Imidazoline behavior as corrosion inhibitor in the electrochemical characterization of SCC behavior of an API X70 steel exposed to brine solution

2.1 Material

Smooth cylindrical tensile samples of X70 steel were made from pipeline steel with external diameter of 36 in (914.4 mm) and a wall thickness of 0.902 in (22.91 mm). Samples were machined from pipeline in cross direction to flow according to the NACE TM-198 standard (NACE 2004). SSRT tests were performed in duplicated in order to assure repeatability. A good reproducibility of tests for each condition was observed. Table 1 shows the chemical composition of X70 steel.

2.2 Test solution

The test solution was brine prepared according to NACE 1D182 standard (NACE 2005b). The synthetic brine is commonly 9.62% sodium chloride (NaCl), 0.305% calcium chloride (CaCl₂), 0.186% magnesium chloride hexahydrate (MgCl₂·6H₂O), and 89.89% distilled water. The solution pH was adjusted by adding HCl to give pH 7. The test solution was employed under aerated conditions at atmospheric pressure at 60 °C. Aerated conditions were used to increase the severity of corrosion and cracking in tests performed; and therefore, reduce testing time on materials that are susceptible to cracking. A commercial imidazoline-based corrosion inhibitor was added in the test solution at concentrations of 0, 50, 100, and 200 ppm.

2.3 Slow strain rate test (SSRT)

The SSRT test were performed in a constant extension rate tests machine (M-CERT) with load capacity of 44 kN and total extension of 50 mm at strain rate of 1 × 10⁻⁶ s⁻¹. A schematic representation of the SSRT setup is illustrated in Figure 1. The SSRT tests and assessment were carried out according to the standard practice of ASTM G129 standard (ASTM 2013). The SSRT tests were carried out both in air and the brine solution at different concentrations of inhibitor (0, 50, 100, and 200 ppm) at 60 °C. Before testing, tensile samples were polished with SiC abrasive paper up to 1200 grit in the gauge section along the direction vertical to avoid small defects and superficial damages. Then, the samples were rinsed with distilled water and degreased with acetone, and finally dried in air. The exposed area of specimen was 2.84 cm².

Figure 1:

Schematic representation of SSRT test and electrochemical cell arrangement for electrochemical measurements.

After SSRT tests, the fracture surface of specimens was cleaned with inhibit acid containing 1000 mL HCl, 20 g Sb₂O₃ and 50 g SnCl₂ as per the guidelines provided in ASTM G1-90 (ASTM 1999). Then, the fracture surface of the specimens was observed by scanning electron microscopy (SEM). Finally, the gage section of the specimens was cut longitudinally, mounted in a polymeric resin and polished to verify the presence of secondary cracking in the gauge section by SEM examination.

2.4 Electrochemical measurements

The electrochemical impedance spectroscopy (EIS) and electrochemical noise (EN) measurements were performed during SSRT test using a potentiostat-galvanostat. As illustrated in Figure 1, a conventional three-electrode arrangement electrochemical cell was implemented, where the X70 steel acted as the working electrode (WE), saturated calomel electrode (SCE) as reference electrode (RE) and a graphite rod as auxiliary electrode (AE). EIS data were collected under free potential over a frequency range from 10,000–0.01 Hz; seven points per decade of frequency were recorded with amplitude of 0.01 V versus E_corr. For EN measurements, a modified arrangement was employed. The second nominally identical working electrode (WE2) was substituted by a Pt microelectrode. This modified EN arrangement, also called electrochemical emission spectroscopy (EES), has been carried out by other authors (Arganis-Juarez et al. 2007; Chen and Bogaerts 1996; Du et al. 2011). EN measurements were registered at a sampling frequency of 1 Hz. EN records were divided in blocks of 1024 readings. For the statistical analysis, the direct current (DC) trend removal of each block was removed using linear regression (Mansfeld et al. 2001). For the time-frequency domain analysis, the Hilbert–Huang transform (HHT) has been implemented using a publicly available MATLAB algorithm developed byRilling et al. (2007). In order to guarantee the reproducibility of electrochemical tests, all experiments were performed in triplicate.

