74 M.M. Page et al. / Construction and Building Materials 16 (2002) 7 81 Table 1 Chemical analsis of ordinar portland cement (OPC) Oxide: CaO SiO 2 Al

Transcription

1 Construction and Building Materials 16 (2002) 7 81 Ion chromatographic analsis of corrosion inhibitors in concrete a, a b c M.M. Page *, C.L. Page, V.T. Ngala, D.J. Anstice a School of Civil Engineering, Universit of Leeds, Leeds LS2 9JT, UK b W.S. Atkins Consultants Ltd., Auchinleck House, Fivewas, Birmingham B15 1DJ, UK c Maunsell Ltd., Imperial House, 1 Temple Street, Birmingham B2 5DB, UK Received 2 April 2001; received in revised form 27 November 2001; accepted 9 Januar 2002 Abstract A method emploing ion chromatograph for identifing and quantifing ions present in the aqueous phase of concrete after surface treatment with corrosion inhibitors is reported in this paper. With this technique, a broad range of ions including ethanolamine, nitrite and monofluorophosphate, present in the inhibitors tested, could be identified in solution. When ethanolamine and nitrite were applied to the surface of concrete, the were readil detected in samples into which the had penetrated. In the case of monofluorophosphate-treated concrete, onl the hdrolsis products, fluoride and phosphate were detected and not the monofluorophosphate ion itself. Furthermore, concentrations of other important ions, such as chloride, were also quantifiable b this technique. One of the main advantages of ion chromatograph for this tpe of application is its abilit to analse a wide range of ions in a given sample Elsevier Science Ltd. All rights reserved. Kewords: Ion chromatograph; Corrosion inhibitors; Surface treatment; Reinforced concrete 1. Introduction In recent ears, various corrosion inhibitors have been applied to the surface of reinforced concrete structures with a view to reducing the rate of corrosion of embedded steel to acceptable levels, in cases where carbonation or chloride contamination of the concrete has occurred w1x. In order to determine whether such substances actuall reach the reinforcing bars and, if so, in what concentration, a reliable method is needed for analsing them. In this paper, a simple technique is described for detecting and determining concentrations of a range of ions present in commerciall available inhibitors, their hdrolsis products and also other ions, which might affect their action. 2. Methods Three different substances used in solution as active components of corrosion inhibitors, viz. ethanolamine, nitrite and monofluorophosphate, were applied to laborator-prepared concrete specimens, both carbonated and *Corresponding author. Fax: q address: m.m.page@leeds.ac.uk (M.M. Page). non-carbonated, and to a naturall carbonated reinforced concrete slab obtained from a 60-ear-old power station in mid-wales. In addition, solutions of inhibitors based on ethanolamine and monofluorophosphate were applied on site, b the suppliers, to the surface of a 25-ear-old concrete structure (carbonated to a depth of between 10 and 15 mm), which was located at a power station in northeast England w2x Preparation of concrete specimens in the laborator The laborator-prepared concrete specimens were made using ordinar portland cement (OPC) of high watercement ratios (0.65 and 0.8) to simulate poor qualit concrete. The composition of the OPC is given in Table 1. The specimens were all 100=100=90 mm, and contained three mild steel bars of uniform diameter of 6.5 mm, at cover depths of 5, 12 and 20 mm from one face. Some were artificiall carbonated b exposing them for several months to an atmosphere of 65% RH in a sealed tank, through which pure CO2 gas was passed for approximatel 0 min twice a da, in order to stimulate carbonation-induced corrosion of the reinforcement. Others were contaminated b the addition of /02/$ - see front matter 2002 Elsevier Science Ltd. All rights reserved. PII: S Ž X

