Surface and Coatings Technology 184 (2004) 322–330

Understanding the inhibiting properties of 3-amino-1,2,4-triazole from fractal analysis a, ´ E. Garcıa-Ochoa *, J. Genescab

´ Molecular, Instituto Mexicano del Petroleo, ´ ´ ´ Programa de Ingenierıa Eje Central Lazaro Cardenas 噛 152, San Bartolo Atepehuacan, Mexico b ´ Metalurgica, ´ ´ Departamento Ingenierıa Facultad Quımica, UNAM. Ciudad Universitaria, 04510 Mexico

a

Received 14 July 2003; accepted in revised form 15 November 2003

Abstract The effect of 3-amino-1,2,4-triazole (ATR) as a corrosion inhibitor for mild steel in acid media has been investigated by means of potentiodynamic polarization curves, electrochemical noise measurements (ENM) and electrochemical impedance spectroscopy (EIS). It appears that this compound inhibits mild steel corrosion by affecting both cathodic and anodic reactions. The experimental results obtained by ENM and EIS shown that a direct relationship can be established between the fractal dimension of the electrode surface and the Hurst exponent calculated from current and potential electrochemical noise time records, being both parameters affected by the presence of the ATR inhibitor. 䊚 2003 Elsevier B.V. All rights reserved. Keywords: Corrosion inhibition; Electrochemical noise; Fractals; Electrochemical impedance spectroscopy (EIS); Hurst exponent; Triazole

1. Introduction Corrosion control can be achieved by many methods, being corrosion inhibitors one of the most effective alternatives for the protection of metallic surfaces against corrosion. Corrosion inhibitors are normally based on organic compounds, which when added to a corrosive environment in small concentrations can produce a significant decrease of the corrosion rate. The inhibition mechanism is carried out through their adsorption on the metallic surfaces, being able to isolate them from the corrosive media w1x. Between the organic compounds of interest, triazoles have been studied extensively because of their corrosion inhibition properties to mild steel in acid media w2–5x. Among the different electrochemical techniques that can be used to study corrosion inhibitors, EIS appears as powerful tool for the information that can provide, as for example, double-layer capacitance, Cdl, and polarization resistance, Rp, values. Changes in these parameters as a function of time or with respect to other *Corresponding author. Tel.: q52-30-03-6240. ´ E-mail address: [email protected] (E. Garcıa-Ochoa).

variables, allow to obtain important information about the kinetics of the corrosion process being involved w6x. In some cases impedance data obtained at the corrosion potential, Ecorr have the shape of depressed semicircles with the center of the semicircle below the real axis. Its Equivalent Circuit, EC, corresponds to a parallel combination of a capacitance and one resistance w7x, Fig. 1. Constant-Phase Elements (CPEs) are widely used in data fitting to allow for depressed semicircles w8x. The capacitor in the EC is replaced with a CPE for better fit quality. The impedance of a CPE is given by ZsZ0Ž jv.yn

(1)

where n can take values between 0 and 1, depending of the circuit element represented.The complex impedance Z (jv) of a depressed semicircle could be expressed as: ZsRsq

Rp 1q( jvCdlRp)n

(2)

The n exponent is a unitless parameter that equals to 1 for an ideal capacitor. In most real systems, ideal capacitive behavior is not observed due to surface roughness, or other effects that causes uneven current

0257-8972/04/$ - see front matter 䊚 2003 Elsevier B.V. All rights reserved. doi:10.1016/j.surfcoat.2003.11.019

´ E. Garcıa-Ochoa, J. Genesca / Surface and Coatings Technology 184 (2004) 322–330

323

Statistical methods were found to be particularly useful for the analysis of electrochemical noise, EN, data. This statistical analysis involves the estimation of the electrochemical noise resistance, Rn, which is calculated as the standard deviation of potential (sv) divided by the standard deviation of the current (sI): R ns

Fig. 1. Equivalent circuit with a constant phase element, CPE.

distributions on the electrode surface. In the case when ns1, the term (jvCdlRp)n reduces to jvCdlRp, where Cdl is the interfacial double layer capacitance. This can be interpreted as an indication of the degree of inhomogeneity of the metal surface w9x, where n take values slightly higher to 0.5 that corresponds to a severe inhomogeneity, to 1, in which case the electrode surface can be taken as smooth. This degree of inhomogeneity has been associated with the fractal dimension of the surface w10,11x. Taking into account the degree of depression of the semicircle in the Nyquist diagram, it is possible to determine the fractal dimension of the electrode surface by means of the following equation w12x ns