3.1 Slow strain rate test (SSRT)

Figure 2 shows the stress–strain curves for X70 steel tested in the inert environment (air) and the brine solution at different concentration of inhibitor. The mechanical parameters including time to failure (TF), yield strength (σ_YS), ultimate strength (σ_UTS), elongation (η), reduction in area (Ψ), and plastic deformation percent (ɛ_p) acquired from SSRT curves are summarized in Table 2. In the blank solution, the steel suffered loss of the mechanical properties, since η and σ_UTS values significantly decreased in comparison with those obtained in air from 14.5% and 613 MPa to 9.8% and 597 MPa, respectively. It is clear that the aggressive species contained in the solution (such as Cl⁻, O_2, and H⁺) took a key role in altering the nature of interatomic bonding and the response of the metal deformation mode (Szklarska-Smialowska et al. 1994). At 50 and 100 ppm of inhibitor, the steel exhibited recovery of ductility of steel, because of some mechanical parameter as η and Ψincreased with respect to the blank solution from 9.8% and 46.5%–12.8% and 72.9%, respectively. This means that inhibiting molecules formed a protective film insulating the metal surface from the aggressive solution. Finally, a further addition of inhibitor (200 ppm) provoked again the loss of ductility, even all mechanical parameter values were lower than those obtained in the blank solution. This implies that the corrosion inhibitor performance was affected by the excess of inhibiting molecules in the solution.

Figure 2:

Stress–strain curves of X70 steel tested in air and the brine solution at several concentrations of corrosion inhibitor.

In order to investigate the SCC susceptibility of X70 steel in the brine solution in the absence and presence of corrosion inhibitor, the SCC susceptibility indexes, including the reduction in area index (I_Ψ) and plastic elongation index (I_ε), were calculated according to NACE TM-0198 (NACE 2004) and ASTM G 129 (ASTM 2013) with the following ratios:

where Ψis the percentage reduction in area, ε_p is the plastic elongation and ɳ the maximum strain (%). The suffix 0 and S indicated that the values were obtained in air and the brine solution, respectively. SCC susceptibility indexes values higher than 0.8 indicates high resistance to SCC whereas SCC indexes lower than 0.5 means susceptibility to SCC (Contreras et al. 2017; Jacobo et al. 2019). From Figure 3, it can be seen that SCC susceptibility of steel showed a complex dependence to inhibitor concentration. It seems that when the inhibitor is added below a critical concentration value, the SCC susceptibility decrease by adsorption of inhibiting molecules, whereas the SCC susceptibility increase when the inhibitor is added above this critical concentration value. This behavior is contrary to what has been reported in other studies (Li et al. 2021; Zhang and Gu 2007), where the SCC susceptibility or embrittlement degree decreased as inhibitor concentration increased. Gonzalez-Rodriguez et al. (2021), also reported a critical concentration of 25 ppm for a gemini surfactant, but unlike this study, the SCC susceptibility slightly decreased at concentrations higher than 25 ppm.

Figure 3:

SCC susceptibility index values as a function of inhibitor concentration.

It is known that imidazoline based inhibitors usually have an amphiphilic molecular structure, which consist of a polar head group (imidazoline ring) and a non-polar hydrophobic tail (alkyl chain), acting as surfactant molecule (Geng et al. 2022). One important surface-activity parameter of surfactant related to their effectiveness as corrosion inhibitor is the critical concentration micelle (CMC). This is because of when the inhibitor concentration is closed to CMC; inhibiting molecules tend to form a compact and protective multilayer film. Instead, at concentrations well above CMC, strong repulsion hydrophobic interactions may lead to desorption of inhibiting molecules, leaving local active sites uncovered on the metal surface, which can contribute to pits or/and cracks nucleation (Hegazy et al. 2010; Mobin et al. 2017).

