1.3 Corrosion rate determination by electrochemical noise analysis (ENA)

Many of the electrochemical techniques, among those described earlier, measure the electrochemical response of the corrosion system following the application of an external disturbance. In the last 50 decades, an original concept has emerged where it was possible to use the inherent noise of the electrochemical system as a stimulus to measure both potential and current changes [31, 32, 37–43]. Broadly, measured inconsistently in corrosion experiments, the electrochemical noise was first considered an unwanted or undesirable artifact that comes from measuring instruments or pickups from the environment. This is why this misleading name was cast. This sort of noise can be easily observed during corrosion potential measurements because the measured corrosion potential always fluctuates slightly, usually randomly. Random fluctuations result from stochastic processes [44], and, considering each chemical process is stochastic in nature, it generates noise.

Since the pioneer work of Iverson [45], who has reported a relation between the frequency or amplitude of the electrochemical noise and the inhibiting power of the environment for a number of metals and alloys (e.g., aluminum alloys, magnesium, mild steel, etc.), there has been a growing interest toward the measurement of electrochemical noise and its peculiar relationship with localized corrosion [11, 12, 31, 32, 41, 46–51]. In this respect, electrochemical noise measurements obtained from the analysis of corrosion potential or current fluctuations provide a new approach to the study of corrosion processes in reactive environments such as aqueous media or hot aggressive gases or even under the effect of mechanical stimuli, e.g., tribocorrosion. Indeed, mechanical friction of solids in contact with a corrosive environment is likely to generate (i) noise due to stochastic contact between randomly distributed surface asperities and (ii) noise due to the synergy of wear-corrosion processes resulting from the activity of the surfaces and controlled by the response of potential-current transients and the configuration of the wear

#### Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods… DOI: http://dx.doi.org/10.5772/intechopen.86743

track area, coordinated by the coupling effects of wear and corrosion in the triboelectrochemical cell. Among the possibilities offered by the measurement of electrochemical noise sources during an electrochemical or a tribo-electrochemical system, the following can be retained: adsorption–desorption processes, e.g., formation and detachment of gas bubbles; fluctuations in the mass transport rate and in temperature; interfacial nucleation and growth processes; degradation processes due dielectric film disruption; kinetics of atom exchange at the surface sites, e.g., Johnson's noise in the interfacial impedance; and so on.

While multiple case studies on electrochemical noise have been regularly reported in recent years, even greater progress is possible, with the scope for increased breakthrough in science and technology (e.g., novel materials, precision tools on macro-to-nanoscopic scales, availability and intelligent use of these materials and tools, and so on). In particular, the main focus of these investigations is to promptly obtain in situ mechanistic information on the repassivation and breakdown of passive films and to monitor any process associated with confined corrosion and/or tribo- or bio-tribocorrosion [46, 47, 51]. It has been, indeed, suggested that the noise is caused by film breakdown and repassivation processes, and given the dynamic competition between these two processes, pitting will initiate. However, the foundation for using electrochemical noise analysis for determining the corrosion rate of an electrode is still a subject of debate within the scientific community. Indeed, the fundamental approach is not as robust as that of other techniques. On the other hand, the advantage of the noise analysis is that it is not necessary to apply any external polarization and the system is in natural corrosion conditions. This renders the technique as nondestructive and nonintrusive, capable of monitoring basic changes in an electrochemically active system. This makes it particularly suitable for online corrosion monitoring in the laboratory, especially for localized corrosion monitoring, detection of general corrosion, crevice investigation, stress corrosion cracking [12, 52, 53], fretting corrosion, or be used in the assessment of anti-corrosive organic coatings, and other surface inhomogeneity case studies [43, 46, 47]. Several approaches extend the use of electrochemical noise measurements in both pilot plant and field facilities, its use is not merely limited to the foregoing phenomena, but its development is justified especially when measurements are performed in systems with very low conductivity, where, for e.g., the impedance technique fails because of the loss of signal in the high resistance of the solution (cf. infra).

#### 1.3.1 Instrumentation for electrochemical noise measurements in corrosion and tribocorrosion systems

Electrochemical noise is a generic term used to describe the naturally occurring fluctuations in potential and current, which is due to spontaneous changes in electrode kinetics and mechanisms [33]. When applied to corrosion studies, electrochemical noise may be redefined as the spontaneous fluctuations observed in potential and current at the free corrosion potential. The electrochemical noise can thus be classified into potential noise and current noise. There are three major possible modes for measuring potential and current noise in a corrosion system, but the most common mode uses two nominally identical working electrodes, WE1 and WE2 (WE1 as the corroding metal and WE2 as a counter electrode), and a noise-free reference noble electrode, RE [33] (see Figure 4a). The current flowing between the two working electrodes is measured by a zero-resistance ammeter (ZRA), and their potential is monitored versus the reference electrode through a voltmeter (V) under free corrosion conditions. The two other leftover modes are two identical working electrodes WE1 and WE2 with a bias potential [54] (not shown here) and one WE coupled to a micro-counter electrode (MC, e.g., Pt wire tip) [29, 46, 47, 55]

experiment is started at the corrosion potential, and the cathodic polarization is first conducted by applying an overpotential of approximately 500 mV or until gas evolution (e.g., hydrogen) occurs at the electrode, at a constant rate of 0.6 V/h. Following, the corrosion potential is measured again (typically for 1 h), and then anodic polarization is conducted by applying an overpotential so that the potential at the end of the anodic polarization reaches +1.6 V versus SCE. Tafel plots are generated by plotting both anodic and cathodic data in a semilog paper as E-log I. From the plot, three values are determined: the anodic Tafel slope, the cathodic Tafel slope, and Icorr (from back-extrapolation of both anodic and cathode curves to Ecorr). The main advantage of this method is that it provides a simple, straightfor-

The disadvantage of the Tafel technique is that large current densities are often required to generate the complete Tafel plots. The use of large current densities can alter the surface conditions of the specimen (e.g., permanent change or surface damage), thereby distorting the results and increasing complications due to mass transport and uncompensated electrolyte resistance. The measurement of current density over a wide potential range may also distort the results if the adsorption of some species is potential dependent. Since this method applies a large overpotential to the metal surface (e.g., anodic polarization), therefore, the technique is rather destructive and can hardly be used for online corrosion monitoring purposes and in particular in the field. An ASTM G5 standard provides a procedure for constructing an anodic polarization plot [36]. However, it does not supply a method to construct a cathodic polarization plot nor a procedure to determine the corrosion current by

1.3 Corrosion rate determination by electrochemical noise analysis (ENA)

chemical process is stochastic in nature, it generates noise.

