Role of Steel and Cement Type on Chloride-Induced Corrosion in Concrete

INTRODUCTION

Chloride-induced reinforcement corrosion is a serious durability problem in reinforced concrete structures. The corrosion of the steel reinforcement embedded in concrete is often due to the result of chloride-induced breakdown of the passive film formed in the high alkaline condition of the concrete pore solution. In hardened concrete, the pH of the concrete pore solution is normally in excess of 13 and the steel reinforcement is generally protected from corrosion at such levels of alkalinity by the formation of a thin oxide layer over the steel surface.1 Even under conditions of high alkalinity, however, the presence of chloride ions may result in corrosion of the steel reinforcement. Chlorides may be introduced into concrete through chloride contaminated aggregates, chloride-containing admixtures, or mixing water.2-5 Chlorides may also enter into hardened concrete from deicing salts in bridge decks and parking structures, from sea water in marine structures, or from soil and ground water containing chloride salts.

The corrosion performance of steel reinforcement embedded in cementitious materials exposed to a chloride environment is a function of both concrete and steel characteristics.6 The addition of alternative cementitious materials such as fly ash, ground granulated blast furnace slag, and silica fume are known to enhance the durability of reinforced concrete structures. It has been reported that the concrete containing these alternative cementitious materials perform excellently in marine environments and in highway structures.7 In addition to improving the durability, the use of these materials in construction reduces the waste deposit and the demand on cement. Fly ash and slag can be added to ordinary portland cement (OPC) at the time of preparation of concrete or can be blended with OPC clinker and intergrinded at the time of manufacturing. The effect of such blending on durability properties of concrete may or may not be similar to that of addition at the time of casting. Where relative control on site is less, especially in remote areas, however, the use of blended cements is generally preferred.

Some research has been done in the past on corrosion performance of different types of steel reinforcement embedded in concrete mixed with chlorides at the time of preparation and subjected to chlorides during the exposure period.6,8,9 Some other works have reported on the corrosion performance of the steel reinforcement in concrete prepared with alternative cementitious materials added to OPC at the time of casting.2,7,10-13 In addition, some studies on corrosion performance of steel in concrete made with blended cements have been reported in the literature.14-17 There are several techniques for the measurement and detection of corrosion of the steel reinforcement embedded in concrete, and it seems that there is no consensus regarding the most accurate method assessing the corrosion performance.13 Thus, techniques used for assessment of corrosion performance of steel reinforcement in concrete vary from one project to another.

The main objective of the current study is to assess the performance of three different types steel reinforcement such as cold twisted deformed (CTD) bars, Thermex thermo-mechanically treated (TMT) bars, Tempcore TMT bars, and three types of cement such as OPC, portland pozzolana cement (PPC), and portland slag cement (PSC) at different levels of sodium chloride in concrete through various performance indicators. For this purpose, a comprehensive experimental investigation was carried out to determine the corrosion current density using an IR-compensated linear polarization resistance technique and half-cell potential values of a centrally embedded steel bar in concrete slab specimens. In addition, the relative resistivity of the cover concrete of all the mixtures has been measured. The free chloride concentration, total chloride concentration, and pH value have been determined. The compressive strength of concrete mixtures has also been obtained.

RESEARCH SIGNIFICANCE

The corrosion resistance of the steel reinforcement in concrete depends on the performance of both the cement and steel type. A suitable combination of the steel and cement type can improve durability of the reinforced concrete structures. In the past, some studies have been done on corrosion performance of different types steel and different types of cement; however, work on corrosion performance of different combinations of steel and cement type is scanty. Therefore, an attempt has been made to evaluate the performance of three different types of steel reinforcement in concrete made with three different types of cement against chloride-induced corrosion so that appropriate combinations of steel and cement type can be selected at the design stage for use in reinforced concrete structures.

EXPERIMENTAL PROGRAM

The experimental program is shown in Fig. 1 for a quick look.