Some works (Prethaler et al. 2015; Shamsa et al. 2022) have found the CMC value of some based imidazoline inhibitors was around 50 ppm. Although the CMC value was not experimentally determinate by superficial tension measurements, the behavior observed in Figure 3 suggested that CMC of the commercial imidazoline evaluated in this study probably was between 50 and 100 ppm.

3.2 Fractography analysys by SEM

In order to elucidate the effect that the inhibitor had on the cracking process, the SEM images of the fracture surface of X70 steel in the brine solution with several concentrations of inhibitor are shown in Figure 4. At the macroscopic level, an elliptical shape fracture area normally is sign of a ductile fracture whereas a circular shape fracture area is indicative of a brittle fracture. In the reference environment (Figure 4a and a1), the top view of the fractured samples showed a typical ductile mode and dimple-type fracture morphology was observed (Figure 4b). The mechanism for this kind of fracture is recognized as microvoid coalescence (Contreras et al. 2012). In the blank solution (Figure 4b and b1), the steel exhibited a brittle fracture with little evidence of plasticity, and cleavage-like features were observed. Despite of the neutral solution pH (pH 7), local acidification inside pit bottom and crack tips can be attributable to the hydrolysis process, furthermore, the synergistic effect between stress and anodic dissolution enhances the absorption of hydrogen (Gui et al. 2010; Li and Cheng 2007). Therefore, hydrogen could take an important role in the SCC process. On the other hand, with addition of 50 ppm (Figure 4c and c1) and 100 ppm of inhibitor (Figure 4d and d1), an increase in the ductility of the material was observed and the fracture morphology changed from the cleavage-like fracture obtained in the blank solution to ductile dimple fracture, very similar to fracture exhibited in air. Qi et al. (2016) reported similar fracture transition their study of SCC inhibition of a stainless steel in a CO₂–H₂S–Cl⁻ solution using 1.65% w/w of methyldiethanolamine (MDEA), attributing this to preferential adsorption of inhibiting molecules rather than corrosive species. At 200 ppm (Figure 4e and e1), the steel showed fracture brittle mode with little necking. Lynch (2012) mentioned that strong bonds between the substrate (metal surface) and the adsorbate (inhibiting molecules), it is envisaged that dislocation nucleation at crack tips is difficult, and that link-up of the crack tip with voids occurs primarily by egress of dislocations around crack tips, producing the ductile dimple fracture morphology. At 50 and 100 ppm, probably the inhibiting molecules formed stronger bonds on the metal surface and stronger bonds between them, forming a compact inhibitor film that limited the adsorption of corrosive species, meanwhile, at 200 ppm, these bonds were weakened by hydrophobic interactions between the inhibiting molecules into the inhibitor film (Mobin et al. 2017).

Figure 4:

SEM micrographs of fracture surfaces of the X70 steel after SSRT tests, (a, a1) in air, (b, b1) in the brine solution without inhibitor, (c, c1) with 50 ppm, (d, d1) with 100 ppm, and (e, e1) with 200 ppm.

In order to confirm if the SCC process occurs in the metal, SEM micrographs of the cross-section of the gage surface of X70 steel samples at concentrations of 0, 100, and 200 ppm of inhibitor are displayed in Figure 5. At all concentrations (Figure 5a, b, c), the secondary cracking cannot be observed in a macroscopic level. However, some pits and cracks were visible at higher magnification (Figure 5a1, b1, c1). Without inhibitor (Figure 5a1), several cracks with approximately 10–25 µm length and pits were observed close to the neck region. These cracks were wide and initiated close to each other.

Figure 5:

SEM micrographs of cross section of the gage length of X70 steel exposed to the brine solution at 60 °C at several inhibitor concentrations: (a, a1) 0 ppm, (b, b1) 100 ppm, and (c, c1) 200 ppm.