2 74 M.M. Page et al. / Construction and Building Materials 16 (2002) 7 81 Table 1 Chemical analsis of ordinar portland cement (OPC) Oxide: CaO SiO 2 Al O 2 Fe O 2 SO MgO Na O 2 KO 2 LOI % of OPC b mass: various quantities of analtical grade NaCl (0. 2.4% chloride b weight of cement) to the concrete mix before casting in order to stimulate chloride-induced corrosion. In the case of the naturall carbonated reinforced concrete slab, core specimens, 75 mm in diameter with a depth of 110 mm, were obtained. One face of each core was cut such that the depth of cover to the reinforcing bars was approximatel 12 mm. The sets of specimens were exposed to a wetting and dring regime which consisted of a period of immersion in deionized water for 48 h, followed b storage in air over saturated magnesium chloride solution (approx. % RH) for at least 28 das. The corrosion activit of the embedded steel bars was monitored as described elsewhere wx. After 6 months, when the corrosion measurements were reproducible, inhibitor treatment was carried out Application of inhibitors Inhibitors were applied to the clean, dr concrete surface as recommended b the manufacturers. The following proprietar inhibitors were investigated: a an ethanolamine-based solution, found b means of ion chromatograph to contain 2. mol dm of ethanolamine and also 1.2 mol dm of phosphate. This was applied five times to the concrete specimens, ensuring that the surfaces were dr before each application. In the case of the 25-ear-old concrete structure in northeast England, the surface was cleaned thoroughl b scrubbing with a wire brush before application of the inhibitor b the supplier on site w2x; b a calcium nitrite-based inhibitor, found b ion chromatograph to contain 2.7 mol dm nitrite. This was applied three times with intervening periods of dring of approximatel 4 h and then overlaid with 25 mm of a cement-based mortar (watercement ratio of 0.4), containing a recommended dosage of 20 dm of approximatel 6 M calcium nitrite solution m of concrete w,4x; and c. sodium monofluorophosphate in aqueous solution applied 10 times with intervening periods of dring, followed b 10 applications of deionized water again with intervening periods of dring. In the case of laborator-treated specimens, a 15% b weight solution (approx. 1 mol dm ) was applied. For the site treated concrete, four applications of a 5% solution (approx. 0.5 mol dm ) followed b six of a 10% solution (approx. 0.7 mol dm the supplier w2x. 2.. Exposure conditions ) were carried out b All specimens treated with inhibitors in the laborator were subsequentl stored over water in a tank for 14 das before being exposed for a further 18 months to the wetting and dring regime described above. The site-treated concrete structure was exposed to natural weather conditions in an unsheltered location on the coast in northeast England for 12 months (October 1998 October 1999), in the case of the ethanolaminebased inhibitor, and 15 months (Jul 1998 October 1999), in the case of sodium monofluorophosphate Extraction of inhibitors After exposure of the specimens, clindrical cores, 75 mm in diameter, were taken so that one end of the clinder was derived from the inhibitor-treated face. A small amount of water, designed to have minimal effect on sample composition, was added as a lubricant during coring. The cores were then ground so that the constituents of the concrete, in 2-mm-thick laers parallel to the treated surface, were collected and stored in sealed polthene bags. In the earl stages of the research, the powdered laers were then dried at 105 8C for 24 h to remove evaporable water. In later work, however, the dring stage was omitted. This was considered preferable as it avoided possible loss of volatile inhibitors, such as ethanolamine, and, although it meant that evaporable water could not be eliminated, the latter generall did not constitute more than approximatel 5 10% of the specimen weight. In practice, it was found that the results obtained for the ethanolamine-treated