INTRODUCTION AND RESEARCH APPROACH
Over the last fifteen years, corrosion-inhibiting admixtures (CIAs) have gained popularity for their long-term protection of reinforced and prestressed concrete in highway bridges and various structures Despite their widespread use, there is ongoing debate regarding the effectiveness and benefits of CIAs in concrete applications.
In 1992, a Joint Committee comprising AASHTO, ACG, and ARTBA developed a Manual for Corrosion Protection of Concrete Components in Bridges, which discusses various corrosion protection systems, including Corrosion Inhibiting Additives (CIAs) However, the manual lacks specifications or guidelines for engineers to effectively evaluate and compare CIAs Consequently, engineers often depend on manufacturer-provided information, which may not adhere to consistent testing procedures This highlights the need for research to establish standardized tests for assessing CIAs, enabling engineers to make informed product comparisons and recommendations.
This research aimed to develop evaluation procedures and performance criteria for corrosion-inhibiting admixtures (CIAs) as defined by the American Concrete Institute The focus was on chemical admixtures added to portland cement concrete in small concentrations to enhance corrosion protection While materials like microsilica, fly ash, and ground-granulated blast-furnace slag can offer some corrosion resistance, they were not classified as CIAs for this study, though they may still be included in concrete mixtures containing CIAs.
The work plan for NCHRP Project 10-45 was divided into the following two Phases.
Task 1 – State-of-the-Art of Corrosion Inhibiting Admixtures.
In Task 1, we conducted a comprehensive review of both domestic and international literature, research findings, performance data, and current practices regarding the use, testing, and evaluation of chemical admixtures (CAs) We compiled a detailed list of available CAs, outlined the mechanisms by which each admixture functions, and summarized their effects on the properties of both fresh and hardened concrete.
Task 2 assessed and analyzed the screening and long-term verification test procedures utilized in the United States and internationally for measuring the effectiveness of corrosion-inhibiting admixtures (CIAs) Key factors considered included performance predictability, practicality, cost, and other relevant aspects, with a particular focus on test duration, the quality of concrete in bridge components, and performance in cracked concrete.
A comprehensive research plan was established to conduct laboratory investigations aimed at evaluating and validating test procedures for the performance of CIAs in Task 3 An interim report was generated, summarizing the research conducted in Tasks 1 through 3 and outlining the work plan for Phase II of the project (Task 4) For detailed information on the experimental work plan and test matrices, please refer to Appendix A.
• Prediction of corrosion rate as a function of chloride concentration and CIA concentration.
• Prediction of chloride threshold concentration.
• Prediction of chloride penetration rate through concrete as a function of CIA concentration.
• Measurement of concrete property data as a function of CIA concentration.
• Measurement of CIA performance in the presence of preformed cracks in the concrete extending down to the steel bar surface.
In Task 5, a comprehensive laboratory test program was conducted to evaluate the corrosion inhibiting effectiveness of the admixture The test procedures specifically examined how the admixture influenced both the fresh and hardened properties of concrete.
In Task 6, a draft standard practice was created, titled “Proposed Method For Qualifying Corrosion Inhibiting Admixtures That Mitigate Corrosion Of Reinforcing Steel In Concrete.” This practice outlines a laboratory testing protocol designed to gather essential data for evaluating corrosion-inhibiting admixtures (CIA) based on established performance criteria.
RESEARCH FINDINGS
REVIEW OF COMMERCIALLY AVAILABLE CIAS
Appendix B provides a comprehensive review of corrosion inhibiting admixtures (CIAs), highlighting five commercially available options: DCI by W.R Grace, Rheocrete 222 from Master Builders, Ferrogard 901 (an enhanced version of Armatec 2000 by Sika), MCI 2000 from Cortec, and Catexol 1000 Cl by Axim Concrete Technologies Notably, Ferrogard, Armatec, and MCI share similar active ingredients, leading to their collective analysis This section briefly outlines the characteristics of the four identified CIAs.