Many of the electrochemical techniques, among those described earlier, measure the electrochemical response of the corrosion system following the application of an external disturbance. In the last 50 decades, an original concept has emerged where it was possible to use the inherent noise of the electrochemical system as a stimulus to measure both potential and current changes [31, 32, 37–43]. Broadly, measured inconsistently in corrosion experiments, the electrochemical noise was first considered an unwanted or undesirable artifact that comes from measuring instruments or pickups from the environment. This is why this misleading name was cast. This sort of noise can be easily observed during corrosion potential measurements because the measured corrosion potential always fluctuates slightly, usually randomly. Random fluctuations result from stochastic processes [44], and, considering each

Since the pioneer work of Iverson [45], who has reported a relation between the frequency or amplitude of the electrochemical noise and the inhibiting power of the environment for a number of metals and alloys (e.g., aluminum alloys, magnesium, mild steel, etc.), there has been a growing interest toward the measurement of electrochemical noise and its peculiar relationship with localized corrosion [11, 12, 31, 32, 41, 46–51]. In this respect, electrochemical noise measurements obtained from the analysis of corrosion potential or current fluctuations provide a new approach to the study of corrosion processes in reactive environments such as aqueous media or hot aggressive gases or even under the effect of mechanical stimuli, e.g., tribocorrosion. Indeed, mechanical friction of solids in contact with a corrosive environment is likely to generate (i) noise due to stochastic contact between randomly distributed surface asperities and (ii) noise due to the synergy of wear-corrosion processes resulting from the activity of the surfaces and controlled by the response of potential-current transients and the configuration of the wear

ward method to determine Tafel parameters, namely, ba and bc.

the Tafel extrapolation method.

Corrosion Inhibitors

174

RN <sup>¼</sup> <sup>σ</sup>Eð Þ<sup>t</sup>

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods…

DOI: http://dx.doi.org/10.5772/intechopen.86743

Eq. (11) implies that in the case where a low-driving force noise produces a high current density noise between the two electrodes (WE<sup>1</sup> and WE2), the yielding noise resistance will be low. Noise resistance, RN, has been shown to correlate well with the polarization resistance, RP, as determined by EIS for certain corrosion systems. This latter being directly related to the corrosion current [14, 29, 60] is using the Stern-Geary equation and Tafel slopes. Notwithstanding, much work has been devoted trying to best match RN or the normalized RN (per unit of exposed surface area) to the corrosion resistance or the corrosion rate [5, 14, 31, 32, 37–45, 48–50, 60]. Although signal analysis in the time domain is well established, an approach based on spectral analysis is gaining more and more importance in research laboratories. It consists of transforming the potential and current noise fluctuations recorded in the frequency

The frequency range for which the FFT is commonly performed extends from 1 mHz up to 1 Hz. The spectral noise plots are similar to those of impedance plots. The spectral noise resistance, RSN, is given by the ratio of the potential and current

RSNð Þ¼ <sup>f</sup> EFFTð Þ<sup>f</sup>

The log–log plot of RSN versus f is similar to the impedance plot, and the spectral

Another approach would be to examine the spectral noise response in terms of power spectral densities (PSD). These latter are calculated from the FFT or using the maximum entropy method (MEM) [62]. RSN is determined from the PSDs by

> IPSDð Þf <sup>1</sup>=<sup>2</sup>

It has been shown that the use of a single data set of potential and current noise [32] would yield identical values of RSN as calculated by either Eqs. (13) or (14). In

effective if the impedance of the two electrodes is identical and much higher than the resistance of the test solution between them [5, 32, 63]. Experiments have validated this relationship for several systems [5, 31, 63]. Nonetheless, there is no agreement on the fundamental basis for the relationship between noise resistance

1.3.4 Illustrative examples of the application of electrochemical noise in tribocorrosion

Investigations into electrochemical kinetics make common point research between tribocorrosion and corrosion. The study of localized phenomena of

RSNð Þ¼ <sup>f</sup> EPSDð Þ<sup>f</sup>

IFFTð Þf

0

domain using the Fast Fourier Transform (FFT) method [61].

<sup>0</sup> is given by

<sup>0</sup> is bound to RN or RP as,

R0 SN ¼ lim f!0

FFTs at each frequency, and the limiting value, RSN

the corrosion resistance:

noise resistance limit RSN

the relation (14):

some cases, RSN

and corrosion rate.

systems

177

<sup>σ</sup>Ið Þ<sup>t</sup> (11)

, can be used as a measure of

(14)

(12)

RSNð Þf (13)

RN ¼ RSNð Þ¼ 0 RP (15)

#### Figure 4.

(a) Schematic view of a tribocorrosion experimental setup. Potential and current are measured on a working electrode (WE1) sliding against a counterbody ball (unidirectional reciprocating sliding, sphere-on-flat) with respect to a RE reference electrode (Ag/AgCl (3 M KCl)) via a V, high-impedance voltmeter, and CE (WE2) via a zero-resistance ammeter (ZRA), respectively. FN, normal force; FT, tangential force; f, sliding frequency; D, displacement amplitude. (b) Schematic view of a tribocorrosion experimental setup. Potential and current are measured on a working electrode (WE) sliding against a counterbody ball (unidirectional reciprocating sliding, sphere-on-flat) with respect to a RE reference electrode (Ag/AgCl (3 M KCl)) via a V, high-impedance voltmeter, and CE micro-cathode (MC) via a zero-resistance ammeter (ZRA), respectively. FN, normal force; FT, tangential force; f, sliding frequency; D, displacement amplitude.

(see Figure 4b). This last mode of electrochemical noise analysis seems to be a promising way to obtain unambiguous estimates of the rate of chemical wear in a tribocorrosion experiment as evidenced by some recent investigations [46, 47, 51] but also to predict the corrosion rate of localized corrosion in a free corroding system [29, 31, 32, 37–45, 47–50, 56].

#### 1.3.2 Electrochemical noise data management process

The overall approach to analyzing noise data is the assessment of mechanistic information from either time-domain analysis, frequency-domain analysis, or both, using statistical methods [44, 57, 58]. If the information in the time-domain records is evident, time-domain analysis is sufficient to distinguish different processes (e.g., different forms of corrosion).

In what follows, one assumes that all various types of noise are excluded from this description, with the exception of the thermal noise. Except for the last noise, all other noise sources can be minimized or eliminated using careful strategy within reasonable limits of materiality. Effective and convenient ways include the removal of unwanted environmental and instrumental noise from the electrochemical noise, e.g., by shielding electrical connections/wires for coupling the electrodes to the experimental apparatus, by using a Faraday cage to exclude electrostatic/electromagnetic influences, even by implementing analogue/digital filters to eliminate systematic noise at frequencies different than the frequency of interest, and so on. Guidelines for the calibration of noise measuring device can be found elsewhere [13, 59].

#### 1.3.3 Noise resistance

The basic quantitative approach is the time-domain analysis of the noise signal. The noise resistance, RN, is defined as the ratio of the standard deviations of potential/current noise signal time dependent, σ(t):

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods… DOI: http://dx.doi.org/10.5772/intechopen.86743

$$R\_N = \frac{\sigma\_E(t)}{\sigma\_I(t)}\tag{11}$$

Eq. (11) implies that in the case where a low-driving force noise produces a high current density noise between the two electrodes (WE<sup>1</sup> and WE2), the yielding noise resistance will be low. Noise resistance, RN, has been shown to correlate well with the polarization resistance, RP, as determined by EIS for certain corrosion systems. This latter being directly related to the corrosion current [14, 29, 60] is using the Stern-Geary equation and Tafel slopes. Notwithstanding, much work has been devoted trying to best match RN or the normalized RN (per unit of exposed surface area) to the corrosion resistance or the corrosion rate [5, 14, 31, 32, 37–45, 48–50, 60]. Although signal analysis in the time domain is well established, an approach based on spectral analysis is gaining more and more importance in research laboratories. It consists of transforming the potential and current noise fluctuations recorded in the frequency domain using the Fast Fourier Transform (FFT) method [61].