samples did not var substantiall irrespective of whether the dring stage was omitted. A 2-g sample was taken from each laer of powdered concrete and placed in a 100-cm beaker. Either 20 cm of 2% HNO with approximatel 60cm of deionized water, in the case of some of the ethanolaminetreated specimens, or approximatel 80cm of deionized water, in the case of all other specimens were added at 20 8C. The mixture was vigorousl agitated and placed in an ultrasonic water-bath for 5 min, after which it was again agitated. After the concrete particles had settled, the extract was filtered through Whatman no. filter paper into a 250-cm volumetric flask. The concrete

3 M.M. Page et al. / Construction and Building Materials 16 (2002) particles were washed with 80-cm portions of deionized water, shaken thoroughl, allowed to settle and the washings poured through the filter paper into the flask. This was repeated until there was approximatel 250 cm of filtrate in the flask. The total volume was then made up to 250 cm with deionized water Extraction efficiencies In order to determine the proportion of inhibitor that could be extracted from concrete, specimens similar to those described above were cast with known quantities of inhibitor added to the concrete mix and cured for 1 month. Representative pieces of the specimens were ground, dried and 2-g samples treated, as described above, in order to extract the inhibitor Stabilit of sodium monofluorophosphate a Previous work has indicated that sodium monofluorophosphate is liable to suffer hdrolsis at rates dependent on solution ph w5x. In order to test whether water (rather than acid) extraction was preferable for obtaining sodium monofluorophosphate samples from concrete, a series of preliminar experiments was undertaken to assess the effect of solution ph on sodium monofluorophosphate stabilit. Aqueous solutions were prepared containing 290 mg of sodium monofluorophosphate per cm (2 mm) at various ph values: (i) ph 1.7 (using NaOH) to simulate the ph of non-carbonated concrete; (ii) ph 9. (using NaOH) to simulate the ph of carbonated concrete; (iii) ph 7 (using deionized water); (iv) ph 1.6 (using HNO ) to reproduce the ph of the acid extracts made from concrete; and (v) ph 0.5 (using HNO ), to represent the concentration of acid added to powdered concrete samples before dilution with water. The quantities of the monofluorophosphate anion, POF, and its hdrolsis products, F and P O4, were monitored over a period of 40 das using ion chromatograph as described in Section 2.7. b Two further experiments were performed in order to check whether the extraction procedure described in Section 2.4 affected the stabilit of sodium monofluorophosphate. It has been suggested that sodium monofluorophosphate ma be unstable under conditions of high ph ()12) in the presence of excess 2q 2q Ca ions w6x. It seems unlikel that excess Ca ions account for the differences claimed to arise between laborator specimens a few months old and site specimens of several ears w6x, since it has generall been observed that the concentration of 2q Ca ions in the pore solutions of a range of cements drops to onl a few micromoles per cm within 1 da of curing, and becomes relativel constant soon thereafter w7,8x. However, it was important to test whether possible exposure of sodium monofluorophosphate to Ca ions or an other species produced 2q during the extraction procedure could be the reason for its absence from extracts. i. Ten samples of cement paste made from OPC with a watercement ratio of 0.6 were cured for 7 weeks at 100% RH ( weeks at 20 8C and 4 weeks at 8 8C), and then carbonated in 100% CO2 for 7 months at 65% RH. Five of the specimens were then placed in a 5 mm depth of 15% (approx. 