Paper No. CORROSION 95 2 8. 4. ‘The NACE International Annual Conference and Corrosion Show ELECTROCHEMICAL AND MICROSTRUCTURAL PERFORMANCE OF STEEL REINFORCED CARBONATED AND NON-CARBONATED MORTARS IN A SALINE ENVIRONMENT A.G. Constantinou, M.A. Sanjuan and K.L. Serivener Department of Materials Imperial College of Science, Technology and Medicine London SW7 2BP United Kingdom ABSTRACT This paper describes work in progress to investigate the combined effect of chloride ions and carbonation in the corrosion of steel in mortar specimens. The corrosion behaviour was investigated by electrochemical techniques, namely linear polarisation resistance. Scanning electron microscopy (SEM) and energy dispersive x-ray microanalysis were used to study the microstructure of the steel/paste interface and the chemistry and distribution of the corrosion products. Eight mortar specimens were cast with water/cement and cement/sand ratios of 0.5 and 1/3, respectively. After 7 days of curing in plastic bags two samples were exposed to a saline environment (immersion in a 0.5M NaCI solution), another five were carbonated (in 100% CO and 65% RH), and the last one was kept immersed in water as reference. After S months, some of the specimens were switched round (ie. the carbonated ones were immersed in 0.5M NaCl), while one specimen from each environment remained in the initial environment. For the specimen which was placed in the CO environment after immersion in NaCl, the exposure time was not long enough to fully carbonate the mortar after the formation of observable amounts of corrosion products, although corrosion can be detected electrochemically. In the case of the specimens in a CO) atmosphere, the corrosion rate was very low until full carbonation. When a specimen was then put in an NaCl solution, the corrosion rate increased almost immediately. The exposure time of the specimens was not long enough to produce any conclusive results, but the samples in a NaC! solution began to corrode soon after exposure. Keywords: Steel corrosion, microstructure, chloride diffusion, marine environment, carbonation, sony nace Publication Right opin Intemational. NACE international hasbeen given st gts of pubcation ofthis manuscript. Requests for permission padi ranaerpin any rm.n arto nwo matte mace wig RACE nretoral Picts Oven PO. S860 ouston, Texas rater rocaniod ana the vows expressed in fe paper are solely tose othe author) and ee ne necessarily endorsed by the Association. Printed in the U.S.A. “ " “we *INTRODUCTION Concrete is the most widely used construction material for sea structures due to its low cost, availability, formability and durability. However, concrete exposed to a marine environment is in the presence of aggressive agents and requires particular care in design and placement in order to achieve the required service life. Despite considerable progress in the understanding of the durability of concrete in marine environments 1-3, concrete degradation remains the foremost problem facing the use of structural reinforced concrete today. For example, tunnels in Hong Kong were designed for a 120 years service life but after less than 10 years they had suffered significant damage and are currently undergoing continuous repairs. The pore solution of the concrete normally has a pH of about 13. In this alkaline environment a thin passivating oxide film of 10-3 - 10! im is formed which protects the embedded steel +. The alkalinity may be reduced either by leaching of the alkaline substances contained in the pore solution of the concrete, or by reaction with CO present in air. The passive oxide film will break down when the pH of the pore solution drops below 9. In addition, chloride ion ingress may also destroy the passive film by pitting attack 5, The service life of marine concrete structures, with regards to corrosion of reinforcement, depends on the time that chloride ions take to reach the steel (initiation period); and the time for corrosion to propagate to an unsafe level of damage (propagation period) 6 Depending on the degree of carbonation and chloride level, any of the electrolytic conditions shown in Figure 1 can occur. Carbonation will proceed on exposed surfaces but below the water line the moisture level is so high