The frequency range for which the FFT is commonly performed extends from 1 mHz up to 1 Hz. The spectral noise plots are similar to those of impedance plots. The spectral noise resistance, RSN, is given by the ratio of the potential and current FFTs at each frequency, and the limiting value, RSN 0 , can be used as a measure of the corrosion resistance:

$$R\_{\rm SN}(f) = \left(\frac{E\_{FFT}(f)}{I\_{FFT}(f)}\right) \tag{12}$$

The log–log plot of RSN versus f is similar to the impedance plot, and the spectral noise resistance limit RSN <sup>0</sup> is given by

$$R\_{\text{SN}}^0 = \lim\_{f \to 0} R\_{\text{SN}}(f) \tag{13}$$

Another approach would be to examine the spectral noise response in terms of power spectral densities (PSD). These latter are calculated from the FFT or using the maximum entropy method (MEM) [62]. RSN is determined from the PSDs by the relation (14):

$$R\_{\rm SN}(f) = \left(\frac{E\_{\rm PSD}(f)}{I\_{\rm PSD}(f)}\right)^{1/2} \tag{14}$$

It has been shown that the use of a single data set of potential and current noise [32] would yield identical values of RSN as calculated by either Eqs. (13) or (14). In some cases, RSN <sup>0</sup> is bound to RN or RP as,

$$R\_N = R\_{\rm SN}(\mathbf{0}) = R\_P \tag{15}$$

effective if the impedance of the two electrodes is identical and much higher than the resistance of the test solution between them [5, 32, 63]. Experiments have validated this relationship for several systems [5, 31, 63]. Nonetheless, there is no agreement on the fundamental basis for the relationship between noise resistance and corrosion rate.

#### 1.3.4 Illustrative examples of the application of electrochemical noise in tribocorrosion systems

Investigations into electrochemical kinetics make common point research between tribocorrosion and corrosion. The study of localized phenomena of

(see Figure 4b). This last mode of electrochemical noise analysis seems to be a promising way to obtain unambiguous estimates of the rate of chemical wear in a tribocorrosion experiment as evidenced by some recent investigations [46, 47, 51] but also to predict the corrosion rate of localized corrosion in a free corroding

(a) Schematic view of a tribocorrosion experimental setup. Potential and current are measured on a working electrode (WE1) sliding against a counterbody ball (unidirectional reciprocating sliding, sphere-on-flat) with respect to a RE reference electrode (Ag/AgCl (3 M KCl)) via a V, high-impedance voltmeter, and CE (WE2) via a zero-resistance ammeter (ZRA), respectively. FN, normal force; FT, tangential force; f, sliding frequency; D, displacement amplitude. (b) Schematic view of a tribocorrosion experimental setup. Potential and current are measured on a working electrode (WE) sliding against a counterbody ball (unidirectional reciprocating sliding, sphere-on-flat) with respect to a RE reference electrode (Ag/AgCl (3 M KCl)) via a V, high-impedance voltmeter, and CE micro-cathode (MC) via a zero-resistance ammeter (ZRA), respectively. FN, normal force;

The overall approach to analyzing noise data is the assessment of mechanistic information from either time-domain analysis, frequency-domain analysis, or both, using statistical methods [44, 57, 58]. If the information in the time-domain records is evident, time-domain analysis is sufficient to distinguish different processes (e.g.,

In what follows, one assumes that all various types of noise are excluded from this description, with the exception of the thermal noise. Except for the last noise, all other noise sources can be minimized or eliminated using careful strategy within reasonable limits of materiality. Effective and convenient ways include the removal of unwanted environmental and instrumental noise from the electrochemical noise, e.g., by shielding electrical connections/wires for coupling the electrodes to the experimental apparatus, by using a Faraday cage to exclude electrostatic/electromagnetic influences, even by implementing analogue/digital filters to eliminate systematic noise at frequencies different than the frequency of interest, and so on. Guidelines for the calibration of noise measuring device can be found elsewhere

The basic quantitative approach is the time-domain analysis of the noise signal.

The noise resistance, RN, is defined as the ratio of the standard deviations of

potential/current noise signal time dependent, σ(t):

system [29, 31, 32, 37–45, 47–50, 56].

different forms of corrosion).

[13, 59].

176

Figure 4.

Corrosion Inhibitors

1.3.3 Noise resistance

1.3.2 Electrochemical noise data management process

FT, tangential force; f, sliding frequency; D, displacement amplitude.

depassivation and repassivation is essential to understand the mechanisms of corrosion-wear as well as to reduce the material loss. The possibility of using the electrochemical noise detection technique as a promising tool to study the electrochemical properties of well-controlled damaged surfaces has been widely considered due to its nondestructive nature and its potential in online corrosion monitoring applications. Time-spatially resolved measurements should provide more reliable data on the electrochemical part of tribocorrosion. The noise analysis in relation to depassivation-repassivation events randomly distributed in time and space can be traced back to Oltra et al. [64]. The power spectral density (PSD) of the noise under the impact of the jet particles was related to the Fourier transform of individual repassivation transients obeying a Poisson distribution. Later, the application of electrochemical noise analysis to tribocorrosion was reviewed. Investigations involving PSD noise analyses on various tribo-electrochemical cells for passivating materials were conducted by Ponthiaux et al. [65], by Déforge et al. [51], and in more details by Berradja et al. [46, 66]. In this latter work, the noise spectra were measured on AISI 304 L stainless steel versus corundum in a Ringer's solution in a pin-on-disk tribometer under stationary sliding-corrosion regime conditions, either at open-circuit potential or at a controlled potential. The PSD of the current noise has been interpreted as resulting from the overlap of the large number of discrete repassivation transients at the contact junctions, including the double-layer charge and the strong dependence of depassivation and repassivation kinetic rates of the oxide surface film on the sliding frequency. This was consistent with the shift in the PSD plots of the current noise fluctuations by about a decade when the sliding frequency was varied from 0.1 to 1 Hz (see Figure 5). Similar findings were obtained via Déforge et al. [51] by dividing open-circuit potential fluctuations to the impedance of the electrode. A 1/f low-frequency noise is explained by a longterm drift of the surface conditions. Only a minor influence of the applied normal load was observed on the PSD plots, recommending reaching the limit rate of the depassivation of the oxide surface film.

Application of the noise analysis to tribocorrosion offers the feasibility to record in parallel the PSD of normal and tangential force fluctuations and their tie-in with the current noise data (see Figure 6). Force fluctuations show an almost flat

spectrum (white noise) as expected following short random mechanical interactions between colliding asperities, whereas the current noise is consistent with finite

1.4 Corrosion forecast by electrochemical impedance spectroscopy (EIS)

PSD record of tangential (gray) and normal (black) force components measured on AISI 304L during continuous sliding-corrosion test in Ringer's solutions at a normal load of 20 N and at 1 Hz frequency.