1 Rheocrete 222 and Rheocrete 222+ (Master Builders) A water-based combination of amines and esters Rheocrete 222+ is a new and, supposedly, improved version.
Ferrogard 901 is a modified formulation of Armatec 2000 and MCI 2000, consisting of a blend of surfactants and amine salts, particularly dimethyl ethanolamine (DMEA), in a water carrier While both Armatec and MCI were previously produced by Cortec, the Armatec variant, which was recently marketed by Sika, features slightly different ingredient proportions compared to the MCI version Ferrogard is exclusively manufactured by Sika.
3 DCI and DCI-S (W.R Grace) Calcium nitrite-based admixture (about 30% concentration of the active ingredient) DCI-S contains a set-retarding admixture.
Catexol 1000 CI, produced by Axim Concrete Technologies, is a water-based solution comprised of amine derivatives However, during the review, there was no specific performance data available for this admixture.
Calcium nitrite, the active ingredient in DCI corrosion inhibitors, protects reinforcing steel by promoting the development of a passive oxide film on its surface, classifying DCI as an anodic inhibitor.
Rheocrete, Ferrogard/Armatec/MCI, and Catexol are organic film-forming inhibitors that function as mixed type inhibitors, effectively hindering both anodic and cathodic reactions These inhibitors are designed to reduce chloride penetration into concrete, providing enhanced protection against corrosion.
Limited information exists on Catexol, as only the manufacturer's data sheets are available Consequently, the report centers on a comprehensive performance analysis of Rheocrete, DCI, and Ferrogard/Armatec/MCI Most performance data for these compounds is derived from laboratory tests, which have shown that these three corrosion-inhibiting admixtures (CIAs) can effectively reduce corrosion of reinforcing steel in concrete under certain conditions However, the protective mechanisms varied, and their effectiveness was influenced by the specific test conditions.
Manufacturers have primarily generated the performance data for corrosion inhibitors (CIs), with calcium nitrite-based DCI being the most extensively tested DCI has been available for a longer duration compared to other commercial CIs Additional data from Federal Highway Administration (FHWA) projects and state departments of transportation (DOTs) indicate that CIs generally reduce the time to corrosion initiation and lower the corrosion rate after initiation compared to control specimens However, the extent of these benefits varies based on the specific type of CI, its concentration, and the testing conditions Overall, while the data suggests that CIs can effectively reduce corrosion, precise benefits of different CIs remain difficult to define.
Appendix C includes the industry survey and findings from limited field studies conducted on Concrete Improvement Agents (CIAs) While CIAs have been utilized across various concrete structures, there is a notable scarcity of field performance data for these agents This deficiency is primarily attributed to the extensive time needed for performance assessment and the absence of follow-up studies after the structures have been established.
The laboratory testing program assessed the corrosion rate of steel in concrete, analyzed chloride penetration into concrete, and evaluated the mechanical properties of concrete Detailed results for each type of testing were conducted for various concentrations of each corrosion-inhibiting agent (CIA) (refer to Appendix A for the experimental plan).
In the conducted tests, two types of concrete were utilized: Concrete 0, which featured a Type I portland cement composition with a water-to-cement ratio of 0.40, included 30% cementitious materials, silica coarse aggregate, quartz sand, and maintained a 6% air content Concrete 1 was designed similarly but incorporated an additional 7% silica fume into the cementitious mix.
The testing program comprised fixed-chloride tests to assess the corrosion rate based on varying chloride concentrations and chloride threshold tests aimed at identifying the minimum chloride level required to initiate corrosion.
This study focuses on several key tests to evaluate the effectiveness of Corrosion Inhibiting Admixtures (CIA) in cracked concrete conditions It includes simulated cracked minibeam tests to assess the benefits of CIA, long-term concrete slab tests to analyze its impact on corrosion, and chloride penetration tests to determine the effective diffusion coefficient for chlorides in concrete Additionally, concrete property tests are conducted to investigate how CIA influences critical parameters such as compressive strength, flexural strength, modulus of elasticity, slump, setting time, and air void content.