that carbonation will slow down and it will not be a significant problem. The rate of chloride penetration is high in a non-saturated concrete because it is promoted by capillary absorption of salty solutions, while in a water-saturated concrete a slower diffusional process takes place When chloride ions reach the reinforcement, localised corrosion is initiated. ‘The chloride penetration in the concrete is affected by the ease with which the chlorides are transported through the pore structure and also by their combination with cement compounds such as the formation of Friedel's salt (3.Ca0.Al;03.CaCly.10H;0). Carbonation of the concrete will affect its pore structure due to the precipitation of carbonates in the pores and will also react with the calcium aluminates. If 2 concrete already contains low levels of chloride ions, subsequent carbonation may release some of the bound chlorides which may in tum lead to corrosion. On the other hand, for a carbonated concrete, the rate at which chloride ions penetrate will be affected, due to changes in permeability and the binding capacity of the concrete paste. The relationship between corrosion rate, chloride content and alkelinity is not completely understood. According to Hausmann, the ratio of chloride to hydroxyl ion concentration should not exceed a threshold value of 0.6 7. Carbonation will affect the chloride to hydroxyl ion ratio but may also affect the threshold value of this ratio. It would appear likely that in carbonated concrete the activation level of chloride ions is reduced significantly. In this work, the interaction of chloride ions and carbonation in the concrete microstructure and on the corrosion of the steel has been investigated electrochemically and microscopically. 24/2EXPERIMENTAL The composition of the ordinary Portland cement used is given in Table 1, The sand grading ranged between 0 and 6 mm, Eight mortar specimens of 20 x 55 x 80 mm’, with a cement/sand ratio of 1/3 and a water/cement ratio of 0.5 were prepared 8, Distilled water was used and the specimens were cast under laboratory conditions. They were removed from the moulds after 24 hours and then cured at 100% RH and 20 - 25°C for 7 days. Two mortar specimens contained no steel bars and were used as control samples to check the progress of carbonation (G and H). Two steel bars were embedded in each of the remaining six specimens, to serve as twin working electrodes while a graphite rod was embedded mid-way between the two steel bars as an auxiliary electrode. The mortar cover of these bars was 7mm. The configuration of the specimens is given in Figure 2. The steel bars were of 6 mm diameter and 80mm length, with a mean chemical composition shown in Table 2. The steel bars were cleaned in a 1:1 water:HCI solution containing 3 grams per litre of hexametilentetramine"’ as corrosion inhibitor, rinsed in acetone, dried and weighed. Their ends were coated with a plastic insulating tape leaving an exposed area of 575 mm? in the middle. The specimens were exposed in different environments water throughout the experiment. in 0,5 M NaCl throughout the experiment. in 0,5 M NaCl for 154 days and then in 100% COp for 40 days In 100% CO> for 154 days and then in water for 40 days. E=In 100% CO» for 154 days and then in 0.5 M NaCl for 40 days. F = In 100% COy throughout the experiment G = In 100% CO, throughout the experiment. H= In 100% CO) throughout the experiment Electrochemical Tests The samples were monitored electrochemically using the linear polarisation resistance technique®!0. The steel reinforcement bars are used as working electrodes, and an external counter electrode as a reference electrode. The corrosion potential (Eso) is first measured with respect to a stable reference electrode (calomel). Four small potential steps, AE, of -5, -10, -15 and -20 mV, are then applied about the Exar value via a counter electrode (graphite rod) and the resulting current AI, is measured. The polarisation resistance, R,, is given by: Q) The overall corrosion current, Ion, flowing between anodic and cathodic regions of the reinforcing steel is given by: © (CHIN 2684/4=2 cor = Be where B is the Stern-Geary constant, for active corrosion = 26 mV. The corrosion rate is therefore: Ff Q) jog = @) where A = the corroding area, 575 mm’. The potentiostat used has a positive feedback in order to eliminate the ohmic drop between the reference and the working electrode, This ohmic drop is due to the resistance of the mortar cover and also the resistance of the passive oxide film surrounding the steel reinforcement. If this electronic resistance compensation is not used, the Rp value will be overestimated and thus Tear will be smaller than the actual value (Equation [2]) sans “Brine * Rg The electrochemical results prior to 200 days are the average of the two steel bars contained in each specimen. After 200 days one of the steel bars plus surrounding mortar was cut off for examination in the SEM in the back-scattering mode, The other steel bar remained under the various conditions and was monitored electrochemically. Hence, measurements at later ages relate to one bar only. Sample Preparation and Electron Microscopy Polished sections of the mortar samples were prepared for SEM examination. After 200 days of testing, one half of each specimen was cut off longitudinally to the rebar for microstructural examination. Sections were then cut transversally to the steel, resin impregnated, lapped with 9m alumina powder and polished from 31m to 1/4um to obtain a flat surface, The samples were then coated with carbon (Figure 2) ‘A JEOL JMS35-CF scanning electron microscope (SEM) (accelerating voltage of 15keV) was used equipped with secondary and backscattered electron detectors, fitted with a Link System AN10000 energy dispersive X-ray spectrometer (EDS). Point analyses were made in the region of the steel/mortar interface, particularly to investigate the presence of chloride ions either as Friedel's salt or unbound. The analyses were corrected by the ZAF procedure. In addition, an X-ray dot map was made of a section of the interfacial area to indicate the relative concentrations of Ca, Al, Fe and Cl. RESULTS AND DISCUSSIO. Electrochemical Results The initial Essa values of specimens A, B and C were very similar after 7 days of curing (Figure 3). At later ages there is a slight difference in the E... measurements between the specimen in water (A) and those in sodium chloride (B and C). However, the ir values indicate that all the specimens are corroding at virtually the same rate. In the case of the sample in water, the rate of corrosion was higher 28413In order to investigate the presence of Friede!’s salt in the specimens in NaCl solution, energy dispersive spectroscopy (EDS) x-ray microanalysis was conducted on the same specimen C in the form of mapping for aluminium and chloride. Figure 10 shows the EDS mapping and the BSE image of the mapped area for sample C. Most of the areas which appear to have a high concentration of Cl (Figure 11) coincide with areas of high Al concentration (Figure 12), suggesting the presence of Friedel’s salt. Point analyses were made at the interfacial area in order to verify this observation, The microanalyses results indicated that virtually all the chloride was probably combined as Friedel’s salt in this specimen, The specimens are still being electrochemically monitored and sections will be removed for examination in the future. The interaction of carbonation and chlorides will be further investigated since the specimens will have more time in the new regimes. Any changes in the electrochemical results will be recorded and they will be related to the microstructure of the mortar when examined in the SEM. It is expected that it will be possible to detect corrosion products in the microstructure and any changes in the distribution of chlorides and the interaction with carbonation CONCLUDING REMARKS The ior and Ev plots for specimens in NaCl solution and CO, atmosphere showed the expected range of values. The specimens in the NaCl environment show a slowly increasing rate of corrosion soon after exposure, The specimens initially in a CO, atmosphere, start to corrode after carbonation irrespective of the environment they are subsequent placed in, However, it was still too early to detect any real changes in the corrosion behaviour of the specimens due to differences in environment. The ohmic drop testing in reinforced mortars and concretes seems to be a good indicator the start of corrosion of the