pubs.acs.org/Langmuir © 2010 American Chemical Society
Quantitative Prediction of the Reduction of Corrosion Inhibitor Effectiveness Due to Parasitic Adsorption onto a Competitor Surface Bernard P. Binks, Paul D. I. Fletcher,* and Ibrahim E. Salama Surfactant & Colloid Group, Department of Chemistry, University of Hull, Hull HU6 7RX, U.K.
David I. Horsup and Jennifer A. Moore Nalco Company, Energy Services Division, 7705 Highway 90A, Sugar Land, Texas 77478, United States Received September 6, 2010. Revised Manuscript Received November 16, 2010 We have investigated how the effectiveness of a corrosion inhibitor added to an aqueous solution to suppress the corrosion rate of steel is reduced by the addition of sand. The equilibrium adsorption isotherms of the inhibitor with respect to both the steel surface (consisting of iron carbonate under the corrosion conditions used here) and the sand surface have been measured. The results enable the quantitative calculation of how the surface concentration of inhibitor at the steel surface is reduced by sand addition. Combining the adsorption information with measurements of how the steel corrosion rate depends on the inhibitor surface concentration enables the quantitative prediction of the inhibitor effectiveness as a function of sand concentration. Excellent agreement is obtained between calculated and measured values of the inhibitor performance as functions of both inhibitor and sand concentrations. This methodology demonstrates how the optimization of a corrosion inhibitor formulation for specific application conditions should take into account the parasitic adsorption of the inhibitor onto the competitor surfaces present.
Introduction Corrosion inhibitors have been used for many years to protect oil and gas pipelines.1,2 The application of small quantities of an inhibitor to production fluids is often one of the most costeffective methods for imparting corrosion protection in a system. For carbon steel pipelines used to transport multiphase fluids consisting of mixtures of oil, water, and gas, a small concentration of the inhibitor is added to the fluids and inhibits corrosion by adsorption at the steel-water surface. Common inhibitor species used include alkyl quaternary ammonium salts and imidazoline derivatives that are strongly surface-active and contain a cationic hydrophilic headgroup coupled to one or more alkyl chains that form the hydrophobic tailgroup. Inhibitor performance is strongly dependent on the composition of the aqueous solution in contact with the steel surface for a variety of factors. First, the steel corrosion rate in the absence of inhibitor depends on the electrolyte concentration, pH, and concentrations of all species appearing in the electrochemical corrosion reaction according to electrochemical theory.1 Second, the nature of the steel surface is controlled by the prevailing solution conditions; for example, under the so-called “sweet” conditions of CO2 saturation, the steel surface layer comprises mainly iron carbonate. Under “sour” conditions of H2S saturation, the surface is mainly iron sulfide, and under air-saturation conditions, the surface consists of mixed iron oxides.3 Inhibitor adsorption is different for these different surfaces. Third, inhibitor adsorption to the relevant surface is also strongly dependent on factors such *Corresponding author. E-mail:
[email protected]. (1) Jones, D. A. Principles and Prevention of Corrosion, 2nd ed.; Prentice Hall: Upper Saddle River, NJ, 1996. (2) Rozenfield, I. L. Corrosion Inhibitors; McGraw Hill: New York, 1981. (3) Binks, B. P.; Fletcher, P. D. I.; Hicks, J. T.; Durnie, W.; Horsup, D. I. Corrosion 2005, NACE, 2005, paper 05307. (4) Atkin, R.; Craig, V. S. J.; Wanless, E. J.; Biggs, S. Adv. Colloid Interface Sci. 2003, 103, 219.
