³) based on costs between 2004 and 2007, updated to July 2008.^(58,72) For corrosion inhibitors, the dosage rates are the rates used in this study and are based on manufacturer recommendations. Rheocrete^® costs $6.08/L ($23.00/gal) and has dosage rate of 5 L/m³ (1 gal/yd³), equal to $30.40/m³ ($23.00/yd³) over the base cost of the concrete. DCI^® S costs $1.32/L ($5.00/gal) and has a dosage rate of 15 L/m³ (3 gal/yd³), equal to $19.80/m³ ($15/yd³). Hycrete™ costs $4.95/L ($18.75/gal) and has a dosage rate of 7.6 L/m³ (1.54 gal/yd³). To counteract the reduction in strength and low freeze-thaw resistance observed in concrete containing Hycrete™, an additional 35.6 kg/m³ (60 lb/yd³) of portland cement at $0.138/kg ($0.0625/lb) is added, for a cost of $42.53/m³ ($32.63/yd³) for Hycrete™ over the in-place cost of conventional concrete.^[4]

Assuming a 216-mm (8.5-inch)-thick bridge deck, 0.216 m³ of concrete are required per 1-m² (0.236 yd³ per 1-yd²) surface area of deck. Concrete costs for all corrosion-protection systems per unit surface area are calculated using the equations in figure 180 through figure 183 and are shown in table 48.

Figure 180. Equation. Costs for concrete placed with conventional reinforcement.

Figure 181. Equation. Costs for concrete placed with Rheocrete^® inhibitor.

Figure 182. Equation. Costs for concrete placed with calcium nitrite inhibitor.

Figure 183. Equation. Costs for concrete placed with Hycrete™inhibitor.

Table 48. Total in-place cost for concrete per unit area of bridge deck.

Steel Designationa	Concrete Cost		Inhibitor Cost		Concrete Use		Total Cost
	$/m 3	$/yd 3	$/m 3	$/yd 3	m 3/m 2	yd 3/yd 2	$/m 2	$/yd 2
Conv.	735.75	562.51	-	-	0.216	0.236	158.92	132.75
RH	735.75	562.51	30.40	23.00	0.216	0.236	165.49	138.18
DCI	735.75	562.51	19.80	15.00	0.216	0.236	163.20	136.29
HYa	740.66	566.29	37.62	28.88	0.216	0.236	168.11	140.46

- Indicates no inhibitor used.
^a Additional 35.9 kg/m³ (60 lb/yd³) cement added to counteract strength reduction.

The total initial cost, equal to the sum of reinforcement and concrete costs, for each system is shown in figure 184 through figure 194 and table 49.

Figure 184. Equation. Total initial cost for decks with conventional steel.

Figure 185. Equation. Total initial cost for decks with conventional steel and Rheocrete^® inhibitor.

Figure 186. Equation. Total initial cost for decks with conventional steel and calcium nitrite inhibitor (DCI).

Figure 187. Equation. Total initial cost for decks with conventional steel and Hycrete™ inhibitor.

Figure 188. Equation. Total initial cost for decks with ECR, ECR(Dupont), and ECR(Valspar).

Figure 189. Equation. Total initial cost for decks with ECR(Chromate).

Figure 190. Equation. Total initial cost for decks with ECR and Rheocrete^® inhibitor.

Figure 191. Equation. Total initial cost for decks with ECR and calcium nitrite inhibitor (DCI).

Figure 192. Equation. Total initial cost for decks with ECR and Hycrete™ inhibitor.

Figure 193. Equation. Total initial cost for decks with ECR with primer.

Figure 194. Equation. Total initial cost for decks with MC reinforcement.

Table 49. Total in-place cost for corrosion protection systems.

Steel Designationa	Reinforcement Cost		Concrete Cost		Total Cost
	$/m 2	$/yd 2	$/m 2	$/yd 2	$/m 2	$/yd 2
Conv.	67.23	56.35	158.92	132.75	226.15	189.10
Conv.(DCI)	67.23	56.35	163.20	136.29	230.43	192.64
Conv.(RH)	67.23	56.35	165.49	138.18	232.72	194.53
Conv.(HY)	67.23	56.35	168.11	140.46	235.34	196.81
ECR	74.98	62.84	158.92	132.75	233.90	195.59
ECR(Chromate)	78.85	66.08	158.92	132.75	237.77	198.83
ECR(DuPont)	74.98	62.84	158.92	132.75	233.90	195.59
ECR(Valspar)	74.98	62.84	158.92	132.75	233.90	195.59
ECR(DCI)	74.98	62.84	163.20	136.29	238.17	199.13
ECR(RH)	74.98	62.84	165.49	138.18	240.46	201.02
ECR(HY)	74.98	62.84	168.11	140.46	243.08	203.30
ECR(primer/Ca(NO2)2)	78.85	66.08	158.92	132.75	237.77	198.83
MC	98.21	82.31	158.92	132.75	257.13	215.06

^a See table 1 for abbreviation definitions.