embedded bars in cases of carbonation. EDS mapping was presented as a quick qualitative technique to assess the possible chloroaluminates formation around the interface mortar/steel. It is expected to be able to observe some differences in the formation and distribution of chloroaluminates at longer exposure times. ACKNOWLEDGEMENTS MAS wants to thank to the Directorate-General for Science, Research and Development of the EC for the receipt of a Fellowship under the Human Capital and Mobility program. AGC wants to thank NACE UK, the Institute of Materials and the Royal Academy of Engineering for financing the attendance to the Conference. REFERENCES 1. OE, Gjorv, "Control of steel corrosion in concrete sea structures", Corrosion of Metals in Concrete, ACI SP-49, (Detroit, MI: American Concrete Institute, 1975): p 1than expected, this may have due to a small part of the specimen being above the level of the water. ‘This observation demonstrates the importance of corrosion rate measurements rather than potential mapping to assess the corrosion state of reinforced concrete structures !!. In all the specimens, the values of ohmic drop are comparatively low due to the saturation of the specimens. Figures 6, 7 and 8 show the Eur, is and Ras Values respectively, corresponding to the bars embedded in the specimens kept in the 100% CO chamber until 161 days of age (D, E and F). In Figure 6, the initial values of Ex. were similar to those of the saturated specimens A, B and C. As usually observed, the values became positive during the carbonation at 65% RH, due to the partial drying of the specimens especially at the surface, At 161 days of age, the Ex, measurements drop suddenly to negative values (Esa ~~450 mV) indicating the onset of steel corrosion. At this stage one of the specimens was submerged in water and one in 0.5MNaCI solution. However, the Eom measurements remained at similar negative values regardless of the environment. These values are also similar to those measured for specimens A, B and C (Figure 3) In the case of the isn measurements of the specimens in the CO, atmosphere (Figure 7), the values are generally lower than the specimens in water or NaCI solution. There is some variation in the ‘measurements during the period of carbonation, In order to take the measurements, electrical contact with the specimen is made by placing a damp sponge between the sample and the reference electrode. This process may alter the humidity profile in the specimen and so affect the ioq readings. When the samples are fully carbonated (at 161 days of age) the corrosion rate increases. ‘The samples subsequently exposed in water or NaCI solution (D and E) have a relatively high corrosion rate relative to the uncarbonated specimens A, B and C. The rate of corrosion for the specimen maintained in the CO2 environment is somewhat lower. The values of the ohmic drop increase throughout of the period of carbonation, probably due to partial dehydration of the mortar and decreasing porosity. At the end of the carbonation period there is a drop in the values which corresponds to the breakdown of the passive film and onset of the active corrosion. The monitoring of ohmic drop values appears to be a good method for detecting the onset of corrosion in this case 12 Microscopical Results A section from specimen C, immersed in NaCl solution up to 161 of age, then placed in CO7 atmosphere, was examined in the SEM. Electrochemical measurements that indicate that this was corroding from fairly early times. However, when this specimen was examined in the SEM, no significant amount of corrosion product was observed in the back-scattered images (Figure 9). The sample shows the expected features of an 200 day old mortar. No signs of carbonation were detected in the microstructure due to the short exposure time of the specimen to the CO, atmosphere. This is in agreement with the observations that the electrochemical readings (Figures 3-5) had not been affected by this change of environment at this time. Large anhydrous cement grains can be recognised by their white appearance with grey hydration rims around them (inner product), Ca(OH), crystals (white) which precipitate in the paste as well as around the sand particles 13 (dark grey) and C-S-H gel which has a light grey appearance, The black space is the porosity which is present in the paste and is more prominent around the steel reinforcement and the sand particles. 