Langmuir 2011, 27(1), 469–473
as electrolyte concentration and temperature.4,5 Fourth, the aqueous phase in contact with the steel surface commonly contains additional components such as sand or other powdered minerals and water-immiscible oil in the form of oil-in-water emulsion drops. In this situation, the inhibitor species concentration in the aqueous phase can be depleted by adsorption to the surfaces of the powdered minerals, partitioning to the oil phase, and adsorption to the oil-water surfaces of the emulsified oil drops.6-12 Because all of these factors vary widely between different application fields, a corrosion inhibitor formulation must be tailored/optimized to suit the particular conditions pertaining to a specific application. There is a large amount of background literature relevant to these types of corrosion inhibitors, including corrosion rates,3,13-21 surface-enhanced Raman studies,22,23 and quartz (5) Zhang, R.; Somasundaran, P. Adv. Colloid Interface Sci. 2006, 123-126, 213. (6) Joosten, M. W.; Kolts, J.; Humble, P. G.; Gough, M. A.; Hannah, I. M. Corrosion 2000, NACE, 2000, paper 00018. (7) McMahon, A. J.; Martin, J. W.; Harris, L. NACE, 2005, paper 05274. (8) Gulbrandsen, E.; Kvarekval, J. Proceedings of the 16th International Corrosion Congress, Sept 19-24, 2005, Beijing, China. (9) Jiang, X.; Zheng, Y. G.; Ke, W. Corros. Sci. 2005, 47, 2636. (10) Horsup, D. I.; Clark, J. C.; Binks, B. P.; Fletcher, P. D. I.; Hicks, J. T. Corrosion 2007 Conference & Expo, NACE, 2007, paper 07617. (11) Liu, X.; Zheng, Y. G.; Okafor, P. C. Mater. Corros. 2009, 60, 507. (12) Horsup, D. I.; Clark, J. C.; Binks, B. P.; Fletcher, P. D. I.; Hicks, J. T. Corrosion 2010, 66, 036001–1. (13) Zvauya, R.; Dawson, J. L. J. Appl. Electrochem. 1994, 24, 943. (14) Vasudevan, T.; Muralidharan, S.; Alwarappan, S.; Iyer, S. V. K. Corros. Sci. 1995, 37, 1235. (15) Khamis, E.; Al-Lohedan, H. A.; Al-Mayouf, A.; Issa, Z. A. Materialwiss. Werkstofftech. 1997, 28, 46. (16) Durnie, W. H.; Kinsella, B. J.; De Marco, R.; Jefferson, A. J. Appl. Electrochem. 2001, 31, 1221. (17) Free, M. L. Corros. Sci. 2002, 44, 2865. (18) Wang, W.-L.; Free, M. L. Anti-Corros. Methods Mater. 2003, 50, 186. (19) Bilkova, K.; Gulbrandsen, E. Electrochim. Acta 2008, 53, 5423. (20) Asefi, D.; Arami, M.; Sarabi, A. A.; Mahmoodi, N. M. Corros. Sci. 2009, 51, 1817. (21) Foss, M.; Diplas, S.; Gulbrandsen, E. Electrochim. Acta 2010, 55, 4851. (22) Oblonsky, L. J.; Chesnut, G. R.; Devine, T. M. Corrosion 1995, 51, 891.
Published on Web 12/03/2010
DOI: 10.1021/la103570e
469
Article
Binks et al.
crystal microbalance studies.24-26 However, despite the large amount of literature, there is a lack of direct information showing how the extent of inhibitor adsorption is related to the corrosion rate under corresponding conditions and how the inhibitor performance is affected by the factors discussed above. Indeed, as pointed out by Free,27 many corrosion researchers do not measure the inhibitor adsorption directly but assume that the extent of inhibition is proportional to the fractional surface coverage of the corroding surface by the inhibitor. In this study, we have examined the inhibition of 1018 carbon steel in CO2-saturated brine solution by hexadecylbenzyldimethylammonium chloride (C16BDMAC) and how this is affected by sand addition. The main aim of this work is to measure directly how sand addition affects the surface concentration of the inhibitor on the surface of iron carbonate (this surface is taken to represent the actual surface layer present for steel under CO2saturated conditions) and combine this with information on how the corrosion rate depends on the inhibitor surface concentration in order to predict inhibitor performance quantitatively under all conditions.