1 Dsy1

(3)

where Ds is the fractal dimension of the surface, which can take values between 2 for a surface completely smooth, to values less than 3 for a roughness surfaces. Mac Rae et al. w13x verified this behavior demonstrating that the fractal dimension of one electrode can be determined by means of electrochemical impedance spectroscopy (EIS) measurements and atomic force microscopy (AFM). Electrochemical noise measurements (ENM) have also been successfully applied to the study of corrosion inhibitor performance w14–16x. These measurements are made without any external perturbation of the system. This technique is very attractive because any information they are able to provide relates to the actual system being studied, with little possibility of artifacts due to the measurement technique w17x.

sV sI

where the value of Rn can be taken as Rp and then inversely proportional to corrosion rate according to Stern–Geary equation, but with the necessary condition that a trend removal is applied over an average baseline, as previously established by Tan et al. w15x. It has been demonstrated w18x that the noise signal contains information about the dynamics that occur on the surface of the electrode. For this reason, many efforts have been realized to analyze the noise signal by different methods, among them the Hurst exponent, H, which has offered excellent results in many systems in which corrosion phenomena exists w18–21x. The aim of this work is to study the performance of a simple organic compound such as triazole, as a corrosion inhibitor by comparing the results obtained by two electrochemical techniques, electrochemical noise measurements (ENM) and EIS. Triazole accounts with two active sites, Fig. 2, where the inhibitor–surface interaction can take place w22x. Special emphasis has been made to determine the morphology of the metal surface expressed as the fractal dimension of the surface as obtained by EIS and comparing this dimension with the Hurst exponent, calculated from ENM. The main goal is to establish if the Hurst exponent can be related with the morphology of the metal surface of a corroding electrode and then can be used as an index of the level of corrosion resistance provided by the inhibitor. 1.1. Estimation of Hurst exponent Many observations of nature consist of time records, which can be analyzed in terms of Hurst’s rescaled

Fig. 2. Molecular structure of 3-amino-1,2,4 – triazole (ATR).

324

´ E. Garcıa-Ochoa, J. Genesca / Surface and Coatings Technology 184 (2004) 322–330

Fig. 3. (a) Current time record for mild steel in 0.5 M HCl. (b) Rescaled range (RyS) analysis result for the time record shown in Fig. 3a.

range analysis w23x. The records are characterized by an exponent H, the Hurst exponent. The trace of the record is a curve with a fractal dimension Dts2yH, where Dt is the fractal dimension of the time record. The Hurst exponent, H, reveals long-term dependence in a time record and can be evaluated from the fluctuations occurring in the data. When the variation in the time record over a specific time interval (the lag time) is proportional to the lag time raised to the power 2H, the time record is said to be fractal w19x. An electrochemical noise time record is a random fractal, where the levels of detail are similar but not identical. The fractal dimension, Dt, describes the structure of a fractal, e.g. the ‘roughness’ of an EN time record w19x. When a time record is a fractal, its average structure is independent of the time (or frequency) scale w19x. One method for analyzing the structure of a time record is by using the Hurst exponent, H. It has been demonstrated that H characterizes the statistical relationship between two values of the time record separated by a time increment. A more detailed explanation can be found in the book of Feder w23x. Horvath and Schiller w18x following the lines of Feder’s w23x work proposed the rescaled range analysis, also called as RyS or Hurst method to calculate the Hurst exponent. The relation between the ratio RyS and the Hurst exponent can be expressed by the following equation: RyS tisA tH

(5)

where R is the range, S is the variance of a discrete time record, xi for the same period, t. Being log (RyS) linearly proportional to log (t), the numerical value of H can be obtained from experimental log (RyS) vs. log (t) plots, called Pox diagram w24x

Fig. 4. Probe sample with four identical electrodes.

Fig. 5. Potentiodynamic polarization curves of mild steel in 0.5 M HCl solution for various concentrations of 3-amino-1,2,4-triazole.