28415RD. Browne, "Mechanisms of corrosion of steel in concrete in relation to design, inspection and repair of offshore and coastal structures", Performance of Concrete in Marine Environments, ACI SP-65, (Detroit, MI: American Concrete Institute, 1980): p 169 PK. Mehta, "Durability of Concrete in Marine Environments" Performance of Concrete in Marine Environments, ACI SP-65, (Detroit, MI: American Concrete Institute, 1980): p 1 KK. Sagoe-Crentsil, F.P. Glasser, "Analysis of the steel:concrete interface", Corrosion of Reinforcement, Society of Chemical Industry, London (Oxford, U.K.: Elsevier Applied Science, 1990): p74 5. CL. Page, K.W.J. Treadaway, Nature, 297 (1982): p 109 10, ul 12. 1B K. Tuutti, Corrosion of Steel in Concrete. CBI Report fo 4-82, Cement and Concrete Institute, Stockholm, 1982. p 468. D.A. Hausmann, Materials Protection, 6,11 (1967): p 19. C. Andrade, Ph.D. Thesis, Complutense University of Madrid. Faculty of Chemistry, Madrid, July, 1973. M. Stern, ALL. Geary, J. Elect. Soc., 104, 1, (1957): p 56. C. Andrade, J.A. Gonzalez, Wekstofife und Korrosion, 29 (1978): p 515. E. Escalante, "Effectiveness of potential measurements for estimating corrosion of steel in concrete" Corrosion of Reinforcement, Society of Chemical Industry, London (Oxford, U.K.: Elsevier Applied Science, 1990) JP. Broomfield, , J. Rodriguez, , L.M. Ortega, AM. Garcia, "Corrosion rate measurements in reinforced concrete structures by a linear polarization device". P.D. Cady Int. Symp. on the Condition Assessment, Protection, Repair and Rehabilitation of Concrete Bridges Exposed to Aggressive environments. ACI Fall Convention, (Detroit, MI: American Concrete Institute, 1993) KL. Scrivener, E. Gartner, “Microstructural Gradients in cement Paste around Ageregate Particles” in Bonding in Cementitious Composites, MRS Symp. Proc. Vol.114, (Pittsburgh, PA: Materials Research Society, 1988): p 77 2847TABLE | CHEMICAL COMPOSITION OF THE ORDINARY PORTLAND CEMENT (% BY WEIGHT * ) SiO2 20.0 AbO3z 5.6 Fe703 2 CaO 647 MgO. 0.9 MnO} 0.06 P205 0.08 TiO 0.23 ‘S03 31 K,0 0.58 NaO 019 F 0.18 Insoluble Residue 0.40 ‘Loss on Ignition 18 Free Lime 17, Surface Area TABLE 2 MEAN CHEMICAL ANALYSES OF THE STEEL REBARS BY ABSORPTION SPECTROSCOPY Element. Percent Cc 0.19 ‘Mn 0.64 Si 0.28 P. 0.021 Ss 0,052 Fe To Balance.SPRAY ZONE TIDAL ZONE SPLASH ZONE 234 FIGURE 1- Possible conditions of concrete in a marine structures: 1) uncarbonated concrete without chlorides; 2) carbonated conerete with chlorides; 3) uncarbonated concretewith chlorides, 4) carbonated concrete with chlorides STEEL RENFORCVENT BARS, scrapure cuecrrooe FoR CORROSION eASUneh NTS FeHSONS Wr Puastc TAFE 10 ist 1 conaooMG ry FIGURE 2- Layout of the mortar specimen and the section cut off for microscopical examination 28419Ecorr (mV) ‘corr (uAlem2) Rohm (Ohms) 7 200 ~ 100 CHANGE OF ENVIRONMENT 100 250 -200 “ -300 -400 -500 -600 TIME (DAYS) FIGURE 3 - Corrosion potential of the steels embedded in the mortars initially immersed in water or a NaCI solution [—e=a) O08 -a-c 10 CHANGE OF ENVIRONMENT ° 50 100 150 200 250 TIME (DAYS) FIGURE 4 - Corrosion rate of the steels embedded in the mortars initially immersed in water or a NaCI solution 350 300 250 200 150 100 CHANGE OF ENVIRONMENT 50 ° 50 100 150 200 250 TIME (DAYS) FIGURE 5 - Ohmic drop of the steels embedded in the mortars initially immersed in water or a NaCl solution 284/10Rohm (Ohms) : 250 -200 E corr (mV) ! CHANGE OF ENVIRONMENT TIME (DAYS) FIGURE 6- Corrosion potential of the steels embedded in the mortars initially placed in 100% CO2 i corr (wAlem2) 0 50 100 150 200 250 TIME (DAYS) FIGURE 7- Corrosion rate of the steels embedded in the mortars initially placed in 100% CO2 70000 60000 50000 40000 30000 20000 10000 ° 50 100 150 200 250 TIME (DAYS) FIGURE 8- Ohmic drop of the steels embedded in the mortars initially placed in 100% CO2 eatFIGURE 9 -Back-scattered electron image in the FIGURE 10 - EDS mapping in the in the interface interface of specimen C: A) Steel; of specimen C B) Aggregate; C) Ca(OH), FIGURE 11 - Chloride map FIGURE 12 - Aluminium map 284/12