1 mol dm ) sodium monofluoro- phosphate solution and all specimens were stored at 100% RH. After 28 das, a 5-mm disc was sliced from the submerged end of each of the five specimens, which had been placed in sodium monofluorophosphate solution. These discs were combined and pore liquid expressed using a pore press as described elsewhere w8x. This avoided an possibilit that sodium monofluorophosphate might be broken down as a consequence of the extraction procedure normall used (Section 2.4). The five specimens of carbonated cement paste, not treated with sodium monofluorophosphate, were similarl sliced and expressed to provide a control sample of pore solution. ii. Two different samples of concrete, one 15-earold naturall carbonated and one non-carbonated, were ground to a powder. Various amounts (10 mg, 100 mg and 1 g) of solid sodium monofluorophosphate were added to 2-g samples of the powdered concretes. A replicate of each of these samples was dried in an oven at 105 8C for 24 h. The dr weights were recorded. To determine whether aspects of the extraction procedure, such as dring or exposure of the extract to powdered concrete, resulted in the breakdown of sodium monofluorophosphate, the dried and undried samples were subjected to the same method of extracting corrosion inhibitors as described in Section 2.4. The solvent used was deionized water Ion chromatograph Solutions and pore liquids, obtained as described in Sections 2.4, 2.5 and 2.6, were diluted as necessar. Twent-five microlitre samples were analsed quantitativel for inhibitor ions, their hdrolsis products and other interacting ions using a Dionex DX500 ion chromatograph sstem fitted with a GP40 gradient pump, a self-regenerating suppression sstem and an ED40 electrochemical detector w9,10x. Samples of anions were analsed in auto-suppression reccle mode after being passed through IonPac AS14 analtical and AG14 guard columns using a solution containing 2 mmol dm

4 76 M.M. Page et al. / Construction and Building Materials 16 (2002) 7 81 Fig. 1. Ion chromatogram showing conductivit peak of ethanolamine. Na2COq2.5 mmol dm NaHCO as eluent at a flow 1 rate of 2 cm min. Cation samples were passed through IonPac CS14 analtical and CG14 guard columns. The were analsed in auto-suppression reccle mode using a 7.5 mmol dm solution of H2SO4 as 1 eluent, at a flow rate of 1 cm min. The conductivit profiles and concentrations of ions were recorded using PeakNet data handling software w11x.. Results and discussion All three inhibitors could be readil detected b means of ion chromatograph (see Figs. 1 )..1. Ethanolamine-based inhibitor Originall, the inhibitor was extracted using dilute nitric acid. Acids are commonl used for the extraction of substances from concrete w12x and do not adversel affect the stabilit of ethanolamine. Penetration profiles, obtained using ion chromatograph, for ethanolamine in various concretes treated with ethanolamine in the laborator are shown in Fig. 4. Later work on concrete, surface-treated with ethanolamine on site, showed similar profiles for both acid-extracted and water-extracted samples, as shown in Fig. 5. Ion chromatograph was also used to show that the recover of ethanolamine from concrete, in which it had been incorporated in the original concrete mix, was approximatel 45% at concentrations of ethanolamine up to 8 mgg concrete. Ethanolamine therefore, is readil detectable b means of ion chromatograph in extracts of concrete obtained using either acid or water as the solvent. It was shown to penetrate to more than 10 mm in all the concretes tested, with the greatest amount entering the laborator non-carbonated concrete and the least found in site concrete (Figs. 4 and 5). Its effect on the corrosion of the reinforcing bars is discussed elsewhere w1x. A further application of ion chromatograph was in the detection of other ions originall present in the commerciall supplied ethanolamine-based solution. For example, the solution was found to contain 1.2 mol q dm PO 4 ions and 1.2 mol dm K ions, suggesting the presence of KH PO as well as ethanolamine. No 2 4 Fig. 2. Ion chromatogram showing conductivit peaks of a range of anions including nitrite.