Li et al. (2020) reported similar crack morphology and crack size in their study of sulfide stress cracking of X65 steel in H₂S environment. The crack width can be associated to severe anodic dissolution at the crack tip and crack walls, which can cause crack tip blunting (Gui et al. 2010). In contrast, in the presence of inhibitor at 100 ppm, there were no cracks, but few semielliptical pits were found in the gage section. Finally, at 200 ppm, a greater number of pits with different sizes were observed in the gage section in comparison to those observed at 100 ppm. It confirms that an excess of addition of inhibitor increase the occurrence of localized corrosion, however, the pit-to-crack transition was not happened in the presence of inhibitor.

Turnbull and Zhou (2004) found that not all pits turn into cracks, because they must attain a critical pit size. Furthermore, Cheng (2013) believes that differences between the electrochemical activity of the pit bottom and pit sides is the mainly factor in the driving force for pit-to-crack transition. Probably, the presence of the inhibiting molecules impeded the pit-to-crack transition interfering with any of the factors above mentioned. It should point out that severe localized corrosion (pitting) caused premature failure of the steel, decrementing the mechanical properties, and obtaining a low SCC index. Cragnolino et al. (1996) also reported decrease of ɳ percent and premature failure of a 316L SS in a solution rich in chloride by pitting corrosion.

3.3 EIS results

For a better understanding of the SCC process during the SSRT test (Figure 2, at T0, YS and BF) in the presence of the corrosion inhibitor, the EIS spectra for X70 steel in brine solution at different concentration of inhibitor at the points T0 (at beginning of SSRT), YS (at yielding strength), and BF (before failure) are shown in Figure 6. At point T0 (Figure 6a), all Nyquist plots exhibited a characteristic of a capacitive behavior, but the number of semicircles associated with the number of time constants of the corrosion processes that occurred at the metal-solution interface was difficult to define. It should be note that no significant effect of the inhibitor concentration on semicircles size of the Nyquist plots was observed at the beginning of the SSRT test. Meanwhile, at point YS (Figure 6c), two well-defined capacitive semicircles at were manly observed at 50 and 100 ppm of inhibitor and the diameters or semicircles, which are associated to corrosion rate, increased in comparison to those at point T0, indicating that the inhibitory action of inhibiting molecules was enhanced from the point T0 to point YS despite the gradual increase of elastic stress. Nevertheless, at the point BF (Figure 6e), the diameter of semicircles decreased. Probably, it is because of when the steel is subjected to an elastic stress, dislocations only oscillate around their equilibrium positions; therefore, the electrochemical behavior at the metal-solution interface is not significantly altered, allowing the protection of the metal through the adsorption of the inhibitor. However, once the local stress exceeds the yield strength of the steel, dislocation movement occurs, resulting in the formation of slip steps on the steel surface, which can introduce active sites and provoking localized corrosion due to the breakdown of the inhibitor film (Liu et al. 2011).

Figure 6:

EIS spectra for X70 steel in the brine solution at 60 °C at different inhibitor concentrations during SSRT test: (a, b), at point T0, (c, d) at YS, and (e, f) BF.

In the Bode plots (Figure 6b, d, and f), it was more distinguishable the presence of one time constant for the blank solution, and two-time constants in the presence of inhibitor. The time constant at high frequencies was associated to formation of inhibitor film and the time constant at intermediate and low frequencies was attributed to transfers charger process at the metal/solution interface.

Although some authors (Lou and Singh 2011; Zheng and Yi 2017) claimed that the fitting of the EIS data with CEE is not adequate in the study of SCC due to the high instability generated by the cracking process in the interface metal-solution, in this study the presence of the inhibitor on the surface, which may have an influence on the EIS spectra. For this reason, the electrical equivalent circuits (EEC) of Figure 7a and b were used for the fitting of experimental EIS data and the physical description of the interface in the absence of inhibitor and presence of inhibitor, respectively. From Figure 7, the electrochemical parameter Rs is the electrolyte resistance, R_ct is the charge-transfer resistance, CPE_dl represents the non-ideal capacitance of electrochemical double layer, and its impedance is given by following equation (Álvarez-Bustamante et al. 2009):

where j is the imaginary number, ω is the angular frequency, and Y₀ and n are the CPE parameters, where the latter has been related to the surface inhomogeneity degree (Wang et al. 2010). The real double layer capacitance (C_dl) values were calculated using the Brug’s equation (Brug’s et al. 1984):

Figure 7:

Equivalent electrical circuit for X70 steel in the brine solution at 60 °C during the SSRT test: (a) without corrosion inhibitor and (b) with corrosion inhibitor.