Experimental Section Materials. Water was purified by an Elga Prima reverse osmosis unit and then treated with a Milli-Q reagent water system. The produced water had a resistivity of 18 MΩ cm or greater. Hexadecylbenzyldimethylammonium chloride (C16BDMAC, >99% pure, Fluka) was used as received. The synthetic brine mixture was prepared according to ref 28 and consisted of sodium chloride (68.08), magnesium chloride (14.43), sodium sulfate (11.35), calcium chloride (3.22), potassium chloride (1.93), sodium hydrogen carbonate (0.55), potassium bromide (0.28), boric acid (0.07), strontium chloride (0.07), and sodium fluoride (0.008), where each number in parentheses represents the weight percentage of the component in the mixture of solids prior to addition to water. All salts were reagent grade and were used as received. The synthetic brine used here consisted of a 4.7 wt % mixture of salts in water. The sand, mainly consisting of SiO2 in the form of quartz, was from Sigma and was supplied as sulfuric acid washed. It was checked that the addition of this sand did not change the pH of the aqueous solutions to which it was added. Iron carbonate (FeCO3) powder (technical grade) was supplied by City Chemical. Figure 1 shows scanning electron micrographs of the two powders. Methods. The specific surface areas of the sand and iron carbonate powders were determined using nitrogen gas adsorption isotherms analyzed according to the Brunauer, Emmett, and Teller (BET) method.29,30 The nitrogen adsorption isotherms were measured at 77 K using a TriStar 3000 Analyzer manufactured by Micromeritics. The adsorption isotherms (not shown) showed no significant hysteresis or porosity and were similar for both powders. The adsorption isotherms of the C16BDMAC corrosion inhibitor from solution in 4.7 wt % brine onto the two powder surfaces were determined by measuring the extent of depletion of solution concentrations following adsorption. Samples containing C16BDMAC in brine solution were first analyzed as described (23) Durnie, W. H.; De Marco, R.; Jefferson, A.; Kinsella, B. J. Surf. Interface Anal. 2003, 35, 536. (24) Knag, M.; Sjoblom, J.; Oye, G.; Gulbrandsen, E. Colloids Surf., A 2004, 250, 269. (25) Knag, M.; Tammelin, T.; Bilkova, K.; Johansson, L.-S.; Gulbrandsen, E.; Sjoblom, J. J. Dispersion Sci. Technol. 2006, 27, 277. (26) Ryu, D. Y.; Free, M. L. J. Colloid Interface Sci. 2003, 264, 402. (27) Free, M. L. Corros. Sci. 2004, 46, 2601. (28) ASTM D-1141-98 (reapproved 2003), ASTM International. (29) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309. (30) Lowell, S. Introduction to Powder Surface Area; John Wiley & Sons: New York, 1979.
470 DOI: 10.1021/la103570e
Figure 1. Scanning electron micrographs of sulfuric acid-washed sand particles (upper image) and iron carbonate particles (lower image). below to obtain the initial C16BDMAC concentration. Weighed amounts of either sand or iron carbonate particles were added to the solutions (60 mL) in polystyrene bottles and equilibrated with stirring in a thermostatted bath for 4 h. Glass bottles were not used to avoid possible complications arising from the strong adsorption of the cationic C16BDMAC onto the glass walls. The pH of all brine solutions was 8.0. Following adsorption equilibration, the samples were decanted and centrifuged using an Eppendorf MiniSpin plus (Centrifuge) at 14 000 rpm for 10 min to remove all solid particles prior to measuring the final depleted concentration of C16BDMAC using high-performance liquid chromatography (HPLC). A Shimadazu liquid chromatograph equipped with an LC-6A pump and a 759A absorbance detector (Applied Biosystems) operating at 262 nm was used for the measurements. A SPHERISORB 5 μm CN (25 0.46 cm2) column was used at a mobile phase flow rate of 2 mL/min and a sample injection volume of 20 μL. The mobile phase was a mixture of 70 vol % acetonitrile and 30 vol % 0.2 M aqueous sodium acetate adjusted to pH 5.0 with acetic acid. The calibration plot of the C16BDMAC response peak integral versus concentration was accurately linear over the C16BDMAC concentration of 1 10-8 to 1 10-2 M with a correlation coefficient r2 value of 0.99. In deriving the final adsorption isotherms, two points were carefully checked. First, samples equilibrated for different times in excess of 4 h produced identical results, demonstrating that equilibrium adsorption was achieved within 4 h. Second, measurements made for samples containing different ratios of C16BDMAC to solid powder produced entirely self-consistent results. Langmuir 2011, 27(1), 469–473
Binks et al.