The project aimed to establish a standardized testing protocol and performance criteria for evaluating three Confidential Information Agents (CIAs), referred to as CIA-A, CIA-B, and CIA-C, without disclosing their identities The CIAs were assessed at four different concentrations: 10%, 50%, 75%, and 100% of the manufacturer’s maximum recommended dosage.
The fixed chloride test aimed to determine the corrosion rate of steel based on chloride and CIA concentration Appendix D presents the corrosion rate and potential maps for each CIA in Concrete 0, reflecting targeted chloride concentrations of 3, 6, and 9 kg/m³ (5, 10, and 15 lb/yd³) For Concrete 0, the measured chloride concentrations closely aligned with the targets (refer to Appendix E for detailed data) However, the Concrete 1 tests did not achieve the targeted chloride concentrations, which hindered the creation of similar plots for Concrete 1.
The mean corrosion rate was calculated for each concentration of CIA, averaging the results across all specimens and varying chloride levels This analysis yields a representative corrosion rate for each CIA concentration, as illustrated in Figures 1 and 2, which compare the corrosion performance of CIA.
Figure 1 Comparison of CIA corrosion performance in Concrete 0 (average for all chloride concentrations) [1 mm/yr = 39 mpy]
Figure 2 Comparison of CIA corrosion performance in Concrete 1 (average for all chloride concentrations) [1 mm/yr = 39 mpy]
INTERPRETATION, APPRAISAL, APPLLICATION
The main goal of this research was to establish a testing procedure to assess the performance of Corrosion Inhibiting Admixtures (CIAs) in short-term laboratory conditions Additionally, it aimed to create performance criteria that would quantitatively qualify CIAs for application in concrete structures.
The first step in this process was to develop a life prediction model The sole purpose of this model was to facilitate the application of the performance criterion for a
The proposed model does not account for all variables essential for accurately predicting the lifespan of a concrete structure in a specific location Furthermore, it fails to address the implications of cracked structures in its predictions.
The following three rate processes govern life of a concrete structure:
1 Rate of chloride diffusion into the concrete.
2 Rate of corrosion, once initiated.
3 Rate of development of damage.
The life of a structure can be divided into the following three phases.
• Phase II – Corrosion Propagation without Damage.
• Phase III – Damage to Structure.
The following information is required for the above model and can be determined by the testing protocol utilized in this research.
1 Effective diffusion coefficients for chloride penetration into concrete.
2 Critical chloride concentration required to initiate corrosion.
• Phase II – Corrosion Propagation without Damage.
1 Corrosion rate as a function of chloride concentration.
2 Chloride concentration as a function of time.
3 Cumulative corrosion necessary to initiate concrete damage.
• Phase III – Damage to Structure.
1 Damage versus cumulative corrosion relationship (at a minimum, cumulative corrosion to cause cracking).
The following provides the predictive modeling analyses for the three CIAs studied in this research.
The conditions affecting actual structures, particularly in non-marine applications, are complex and difficult to define due to the variability in salting events The duration and frequency of chloride solution applications on concrete surfaces fluctuate significantly, influenced by weather conditions Consequently, multiple salt applications occur over time, each lasting an indeterminate period, allowing additional salt to penetrate the concrete during these intervals When no salt is present on the surface, the existing salt within the concrete can redistribute Therefore, any model aimed at establishing a general performance criterion must account for these variable conditions and prioritize ease of calculation for practical application.
In this model, it is assumed that salt applications begin at the start of the structure's lifespan, maintaining a consistent surface salt concentration of 18 kg/m³ (30 lb/yd³) after the initial application This assumption greatly simplifies the calculation of chloride concentration over time Additionally, it is assumed that once cracking occurs, the corrosion rate remains unaffected by the presence of these cracks.
Effective diffusion coefficients for chloride penetration into concrete are crucial for predicting chloride concentration at the steel surface over time under specific conditions This information is essential for determining the transition between Phase I and Phase II, as well as for calculating corrosion rates in both Phase II and III.