5 M.M. Page et al. / Construction and Building Materials 16 (2002) Fig.. Ion chromatogram showing conductivit peaks of monofluorophosphate (with fluoride and phosphate). Table 2 Amounts of nitrite detected at 12-mm cover depth in various concretes Tpe of concrete Laborator non-carbonated 1.6 Laborator carbonated 0.21 Site carbonated 0.24 Concentration of nitrite (mgg concrete) O 4 ions such as P, which are potential corrosion inhibitors w1x, in addition to the specified inhibitor. Fig. 4. Penetration profiles of ethanolamine in concrete surface-treated with ethanolamine in the laborator. Fig. 5. Penetration profiles of ethanolamine in concrete surface-treated with ethanolamine on site. PO 4 ions were detected in water-extracted samples of inhibitor-treated concrete. In acid extracts, however, some PO 4 was detected in the carbonated samples of concrete but onl within the first few millimetres and not to the level of the reinforcing bars. Thus, ion chromatograph can be used to detect the presence of.2. Calcium nitrite-based inhibitor In the cases of nitrite-based inhibitors, the use of acid in their extraction from concrete was not feasible as the are unstable at low ph, NO 2 being converted to N O. Therefore, extraction of the calcium nitrite-based inhibitor was performed using deionized water. Extraction efficiencies in the region of 90%, using water as a solvent, have been obtained for NO 2 incorporated into the concrete mix w14x. As in the case of ethanolamine, ion chromatograph has shown that NO 2 has penetrated to more than 10 mm in all the concrete specimens tested, with significantl greater amounts found in the laborator noncarbonated (chloride-contaminated) specimens than in either the site or laborator carbonated samples. For example, at 12 mm cover depth, the amounts of nitrite detected (in mgg concrete) are shown in Table 2. The larger degree of penetration of nitrite into non-carbonated concrete might be due to the counter-diffusion of chloride and hdroxl ions present at higher concentrations in this concrete. Ion chromatograph can also be used to detect and quantif chloride ions present in the same samples (Fig. 6). This is a useful application since it has been suggested that the ratio of NO 2 Cl ions is important in determining the effectiveness of N as an inhibitor O 2

6 78 M.M. Page et al. / Construction and Building Materials 16 (2002) 7 81 Fig. 6. Ratio of water-soluble nitrite to chloride ion concentrations in non-carbonated concrete. Fig. 9. Changes in concentrations of monofluorophosphate, fluoride and phosphate ions with time in deionized water. Fig. 7. Changes in concentrations of monofluorophosphate, fluoride and phosphate ions with time in 0.5 M NaOH. Fig. 10. Changes in concentrations of monofluorophosphate, fluoride and phosphate ions with time in 0.16% HNO. Fig. 8. Changes in concentrations of monofluorophosphate, fluoride and phosphate ions with time in 20 mm NaOH. w15x. Further discussion of the effect of NO 2 on the corrosion of embedded steel is given elsewhere w1,x... Sodium monofluorophosphate-based inhibitor 1. The results of experiments designed to test the stabilit of sodium monofluorophosphate under various Fig. 11. Changes in concentrations of monofluorophosphate, fluoride and phosphate ions with time in 2% HNO. conditions, as described in Section 2.6 (a) and (b), are discussed below. a In order to investigate the stabilit of sodium monofluorophosphate in solutions of different ph values, ion chromatograph was carried out over a period of 40 das to detect the amounts of PO F, F and PO ions. The results are shown 4

7 M.M. Page et al. / Construction and Building Materials 16 (2002) Table The amounts of monofluorophosphate detected in extracts compared with the amounts added to concrete powder State of With or Conc. Conc. carbonation without of POF of POF of concrete dring of added detected concrete (ppm of (ppm of powder extract) extract) Carbonated Undried 0 0 Carbonated Undried Carbonated Undried Carbonated Undried Carbonated Dried 0 0 Carbonated Dried Carbonated Dried Carbonated Dried Non-carbonated Undried 0 0 Non-carbonated Undried Non-carbonated Undried Non-carbonated Undried Non-carbonated Dried 0 0 Non-carbonated Dried Non-carbonated Dried Non-carbonated Dried in Figs It ma be seen that the POF ion remains relativel stable over 40 das in alkaline and neutral solutions (covering the ph range found in the concrete samples and in the solutions