´ E. Garcıa-Ochoa, J. Genesca / Surface and Coatings Technology 184 (2004) 322–330

325

Table 1 Corrosion potential, Ecorr, corrosion current density, icorr and inhibitor efficiency as a function of 3-amino-1,2,4 – triazole (ATR) concentration. Mild steel in 0.5 M HCl at room temperature wATRx, mM

Ecorr, mV (SCE)

icorr, mA cmy2

Inhibition efficiency, %

0 1.0 1.5 2.0

y474 y500 y493 y495

2.130 0.481 0.363 0.120

77 82 94

via statistical regression as the slope of a straight-line fit. This analysis is presented in Fig. 3. Fig. 3a is the experimentally obtained current time record for mild steel in 0.5 M HCl solution, and Fig. 3b shows the corresponding rescaled range analysis. The proportionality constant, called the Hurst exponent, H, is a measure of the memory effect of the system, which can be related with the corrosion form w20x. The Hurst exponent, H, was seen to be between 0 and 1 w18x. The value Hs 0.5 has a special significance, because this reflects that the observations are statistically independent of each other. When these values are between 0-H-0.5, called antipersistence zone, they represent a phenomena with a short memory effect; if 0.5-H-1, called persistence zone, then represents a phenomena with large memory effect w20,23,24x. 2. Experimental The experimental cell consisted of four identical carbon steel electrodes of 1 cm2 active surface area each one, as shown in Fig. 4. Two of these electrodes, W1 and W2, respectively, were used as working electrodes for the electrochemical current noise (ECN) measurement. Another electrode, R, was used as a reference electrode for electrochemical potential noise (EPN) measurements, with respect to W1. Impedance measurements were carried out in a threeelectrode arrangement, being W1, R and C, the working, reference and counter-electrode, respectively. The same experimental arrangement was used to obtain the polarization curves at the end of the experiment. Samples of mild steel (SAE 1810) were used as working electrodes. All specimens were wet polished down to 600 grade with SiC paper just before each test, degreased with acetone, rinsed in water, then in distilled water. As electrolyte a 0.5 M HCl solution was used with different concentrations of triazole. All these substances are 99% pure and were provided by Fluka. Oxygen was removed from the solutions by bubbling with nitrogen gas during 45 min, after which electrochemical test initiated. Tests were done at room temperature and under static conditions. Potentiodynamic polarization curves were obtained by polarizing the working electrode from y300 to q300

mV with respect to the corrosion potential at a scanning rate of 1 mV sy1. This value allows to obtain the quasistationary state measurements. The experimental set-up used for EIS was a classical three-electrode arrangement. The main electronic equipment used was an electrochemical interface Solartron model 1287 coupled to a frequency response analyzer (FRA) model 1255B, also from Solartron, controlled by a PC through the software CorrWare and Zplot, both from Schribner. The employed procedure was as follows: the mounted samples having been immersed in the solution, after 5 min the impedance was measured from an initial frequency of 100 000 Hz to a final frequency of 10 mHz, an amplitude of 10 mV (rms) from the Ecorr being used. 3. Experimental results and discussion Polarization curves of carbon steel obtained after 6 h of immersion in 0.5 M HCl with different ATR concentrations are shown in Fig. 5. The triazole effectively reduced the corrosion current density (corrosion rate)

Fig. 6. Nyquist plots for mild steel in 0.5 M HCl solution with different concentrations of 3-amino-1,2,4-triazole.

326

´ E. Garcıa-Ochoa, J. Genesca / Surface and Coatings Technology 184 (2004) 322–330

Fig. 7. Variation of polarization resistance, Rp, with time for mild steel in 0.5 M HCl for various concentrations of 3-amino-1,2,4-triazole.

Fig. 8. Potential time records for mild steel in 0.5 M HCl with different concentrations of 3-amino-1,2,4-triazole. (a) 0.0 mM, (b) 1.0 mM, (c) 1.5 mM and (d) 2 mM.

´ E. Garcıa-Ochoa, J. Genesca / Surface and Coatings Technology 184 (2004) 322–330

327

Fig. 9. Current time records for mild steel in 0.5 M HCl with different concentrations of 3-amino-1,2,4-triazole. (a) 0.0 mM, (b) 1.0 mM, (c) 1.5 mM and (d) 2 mM.