Table 5 presents the diffusion coefficients for the Controlled Infiltration Additives (CIAs) at their maximum recommended dosage, alongside the control samples without CIAs, for the two types of concrete analyzed in this study The chloride concentration at a specific depth (x) and time (t) is detailed in the findings.
− where Co is the concentration of chlorides at the concrete surface (assumed to be
The effective diffusion coefficient (Deff) and the error function (erf) are key components in this model, which calculates chloride concentrations at a reinforcement bar depth of 64 mm (2.5 in) It's important to highlight that the value of 18 kg/m³ (30 lb/yd³) used in this model is derived from field experience rather than laboratory tests.
Figures 18 through 20 show the predicted chloride concentrations versus time for the effective diffusion coefficients provided in Table 5.
The addition of silica fume to Concrete 1 notably reduces the diffusion coefficient, significantly affecting chloride concentrations at a depth of 2.5 inches (6.4 cm) from the reinforcement bars This enhancement indicates a marked improvement in the performance of the concrete regarding chloride penetration, which is crucial for the durability of steel reinforcement.
The addition of CIA-B and CIA-C to Concrete 0 significantly reduces chloride concentration at the steel level compared to concrete without CIA In contrast, CIA-A exhibits a higher diffusion coefficient, leading to increased chloride levels at the steel interface over time.
The addition of CIA to Concrete 1 significantly influences chloride concentration, as illustrated in Figure 20 Both CIA-B and CIA-C effectively reduced the diffusion coefficient, leading to lower chloride levels over time In contrast, CIA-A demonstrated only a minimal impact compared to the no-CIA scenario.
The data input necessary to establish the time prior to corrosion initiation is (1) the chloride concentration as a function of time (see Figures 18 through 20, above) and (2)
Table 5 Effective diffusion coefficients for concretes with and without CIAs.
No CIA 1.86E-08 5.21E-09 CIA-A 2.92E-08 5.92E-09 CIA-B 1.29E-08 2.54E-09 CIA-C 1.51E-08 3.25E-09
Time (years) Chloride at Rebar Level (lb/yd 3 )
Figure 18 Chloride concentration versus time for Concretes 0 and 1 (no CIA).
Time (years) Chloride at Rebar Level (lb/yd 3 )
Concrete 0 CIA-A Concrete 0 CIA-B Concrete 0 CIA-C Concrete 0
Figure 19 Chloride concentration versus time for the three CIAs in Concrete 0.
Time (years) Chloride at Rebar Level (lb/yd 3 )
Concrete 1 CIA-A Concrete 1 CIA-B Concrete 1 CIA-C Concrete 1
Figure 20 Chloride concentration versus time for the three CIAs in Concrete 1.
Chloride threshold tests were conducted to determine the corrosion initiation threshold, with results summarized in Table 1 However, due to the absence of chloride threshold data for all Corrosion Initiation Areas (CIAs), these findings were not incorporated into the model.
The chloride threshold for corrosion initiation can be estimated using fixed chloride tests, where linear regression is applied to extrapolate data to identify negligible corrosion rates It's important to note that these tests were conducted at 98% relative humidity without ponding after chloride introduction, in contrast to threshold tests that involved continuous ponding with chloride Table 6 presents equations relating corrosion rates to chloride levels for the tested conditions in Concrete 0, while insufficient data prevented the calculation of equations for Concrete 1.
The linear regression equations for Control, 100% CIA-A, and 100% CIA-C, as shown in Table 6, yielded R² values of 0.20, 0.22, and 0.24, indicating significant data scatter The highest R² value of 0.83 was found in the 75% dosage CIA-A data For 100% CIA-B dosage, the regression equation was y = 0.0190X + 0.0521 with an R² of 0.13, while the 50% CIA-B data produced a better fit with y = 0.0309X - 0.0873 and an R² of 0.45 The 100% CIA-B data yielded unrealistic chloride threshold predictions at low corrosion rates Notably, corrosion rate versus chloride concentration data for both 100% and 50% CIA-B dosages were similar at intermediate chloride levels, prompting the use of the 50% CIA-B equation for analysis, which demonstrated slightly better corrosion performance than the 100% dosage.