used for extraction), and indeed was still stable 1 ear later. Under acid conditions such as those used in the extraction of ethanolamine, the sodium monofluorophosphate undergoes rapid hdrolsis at ph 0.5 (2% HNO ), the POF ion becoming undetectable after 4 das; at ph 1.6 (0.16% HNO ), hdrolsis is slower with approximatel 19% of the original POF ion concentration remaining after 40 das. Thus, in order to avoid hdrolsis of sodium monofluorophosphate during extraction, water was used as the solvent. i. Analsis of pore liquid expressed from specimens of cement paste, which had been stored in sodium monofluorophosphate solution, revealed the presence of a little additional dissolved fluoride (5.1 mg F g cement paste) compared with untreated specimens (1.7 mg F g cement paste). However, neither PO 4 nor POF was detected in the pore solution. Since the extraction procedure described in Section 2.4 was not used to produce this pore liquid, the absence of sodium monofluorophosphate cannot be explained b its breakdown during the extraction procedure. ii. The amounts of POF extracted from the dried and undried samples of powdered concrete, with various quantities of POF added to them and subjected to the same method of extracting corrosion inhibitors as described in Section 2.4, are shown in Table. The concentrations of POF detected in the solutions after extraction ma be compared directl with the concentrations calculated from the quantities of POF added before extraction. In all cases where sodium monofluorophosphate was added to the concrete powders (carbonated and noncarbonated, dried and undried), the POF ion was clearl detectable in the extracts made from them. There was no detectable loss of sodium monofluorophosphate in carbonated samples. Although some loss was observed in non-carbonated samples, significant quantities of sodium monofluorophosphate were detectable even when the specimens were left in contact with water for more than a week. It therefore appears that the dring regime and the normal extraction procedure, which involved contact between powdered concrete and water for 24 h, are not the causes of failure to detect sodium monofluorophosphate in extracts made from sodium monofluorophosphate-treated concrete. 2. Although POF is stable at the ph of the concrete specimens tested, easil extractable when added directl to concrete powder (as shown in the previous section) and readil detectable b means of ion chromatograph (see Fig. ), no trace of the ion was found in the samples extracted, using water as the solvent, from an of the concrete specimens tested, which had been surface-treated with the compound. However, F, a hdrolsis product of POF, was readil detectable and shown to penetrate to depths in excess of 10 mm in both carbonated and noncarbonated concrete samples (Fig. 12). The other hdrolsis product of sodium monofluorophosphate, the PO 4 ion, was also found to penetrate carbonated concrete (Fig. 1), but was not detected in soluble form beond 4 mm depth in non-carbonated concrete. Fig. 12. Penetration profiles of water-soluble fluoride in concrete surface-treated with sodium monofluorophosphate.

8 80 M.M. Page et al. / Construction and Building Materials 16 (2002) Conclusions Fig. 1. Penetration profiles of water-soluble phosphate in concrete surface-treated with sodium monofluorophosphate. This might be due to the removal from solution of various low solubilit phosphate-containing compounds, such as fluorapatite (Ca 5(PO 4) F), under the highl alkaline conditions of non-carbonated concrete. Water-soluble POF and PO4 could not be detected in extracts of concrete made with up to 60 mg sodium monofluorophosphate per gram of cement incorporated into the original concrete mix. F was found to be present at concentrations equivalent to approximatel 25% of that added as sodium monofluorophosphate. Thus, there is evidence for hdrolsis of sodium monofluorophosphate and significant penetration of water-soluble F and PO 4 derived from sodium mon- ofluorophosphate into carbonated concrete, and of F into non-carbonated concrete. Similar degrees of penetration of F and PO 4 were observed for laborator and site-carbonated concrete samples treated in the laborator. There was less penetration for laborator non-carbonated concrete, in which the solubilities of various fluoride and phosphate-containing compounds are ver low, and site-treated concrete, to which less sodium monofluorophosphate