Materials used

The major test variables for the present experimental investigation are steel type, cement type, water-cement ratio (w/c), and admixed chloride content. Three types of steel bars with a diameter of 0.47 in. (12 mm), that is, namely CTD bars, and two varieties of TMT bars were used. One variety of TMT bar is of a particular make that uses the Thermex process for manufacturing while the other variety is of a another particular make that uses the Tempcore process. The carbon (C) content, phosphorus (P) content, and sulfur (S) content of CTD steel used in the present investigation are 0.22%, 0.044%, and 0.037%, respectively. The contents of the three elements in the same sequence as previously mentioned in Thermex TMT and Tempcore TMT steel are 0.19%, 0.046%, and 0.029%, and 0.20%, 0.021%, and 0.044%, respectively. Three types of cement such as OPC satisfying ASTM Type I and Indian standards IS: 8112-1989,18 PPC having 20% pozzolana (fly ash) content and satisfying ASTM Type IP and IS:1489-199119 (Part I), and PSC satisfying ASTM Type IS and IS:455-198920 were used. The specific surface of OPC, PPC, and PSC are 1298.73 ft^sup 2^/lb (266 m^sup 2^/kg), 1728.38 ft^sup 2^/lb (354 m^sup 2^/kg), and 1723.50 ft^sup 2^/lb (353 m^sup 2^/kg), respectively. The three w/c used were 0.45, 0.50, and 0.55. The chloride was added to concrete as sodium chloride of an analytical reagent grade. The concentrations of admixed sodium chloride used were 0, 1.5, 3, and 4.5% by mass of cement. The corresponding admixed chloride contents calculated by dividing the sodium chloride contents with 1.648 (58.44/35.45) were 0, 0.91, 1.82, and 2.73% by mass of cement, respectively. Coarse aggregates with a size of 0.79 in. (20 mm) maximum size of aggregate (MSA) and 0.39 in. (10 mm) MSA of quartzite origin were used in the ratio of 1.78:1 to satisfy the overall grading requirement of coarse aggregate as per ASTM C33-92a21 and IS:383-1970.22 Land-quarried sand conforming to Zone II classification of British standards was used as fine aggregate. Tap water from the laboratory of a deep ground water source was used for the experiment. The composition of the tap water is given in Table 1. The w/c, cement content, fine aggregate content, and coarse aggregate content of the concrete mixtures are presented in Table 2. All the concrete mixtures have been designed for similar workability with a slump in the range of 1.2 to 2.4 in. (30 to 60 mm). Thus, water content in all the mixtures was kept constant to 353.99 lb/yd^sup 3^ (210 kg/m^sup 3^). To vary the strength, the w/c was varied by changing the cement content.

Preconditioning of steel bars

Steel bars were cut to the required length of 14.17 in. (360 mm). These bars were drilled and threaded at one end to be fitted with the coarse threaded stainless steel screws. Each bar was then wire brushed to remove any surface scale. These were then cleaned by soaking in analytical reagent grade hexane and allowed to air dry. One stainless steel screw was attached to each cleaned steel bar. Insulating tap was then applied on each end of the bar so that the central 9.84 in. (250 mm) portion of the steel bar remained bare. Neoprene tubes with internal diameters of 0.47 in. (12 mm) and 0.12 in. (3 mm) thickness was applied over the insulating tap for a length of 2.16 in. (55 mm) at each end of the steel bar. Epoxy was then filled in the protruded length of the neoprene tubes beyond the steel bar and also applied at the inner end of the neoprene tubes over the reinforcing bar, as shown in Fig. 2. This steel specimen preparation is similar to that specified in ASTM G109-99a.23