Finally, R_f and CPE_f are the resistance and non-ideal capacitance of film of corrosion inhibitor.

The electrochemical parameters derived from the fitting of the impedance response are illustrated in Table 3. At concentrations of inhibitor of 0 and 200 ppm, the R_ct values were close to each other, and no significant variation was exhibited during the SSRT test. Meanwhile, at concentrations of inhibitor of 50 and 100 ppm, the R_ct values increased achieving a maximum value at YS point and then the R_ct decreased at the point BF.

Some authors (Porcayo-Calderon et al. 2015; Shamsa et al. 2022) have found that with addition of based imidazoline in CO₂ saturated brine solution, the corrosion rate sharply decrease, reaching a stable value between 6 and 10 h of exposure, maintaining its inhibition efficiency throughout the evaluation period. In the present study, the stress conditions can difficult the inhibitor persistence on the metal surface. In addition, it has been reported that imidazoline group can undergo hydrolysis at higher pH (pH > 7) having the hydrolysis products little anticorrosion properties (Porcayo-Calderon et al. 2015; Zheng et al. 2022). Therefore, the partial hydrolysis of imidazoline along with the presence of stress could have affected the performance of inhibition.

On the other hand, the general trend of C_dl values was increasing during the SSRT test, which can be attributed to the increase of active surface area of metal during the crack propagation and the formation of non-protective, porous, and conductive corrosion products (Galván-Martinez et al. 2019a, 2019b).

Despite of the tendency of increment of C_dl values with inhibitor, the SEM fractography showed that the presence of inhibitor decreased the degree of brittleness on the steel at 50 and 100 ppm (Figure 4c and 4d). McCafferty (2010) mentioned that the corrosion inhibitor should reach the inside of the crack to be effective in SCC and corrosion fatigue inhibition. Therefore, the inhibitor molecules could have diffused into the crack tip, retarding the crack propagation and consequently, inhibiting the SCC process.

3.4 EN measurements

The electrochemical current noise (ECN) records for X70 steel exposed to brine solution without corrosion inhibitor (0-ppm) and with corrosion inhibitor (50-ppm) during SSRT test (Figure 2, at points T0, YS, and BF) are shown in Figure 8 and Figure 9. In the blank solution (Figure 8a, c, and e), overlapped peak shaped transients were observed during SSRT test. Similar behavior (overlapping current transients) was observed by several authors  (Cottis et al. 2016; Wang and Cheng 2016) which makes ECN analysis difficult in the time domain. Apparently, the amplitude of fluctuations (from 80 to 200 nA) and lifetime of this transients (from 100 to 200 s) increased from T0 to BF point, indicative of the occurrence of localized corrosion events, such as formation of microcrack, but this fact can be produced by the current transient overlap.

Figure 8:

ECN for X70 steel exposed to brine solution at 60 °C at different concentrations of inhibitor: (a, c, e) 0 ppm and (b, d, f) 50 ppm during SSRT test at points T0 (a, b), YS (c, d), and BF (e, f).

Figure 9:

ECN of X70 steel immersed in the brine solution at 60 °C at concentrations of inhibitor: (a, c, e) 100 ppm and (b, d, f) 200 ppm during SSRT test at points T0 (a, b), YS (c, d), and BF (e, f) points.

At 50 ppm (Figure 8b, d, and f), the features of ECN signals were like those obtained in the blank solution, with the difference of the lifetime of transients were lower (∼50 s) and their amplitude decreased during the SSRT test (from 200 to 100 nA). This behavior has been attributed to increase of inhibitory effect of corrosion inhibitor as time elapses (Zhao et al. 2021).