was applied. It therefore appears that sodium monofluorophosphate was not stable under an of the conditions of surface application investigated in the work reported here. Discussion of the effect of sodium monofluorophosphate on corrosion of the reinforcing bars is presented elsewhere w1x. It should be pointed out that ion chromatograph has an advantage over a number of commonl used techniques, such as EDXA and XRF, which involve the identification of a constituent element of the inhibitor and not the complete ion. These latter methods ma lead to erroneous conclusions about the presence of ions in concrete. For example, the detection of phosphorus has been taken to indicate that sodium monofluorophosphate is present w6,16,17x when an other phosphoruscontaining species, such as phosphate, might be the actual source of the element. The research described in this paper has demonstrated that, when used with appropriate extraction techniques, ion chromatograph provides a reliable means of characterising the distribution within concrete of certain soluble corrosion inhibitors applied in concrete repair sstems. One of the main advantages of ion chromatograph for this tpe of application is its abilit to analse a wide range of ions in a given sample. When ethanolamine and nitrite were applied to the surface of concrete, their penetration into concrete samples was readil detected b means of ion chromatograph. In the case of monofluorophosphate-treated concrete, onl the hdrolsis products, fluoride and phosphate, were detected in solution and not the monofluorophosphate ion itself. Further work is being undertaken to extend the ion chromatographic techniques to the analsis of a wider range of inhibitors. Acknowledgments Most of the work described in this paper was undertaken while the authors were based at Aston Universit. The research was supported b the EPSRC through a grant (GRK 5268) under the Materials for Better Construction Programme, and b the Electricit Research Co-funding Scheme (ERCOS) with a contribution from British Energ plc. The authors wish to acknowledge this financial support, and are grateful to Dr S. Khan (British Energ) and Mr D. Warne (ERCOS) for their advice. References w1x Page CL, Ngala VT, Page MM. Corrosion inhibitors in concrete repair sstems. Mag Concr Res 2000;52(1):25 7. w2x Anstice DJ. Corrosion inhibitors for the rehabilitation of reinforced concrete. Ph.D. thesis. Aston Universit, wx Ngala VT, Page CL, Page MM. Corrosion inhibitor sstems for remedial treatment of reinforced concrete. Part 1: calcium nitrite. Corros Sci (in press). w4x Grace Construction Products. Application guide: Postrite. Cambridge, MA, USA: W.R. Grace & Co, w5x Van Wazer JR. Phosphorus and its compounds, vol. 1. New York: Interscience, w6x Raharinaivo A, Bouzanne M, Malric B. Influence on concrete ageing on the effectiveness of monofluorophosphate for mitigating the corrosion of embedded steel. Proc EuroCorr ; w7x Diamond S. Effects of two Danish flashes on alkali contents of pore solutions of cement flash pastes. Cem Concr Res 1981;11:8 94. w8x Page CL, Vennesland Ø. Pore solution composition and chloride binding capacit of silica-fume cement pastes. Mater Struct 198;16:19 25.

9 M.M. Page et al. / Construction and Building Materials 16 (2002) w9x GP50 gradient pump operator s manual. Document no. 0177, Dionex Corporation, w10x ED40 electrochemical detector operator s manual. Document no , Dionex Corporation w11x PeakNet software user s guide. Document no , Dionex Corporation, w12x British Standard: testing concrete methods for analsis of hardened concrete. BS , w1x Mane JEO, Menter JW. The mechanism of inhibition of corrosion of iron b solutions of sodium phosphate, borate and carbonate. J Chem Soc 1954; w14x Jeknavorian AA, Chin D, Saidha L. Determination of a nitritebased corrosion inhibitor in plastic and hardened concrete. Cem Concr Aggregates 1995;17: w15x Gaidis JM, Rosenberg AM. The inhibition of chloride-induced corrosion in reinforced concrete b calcium nitrite. Cem Concr Aggr 1987;9:0. w16x Raharinaivo A. Action des monofluorophosphates sur la corrosion des armatures dans le beton. Laboratoire Central des Ponts et Chaussees: rapport DTOAMAR 81 96, 1996, 16 pp. w17x MFP technical information. Balvac, BICC Group, 1998, 8 pp.

10 本文献由 学霸图书馆 - 文献云下载 收集自网络, 仅供学习交流使用 学霸图书馆 ( 是一个 整合众多图书馆数据库资源, 提供一站式文献检索和下载服务 的 24 小时在线不限 IP 图书馆 图书馆致力于便利 促进学习与科研, 提供最强文献下载服务 图书馆导航 : 图书馆首页文献云下载图书馆入口外文数据库大全疑难文献辅助工具