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods…

DOI: http://dx.doi.org/10.5772/intechopen.86743

The EIS has matured greatly over the past 25 years as a tool in corrosion protection research and has proven to be one of the most useful electrochemical characterization techniques presently available. In practice, EIS has become a standardized research tool for corrosion prediction [15] and found wide applications in both fundamental and applied laboratory researches [67]. Recent applications in tribocorrosion reflect the steady progress of the EIS method in terms of research and development [62]. Compared with the LPR technique, the EIS technique is considered more advanced, since it has the ability to study high-impedance systems, in which the conventional LPR technique has failed, such as coatings and linings [16, 68], high pure water, and organic coating/metal systems [69] or corrosion in a low conductive solution [70]. This technique is especially useful for

time-constant transient responses to the depassivation events.

Reproduced with permission from IOPScience [46].

Figure 6.

179

#### Figure 5.

PSD record of current fluctuations measured on AISI 304L during continuous sliding-corrosion test in Ringer's solution at 0.1 Hz (gray) and 1 Hz (black) and at a constant normal load of 5 N. Reproduced with permission from IOPScience [46].

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods… DOI: http://dx.doi.org/10.5772/intechopen.86743

Figure 6.

depassivation and repassivation is essential to understand the mechanisms of corrosion-wear as well as to reduce the material loss. The possibility of using the electrochemical noise detection technique as a promising tool to study the electrochemical properties of well-controlled damaged surfaces has been widely considered due to its nondestructive nature and its potential in online corrosion monitoring applications. Time-spatially resolved measurements should provide more reliable data on the electrochemical part of tribocorrosion. The noise analysis in relation to depassivation-repassivation events randomly distributed in time and space can be traced back to Oltra et al. [64]. The power spectral density (PSD) of the noise under the impact of the jet particles was related to the Fourier transform of individual repassivation transients obeying a Poisson distribution. Later, the application of electrochemical noise analysis to tribocorrosion was reviewed. Investigations involving PSD noise analyses on various tribo-electrochemical cells for passivating materials were conducted by Ponthiaux et al. [65], by Déforge et al. [51], and in more details by Berradja et al. [46, 66]. In this latter work, the noise spectra were measured on AISI 304 L stainless steel versus corundum in a Ringer's solution in a pin-on-disk tribometer under stationary sliding-corrosion regime conditions, either at open-circuit potential or at a controlled potential. The PSD of the current noise has been interpreted as resulting from the overlap of the large number of discrete repassivation transients at the contact junctions, including the double-layer charge and the strong dependence of depassivation and repassivation kinetic rates of the oxide surface film on the sliding frequency. This was consistent with the shift in the PSD plots of the current noise fluctuations by about a decade when the sliding frequency was varied from 0.1 to 1 Hz (see Figure 5). Similar findings were obtained via Déforge et al. [51] by dividing open-circuit potential fluctuations to the impedance of the electrode. A 1/f low-frequency noise is explained by a longterm drift of the surface conditions. Only a minor influence of the applied normal load was observed on the PSD plots, recommending reaching the limit rate of the

Application of the noise analysis to tribocorrosion offers the feasibility to record in parallel the PSD of normal and tangential force fluctuations and their tie-in with the current noise data (see Figure 6). Force fluctuations show an almost flat

PSD record of current fluctuations measured on AISI 304L during continuous sliding-corrosion test in Ringer's solution at 0.1 Hz (gray) and 1 Hz (black) and at a constant normal load of 5 N. Reproduced with permission

depassivation of the oxide surface film.

Corrosion Inhibitors

Figure 5.

178

from IOPScience [46].

PSD record of tangential (gray) and normal (black) force components measured on AISI 304L during continuous sliding-corrosion test in Ringer's solutions at a normal load of 20 N and at 1 Hz frequency. Reproduced with permission from IOPScience [46].

spectrum (white noise) as expected following short random mechanical interactions between colliding asperities, whereas the current noise is consistent with finite time-constant transient responses to the depassivation events.

#### 1.4 Corrosion forecast by electrochemical impedance spectroscopy (EIS)

The EIS has matured greatly over the past 25 years as a tool in corrosion protection research and has proven to be one of the most useful electrochemical characterization techniques presently available. In practice, EIS has become a standardized research tool for corrosion prediction [15] and found wide applications in both fundamental and applied laboratory researches [67]. Recent applications in tribocorrosion reflect the steady progress of the EIS method in terms of research and development [62]. Compared with the LPR technique, the EIS technique is considered more advanced, since it has the ability to study high-impedance systems, in which the conventional LPR technique has failed, such as coatings and linings [16, 68], high pure water, and organic coating/metal systems [69] or corrosion in a low conductive solution [70]. This technique is especially useful for

evaluating corrosion inhibitors [24, 71], analyzing the corrosion mechanisms [72, 73], and so on.

A significant number of tutorials have been addressed on the EIS experimental setup, the measurement methodology, and data analysis methods [27, 74–78]. The technique has been of a great deal of concern to the extent that an ASTM standard, i.e., ASTM G106-89, has been produced to provide the practitioner with a test method to verify that the electronic equipment, the electrochemical cell, and the spectrum generation algorithm impedance work properly [15].

### 1.4.1 Principle of the EIS technique

The EIS technique normally uses a typical three-electrode cell system controlled by a potentiostat, similar to that used in the LPR technique. Unlike the previous time-resolved techniques, where the current system response is either the consequence of a large voltage perturbation from the steady-state condition (case of Tafel extrapolation) or from a smaller perturbation (case of LPR method), in the EIS approach, however, by applying a small varying perturbation over a range of frequency, it is possible to probe the full response of the electrochemical system, and not just the resistive components. In that respect, a small AC signal, i.e., alternating potential or voltage V(ω) typically a sine wave of amplitude �10 mV of the corrosion potential, is applied over a wide range of frequency (typically from 10<sup>5</sup> down to 10�<sup>2</sup> or 10�<sup>3</sup> Hz) at a number of discrete frequencies (typically 5–10 frequencies per decade), and the alternating current response, i(ω), is measured at each frequency, ω (i.e., the ac polarization or angular frequency, ω = 2πf). For a linear system, the current response signal will be a sine wave of the same frequency as the excitation signal (voltage) but shifted in phase. This is transmitted to a frequency response analyzer or a lock-in amplifier to calculate the impedance and phase shift. Full frequency sweeps provide phase-shift information that can be used in combination with equivalent circuit models to gain valuable information from the complex interface of the corrosion system. The frequency-dependent impedance is determined by the relation: Zð Þ¼ ω Vð Þ ω =ið Þ ω .

and CDL represent labels for the solution resistor, the Faradaic charge transfer resistor, and the double-layer capacitance, respectively. The capacity is associated with the separation of charges at the electrode/electrolyte interface as in the case of a working electrode having a surface film (e.g., AISI 304 stainless steel immersed in a 0.5 M H2SO4 electrolyte), in which case the capacity of the equivalent circuit can be associated with the capacity of the passive oxide surface film and the resistor in parallel with the capacitor is considered as the charge transfer resistance, RCT (or the polarization resistance, RP, under EIS-free corrosion conditions), while the ohmic resistance in solution, RS, between the working electrode and the reference electrode is in series with the parallel resistor and the capacitor. If the amplitude of the perturbation signal is small enough (e.g., a voltage less than 20 mV), RCT can be

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods…

The behavior of such an electrochemical interface can be described by Eq. (16):

RCT or RP can be determined in several ways. A convenient way is to use the Nyquist diagram. For the simple Randles-type equivalent circuit as shown in Figure 7, the corresponding Nyquist diagram is displayed in Figure 8, in which a perfect semicircle is observed. The high-frequency response is used to determine the component of RS involved in the measurement. RS can be read directly from the abscissa when the angular frequency ω (ω = 2πf) tends to be infinite (fmax or f ! ∞). The total resistance (RP + RS) can also be read from the abscissa when ω approaches zero (fmin or f ! 0). So, RP can be determined by subtracting the RS value from the

Rp 1 þ ð Þ jωRPCDL

<sup>β</sup> (16)

regarded as equivalent to the linear polarization resistance (RP).