A deck with conventional reinforcement has the lowest initial in-place cost, $226.15/m² ($189.10/yd²). This increases to $233.90/m² ($195.59/yd²) for ECR, which exceeds the cost of a deck with conventional steel and calcium nitrite or Rheocrete^® at $230.43/m² ($192.64/yd²) or $232.72/m² ($194.53/yd²) but is less than the cost of a deck with conventional steel and Hycrete™ at $235.34/m² ($196.81/yd²). At $257.13/m² ($215.06/yd²), a deck with MC reinforcement has the highest in-place cost of all systems with an epoxy coating.

Repair Costs

Repair costs for a typical 216-mm (8.5-inch) bridge deck were obtained from KDOT. Current data include repair of bridge decks with conventional reinforcement only because bridge decks constructed since the late 1970s have been constructed using ECR and have not needed repair as of the date of this report. It is estimated that repair costs of bridge decks with ECR will be similar to those for decks with conventional reinforcement. Based on experience in Kansas and South Dakota, repairs are assumed to last for 25 years.⁽²³⁾

In Kansas, repair consists of applying either a silica fume or polymer overlay to the deck. Repair costs include a unit cost for the overlay and machine preparation, costs for mobilization and traffic control, and patching costs based on the percentage of decks that received partial or full depth repairs, 5 and 17 percent, respectively. Based on an analysis of bid costs from 2008 through 2010, total repair costs are given in figure 195 and figure 196.

Figure 195. Equation. Total repair costs with silica fume overlay.

Figure 196. Equation. Total repair costs with polymer overlay.

The current KDOT repair costs are compared with a previous analysis based on costs obtained from SDDOT, which are based on an average of costs for bridge deck repair projects for the year 2006.⁽⁷²⁾ A typical repair project includes costs for removing deleterious concrete and replacing with a low-slump dense concrete overlay, bridge rail modifications, approach guard rail replacement, approach pavement work, mobilization, traffic control, and other miscellaneous costs. Costs were determined per square yard for the 46-m (150-ft)-long deck described at the beginning of this section. A summary of the repair costs is shown in table 50 and described in figure 197 through figure 202.

Table 50. Repair costs for bridge decks in South Dakota.⁽⁷⁵⁾

Item	Unit	Cost	Cost/yd2	Unit	Cost	Cost/m2
Low slump dense concrete overlay	Per yd2	$130.00	$130	Per m2	$155.00	$155
Bridge rail modification	Per linear ft	$62.00	$31	Per linear m	$814.00	$37
Approach guard rail	Lump sum	$16,500.00	$28	Lump sum	$16,500.00	$34
Approach pavement work	Lump sum	$17,000.00	$28	Lump sum	$17,000.00	$34
Mobilization	Lump sum	$25,000.00	$42	Lump sum	$25,000.00	$50
Traffic control and misc.	Lump sum	$20,000.00	$33	Lump sum	$20,000.00	$39
Total repair costs			$292			$349

1 m = 1.09 yd = 3.28 ft

Figure 197. Equation. Bridge rail modification cost.

Figure 198. Equation. Approach guard rail cost.

Figure 199. Equation. Approach pavement work cost.

Figure 200. Equation. Mobilization cost.

Figure 201. Equation. Traffic control and miscellaneous costs.

Figure 202. Equation. Total repair costs.

A comparison of the repair costs provided by KDOT for 2008 through 2010 ($224/m² or $175/m²($188/yd² or $147/yd²)) and those provided by SDDOT for 2006 ($349/m² ($292/yd²)) shows that costs have significantly decreased in the current economic climate. Because the current highly competitive environment in the construction industry is not expected to be long term, the higher costs for 2006 analysis are used in this study.

Present Value

The total life cycle cost of each corrosion-protection system is calculated using the times to first repair for systems in cracked concrete listed in table 46. Cost effectiveness is based on the initial cost of the deck and the present value of future repair costs. The present value is calculated as shown in figure 203, where P is the present value, F is the future cost of a repair ($349/m² ($292/yd²)), i is the discount rate, and n is the time to repair.

Figure 203. Equation. Present value.

For this study, discount rates of 2, 4, and 6 percent are assumed. As the most realistic, the value 2 percent is used for most of the discussion that follows.