Using equations from Table 6, the chloride threshold can be estimated as the chloride concentration at negligible corrosion rate Assuming the same value for
Table 6 Equations for linear fit to corrosion rate versus chloride concentration.
[1 mm/yr = 39 mpy], [1 kg/m 3 = 1.67 lb/yd 3 ]
0 CIA-C 100 0.0127 -0.0118 0.24 y = Corrosion rate in mpy. x = Chloride concentration in lb/yd 3 m = slope b = y intercept
R 2 = Percent of variance explained by regression fit.
The study utilized 50% CIA data, which exhibited similarities in magnitude to the 100% CIA data, due to inconsistencies observed in the 100% and 75% CIA datasets The corrosion rate at initiation was defined using the equation y = mx + b, with a threshold of 0.1 mpy (0.0025 mm/yr) indicating the onset of corrosion Table 7 compares the predicted chloride threshold from the "Fixed Chloride Tests" with the "Chloride Threshold Tests" for Concrete 0, as no chloride threshold data was available for Concrete 1 Although the chloride threshold values obtained are higher than those typically reported in the literature, there is a reasonable alignment between the two testing methods The elevated chloride threshold values may be attributed to the selected corrosion rate threshold, and it is suggested that the LPR measurements conducted on these small test cells are significantly more sensitive than those taken in concrete field applications or larger slab tests.
If the level of corrosion rate for establishing the threshold is lowered (either 0.05 or 0.025 mpy [0.0013 or 0.00063 mm/yr]), the threshold concentration is closer to expected values (see Table 8).
The chloride threshold for the life prediction model is established through “Fixed Chloride Tests” regression analyses, indicating a necessary chloride concentration for a corrosion rate of 0.05 mpy (0.0013 mm/yr) The results show that under no-CIA, CIA-B, and CIA-C conditions, the chloride threshold values range from 2.4 to 3.0 kg/m³ (4 to 5 lb/yd³), while CIA-A demonstrates a higher threshold value of 4.2 kg/m³ (7.0 lb/yd³) Consequently, CIA-A is identified as the only inhibitor that significantly enhances the chloride threshold for Concrete 0.
CONCLUSIONS AND SUGGESTED RESEARCH 66 APPENDIX A – Experimental Work Plan A-1 APPENDIX B – Literature Review B-1 APPENDIX C – Transportation Agency Surveys C-1
The following conclusions are based on the results of the testing program presented above.
A literature review has identified three primary commercial inhibitors that have demonstrated effective performance in various studies conducted before this project.
•Rheocrete 222 and Rheocrete 222+ (Master Builders) A water-based combination of amines and esters.
•DCI and DCI-S (W.R Grace) Calcium nitrite-based admixture.
Ferrogard 901 is an advanced formulation derived from Armatec 2000 (SIKA) and MCI 2000 (Cortec), consisting of a unique blend of surfactants and amine salts This product features dimethyl ethanolamine (DMEA), commonly known as alkanolamines or amino alcohols, all suspended in a water carrier for optimal performance.
The addition of three tested Chemical Inhibiting Agents (CIAs) did not negatively impact the concrete properties, including compressive strength, flexural strength, modulus of elasticity, electrical resistivity, slump, time-to-set, and air content Standard concrete admixtures effectively mitigated any potential effects, ensuring optimal performance.
3 In general, the effect of CIAs on the diffusion coefficient was consistent for both the standard and silica fume containing concrete.
4 The effect of CIA on corrosion of steel bar in concrete is dependent on the concrete mix, concentration of the CIA, chloride concentration, and exposure conditions.
A predictive model was created to estimate the lifespan of structures using short-term laboratory test results, enabling a quantitative assessment of the benefits of using corrosion-inhibiting admixtures (CIAs) This model was specifically designed for life prediction to apply performance criteria rather than to forecast the actual lifespan of a structure When applied to both control concrete without CIAs and concrete containing CIAs, the model indicated a significant extension of life for all three CIAs analyzed in the study Additionally, the model allows for the relative benefits to be attributed to three main factors: an increase in the chloride threshold for corrosion initiation, a reduction in the slope of the corrosion rate versus chloride concentration equation, and a decrease in the effective diffusion coefficient.