Figure 7.

Figure 8.

181

A simple Randles-type equivalent circuit (RC).

DOI: http://dx.doi.org/10.5772/intechopen.86743

Zð Þ¼ ω Rs þ

The Nyquist diagram responding to the simple Randels-type equivalent circuit.

#### 1.4.2 Electrode/electrolyte electrochemical interface circuit

Basically, the electrode/electrolyte interface is characterized by a separation of charges resulting in the creation of parallel planes of electrical charges whose behavior is similar to a circuit consisting of a capacitor and a resistor in parallel and certainly not to a perfect capacitor. Indeed, the current flowing in a perfect capacitor would cease when the latter would be fully charged, hence the need to add a resistor in parallel to let a weak current flow. An electrochemical interface can be viewed as an electrical circuit, or called the equivalent circuit, composed of a number of elements such as resistances (R), capacitances (C), and inductances (L) [26]. Explanations of the EIS results are usually based on the equivalent circuit used. Many software programs and packages are now available for fitting the impedance spectra to analogous circuits [15], a strategy often used to analyze data. Further information on the EIS measurements and instrumentation can be found elsewhere [15, 17, 79, 80].

Not all the available proposed equivalent circuits to model electrochemical interfaces can actually satisfy what is applied to a freely corroding system. In most cases, the impedance corresponding to a simple corrosion process, under activation control, can be represented by the well-known Randles' [81] equivalent circuit (RC circuit) which allows to describe the behavior of many electrochemical electrode/ electrolyte interfaces. A typical example is shown in Figure 7, where RS, RCT,

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods… DOI: http://dx.doi.org/10.5772/intechopen.86743

Figure 7. A simple Randles-type equivalent circuit (RC).

evaluating corrosion inhibitors [24, 71], analyzing the corrosion mechanisms

spectrum generation algorithm impedance work properly [15].

A significant number of tutorials have been addressed on the EIS experimental setup, the measurement methodology, and data analysis methods [27, 74–78]. The technique has been of a great deal of concern to the extent that an ASTM standard, i.e., ASTM G106-89, has been produced to provide the practitioner with a test method to verify that the electronic equipment, the electrochemical cell, and the

The EIS technique normally uses a typical three-electrode cell system controlled by a potentiostat, similar to that used in the LPR technique. Unlike the previous time-resolved techniques, where the current system response is either the consequence of a large voltage perturbation from the steady-state condition (case of Tafel extrapolation) or from a smaller perturbation (case of LPR method), in the EIS approach, however, by applying a small varying perturbation over a range of frequency, it is possible to probe the full response of the electrochemical system, and not just the resistive components. In that respect, a small AC signal, i.e., alternating potential or voltage V(ω) typically a sine wave of amplitude �10 mV of the corrosion potential, is applied over a wide range of frequency (typically from 10<sup>5</sup> down to 10�<sup>2</sup> or 10�<sup>3</sup> Hz) at a number of discrete frequencies (typically 5–10 frequencies per decade), and the alternating current response, i(ω), is measured at each frequency, ω (i.e., the ac polarization or angular frequency, ω = 2πf). For a linear system, the current response signal will be a sine wave of the same frequency as the excitation signal (voltage) but shifted in phase. This is transmitted to a frequency response analyzer or a lock-in amplifier to calculate the impedance and phase shift. Full frequency sweeps provide phase-shift information that can be used in combination with equivalent circuit models to gain valuable information from the complex interface of the corrosion system. The frequency-dependent impedance is deter-

Basically, the electrode/electrolyte interface is characterized by a separation of

charges resulting in the creation of parallel planes of electrical charges whose behavior is similar to a circuit consisting of a capacitor and a resistor in parallel and certainly not to a perfect capacitor. Indeed, the current flowing in a perfect capacitor would cease when the latter would be fully charged, hence the need to add a resistor in parallel to let a weak current flow. An electrochemical interface can be viewed as an electrical circuit, or called the equivalent circuit, composed of a number of elements such as resistances (R), capacitances (C), and inductances (L) [26]. Explanations of the EIS results are usually based on the equivalent circuit used. Many software programs and packages are now available for fitting the impedance spectra to analogous circuits [15], a strategy often used to analyze data. Further information on the EIS measurements and instrumentation can be found elsewhere

Not all the available proposed equivalent circuits to model electrochemical interfaces can actually satisfy what is applied to a freely corroding system. In most cases, the impedance corresponding to a simple corrosion process, under activation control, can be represented by the well-known Randles' [81] equivalent circuit (RC circuit) which allows to describe the behavior of many electrochemical electrode/ electrolyte interfaces. A typical example is shown in Figure 7, where RS, RCT,

[72, 73], and so on.

Corrosion Inhibitors

1.4.1 Principle of the EIS technique

mined by the relation: Zð Þ¼ ω Vð Þ ω =ið Þ ω .

[15, 17, 79, 80].

180

1.4.2 Electrode/electrolyte electrochemical interface circuit

and CDL represent labels for the solution resistor, the Faradaic charge transfer resistor, and the double-layer capacitance, respectively. The capacity is associated with the separation of charges at the electrode/electrolyte interface as in the case of a working electrode having a surface film (e.g., AISI 304 stainless steel immersed in a 0.5 M H2SO4 electrolyte), in which case the capacity of the equivalent circuit can be associated with the capacity of the passive oxide surface film and the resistor in parallel with the capacitor is considered as the charge transfer resistance, RCT (or the polarization resistance, RP, under EIS-free corrosion conditions), while the ohmic resistance in solution, RS, between the working electrode and the reference electrode is in series with the parallel resistor and the capacitor. If the amplitude of the perturbation signal is small enough (e.g., a voltage less than 20 mV), RCT can be regarded as equivalent to the linear polarization resistance (RP).