Article
The corrosion rates of 1018 carbon steel in 4.7 wt % brine solution under CO2-saturated conditions (so-called sweet corrosion) were determined using linear polarization resistance measurements made with a GillAC 12 potentiostat (ACM Instruments). 1018 carbon steel electrodes (with the following elemental composition:3 98.898% Fe, 0.72% Mn, 0.19% Si, 0.18% C, 0.009% P, and 0.003% N) of 4 mm diameter and 30 mm length were supplied by European Corrosion Supplies. The electrodes were received in vapor corrosion inhibitor (VCI)-impregnated envelopes. They were washed with xylene and acetone to remove the VCI and then dried in an oven at 60 °C prior to use. For the measurements, 800 mL of an aqueous phase was added to the thermostatted measurement cell fitted with temperature and electrochemical probes, gas dispersion tubes, condensers, and air traps and sealed against the ingress of air. The electrochemical probe consisted of a stainless steel rod fitted with three 1018 carbon steel electrodes: the working, counter, and reference electrodes. The system was thermostatted at 25 ( 0.2 °C using a hot plate and stirred at a speed of 200 rpm with a magnetic stirrer. The aqueous phase in the cell was sparged with CO2 for 2 h prior to the start of the experiment to ensure CO2 saturation. In a typical experimental run, corrosion rates in the absence of added C16BDMAC inhibitor were monitored for 1 to 2 h to determine the baseline corrosion rate. The required concentration of C16BDMAC inhibitor was then added, and the corrosion rate was continuously monitored for 15-20 h, which was long enough to ensure that the rate had reached a steady value. The critical micelle concentration (cmc) of C16BDMAC in a 4.7 wt % brine solution was determined by measuring the surface tension of the aqueous solution-air surface as a function of C16BDMAC concentration. Surface tensions were measured by the static maximum pull method using a du No€ uy ring and a K12 tensiometer (Kr€ uss). Scanning electron microscopy (SEM) images of the iron carbonate and sand powders were recorded using an EVO 60 SEM (Carl Zeiss SMT AG) instrument equipped with an INCAEnergy 350 EDX detector (Oxford Instruments).
Results and Discussion Figure 2 shows the cmc value and equilibrium adsorption isotherms for C16BDMAC adsorbing from a 4.7 wt % brine solution onto sand or iron carbonate. Both isotherms are fitted to eq 1 in which the total adsorbed surface concentration is given by the sum of two Langmuir-type isotherms—a two-stage Langmuir isotherm Γ ¼
Γ1, max k1 C Γ2, max k2 C þ 1 þ k1 C 1 þ k2 C
ð1Þ
where Γ is the total adsorbed amount, Γ1,max and Γ2,max are the maximum surface concentrations of adsorbate in stages 1 and 2, k1 and k2 are constants reflecting the affinity of the adsorbate for the surface in stages 1 and 2, and C is the equilibrium concentration of “free” (i.e., nonadsorbed) adsorbate. Although a Langmuir isotherm corresponds to a well-defined model in which adsorbate molecules bind to defined surface sites as a monolayer with no lateral interactions, the use of eq 1 is not meant to imply that the model assumptions of a Langmuir isotherm are valid for the adsorption of C16BDMAC onto either sand or iron carbonate or to infer any particular configuration of the molecules in the adsorbed film. For this type of information, the reader is referred to the extensive literature on the adsorption and structures of adsorbed films of quaternary ammonium surfactant (31) (32) (33) (34) 365.
Zajac, J.; Trompette, J. L.; Partyka, S. J. Therm. Anal. 1994, 41, 1277. Trompette, J. L.; Zajac, J.; Keh, E.; Partyka, S. Langmuir 1994, 10, 812. W€angnerud, P.; J€onsson, B. Langmuir 1994, 10, 3268. W€angnerud, P.; Berling, D.; Olofsson, G. J. Colloid Interface Sci. 1995, 169,
Langmuir 2011, 27(1), 469–473
Figure 2. Adsorption isotherms at 25 °C for C16BDMAC from 4.7 wt % brine onto iron carbonate and sand. The vertical dashed line corresponds to the cmc of C16BDMAC. Table 1. Powder Surface Area Values and Fitting Parameters for the Two-Stage Langmuir Adsorption Isotherms of C16BDMAC for 4.7% Brine Solution onto Iron Carbonate and Sand Particles at 25 °C parameters 2
-1
specific surface area/m g k1/M-1 k2/M-1 Γ1 (max)/molecules nm-2 Γ2 (max)/molecules nm-2 R2
sand
FeCO3
0.22 1.30 106 4.99 104 0.80 3.99 0.99
0.198 3.24 106 2.21 102 0.05 2.66 0.99
species onto (sandlike) silica surfaces4,5,31-39 and a single study of adsorption onto iron carbonate using zeta potential measurements.40 However, for the purpose of the present study, the excellent fits to eq 1 are primarily used as a convenient means to parametrize the extent of C16BDMAC adsorption to the two different surfaces. The values of the best-fit parameters are summarized in Table 1. Figure 2 shows that C16BDMAC adsorbs more strongly to sand than to iron carbonate over the entire concentration range. As discussed in the Introduction, we expect that the effect of C16BDMAC on inhibiting steel corrosion under CO2 to be governed by its extent of adsorption on iron carbonate. If sand is added to a system in which free C16BDMAC is in equilibrium with C16BDMAC adsorbed on iron carbonate, then competing adsorption to the sand will reduce the extent of adsorption on iron carbonate, resulting in a reduced corrosion inhibition effect. We can use the isotherm data to calculate this effect using the following steps. (1) From the experimental geometry and sand concentration, we first calculate the surface areas of steel (i.e., iron carbonate for CO2 conditions), Aic, and sand As per unit volume of the aqueous solution. (2) For a set value of [C16BDMAC]free, we then calculate the surface concentrations on both sand (Γs) and iron carbonate (Γic) using eq 1. (3) We then calculate the overall concentration of C16BDMAC using ½C16BDMACtotal ¼ ½C16BDMACfree þ Γic Aic þ Γs As (35) (36) (37) (38) (39) (40)
ð2Þ
Zajac, J.; Trompette, J. L.; Partyka, S. Langmuir 1996, 12, 1357. Atkin, R.; Craig, V. S. J.; Biggs, S. Langmuir 2000, 16, 9374. Paria, S.; Yuet, P. K. Ind. Eng. Chem. Res. 2006, 45, 712. Macakova, L.; Blomberg, E.; Claesson, P. M. Langmuir 2007, 23, 12436. Gutig, C.; Grady, B. P.; Striolo, A. Langmuir 2008, 24, 4806. Foss, M.; Gulbrandsen; Sjoblom, J. J. Disp. Sci. Technol. 2010, 31, 200.
