Corrosion Performance of Steel Reinforcement in GFRP Strengthened Concrete Cylinders

The long-term durability of FRP composites is a crucial factor in their successful application as repair materials or as reinforcement for concrete. Extensive research has been carried out on using these FRP composites for repair and strengthening, however information about their performance in environments simulating hostile service conditions and their long-term durability are only beginning to be studied. Little is known about the long-term performance of FRP composites in corrosion prevention. In the present study, a detailed experimental study was carried out to investigate the corrosion performance of embedded steel reinforcement in Glass Fibre Reinforced Polymer (GFRP) wrapped cylindrical reinforced concrete specimens subjected to an impressed current and a high salinity solution. Test variables include types of resins, Configuration of fibre mat and number of wrap layers. Samples were evaluated for corrosion activity by monitoring impressed current flow levels, and by examining reinforcement bar mass loss and concrete chloride content among samples. Test results indicated that FRP wrapped specimens had prolonged test life, decreased reinforcement mass loss, and reduced concrete chloride content .The performance of wrapped specimens was superior to that of control samples. It was concluded that GFRP wraps were able to confine concrete, slowing deterioration from cracking and spalling and inhibiting the passage of salt water.

Dr. R.Kumutha, Dean & Head, Mr.K.Vijai, Associate Professor, Department of Civil Engineering, Sethu Institute of Technology, Pulloor

Introduction

Concrete structures in an aggressive environment, such as coastal areas, marine environments and regions where deicing salts are used, are specifically prone to premature deterioration. The ingress of chlorides present in seawater, salt spray, and deicing compounds into concrete promotes reinforcement corrosion and subsequent deterioration of the entire concrete member. As reinforcement corrosion intensifies, not only the expansive products of corrosion cause failures in the concrete surrounding the reinforcement frequently evidenced by cracking and spalling of the concrete, but may also lead to a loss in the structural integrity of the reinforcing steel. When bridges and structures are built in coastal areas, corrosion related problems are especially evident.

Overall, developing innovative ways to prevent corrosion from taking place and implementing long-term solutions to repair chloride contaminated concrete are necessary endeavors. A recent solution for repairing damages due to corrosion in reinforced concrete is to use fiber reinforced plastic (FRP) composite wrap. Several works have focused on the use of FRP composites for repairing and strengthening of structures, however, information about the corrosion performance of systems using these advanced materials is still lacking. Little is known about the long-term performance of FRP composites in corrosion prevention. This research is directed towards this endeavor.

Materials Used

Ordinary locally available Portland cement was used for the casting of the specimens. The fine aggregate (sand) used is clean dry river sand and hard granite broken stones were used as coarse aggregate. Aggregates passing through 12.5mm sieve and retaining on 4.75mm sieve were used. Three concrete cubes were cast as control samples and the average standard 28 days characteristic compressive strength of concrete cubes was found out to be 32.51 N/mm² with a mix ratio of cement: sand: gravel: water 1:1.3:3.29:0.47. Concrete cylinders were confined by wrapping them with Glass Fiber Reinforced Plastics (GFRP). Two types of GFRP sheets namely Chopped Strand Mat (CSM) having a density of 300g/m2 and Woven Roving Mat (WRM) having a density of 610g/m2 were used. Two types of resins namely General Purpose Polyester resin and epoxy resin systems were used.

Experimental Programme

Thirty-nine test specimens (lollipop samples) were cast, consisting of 51 mm diameter, 102 mm in height concrete cylinders, in which a single 12 mm diameter steel reinforcing bar protruding 19mm(0.75in) from the top centre of the specimen. A typical wrapped specimen, together with reinforcement is shown in Fig.1. Prior to casting, reinforcing bars were cleaned with a wire brush to remove all rust from the surface. As concrete was placed in the moulds, a bar was used to consolidate the concrete by rodding ten times. The reinforcement was secured such that it protruded from the top of the mold by 19 mm thus providing a uniform concrete cover at the sides and bottom. Test specimens were allowed to cure for 24 hours before the moulds were removed. Thirty- nine concrete cylinders were then moist cured in a saturated calcium hydroxide bath at a room temperature of 24ºC for 28 days.