To qualify as a corrosion-inhibiting admixture (CIA), a candidate formulation must meet four established performance criteria.
Criterion 1 The CIA should provide an improvement over the base (no-CIA) condition with respect to the predicted life by a minimum of 25 percent.
To enhance lifespan, it is essential to achieve either an increased chloride threshold for corrosion initiation or a reduced slope in the regression of corrosion rate against chloride concentration.
Criterion 3 The CIA should provide some improvement in corrosion performance for cracked concrete.
Criterion 4 mandates that the CIA must not negatively impact the essential properties of concrete, ensuring compliance with key specifications such as compressive strength, flexural strength, modulus of elasticity, slump, time-to-set, and air content Additionally, while not directly examined in this project, factors like air distribution and shrinkage may also significantly influence concrete performance.
Both the life prediction model, which relies on short-term tests, and the long-term slab tests demonstrated that all three CIAs outperformed the control condition (no-CIA) However, there are some inconsistencies in understanding the reasons behind the enhanced performance when comparing the findings from these two methodologies.
8 A recommended practice was developed for testing and qualifying a proposed admixture as a CIA The testing protocol includes a modified procedure of the
68 fixed-chloride tests performed in this program and the performance criteria discussed above.
The discrepancies highlighted in Conclusion 7 necessitate modifications to the testing protocol to address the identified causes While a preliminary practice has been suggested in Appendix O, additional testing and validation of these procedures are essential before they can be established as standard practice.
The challenges in validating recommended practices are significantly impacted by the low statistical reproducibility of the "Fixed Chloride Tests." This issue arises from difficulties in integrating chlorides into specimens, likely due to CIA interactions, as well as complications in accurately measuring chloride levels.
Follow-on work is required to finalize the testing protocol proposed in this study The following discusses the reasons why additional work is required.
The testing protocol underwent significant modifications based on project test results, with the main change being the implementation of cyclic exposures across all methodologies This adjustment aims to yield more consistent results, as it is believed that critical infrastructure assets (CIAs) react differently to wet-dry cyclic exposures compared to continuous ponding or humidity conditions Consequently, it is essential to evaluate the new testing protocol to determine its effectiveness.
The reproducibility of fixed chloride tests was found to be inadequate, prompting the need for methodological adjustments due to interference from CIAs Key criteria for effective chloride incorporation include ensuring concrete is cured before chloride application, maintaining controllable and consistent chloride concentrations, and achieving uniformity across the steel surface Minor modifications to the incorporation method, such as increasing the temperature for chloride migration, extending migration time, and employing multiple drying cycles, could significantly enhance the reliability and reproducibility of the results An improved methodology is essential for achieving the desired chloride concentration effectively.
The chloride sampling method aimed to collect minimal samples close to the steel surface; however, this approach often led to misleading results due to the heterogeneous nature of concrete Samples tended to have varying ratios of cement to aggregate, resulting in inconsistent chloride concentration measurements when expressed as total weight Despite repeated sampling across multiple specimens, these inconsistencies contributed to the poor reproducibility of the test results.
Comparative analysis of chloride measurement methodologies suggests that the titration method (AASHTO T260-82) may yield more consistent results than the ion-probe method (Germann).
“Rapid Chloride Test”) primarily used in this study It is proposed to use the titration method in any future work.
The laboratory test plan aimed to establish performance criteria for evaluating the service life of a structure up to 75 years, necessitating a testing protocol to gather relevant data To define these criteria, a predictive model was developed, focusing on experimentally measurable parameters The chosen model centers on cumulative corrosion, specifically the cumulative metal loss of reinforcing steel, as corrosion rate can be accurately measured in both laboratory and field settings This article outlines the model and the essential information needed for its application.