Table 51 and table 52 list the estimated costs over a 75-year design life using the time to first repair based on the corrosion rate in cracked concrete. Under this scenario, all of the corrosion-protection systems must be repaired at least once during the 75-year design life. Conventional reinforcement in concrete without a corrosion inhibitor and conventional reinforcement in concrete with the inhibitors calcium nitrite (DCI) and Rheocrete^® must be repaired three times, and conventional reinforcement in concrete with the inhibitor Hycrete™ must be repaired twice. All of the coated-bar systems must be repaired once during the 75-year design life of the deck.

Table 51. Total costs ($/m²) over 75-year design life for corrosion-protection systems using time to first repair based on corrosion rates in cracked concrete.

Steel Designationa	Initial Cost, $/m2	Time to Repair, years			Repair Cost, $/m2	Present Cost, $/m2
		1	2	3		i=2%	i=4%	i=6%
Conv.	226	19	44	69	349	700	477	374
Conv.*(DCI)	230	23	48	73	349	670	446	349
Conv.*(RH)	233	21	46	71	349	690	466	366
Conv.*(HY)	235	38	63	-	349	502	345	283
ECR	234	55	-	-	349	351	274	248
ECR(Chromate)	238	55	-	-	349	355	278	252
ECR(DuPont)	234	55	-	-	349	351	274	248
ECR(Valspar)	234	55	-	-	349	351	274	248
ECR(DCI)	238	64	-	-	349	336	266	246
ECR(RH)	240	55	-	-	349	358	281	255
ECR(HY)	243	58	-	-	349	353	278	255
ECR(primer/Ca(NO2)2)	238	69	-	-	349	327	261	244
MC	257	56	-	-	349	373	296	271

- No repair required.
^a See table 1 for abbreviation definitions.

Table 52. Total costs ($/yd²) over 75-year design life for corrosion protection systems using time to first repair based on corrosion rates in cracked concrete.

Steel Designationa	Initial Cost, $/yd2	Time to Repair, years			Repair Cost, $/yd2	Present Cost, $/yd2
		1	2	3		i=2%	i=4%	i=6%
Conv.	189	19	44	69	292	586	399	313
Conv.*(DCI)	193	23	48	73	292	560	373	291
Conv.*(RH)	195	21	46	71	292	577	390	306
Conv.*(HY)	197	38	63	-	292	420	288	237
ECR	196	55	-	-	292	294	229	207
ECR(Chromate)	199	55	-	-	292	297	232	211
ECR(DuPont)	196	55	-	-	292	294	229	207
ECR(Valspar)	196	55	-	-	292	294	229	207
ECR(DCI)	199	64	-	-	292	281	223	206
ECR(RH)	201	55	-	-	292	299	235	213
ECR(HY)	203	58	-	-	292	295	233	213
ECR(primer/Ca(NO2)2)	199	64	-	-	292	274	219	204
MC	215	56	-	-	292	312	248	226

- No repair required.
^a See table 1 for abbreviation definitions.

Because it requires three repairs and has the lowest time to first repair, conventional reinforcement without corrosion inhibitors in the concrete has the highest present cost, $700/m² ($586/yd²) at a 2 percent discount rate. Conventional reinforcement used in conjunction with concrete containing calcium nitrite and Rheocrete^® has present costs of $670/m² and $690/m² ($560/yd² and $577/yd²), respectively. Conventional reinforcement used in conjunction with concrete containing Hycrete has the lowest present cost among systems with conventional reinforcement, $502/m² ($420/yd²), but is less cost effective than any of the coated bar systems.

ECR in concrete with the calcium nitrite primer is the most cost-effective protection system, with a present cost of $327/m² ($274/yd²) at a discount rate of 2 percent. ECR with calcium nitrite is the next most efficient system with a present cost of $336 m² ($281/yd²) at a discount rate of 2 percent. Conventional ECR, as well as increased adhesion epoxies from DuPont™ and Valspar^®, have present costs of $351/m² ($294/yd²) at a discount rate of 2 percent. ECR(HY)
and ECR(Chromate) have present costs of $353/m² and $355/m² ($295/yd² and $297/yd²) at a discount rate of 2 percent; however, at a discount rate of 6 percent, ECR(Chromate) is more cost effective than ECR(HY). MC reinforcement has a present cost of $373/m² ($312/yd²) at a 2 percent discount rate. The total spread in cost for the coated bar systems is $46/m² ($38/yd²), or about 12 percent of the highest price. Considering the level of uncertainty inherent in the analysis, the differences in present costs for coated bar systems are not significant. However, the differences in the number of projected repairs and the differences in present costs between uncoated and coated bars systems are significant.