The behavior of such an electrochemical interface can be described by Eq. (16):

$$Z(\omega) = R\_s + \frac{R\_p}{1 + \left(j o R\_P C\_{DL}\right)^\beta} \tag{16}$$

RCT or RP can be determined in several ways. A convenient way is to use the Nyquist diagram. For the simple Randles-type equivalent circuit as shown in Figure 7, the corresponding Nyquist diagram is displayed in Figure 8, in which a perfect semicircle is observed. The high-frequency response is used to determine the component of RS involved in the measurement. RS can be read directly from the abscissa when the angular frequency ω (ω = 2πf) tends to be infinite (fmax or f ! ∞). The total resistance (RP + RS) can also be read from the abscissa when ω approaches zero (fmin or f ! 0). So, RP can be determined by subtracting the RS value from the

Figure 8. The Nyquist diagram responding to the simple Randels-type equivalent circuit.

low-frequency measurement. The conversion of the polarization resistance into a corrosion rate requires an independent empirical measurement of the Tafel slopes using a potentiodynamic polarization method and/or harmonic distortion analysis or otherwise taken from the literature. The double-layer capacitance, CDL, can also be determined for a system exhibiting a behavior similar to that of a perfect RC circuit from the values of RP and the maximum frequency, fmax, that corresponds to the frequency of the point at which the imaginary component has a maximum value, viz.:

$$C\_{DL} = \frac{1}{2\pi f\_{\text{max}} R\_p} \tag{17}$$

It is worth of note that in practice, f cannot really go as high as infinite; it is inevitable that some extrapolation has to be made. Extrapolation at the highfrequency limit usually presents few issues because the impedance becomes nonreactive at frequencies as low as 10 kHz in most cases [82]. On the other hand, reactance is still commonly observed at frequencies as low as 10�<sup>3</sup> Hz [82]. Therefore, special precautions must be taken to obtain reliable data and to avoid possible artifacts [17, 83]. Furthermore, the measurement cycle time depends on the frequency range used, in particular the low frequencies. For instance, a single-frequency cycle at 10�<sup>3</sup> Hz needs about 15 min of testing time. A high-to-low-frequency analysis moving down to 10�<sup>3</sup> Hz frequency likely requires more than 2 hours of scan time. In order to perform a normal standard corrosion monitoring with the EIS technique, assistance is needed to optimize the use of the high-frequency data and reduce measurement time. There is a constant need to improve data processing and analysis in order to minimize uncertainties and to allow the EIS technique becoming userfriendly for corrosion monitoring in both laboratory and field facilities, though it must be emphasized that the need for an easy-to-deploy field instrument has always been an obstacle to online corrosion monitoring with the EIS technique.

An alternative to the impedance model in the Nyquist diagram involves the conversion of the impedance into a complex number. The impedance can thus be designated by an amplitude, |Z|, and a phase shift, ϕ, or by the sum of the real (Z<sup>0</sup> ) and imaginary (Z″) components, such that,

$$Z(\alpha) = Z'(\alpha) + jZ''(\alpha) \tag{18}$$

and RS. By extrapolating from A toward the central linear portion of the |Z| curve, a linear line can be determined. On this line, point B is obtained at log ωL. With the projection of point B to the log|Z| axis, the total resistance (RS + RCT) can be measured. In this way, RP can be determined. At intermediate frequencies, the

The situation struggles when diffusion processes govern the corrosion behavior. A convenient way to deal with this complication is to add a Warburg impedance. The latter describes the impedance of the concentration and diffusion processes in

> ffiffiffi <sup>ω</sup> <sup>p</sup> � <sup>j</sup>

Eq. (19) implies that, whatever the frequency, the real and imaginary parts of

plot, this impedance will result in a straight line at a constant phase angle at 45°, as shown in Figure 10. However, the effect of the Warburg impedance can complicate the correct estimate of the RP value in certain cases. Therefore, the impedance data must be numerically adjusted to fit with the correct model to facilitate the extraction of the total resistance (RS + RP) from the abscissa or by using an appropriate

σw ffiffiffi

<sup>ω</sup> <sup>p</sup> (19)

½. In the Nyquist

capacitor affects the response of the overall RC circuit.

Bode diagram with respect to the Randels-type equivalent circuit in Figure 7.

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods…

DOI: http://dx.doi.org/10.5772/intechopen.86743

The Warburg impedance, ZW, is given by the equation

A Randles-type equivalent circuit including Warburg impedance component, ZW.

the Warburg impedance are equal and inversely proportional to σ<sup>w</sup>

ZW <sup>¼</sup> <sup>σ</sup><sup>w</sup>

the equivalent circuit as shown in Figure 10.

Figure 9.

Figure 10.

183

where σ<sup>w</sup> is the Warburg coefficient.

Both the log|Z| data and the phase angle ϕ are plotted against the angular frequency, log ω, of the excitation signal, a format which substitutes for the Nyquist diagram, i.e., the so-called Bode diagram. Figure 9 shows how the same data (Nyquist plot) appears in a Bode plot format with respect to the equivalent circuit of Figure 7.

Highest (ωH) and lowest frequencies (ωL) can be readily determined. As shown in Figure 9, Z is independent of the frequency at ω<sup>H</sup> and ωL, limit values represented by horizontal lines. From these lines, values of RS and (RS + RCT) can be measured. This analysis forms the basis of the corrosion monitoring as proposed by Tsuru et al. [74] to allow the determination of |Z| at each frequency in the horizontal portions of the Bode diagram.

Sometimes, it is not convenient to perform impedance measurements at very low frequencies (as in DC techniques such as linear polarization). However, it is still possible to extrapolate the polarization resistance, RP, from the Bode diagram. In Figure 9, the low- and high-frequency breakpoints (i.e., ω<sup>L</sup> and ωH, respectively) can be determined from the 45° phase angle Bode diagram (see the pseudo-Gaussian curve). The intersection point A can be determined from the log ω<sup>H</sup>

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods… DOI: http://dx.doi.org/10.5772/intechopen.86743

Figure 9. Bode diagram with respect to the Randels-type equivalent circuit in Figure 7.

Figure 10. A Randles-type equivalent circuit including Warburg impedance component, ZW.

and RS. By extrapolating from A toward the central linear portion of the |Z| curve, a linear line can be determined. On this line, point B is obtained at log ωL. With the projection of point B to the log|Z| axis, the total resistance (RS + RCT) can be measured. In this way, RP can be determined. At intermediate frequencies, the capacitor affects the response of the overall RC circuit.

The situation struggles when diffusion processes govern the corrosion behavior. A convenient way to deal with this complication is to add a Warburg impedance. The latter describes the impedance of the concentration and diffusion processes in the equivalent circuit as shown in Figure 10.

The Warburg impedance, ZW, is given by the equation

$$Z\_W = \frac{\sigma\_w}{\sqrt{w}} - j\frac{\sigma\_w}{\sqrt{w}}\tag{19}$$

where σ<sup>w</sup> is the Warburg coefficient.

Eq. (19) implies that, whatever the frequency, the real and imaginary parts of the Warburg impedance are equal and inversely proportional to σ<sup>w</sup> ½. In the Nyquist plot, this impedance will result in a straight line at a constant phase angle at 45°, as shown in Figure 10. However, the effect of the Warburg impedance can complicate the correct estimate of the RP value in certain cases. Therefore, the impedance data must be numerically adjusted to fit with the correct model to facilitate the extraction of the total resistance (RS + RP) from the abscissa or by using an appropriate

low-frequency measurement. The conversion of the polarization resistance into a corrosion rate requires an independent empirical measurement of the Tafel slopes using a potentiodynamic polarization method and/or harmonic distortion analysis or otherwise taken from the literature. The double-layer capacitance, CDL, can also be determined for a system exhibiting a behavior similar to that of a perfect RC circuit from the values of RP and the maximum frequency, fmax, that corresponds to the frequency of the point at which the imaginary component has a maximum

CDL <sup>¼</sup> <sup>1</sup>

It is worth of note that in practice, f cannot really go as high as infinite; it is inevitable that some extrapolation has to be made. Extrapolation at the highfrequency limit usually presents few issues because the impedance becomes nonreactive at frequencies as low as 10 kHz in most cases [82]. On the other hand, reactance is still commonly observed at frequencies as low as 10�<sup>3</sup> Hz [82]. Therefore, special precautions must be taken to obtain reliable data and to avoid possible artifacts [17, 83]. Furthermore, the measurement cycle time depends on the frequency range used, in particular the low frequencies. For instance, a single-frequency cycle at 10�<sup>3</sup> Hz needs about 15 min of testing time. A high-to-low-frequency analysis moving down to 10�<sup>3</sup> Hz frequency likely requires more than 2 hours of scan time. In order to perform a normal standard corrosion monitoring with the EIS technique, assistance is needed to optimize the use of the high-frequency data and reduce measurement time. There is a constant need to improve data processing and analysis in order to minimize uncertainties and to allow the EIS technique becoming userfriendly for corrosion monitoring in both laboratory and field facilities, though it must be emphasized that the need for an easy-to-deploy field instrument has always

been an obstacle to online corrosion monitoring with the EIS technique.