At 100 and 200 ppm of inhibitor, at the beginning of the test (Figure 9a and b, respectively) the pattern of ECN signal was similar with respect to the lower concentrations, nevertheless, the features of the signals changed in the YS and BF points.

From Figure 9c and e, it can be seen that an undulatory behavior predominated in the ECN signal with decrease of noise intensity. It can be attributed to the presence of a protective film of inhibitor that inhibited the cracking process at 100 ppm. In contrast, at 200 ppm, many fast transients with amplitudes around 0.1 µA were observed at the points YS (Figure 9d) and BF (Figure 9f). This type of transients has been reported for the nucleation of many small pits and the occurrence of metastable pitting (Jiao et al. 2021; Stewart and Williams 1992).

Ortiz-Alonso et al. (2014) obtained similar transients in their EN measurements during SSC in a super martensitic stainless steel. Moreover, Zhang et al. (2007) say that many fast transients can reflect the competition process of metastable pitting and repassivation. In the same way, the breakdown of the corrosion inhibitor film by the action of stress during the main crack propagation produced a competition process between desorption and adsorption of inhibitor molecules, contributing to the nucleation of anodic sites where localized corrosion events can take place.

3.5 Statistical EN analysis

After the DC trend of the EN data was removed by linear regression, statistical parameters such as noise resistance (Rn) and localization index (LI) derived from statistical analysis of electrochemical potential noise (EPN), and ECN records during the SSRT test, were calculated according to the following equations (Cottis 2001):

where σE and σI are the standard deviation of potential and current, respectively, and IRMS is the root mean square of current. R_n is a parameter inversely proportional to corrosion rate, whereas LI have been successfully related to corrosion mechanism, despite of its theoretical limitations (Mansfeld and Sun 1999). Rn and LI values of X70 steel in the brine solution at several concentrations of inhibitor during SSRT test (Figure 2, T0, EZ, YS, UTS, and BF) are displayed in Figure 10. Like EIS results from R_ct value, the lower R_n values were obtained for the steel in the solution blank and the higher R_n values were presented in the brine solution with 100 ppm of inhibitor. This meant that higher corrosion (anodic dissolution) rate, represented higher SCC susceptibility, indicative that anodic dissolution mechanism dominated the SCC process of the steel in the brine solution without inhibitor. The trend of R_n values was to decrease from the point T0 to point YS and slightly increase at the points UTS and BF. Ortíz-Alonso et al. (2014) claimed that R_n decrement can be due to more localized type of corrosion events, such as pits or cracks taking place on the metal surface. It should be noted that Rn values were more than two orders of magnitude with respect to those obtained from R_ct. Several authors (Hernández et al. 2009; Hoseinieh et al. 2016) have reported this lack of correlation between EN and EIS to a higher sensitivity to localized corrosion by EN and the analyzed frequency range. On the other hand, as shown in Figure 10b, most LI values were within the range (0.01–0.1) corresponding to mixed corrosion (general and localized corrosion). Localized corrosion can mainly take place at the crack tip, since, being sites of stress concentration, anodic dissolution is faster, and a localized environment can be generated due to local acidification caused by hydrolysis of dissolving metal cations (Gui et al. 2010). Moreover, the presence of the inhibitor film can also provoke localized corrosion events due to defects of the film and incomplete coverage surface. Tan et al. (2011) performed locally galvanic measurements by a wire beam electrode, and they found that imidazoline can aggravate localized corrosion by creating a small number of anodic sites with highly concentrated anodic dissolution. In addition, general corrosion can occur on large zones on the gage section surface. It should be noted that LI is a punctual value that does not provide information related on the contribution of the corrosion process to the ECN signal. The time-frequency analysis of ECN using by the Hilbert–Huang transform (Calabrese et al. 2018; Homborg et al. 2013, 2014) can overcome these limitations.

Figure 10:

(a) Noise resistance (Rn) and (b) LI values for X70 steel in brine solution at 60 °C at different concentrations of corrosion inhibitor during SSRT test.