and this inhibition effect is significantly affected by the increase in the inhibitor concentration. Anodic and cathodic branch show a clear tendency to increase its polarization. This behavior is typical of a mixed-control inhibitor, where the inhibitor controls both anodic and cathodic reactions. Corrosion potential, Ecorr, corrosion current density (corrosion rate), icorr as well as corrosion inhibition efficiencies have been calculated from the obtained potentiodynamic polarization curves. These experimental results are showed in Table 1. Ecorr shows a clear tendency to be more negative as the inhibitor concentration increases and correspondingly the icorr values are significantly lower, in good agreement with a mixedcontrol mechanism. Nyquist impedance plots for the same experimental conditions are shown in Fig. 6. For the four studied concentrations of ATR a single semicircle was obtained. The impedance spectra were similar, exhibiting in all the cases a depressed semicircle with only one time constant. This impedance spectrum can be modeled as an electric equivalent circuit with a parallel combination of a CPE and a polarization resistance, Rp, in series with the solution resistance, Rs (Fig. 1), as has been reported previously for carbon steel studies in acid media

w25–27x. With the addition of triazole inhibitor, it is clearly seen that the diameter of the semicircle increases with increasing inhibitor concentration. It can be remembered that for a typically Randles equivalent circuit the value of Rp can be obtained by extrapolation of the Nyquist impedance plot at the low frequency region and corresponds to the diameter of the semicircle. The impedance plots obtained at the corrosion potential, Ecorr having the shape of depressed semicircles, with the center of the circle below the real axis, can be modeled by the EC proposed in Fig. 1, with a CPE representing the surface roughness and being the measured depression angle used to calculate the fractal dimension of the metal surface, as can be seen further. Fig. 7 resume the Rp values obtained from EIS experiments as a function of immersion time and inhibitor concentration. Greater the value of Rp, better is the inhibiting power of the inhibitor. A monotonous dependence of Rp with triazole concentration can be observed. Time records of voltage and current for each triazole concentration are shown in Figs. 8 and 9. While the result of inhibitor is manifested in both time records, it is more notable in the current noise signal, Fig. 9, where the signal amplitude is affected remarkably, decreasing as the inhibitor concentration increase. The combined

328

´ E. Garcıa-Ochoa, J. Genesca / Surface and Coatings Technology 184 (2004) 322–330

Fig. 10. Variation of electrochemical noise resistance, Rn , with time for mild steel in 0.5 M HCl for various concentrations of 3-amino-1,2,4triazole.

effect on the amplitude of voltage and current noise time records determine the noise resistance, Rn, value, which is inversely proportional to Icorr. As previously made with the values of Rp obtained from EIS, an estimation of the corrosion rate can be completed through the variation of Rn values with time, as shown in Fig. 10. The behavior is similar to the obtained for Rp values, Fig. 7. Fractal geometry of metal surface as a function of time and inhibitor concentration is presented in Table 2. The fractal dimension w19x describes the structure of a fractal, e.g. the irregularity of an EN time record. It can be noted that this fractal dimension is sensitive to inhibitor concentration, modifying its value with the time. The higher values correspond in all the cases at the beginning of the experiment, 0 h time, after that the metal surface has been polished and immersed into the fresh solution. Without inhibitor, the fractal dimension value decreases as a function of immersion time by the effect of the corrosion process itself. The effect of

inhibitor could be clearly seen for the higher concentration used, 2.0 mM. During the first hour of experiment, the fractal dimension changes from 2.53 to 2.32, while without inhibitor the values are 2.54 to 2.28 Fig. 11 shows a comparison between the values of the fractal dimension calculated from the depression angle of impedance plots and the Hurst exponent, H, determined from EN voltage and current time records. A clear relationship between these two parameters can be observed, then showing that the morphology of the corrosion attack can be followed by means of EN. 4. Conclusions 3-amino-1,2,4-triazole (ATR) has been found to inhibit corrosion of mild steel in 0.5 M HCl solution by polarizing both anodic and cathodic reactions, which corresponds with a mixed-control inhibition mechanism. The processes on the interface mild steely0.5 M HCl are described by a simple equivalent circuit including a polarization resistance, Rp, a parallel double-layer capac-

Table 2 Fractal geometry of metal surface as a function of time and inhibitor concentration Time, h

wATRx, 0 mM

wATRx, 1.0 mM

wATRx, 1.5 mM

wATRx, 2.0 mM

0 1 2 3 4 5 6

2.5466 2.2864 2.2379 2.2215 2.2134 2.2054 2.2006

2.3181 2.2261 2.2261 2.2269 2.2206 2.2163 2.2165

2.3351 2.2513 2.2183 2.2122 2.2062 2.1988 2.2022

2.5345 2.3245 2.2524 2.2529 2.2522 2.2400 2.2405

´ E. Garcıa-Ochoa, J. Genesca / Surface and Coatings Technology 184 (2004) 322–330