Preparation of slab specimens and cubes

The slab specimens prepared were 11.81 x 11.81 x 2.04 in. (300 x 300 x 52 mm), having a centrally embedded steel specimen mentioned in the previous section, and is shown in Fig. 3. The cover of the reinforcement bar was 0.79 in. (20 mm). With three types of steel, three cement types, three w/c, four admixed chloride contents, and three replicates, in total 324 slab specimens were prepared. The specimens were demolded after 24 hours of preparation and were moist-cured for 27 days in a curing tank. On completion of moist curing, the slab specimens were kept in ambient laboratory conditions (with temperatures ranging from 25 to 35 °C (77 to 95 °F) and relative humidity ranging from 60 to 80%). The linear polarization measurement with IR compensation using guard ring arrangement was conducted on slab specimens at the age of 60 days. The concrete cubes with a size of 5.90 x 5.90 x 5.90 in. (150 x 150 x 150 mm) were prepared and demolded after 24 hours. Then the cubes were moist-cured in a curing tank for 27 days after demolding. The 28-day cube compressive strength has been determined for all the concrete mixtures.

Electrochemical tests

The electrochemical test was performed on slab specimens using an instrument capable of performing several measurements, namely half-cell potential and corrosion rate with IR compensation. The details about the instrument are provided elsewhere.24 Linear polarization resistance test with IR compensation was conducted with guard ring arrangement on all 324 slab specimens of all the concrete mixtures at the age of 60 days. As concrete is a high resistive media, the IR drop in the cover concrete is significant and may vary from specimen to specimen. Therefore, the IR drop value of the cover concrete needs to be determined and compensated for determining the corrosion current density. The instrument automatically compensates for the aforementioned IR drop and also records the IR drop value.

For linear polarization resistance measurement, the working electrode, that is, the steel reinforcement in the slab specimen, was polarized to ±20 mV from the equilibrium potential at a scan rate of 0.1 mV/second. The linear polarization resistance (LPR) measurement with an IR compensation technique used with the corrosion instrument automatically calculates the equivalent IR value of the cover concrete and compensates while determining the corrosion current density. Before performing the test, the conducting sponge was wetted with soap solution and placed on the surface of the slab specimen to have proper electrical contact with the guard ring. Assembly of guard ring and other electrodes was then placed above the wetted sponge. The electrical connections were made to the steel reinforcement. The surface area of the steel reinforcement polarized was taken to be that lying under a circle intersecting the midpoint between the two sensor electrodes25 and only the top half surface area of the steel reinforcement was assumed to be polarized.26 For calculation of the corrosion current density I^sub corr^, the value of the Stern-Geary constant B was taken as 26 mV considering steel in active condition.10,27 The half-cell potential of the steel reinforcement was also measured separately with reference to a saturated calomel reference electrode (SCE).

Determination of free chloride, total chloride, and pH value of concrete

Concrete cubes were crushed in a compression testing machine at the age of 60 days. The crushed material was then passed through a square sieve with a mesh size of 150 microns and the collected powder was then stored in an air-tight condition. The powder sample was thus a mixture of cement hydrates, fine and coarse aggregates, and admixed chloride contents in their respective proportions.24 This powder was then used for chemical testing of the concrete. The free chloride and total chloride contents were determined by potentiometric titration using an automatic titrator equipped with a screen display that can display the chloride content automatically when appropriate inputs are fed.

For the determination of free chloride content, 3 g (0.1058 oz) of powdered concrete sample prepared as mentioned previously was transferred to a 100 mL (0.0264 gal.) beaker and 50 mL (0.0132 gal.) of distilled water was added. The sample was heated gently and thoroughly mixed by a stirrer. The solution was then cooled and filtered using filter paper and the free chloride content was then determined by titrating against 0.1 M AgNO3 solution in the automatic titrator. For the determination of total chloride, concentrated nitric acid (6N) was used in place of distilled water as the solvent and the same titration procedure was adopted.24 The determined free and total chloride ion concentrations were expressed as a percentage by mass of cement/concrete as appropriate. The pH value of the aqueous solution of the concrete powder sample was measured using a digital pH meter.