Zð Þ¼ ω Z<sup>0</sup>

in Figure 9, Z is independent of the frequency at ω<sup>H</sup> and ωL, limit values

and imaginary (Z″) components, such that,

An alternative to the impedance model in the Nyquist diagram involves the conversion of the impedance into a complex number. The impedance can thus be designated by an amplitude, |Z|, and a phase shift, ϕ, or by the sum of the real (Z<sup>0</sup>

Both the log|Z| data and the phase angle ϕ are plotted against the angular frequency, log ω, of the excitation signal, a format which substitutes for the Nyquist diagram, i.e., the so-called Bode diagram. Figure 9 shows how the same data (Nyquist plot) appears in a Bode plot format with respect to the equivalent circuit

Highest (ωH) and lowest frequencies (ωL) can be readily determined. As shown

represented by horizontal lines. From these lines, values of RS and (RS + RCT) can be measured. This analysis forms the basis of the corrosion monitoring as proposed by Tsuru et al. [74] to allow the determination of |Z| at each frequency in the horizontal

Sometimes, it is not convenient to perform impedance measurements at very low frequencies (as in DC techniques such as linear polarization). However, it is still possible to extrapolate the polarization resistance, RP, from the Bode diagram. In Figure 9, the low- and high-frequency breakpoints (i.e., ω<sup>L</sup> and ωH, respectively) can be determined from the 45° phase angle Bode diagram (see the pseudo-Gaussian curve). The intersection point A can be determined from the log ω<sup>H</sup>

ð Þþ ω jZ"ð Þ ω (18)

2πf maxRp

(17)

)

value, viz.:

Corrosion Inhibitors

of Figure 7.

182

portions of the Bode diagram.

modeling software. However, the situation can readily become more complicated if other effects, such as time-constant dispersion, adsorption processes, and so on, are taken into account; the time-constant dispersion, which can be caused by inhomogeneities in the corroded surface, results in a depression of the semicircle [75, 76]. Adsorption, on the other hand, can reveal a second semicircle at low frequencies [77]. All these effects can occur simultaneously [27], making the interpretation of impedance data rather more difficult [78, 84] (Figure 11).

There is a need for an appropriate model equivalent circuit beyond the existing model standards to remedy that shortcoming. An "appropriate" model is understood not only as a good fit of the impedance data but also as a rational explanation of the underlying corrosion mechanism. Moreover, the requirement of sophisticated AC frequency generator and analyzer and the time needed to acquire the complete impedance diagram (particularly in the range of low frequency) impose a serious limitation in real-time corrosion monitoring applications. Other disadvantages include a priori knowledge of the Tafel parameters in order to convert the polarization resistance into a corrosion rate and the fact that it is too difficult to detect and monitor localized corrosion, even if such applications have been explored [85, 86].

> sliding conditions, the size of the capacitive arc increases, suggesting an increase in the transfer resistance and a decrease in the reactivity of the surface, consistent with the effect of mechanical straining of the worn surface. At low frequency, however, the inductive arc indicates the relaxation of the surface concentration of adsorbed intermediate species involved in the dissolution mechanism. Under corrosion-wear conditions, the kinetics of the dissolution process is apparently modified, as revealed by the second inductive loop in the diagram. Given that not all of these investigations have been concluded, a detailed explanation is not straightforward, and further research is recommended. Although these impedance measurements provide a convenient way to study the mechanism of electrochemical reactions involved in tribocorrosion processes, still the interpretation of impedance records during sliding-corrosion experiments is rather difficult because of the heterogeneous surface-state condition. Actually, a nonuniform distribution of the electrochemical impedance on the steel surface must be taken into account. The action of friction can be analyzed thoroughly if this distribution is known. Equivalent electrical circuit models or finite element models could be used to obtain impedance

Nyquist plots recorded at E = 675 mV/SSE (I = 20 mA) on Fe—31% Ni in 0.5 M H2SO4 under free (unloaded) and sliding conditions (against a corundum counterbody pin; 60 N normal force, sliding speed

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods…

DOI: http://dx.doi.org/10.5772/intechopen.86743

). Reproduced from [8] with permission from Elsevier.

2. Comparison of the techniques for the assessment of corrosion rate

ations is illustrated in [63] for the assessment of corrosion rate. The results

Fe electrodes in 0.5 M H2SO4 under well-controlled conditions and their corresponding corrosion current densities, resistances, and required parameters

involved in the comparative decision of which is the best to use.

The transposition of the foregoing electrochemical techniques to corrosion situ-

All these techniques monitor the electrode response following the stimulation by a potential variation in time or frequency domain with the exception of the electrochemical noise analysis technique. The extent of the potential stimulation and the current response decreases in the order from Tafel extrapolation method, linear polarization, EIS, to electrochemical noise. Each of these techniques provides the necessary information for a given corroding system, and there are trade-offs

presented in Table 1 summarize the data generated by the different techniques for

distributions and to calculate the overall impedance.

which determined those data.

185

Figure 12.

0.031 m s<sup>1</sup>

Figure 11. The Nyquist diagram responding to the equivalent circuit of Figure 10.

### 1.4.3 Illustrative examples of the application of EIS in corrosion and tribocorrosion systems

Attempts were made to use the EIS technique in corrosion and corrosion-wear monitoring of Fe-31% Ni electrode immersed in 0.5 M H2SO4 [87]. The corresponding Nyquist impedance diagrams were recorded at an anodic potential of 675 mV/SSE (+100 mV/open-circuit potential) before and during slidingcorrosion as shown in Figure 12. At this potential, the prevailing reaction is dissolution. At high frequency, under free corrosion and unloaded conditions, the capacitive arc reveals the influence of the dielectric properties of the electrochemical double layer and the charge transfer due to electrochemical reactions. Under

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods… DOI: http://dx.doi.org/10.5772/intechopen.86743

#### Figure 12.

modeling software. However, the situation can readily become more complicated if other effects, such as time-constant dispersion, adsorption processes, and so on, are taken into account; the time-constant dispersion, which can be caused by inhomogeneities in the corroded surface, results in a depression of the semicircle [75, 76]. Adsorption, on the other hand, can reveal a second semicircle at low frequencies [77]. All these effects can occur simultaneously [27], making the interpretation of

There is a need for an appropriate model equivalent circuit beyond the existing model standards to remedy that shortcoming. An "appropriate" model is understood not only as a good fit of the impedance data but also as a rational explanation of the underlying corrosion mechanism. Moreover, the requirement of sophisticated AC frequency generator and analyzer and the time needed to acquire the complete impedance diagram (particularly in the range of low frequency) impose a serious limitation in real-time corrosion monitoring applications. Other disadvantages include a priori knowledge of the Tafel parameters in order to convert the polarization resistance into a corrosion rate and the fact that it is too difficult to detect and monitor localized corrosion, even if such applications have been

1.4.3 Illustrative examples of the application of EIS in corrosion and tribocorrosion

monitoring of Fe-31% Ni electrode immersed in 0.5 M H2SO4 [87]. The

The Nyquist diagram responding to the equivalent circuit of Figure 10.