3.6 Hilbert–Huang transform (HHT) analysis

HHT is a novel time–frequency analysis method for nonstationary signals, which has showed several advantages on other analysis methods such as fast Fourier transform and discrete wavelet, mainly when EN records consist in overlapped transients, which can hind information of each electrochemical process on the obtained EN signal (Cottis et al. 2016; Homborg et al. 2013). In general terms, the HHT decompose the EN signal into instantaneous frequencies and variable amplitudes, which are displayed in a Hilbert spectrum, which enable a determination of the instantaneous frequency composition of individual corrosion process observed in the EN signals at any given moment in time and in turn allows to identify and distinguish between different corrosion mechanisms (Homborg et al. 2014; Xia et al. 2016). A detailed description of the mathematical procedure of HHT is reported by Huang et al. (1998).

In order to obtain further mechanistic information on SCC, the Hilbert spectra of the ECN records of steel in the brine solution at different concentration of inhibitor at the YS point of the SSRT curve and the relative amplitudes of each original ECN record at the back of each Hilbert spectrum are shown in Figure 11. The YS point was chosen because several works (Galvan-Martinez et al. 2019a, 2019b; Jiao et al. 2021; Ritter and Seifert 2013) have found by EN measurements that the SCC crack initiation occurs in a point close to YS of metal. Depending on the frequency range where the highest relative contribution of instantaneous frequencies, the corrosion process nature can be determinate. According to literature (Calabrese et al. 2018; Homborg et al. 2016), fast events with short timescales and high frequency have been attributed to localized corrosion process and slow corrosion process corresponds to larger timescales in the lower frequency domain (general corrosion or diffusion). In the absence of inhibitor (Figure 11a), the Hilbert spectrum exhibited the higher relative contribution of instantaneous frequencies in intermediate frequencies between 10⁻² and 10⁻¹ Hz and slight contribution of instantaneous frequency at low frequencies above 10⁻¹ Hz. This is indicative that signal is dominate for localized corrosion events at intermediate frequencies attributable to crack initiation and propagation and the occurrence of pits at higher frequencies. At 50 ppm (Figure 11b), as in the blank solution, transients located in intermediate frequencies (10⁻² and 10⁻¹ Hz) had the highest relative contribution of instantaneous frequencies, but and increment of contribution of instantaneous frequencies below 10⁻² Hz along the entire ECN signal was observed. This behavior is more remarkable at 100 ppm (Figure 11c), even at this concentration, the low frequency component in the ECN signal dominated along the entire time axis. Homborg et al. (2013) associated this fact to a (oxygen) diffusion-controlled process. Therefore, this increase at the low frequency range can be associated to formation of inhibitor film enough protective to inhibit the occurrence of localized corrosion events, such as pits and crevices which can function as active sites of nucleation of cracks, and in turn, this inhibitor film limit both the oxygen reduction reaction and anodic dissolution. This was confirmed with the absence of corrosion pits in the gage section observed by SEM (Figure 5b). In contrast, the Hilbert spectrum at 200 ppm (Figure 11d) showed negligible contribution of instantaneous frequencies at low frequencies below 10⁻² Hz and many transients with high contribution of instantaneous frequencies at the high frequency region (above 10⁻¹ Hz). This behavior is indicative that short timescale processes (localized corrosion events) dominated the ECN signal. Some works (Homborg et al. 2014; Xie et al. 2021) have associated this fact with fast metastable pitting events, which suggest that the inhibitor film is constituted of multiples defects produced by desorption of inhibitor molecules, where localized corrosion process can take place on the surface. This was corroborated by SEM micrographs where several corrosion pits were found in the gage section at this concentration (Figure 5c).

Figure 11:

Hilbert spectra of the ECN record of X70 steel in brine solution at 60 °C at YS point during the SSRT test at different concentration of inhibitor: (a) 0 ppm, (b) 50 ppm, (c) 100 ppm and (d) 200 ppm.