RESULTS

Compressive strength

The 28-day cube compressive strength of the concrete mixtures was determined by a compression testing machine and is given in Table 3. Each value shown in the table is the average value of the test result of three cubes. It was found that the concrete cubes made with PPC exhibited higher compressive strength than those made with OPC followed by PSC. The compressive strength ranged from 4570.03 to 6798.50 psi (31.52 to 46.89 MPa), 5316.72 to 7185.62 psi (36.67 to 49.56 MPa), and 4274.25 to 6293.94 psi (29.48 to 43.41 MPa), respectively, for the concrete mixtures made with OPC, PPC, and PSC. The 28-day cube compressive strength of the concrete mixtures was apparently independent of admixed chloride content for all types of cement and for all w/c.

Free chloride, total chloride, and pH value

The measured free chloride concentration, total chloride concentration, and pH of the concrete mixtures made with OPC, PPC, and PSC are presented in Table 4. In all the concrete mixtures, both free and total chloride concentrations increased with an increase in the admixed chloride content. It is observed that, at the highest level of admixed chloride, that is, 2.73%, the ratio of total chloride to free chloride is least indicating lower chloride binding. The chloride binding indicated by the ratio of total chloride to free chloride concentration for different concrete mixtures at similar admixed chloride levels is presented in Table 4. In all the three types of cement, the w/c does not significantly affect the chloride binding and the variation in chloride binding with the w/c is inconsistent. The PSC concrete exhibited the highest average chloride binding indicated by the average ratio of total chloride to free chloride concentration as 1.45 compared with 1.31 and 1.29 for OPC and PPC concrete mixtures, respectively. Plots of free chloride versus total chloride (percent by mass of concrete) for all three types of cement at each w/c is shown in Fig. 4, and from this figure it is observed that free chloride bears a linear relationship with total chloride. The relationships together with regression coefficient R are shown in this figure and have a form; FC = k^sub 1^*(TC) - k^sub 2^, where FC is the free chloride concentration and TC is the total chloride concentration. The values of k^sub 1^ and k^sub 2^ are two constants. The ratio k^sub 2^/k^sub 1^ represents a total chloride content that is equal to bound chloride at zero free chloride content. The value k^sub 2^/k^sub 1^ is 0.026, 0.022, and 0.018 for PSC, OPC, and PPC, respectively. The term k^sub 1^ represents the rate of increase of free chloride with total chloride and, as seen from Fig. 4, k^sub 1^ is least for PSC (0.7856) followed by PPC (0.8309) and OPC (0.856). From Table 4, it is observed that the pH values lie in the range of 12.35 to 12.72 for all the concrete mixtures and it is apparent that pH values do not vary with the cement type, w/c, and admixed chloride dosage. It is worth mentioning herein that the PPC had a 20% pozzolana content and may not be sufficient to consume all the calcium hydroxide (CH) released by the OPC hydrates in the blend, hence the pH value remained almost same as that of OPC. Similarly, the effective pozzolana present in PSC may not be sufficient to consume all the CH to cause any significant reduction in reserve alkalinity of concrete, hence leading to no pH alteration.

Half-cell potential

The half-cell potential values of three types of steel reinforcement embedded in the slab specimens made with three types of cement, three w/c, and varying dosages of sodium chloride have been measured with reference to the SCE. From the results, it is observed that the variation in half-cell potential values with w/c is inconsistent. The variation in half-cell potential with steel type, cement type, and chloride content, however, is systematic and relatively more significant. Hence, plots of absolute value of half-cell potential versus free chloride concentration (percent by mass of concrete) for different types of steel and cement are presented in Fig. 5 through 7. Each value shown in these figures is the average value of the results of three replicates. From Fig. 5 through 7, it is found that all the three types of steel (Tempcore TMT steel, Thermex TMT steel, and CTD steel) exhibited less negative potential in PSC than PPC and OPC. The difference in potential values of the steel reinforcement between PSC and PPC, and that between PSC and OPC, is significant and the difference being more pronounced as the free chloride content increases. The difference in potential values between PPC and OPC, however, is marginal as observed from the plots. The half-cell potential value of all three types of steel reinforcement in all the concrete mixtures decreased with an increase in free chloride concentration. Furthermore, the half-cell potential values generally bear a linear relationship with free chloride concentration irrespective of the cement type. The relationships are given in the plots together with respective R values. As per ASTM C876,28 potential values more negative than -350 mV (Cu/CuSO4)/-270 mV (SCE) correspond to a chance of corrosion to be more than 90%. Therefore, in Fig. 8, a horizontal line indicating -270 mV potential is drawn intersecting the line representing the relationship between the half-cell potential and free chloride concentration (percent by mass of concrete) for OPC at a w/c of 0.45 for all the three types of steel. The free chloride concentration values corresponding to the potential value of -270 mV (SCE) are obtained at the intersection points. By plotting similar plots for all the steels for different types of cement and w/c, the values of free chloride concentration (percent by mass of concrete) corresponding to -270 mV (SCE) are obtained and presented in Table 5. The standard deviations of chloride tolerance for three types of cement, three types of steel, and three w/c are also presented in Table 5.