DOI: 10.1021/la103570e
471
Article
Figure 3. Variation of the surface concentration of C16BDMAC on an iron carbonate surface with the wt % of added sand for different fixed total concentrations of C16BDMAC (indicated in the key). The calculations refer to a surface area of the iron carbonate per liter of solution of 4 10-4 m2.
Using the adsorption isotherm fitting parameters and specific surface areas given in Table 1, we have calculated how Γic changes with the addition of sand for various fixed values of [C16BDMAC]total. For the calculations shown in Figure 3, we have taken Aic to be 4 10-4 m2/L, which corresponds to the surface area per unit solution volume of corroding steel electrode in the corrosion rate apparatus used in this work. In a practical application such as a solution flowing down a pipe, Aic is given by the internal surface area, including surface roughness effects, of the pipe divided by its volume. The specific surface area of the sand used here is 0.22 m2 g-1, hence a sand concentration of 1 wt % corresponds to As = 2.2 m2/L. From Figure 3, we can see that, at a fixed value of [C16BDMAC]total, Γic is reduced by sand addition with the effect being most pronounced at low overall inhibitor concentrations. We next examine how the steel corrosion rate is related to the extent of inhibitor adsorption (i.e., Γic). Figure 4 shows examples of typical corrosion rate experimental runs from which the rates with and without inhibitor with different concentrations of sand were determined. In the absence of inhibitor, 1018 steel in CO2-saturated brine solution corrodes at a rate of approximately 1 mm year-1. Following the addition of a total concentration of 4 ppm (1.01 10-5 M) C16BDMAC, the rate drops over 1 or 2 h to an inhibited steady value of about 0.03 mm year-1. The presence of increasing concentrations of sand does not affect the corrosion rate in the absence of inhibitor but does progressively reduce the inhibitory effect of the C16BDMAC (i.e., sand addition increases the corrosion rate at a fixed total concentration of inhibitor). Hence, we can see qualitatively that the effects discussed above are observed experimentally. To model the effects of sand addition quantitatively, we need to understand how the corrosion rate is related to the extent of C16BDMAC adsorption in the absence of added sand. Figure 5 shows how the steady corrosion rate depends on the extent of C16BDMAC adsorption on the iron carbonate surface Γic. The corrosion rate is reduced by about 2 orders of magnitude as Γic increases from 0 to a very low value of only 0.08 molecules nm-2. It is interesting to estimate how this actual coverage by inhibitor compares with the coverage corresponding to a close-packed monolayer. C16BDMAC (with a molecular weight of 395 g mol-1) has a molecular volume of about 0.7 nm3. If it is assumed that the adsorbed C16BDMAC molecules lie flat (i.e., with the long molecular axis oriented 472 DOI: 10.1021/la103570e
Binks et al.
Figure 4. Example runs showing the corrosion rates of mild steel before and after the addition of 1.01 10-5 M (4 ppm) C16BDMAC in the presence of various concentrations of sand. Measurements were made under CO2 in 4.7 wt % brine at 25 °C. The vertical dashed lines mark the times at which the C16BDMAC was added.