675 mV/SSE (+100 mV/open-circuit potential) before and during slidingcorrosion as shown in Figure 12. At this potential, the prevailing reaction is dissolution. At high frequency, under free corrosion and unloaded conditions, the capacitive arc reveals the influence of the dielectric properties of the electrochemical double layer and the charge transfer due to electrochemical reactions. Under

Attempts were made to use the EIS technique in corrosion and corrosion-wear

corresponding Nyquist impedance diagrams were recorded at an anodic potential of

impedance data rather more difficult [78, 84] (Figure 11).

explored [85, 86].

Corrosion Inhibitors

systems

Figure 11.

184

Nyquist plots recorded at E = 675 mV/SSE (I = 20 mA) on Fe—31% Ni in 0.5 M H2SO4 under free (unloaded) and sliding conditions (against a corundum counterbody pin; 60 N normal force, sliding speed 0.031 m s<sup>1</sup> ). Reproduced from [8] with permission from Elsevier.

sliding conditions, the size of the capacitive arc increases, suggesting an increase in the transfer resistance and a decrease in the reactivity of the surface, consistent with the effect of mechanical straining of the worn surface. At low frequency, however, the inductive arc indicates the relaxation of the surface concentration of adsorbed intermediate species involved in the dissolution mechanism. Under corrosion-wear conditions, the kinetics of the dissolution process is apparently modified, as revealed by the second inductive loop in the diagram. Given that not all of these investigations have been concluded, a detailed explanation is not straightforward, and further research is recommended. Although these impedance measurements provide a convenient way to study the mechanism of electrochemical reactions involved in tribocorrosion processes, still the interpretation of impedance records during sliding-corrosion experiments is rather difficult because of the heterogeneous surface-state condition. Actually, a nonuniform distribution of the electrochemical impedance on the steel surface must be taken into account. The action of friction can be analyzed thoroughly if this distribution is known. Equivalent electrical circuit models or finite element models could be used to obtain impedance distributions and to calculate the overall impedance.

#### 2. Comparison of the techniques for the assessment of corrosion rate

The transposition of the foregoing electrochemical techniques to corrosion situations is illustrated in [63] for the assessment of corrosion rate. The results presented in Table 1 summarize the data generated by the different techniques for Fe electrodes in 0.5 M H2SO4 under well-controlled conditions and their corresponding corrosion current densities, resistances, and required parameters which determined those data.

All these techniques monitor the electrode response following the stimulation by a potential variation in time or frequency domain with the exception of the electrochemical noise analysis technique. The extent of the potential stimulation and the current response decreases in the order from Tafel extrapolation method, linear polarization, EIS, to electrochemical noise. Each of these techniques provides the necessary information for a given corroding system, and there are trade-offs involved in the comparative decision of which is the best to use.


References

[1] Landolt D, Mischler S.

ISBN: 978-1-84569-966-6

Tribocorrosion of Passive Metals and Coatings. Sawston, Cambridge: Woodhead Publishing Limited; 2011.

DOI: http://dx.doi.org/10.5772/intechopen.86743

Society for Testing and Materials; 1981.

[9] Epelboin I, Gabrielli C, Keddam M, Takenouti H. Alternating-current impedance measurements applied to corrosion studies and corrosion-rate determination. In: Mansfeld F, Bertocci U, editors. ASTM STP 727. Barr Harbor Drive, West Conshohocken, PA: American Society for Testing and

[10] ASTM G59-97(2014). Standard Test

[11] Mansfeld F, Xiao H. Electrochemical Noise Measurements for Corrosion Applications, ASTM STP 1277.

Philadelphia, PA: American Society for

[12] Dawson JL. Electrochemical noise measurement: The definitive in-situ technique for corrosion applications? In: Kearns JR, Scully JR, Roberge PR, Reichert DL, Dawson JL, editors. Electrochemical Noise Measurement for Corrosion Applications. ASTM STP

[13] Kearns JR, Eden DA, Yaffe MR, Fahey JV, Reichert DL, Silverman DC. ASTM standardization of electrochemical noise measurement, ASTM STP 1277. In: Kearns JR, Scully JR, Roberge PR, Reichert DL, Dawson JL, editors. Electrochem. Noise Meas. Corros. Appl. Philadelphia, PA:

[14] Reichert DL. Electrochemical noise measurement for determining corrosion rates. In: Kearns JR, Scully JR, Roberge PR, Reichert DL, Dawson JL, editors. Electrochem. Noise Meas. Corros. Appl. Philadelphia, PA: ASTM; 1996. pp. 79-92

Testing and Materials; 1996

1277. pp. 3-35

ASTM; 1996. pp. 446-470

pp. 167-186

Electrochemical Techniques for Corrosion and Tribocorrosion Monitoring: Methods…

Materials; 1981. p. 110

Method for Conducting Potentiodynamic Polarization Resistance Measurements. West Conshohocken, PA: ASTM International; 2014

[2] Celis JP, Ponthiaux P, editors. Testing tribocorrosion of passivating materials supporting research and industrial innovation. In: Handbook EFC 62. UK, France, Germany, Belgium:

Maney Publishing; 2012. ISBN:

[3] Berradja A. Metallic glasses for triboelectrochemistry systems, Chap 5. In: Huang H, editor. Metallic Glasses— Properties and Processing. London, UK: Intech Open; 2018. ISBN: 978-953-51-

[4] Landolt D. Corrosion and Surface Chemistry of Metals. EFPL Press; 2007.

limitations, and needs. Journal of ASTM International. 2008;5:101241. DOI:

[7] Roberge PR. Handbook of Corrosion Engineering. New York, Chicago, San Francisco, Lisbon, London, Madrid, Mexico City, Milan, New Delhi, San Juan, Seoul, Singapore, Sydney,

Toronto: McGraw-Hill Education; 2000.

[8] Haruyama S, Tsuru T. A corrosion monitor based on impedance method. In: Mansfeld F, Bertocci U, editors. Electrochemical Corrosion Testing, ASTM STP 727. Barr Harbor Drive, West Conshohocken, PA: American

[5] Frankel GS. Electrochemical techniques in corrosion: Status,

[6] Trenthewey KR, Camberlain J. Corrosion for Science and Engineering. 2nd ed. United States: Longman Group Limited; 1995. ISBN-13: 978-0582238695

pp. 415-460. Chap 10

10.1520/JAI101241

ISBN: 007-076516-2

187

9781907625202

6208-7

#### Table 1.

Data outcomes determined by different electrochemical techniques on Fe in 0.5 M H2SO4.