IR drop and relative resistivity value

The instrument measures the IR drop value across the cover concrete within the confined area of the guard ring and reports the IR compensation value in terms of Ohm-cm^sup 2^ obtained by dividing the IR drop value with the corrosion current density. Because the cover depth for all specimens is the same and the same guard ring was used throughout, the area of the cross section of concrete through which the applied auxiliary current flows and the length of flow path are the same in all cases. Therefore, the IR compensation value in Ohm-cm^sup 2^ represents the relative resistivity value of the cover concrete. As all the specimens are subjected to the same exposure condition, all specimens have similar moisture conditions, and the same was confirmed through a moisture meter measurement. The IR compensation values of all the concrete mixtures have been determined at the age of 60 days. The IR compensation values for the three types of steel in the concrete mixtures mostly vary within ±5% of their mean. Thus, relative resistivity is independent of steel type as expected, being the property of the cover concrete. Therefore, the histogram plot of the mean of the relative resistivity values of the three types of steel at various admixed chloride contents for different types of cement and for different w/c is shown in Fig. 9. Each value shown in this figure is the average value of the three replicates, making an average of nine (three steels × thee replicates) values. From this plot, it is observed that the resistivity of the cover concrete in the slab specimens made with OPC is much lower than the those made with PPC and PSC at all w/c, that is, 0.45, 0.5, and 0.55. It is further observed from this figure that, for all three types of cement, there is no systematic variation in relative resistivity value with the admixed chloride content. It is also observed that the relative resistivity value decreased with an increase in the w/c irrespective of the cement type and admixed chloride content. This is expected as a higher w/c results in a higher capillary porosity, thereby a less dense structure. Therefore, from a resistivity point of view, the blended cements performed better than OPC.

Corrosion current density

From the determined corrosion current density values, it was observed that the variation in corrosion current density with a w/c is inconsistent. The variation in values of corrosion current density for all the concrete mixtures at all w/c is mostly less than 1 µA/cm^sup 2^ from the respective mean values. However, the variation in corrosion current density with chloride content, cement type, and steel type is systematic and relatively more significant. Therefore, the plots of the log of the determined corrosion current density in µA/cm^sup 2^ versus free chloride concentration (percent by mass of concrete) for Tempcore TMT steel, Thermex TMT steel, and CTD steel in three types of cement are shown in Fig. 10, 11, and 12, respectively, and those of PSC, OPC, and PPC for three types of steel are shown in Fig. 13, 14, and 15, respectively. Each value shown in these figures is the average value of three replicates. The relationship between the log of the corrosion current density (I^sub corr^) and free chloride concentration are also shown in these figures together with their respective R values. From the plots, it is observed that the corrosion current density increased with an increase in free chloride concentration in all the concrete mixtures. The highest value of corrosion current density occurred at the maximum admixed chloride content, that is, 2.73% by mass of cement in all the concrete mixtures. From Fig. 10 through 12, it is observed that the corrosion current density values are higher in almost all the concrete mixtures made with OPC than PSC and PPC for all the three types of steel. In addition, PPC concrete exhibited lower corrosion current density than PSC concrete for all the three types of steel as observed from these plots.