Figure 5. Variation of the corrosion rate of mild steel with the surface concentration of adsorbed C16BDMAC under CO2 in 4.7 wt % brine at 25 °C. The upper plot (with the corrosion rate on a logarithmic scale) shows the fit to an exponential function, and the lower plot shows a linear fit to the initial part of the plot.
parallel to the surface) on the iron carbonate surface and that the thickness of the molecule is about 0.4 nm, then the area occupied by a single adsorbed molecule (= volume/thickness) is estimated to be approximately 1.8 nm2. Hence, the value of Γic Langmuir 2011, 27(1), 469–473
Binks et al.
Figure 6. Percentage decrease in the corrosion protection efficiency of C16BDMAC as a function of the total inhibitor concentration caused by the addition of 1 (filled symbols and dashed line) and 2 (open symbols and solid line) wt % sand. The solid and dashed lines show the predicted changes based on the corrosion rate decreasing linearly with the amount of C16BDMAC adsorbed on the iron carbonate surface. Measurements were made under CO2 in 4.7 wt % brine at 25 °C.
corresponding to a close-packed monolayer of C16BDMAC molecules lying flat on the surface is approximately 0.6 molecules nm-2. From this crude calculation, it can be seen that the corrosion rate is reduced approximately 100-fold by an extent of adsorption that corresponds to only about 13% of a hypothetical state of complete coverage by flat C16BDMAC molecules. The fraction of the surface covered by adsorbed inhibitor molecules oriented perpendicular to the surface would be even lower than 13%. Overall, this crude comparison suggests that the coverage of only a small fraction of the total carbon steel (iron carbonate) surface controls the corrosion inhibition. We speculate here that these critical surface sites may, for example, correspond to the sites of cathodic corrosion where the negative surface charge is likely to cause the strong adsorption of this cationic inhibitor. The data in Figure 5 has been fitted in two ways. For the entire data set, the corrosion rate very approximately follows an exponential decrease with Γic, but we note that the measured rates do show significant deviation from the fitted line. For values of Γic of less than about 0.05 molecules nm-2, the corrosion rate decreases linearly with Γic. We have used this latter linear fit to
Langmuir 2011, 27(1), 469–473
Article
calculate the percentage decrease in corrosion protection efficiency P (defined as P=(Ro - R)/R where Ro is the corrosion rate in the absence of inhibitor and R is the rate in the presence of inhibitor) caused by the addition of either 1 or 2 wt % sand. Figure 6 compares the calculated changes in P with the measured values where it can be seen that the addition of sand decreases the corrosion inhibitor efficiency at low total concentrations of C16BDMAC but has a negligible effect at higher inhibitor concentrations. The agreement between measured and calculated values over a range of total C16BDMAC concentration is excellent. The conditions used for this experiment correspond to those for which the corrosion rate decreases linearly with Γic (Figure 5), consistent with the good quality of fit seen in Figure 6. Using the exponential function to describe the corrosion rate decrease with Γic produces a significantly poorer fit to the data of Figure 6. Some aspects of the relationship between corrosion rate inhibition and inhibitor surface concentration have been discussed by Free27 (although the arguments presented do not appear to be based on direct measurements of the relevant surface concentrations). Finally, we note here that we have considered only the effects due to parasitic inhibitor adsorption caused by sand addition. Under strong flow conditions, sand addition can additionally affect corrosion rates through erosion processes (e.g., ref 9). Under our experimental conditions of relatively low liquid flow/ stirring, we note (i) that the corrosion rate in the absence of inhibitor is unaffected by sand addition (Figure 4) and (ii) that the effects of sand addition are fully accounted for using only considerations of the parasitic adsorption of inhibitor (Figure 6). Hence, we conclude that erosion effects by sand particles are not significant under our conditions.
Conclusions We have shown how adsorption isotherms for the adsorption of a corrosion inhibitor species onto the relevant surfaces can be used to predict quantitatively the conditions under which the inhibitor efficiency will decrease because of loss through adsorption to a competitive surface. Optimization of the overall performance of a corrosion inhibitor for a particular set of operating conditions needs to take account of the affinity of the inhibitor for competitor surfaces present in addition to that for the primary target surface undergoing corrosion. The approach described in this work is easily extendable to cover inhibitor loss to multiple competitor surfaces; for example, in addition to inhibitor loss to sand surfaces considered here, eq 2 is easily modified to include additional loss by adsorption to the oil-water interface present for a system containing emulsion oil drops in addition to sand. Acknowledgment. We thank Nalco Company for their funding of this work. Nalco is a trademark of Nalco Company.
DOI: 10.1021/la103570e
473