329

Fig. 11. Fractal dimension of the surface, Ds , and Hurst exponent, H, for mild steel in 0.5 M HCl with different concentrations of 3-amino-1,2,4triazole. (a) 0.0 mM, (b) 1.0 mM, (c) 1.5 mM and (d) 2 mM.

itance, Cdl which is distributed and modeled by a CPE and an ohmic resistance, Rs in series with the other elements. The intensity of the corrosive attack clearly decreases with the addition of ATR to the solution, modifying the metal surface morphology. The Hurst exponent, H, obtained from the time records of voltage and current EN seems to be a promising tool for the analysis of corrosion morphology because of their direct relationship with the fractal dimension of the metal surface. Acknowledgments The authors wish to thank Instituto Mexicano del Petroleo, IMP, for their financial support (FIES 98-01VI) and to the technician Pedro Rebollar for carrying out the experiments. References w1x V.S. Sastri, Corrosion Inhibitors: Principles and Applications, John Wiley and Sons, New York, 1998, p. 25. w2x M.A. Quraishi, S. Ahmad, M.Q. Ansari, Br. Corros. J. 32 (4) (1997) 297.

w3x B. Mernari, H.El Attari, M. Traisnel, F. Bentiss, M. Lagrenee, ´ Corros. Sci. 40 (2–3) (1998) 391. w4x F. Bentiss, M. Traisnel, H. Vezin, M. Lagrenee, ´ Ind. Eng. Chem. Res. 39 (2000) 3732. w5x M.A. Quraishi, R. Sardar, Corrosion 58 (9) (2002) 748. w6x R. Cottis, S. Turgoose, Electrochemical Impedance and Noise, NACE International, Houston, 1999. w7x W.H. Mulder, J.H Sluyters, Electrochim. Acta 33 (1988) 303. w8x G.J Brug, A.L.G. Van Den Eeden, M. Sluyters-Rehbach, J.H. Sluyters, J. Electroanal. Chem. 176 (1984) 275. w9x R. de Levie, J. Electroanal. Chem. 261 (1989) 1. w10x A. Popova, S. Raicheva, E. Sokolova, M. Christov, Langmuir 12 (1996) 2083. w11x L. Nykos, T. Pajkossy, Electrochim. Acta 30 (11) (1985) 1533. w12x W. H. Mulder, et al., J. Electroanal. Chem. 285 (1990) 103. w13x G.A. McRae, M.A. Maguire, C.A. Jeffrey, D.A. Guzonas, C.A Brown, Appl. Surf. Sci. 191 (2002) 94. w14x C. Monticelli, G. Brunoro, A. Frignani, G. Trabanelli, J. Electrochem. Soc. 139 (3) (1992) 706. w15x Y.J. Tan, S. Bailey, B. Kinsella, Corros. Sci. 38 (10) (1996) 1681. w16x A. Aballe, M. Bethencourt, F.J. Botana, M. Marcos, R.M. Osuna, Electrochim. Acta 47 (2002) 1415. w17x R.A. Cottis, Corrosion 57 (3) (2001) 265. ´ Horvath, w18x A. ´ R. Schiller, Corros. Sci. 45 (2003) 597. w19x M. Moon, B. Skerry, J. Coat. Technol. 67 (843) (1995) 35. w20x M. Saitou, J. Electrochem. Soc. 147 (6) (2000) 2133. w21x P.R. Roberge, D. Lenard, J. Appl. Electrochem. 28 (1998) 405.

330

´ E. Garcıa-Ochoa, J. Genesca / Surface and Coatings Technology 184 (2004) 322–330

w22x J. Cruz, E. Garcia-Ochoa, M. Castro, J. Electrochem. Soc. 150 (1) (2002) B26. w23x J. Feder, Fractals, Plenum Press, New York, 1989, p. 149. w24x Z.H. Gu, T.Z. Fahidy, J.P. Chopart, Electrochim. Acta 37 (1) (1992) 97.

w25x L. Elkadi, B. Mernari, M. Traisnel, F. Bentiss, M. Lagrenee, ´ Corros. Sci. 42 (2000) 703. w26x F. Bentiss, M. Traisnel, M. Lagrenee, ´ J. Appl. Electrochem. 31 (2001) 41. w27x K.F. Khaled, Electrochim. Acta 48 (2003) 2493.