DISCUSSION

The test results presented in the previous sections reveal that the use of intergrinded blended cements, namely PPC and PSC, does not alter the pH value of the concrete significantly vis-à-vis OPC irrespective of w/c and admixed chloride content. Hence, the protection provided by concrete made with PSC and PPC against corrosion due to passivation of the reinforcing bar at a high pH is likely to be similar to that provided by OPC. Chloride binding was found to be more evident in PSC than OPC and PPC. Therefore, slag blended cement is likely to enhance the corrosion initiation period due to lack of available free chloride in the pore solution of the concrete compared with PPC and OPC. It is observed that in all the three types of cement used for the present work, there exists a linear relationship between free chloride and total chloride concentrations. Therefore, all the relationships that the free chloride bears with corrosion parameters can be applied to total chloride also.

From results of half-cell potential measurements presented in the previous section, it is observed that the blended cements exhibited less negative potential than OPC. The measured half-cell potential is the potential drop across the electrical double layer between the metal surface and the neighboring concrete solution phase. The less negative potential exhibited in blended cements may be because of passivation of the steel surface exposed to the concrete pore solution through the formation of a film of oxides and resulting in the reduction in concentration of iron [Fe2+] in the solution.29 Tempcore TMT steel showed less negative potential values than CTD steel followed by Thermex TMT steel in all types of cement, all w/c and at all levels of free chloride as evident from Fig. 5 to 7. From Table 5, it is observed that Tempcore TMT steel exhibited a higher tolerance limit for corrosion initiation than CTD steel followed by Thermex TMT steel in terms of free chloride concentration corresponding to -270 mV (SCE) in all types of cement and w/c. Therefore, Tempcore TMT steel is likely to have enhanced chloride tolerance. While considering the cement type, PSC exhibited higher values of tolerance limit for corrosion initiation than PPC and OPC in all w/c and for all types of steel. Mostly, PPC exhibited slightly higher values of chloride tolerance than OPC. Therefore, PSC is likely to enhance the chloride tolerance of the reinforcing bar; in other words, is likely to extend the time required for corrosion initiation. So, Tempcore TMT steel along with PSC performed better than other types of steel and cement in all w/c values with respect to increased time for corrosion initiation.

The resistivity values of PPC and PSC are three to six times higher compared with that of OPC. There is very little change in resistivity between PPC and PSC concretes, however, in all w/c values. The increase in resistivity value of the cover concrete with a decrease in w/c is found to be more in PPC and PSC concrete than that in OPC and this is due to the more dense structure provided by PPC and PSC concrete. Hence, the corrosion current density values are found to be more in OPC concrete than PPC and PSC concrete at all w/c, almost all levels of admixed chloride dosages. PPC exhibited lower corrosion current density values for all the three types of steel than both PSC and OPC and is likely to exhibit longer propagation period. While considering the steel type, Tempcore TMT exhibited lower corrosion current density values in all the concrete mixtures than both Thermex TMT steel and CTD steel as evident from Fig. 13 through 15 and is likely to show a longer propagation period.

CONCLUSIONS

Based on the study results, the following conclusions can be made:

1. From the results of the LPR test, relative resistivity, half-cell potential measurement, chloride content, and pH measurements carried out, it has been found that the blended cements, that is, PPC and PSC, performed better than OPC in the corrosion performance of steel reinforcement in concrete in the present investigation;

2. PSC is likely to improve the corrosion performance of reinforcing bar in concrete in the corrosion initiation period compared with PPC and OPC, whereas PPC is likely to improve the corrosion performance of reinforcing bar in the propagation period compared with PSC and OPC in chloride contaminated concrete;

3. The particular type of Tempcore TMT steel used showed higher chloride tolerance and lower corrosion current density values than Thermex TMT steel and CTD steel and hence is likely to exhibit longer initiation as well as propagation period in chloride contaminated concrete; and

4. The combination of PSC and Tempcore TMT steel is likely to perform best against chloride-induced corrosion in the initiation phase, whereas the combination of PPC and Tempcore TMT steel is likely to perform best during the propagation period.