Procedure to Bond FRP

Cylinders were confined by wrapping them with Glass Fiber Reinforced Plastics (GFRP) with a hand lay-up procedure. The surface was first cleaned to remove any dust particles. Then the surface was applied with the mixed solution of resin, accelerator and catalyst for General purpose polyester and resin and hardener for epoxy resins respectively in the proportions as suggested by the manufacturer. Then the properly cut fibre mat was placed over the surface and another coating of the mix was applied. The procedure was repeated for successive layers of FRP. Then the specimens were left to dry for 10 hours at room temperature. Approximately 25mm (1in) of overlap was maintained per layer to allow confinement to develop. Samples were grouped into 13 categories, or styles, each receiving a different type of surface treatment as shown in Table1.

Test Configuration

After test parameters were established, specimens were placed in a tank and immersed 89 mm in a 5% NaCl by weight solution, approximately double that of typical seawater, at a room temperature of approximately 24º C. In this experimental investigation, samples were connected to a single 12-volt DC power supply, which impressed a current such that the reinforcing bars are anodic. Fig.2 shows a schematic of test configuration. Samples were wired using parallel circuitry so that the gradual removal of samples would not influence the corrosion current of individual specimens. Electrodes, serving as cathodes, consisted of 25 mm wide, 3 mm thick steel bars, distributed throughout the tank bottom in such a way to ensure a consistent environment for all samples. The high salinity and the impressed current were both used to create an especially aggressive environment by providing an abundance of chloride ions and by stimulating an increased flow of electrons, respectively.

Current Monitoring

In an attempt to characterize the impressed current, measurements were taken with an ammeter in each of the parallel electrical circuits twice a day. A current spike or rapid increase in current flow, indicating the short circuit, occurred after cracks had formed in a concrete that failure was imminent. Due to space limitations, current measurements are not presented in this paper, however the data is consistent with other results. For most of the sample styles, type of failure mechanism commonly encountered was excessive current with concrete failure which is a type of failure whereby the concrete cracked and a current spike was observed. Failure modes of some of the sample styles are shown in Fig. 3. Days to failure for all the sample styles are summarized in the Table 2.

Physical Techniques

Samples were visually inspected for cracks daily and removed from the tank when the concrete cracked, the wrap failed, and/or the current flow in the reinforcement spiked. During sample testing, it was noticed that, corrosion products in the form of dark green to black paste, leached out of the concrete from around the steel reinforcement. When the buildup of corrosion products at this interface zone became excessive, it was removed. After each sample was removed from the tank, the reinforcing bar was extracted from the concrete and placed in a 10% solution of Hydrochloric Acid (HCl) for a week to remove all corrosion products and remaining concrete. The acid etched all contaminants away, leaving only the steel bars. The bars, or pieces of bars, were then weighed. Table 3 presents the mass loss data for all samples that were tested until failure.

Chloride Ingress Measurements

Chloride penetration into the concrete was established by the chloride ion concentration as obtained from a chemical analysis of concrete samples. After failed specimens had been removed from the tank, powder samples for chloride analysis were obtained by drilling a 19 mm deep hole in a perpendicular direction from the bottom end surface of each concrete specimens directly below the embedded reinforcing bar. Powder obtained from the first 6.4mm was discarded and a powder sample of at least 5 g was collected for analysis. Because the bottoms of the samples were in continual contact with the bottom of the tank, this location was deemed to be suitable for chloride testing. The ingress of chlorides ions into the concrete surrounding reinforced steel was used as a relative measure to compare and evaluate the effectiveness of various surface treatments. Powder samples were analyzed for their soluble chloride content using Nordtest method5 by dissolving the concrete dust in a solution of nitric acid followed by Volhard titration. Table 4 presents the chloride content for all samples.

Experimental Results and Discussion

It is presumed that the electrical current applied to the reinforcement attracted negatively charged chloride ions from the solution into the concrete towards the positively charged steel bars. As the chloride ion reached the steel-concrete interface, the steel surface began to corrode. The expansive products of corrosion imposed tensile stresses on the concrete cover, and resulted in cracking when the tensile stresses became too high. Cracking especially large cracks, would allow the conductive chloride solution to come in immediate contact with the steel surface thus providing a direct current path between the reinforcement and the electrodes in solution. Therefore a current spike, or a dramatic increase in current flow, followed cracking in the concrete.

From the data and from observations made during testing, it was found that in general, the chloride ions migrated into the concrete quickly in control samples, as evidenced by the rapid onset of cracks or concrete failure and the subsequent current spikes. It was also found the samples treated with polyester generally had longer test lives and had fewer current spikes than treated with epoxy. Fig.4 shows the number of days to failure for the lollypop specimens. Increasing number of wraps from one to two proved to be effective, likely because of increased confinement; however, three wrap layers was not clearly shown to be more effective than two layers. It implies that due to their sequencing or the chronology of spiking, both polyester and epoxy type and number of wraps, affected performance.

Reinforcing bar mass loss

In this study, the reinforcing bar mass loss ratio (percentage per day) was calculated, as it would yield a relative figure that could be used to compare the general performance of samples and treatment options. It was found that, in control samples, chloride ions are generally migrated into the concrete very quickly than the wrapped samples. The unconfined samples and wrapped specimens coated with epoxy experienced the highest mass loss ratios indicating the most severe levels of corrosion. The results from the study of the ratio of the percent mass loss per day shown in Fig.5 reveals that sample CSM-P3 and WRM-P3 showed 3.664% and 3.517% less mass loss per day respectively than control samples. In all the samples epoxy appears to allow free migration of chlorides into the concrete while polyester offers much more impervious protection.

Using Chopped strand mat with polyester resin, wrapped samples lost 0.38%, 0.46% and 0.659% less mass loss per day, for one, two, three layers respectively as compared with those of epoxy resin. This reduction in the rate of mass loss was about 40%, 55%, and 64%. The maximum reduction in the rate of mass loss was observed in CSM-P3 specimens, which was about 91% as compared with that of unconfined specimens. Using Woven roving mat with polyester resin, wrapped samples lost 0.23%, 0.26% and 0.52% less mass loss per day, for one, two, three layers respectively as compared with those of epoxy resin. This reduction in the rate of mass loss was about 24%, 32% and 50%.The maximum reduction in the rate of mass loss was observed in WRM-P3 specimens, which was about 87% as compared with that of unconfined specimens. While comparing the results of ratio of mass loss per day between both types of resins, the wrapped samples coated with general purpose polyester proves to be more effective than the wrap- ped samples coated with epoxy.

Chloride content

Results from chloride ingress measurements are shown in Fig.6. It was found that the unwrapped samples had a higher chloride concentration than the wrapped samples. It was also noticed that, samples treated with polyester performed better than those treated with epoxy. The content of chloride in unconfined sample was found to be 0.141% whereas the sample CSM-P3 performed better with a chloride content of 0.105%. Maximum reduction of chloride content was observed in CSM-P3 and it was about 25% as compared to unconfined samples. The effect of an additional wrap beyond the second layer did not prove to be any more effective in reducing the chloride content or rate of migration of chlorides into the concrete. As evident in the figure, the results of the concrete chloride contents provide strong evidence that the number of wraps and type of resins influence the ingress of chlorides.

Conclusion

Based on the results of this experimental investigation, the following conclusions are drawn:

• Glass fiber-reinforced polymer wraps increase a reinforced concrete sample’s resistance to accelerated corrosion in a submerged environment as evidenced by prolonged life, decreased overall rate of reinforcement mass loss, and the reduced concrete chloride content.
• It was also found that the samples treated with polyester generally had longer test lives and had fewer current spikes than treated with epoxy. FRP wrapping of samples increased the life span by about six fold compared with control samples.
• Examination of the experimental data reveals that the type of resins used as a surface treatment has a significant effect on the corrosion resistance.
• Increasing number of the wraps from one to two proves to be effective, likely because of increased confinement; however, three wrap layers is not clearly shown to be more effective than two wrap layers. This could be because the confining strength of two wraps was sufficient to restrict the expansion of the corrosion residual.
• FRP wrapping is potentially effective in reducing corrosion in reinforced concrete structures in marine environments and this improved performance is likely due to the establishment of confining stresses in the concrete and the added resistance to the permeation of moisture and chlorides, both provided by the composite wraps.

References

• Houssam A.Toutanji., "Durability characteristics of concrete columns confined with advanced composite materials", Composite Structures, 44, 1999, pp. 155-161.
• Houssam Toutanji., and Yong Teng "Strength and Durability performance of concrte axially loaded members confined with AFRP composite sheets", Composites: Part B, 33, 2002, pp. 255-261.
• Saadatmanesh, H., Eshani, MR., and Li, MW., "Strength and Ductility of concrete columns externally reinforced with fibre composite straps", ACI Structural Journal, V. 91, No. 4, 1994, pp. 434-447.
• Isaac A Wootton., Lisa K. Spainhour., and Nur Yazdani., "Corrosion of steel reinforcement in carbon fibre-reinforced polymer wrapped concrete cylinders," Journal of Composites for Construction, November 2003, pp. 339-346.
• NT BUILD 208: Concrete, Hardened: sampling and treatment of cores for strength tests- Chloride content by volhard titration, edition-3.