Effect of pH on the Electrochemical Behaviour and Passive Film Composition of 316L Stainless Steel

1 Introduction

Amine solutions were widely used in the chemical industry to absorb tail gases, such as CO₂ and SO₂ [1, 2]. The pH of the amine solution typically ranges from 11 to 13. Austenitic stainless steels, especially 316L stainless steel, were widely used as structural materials in absorption environments due to their good corrosion resistance and mechanical properties. In alkaline environments, the steel surface is covered with a protective oxide film, which is strongly adherent and chemically stable. The passive film is very thin and has been believed to be only several nanometres in thickness [3]. The passive film typically had a bilayer structure with a varying composition that depends on the environment [4]. For example, the presence of CO₂ could stimulate the dehydroxylation of the outer part of a passive film [5] and the addition of sodium sulphide could affect the Cr oxide in the inner passive film [6]. Commonly, the oxide film exhibits a Cr-enriched inner layer and a Fe-enriched outer layer in the literature [7, 8]. Nevertheless, the structure of the passive film could be strongly influenced by pH. Li et al. [9] found that under strong acid conditions, the passive film might exhibit the opposite structure, where the outer film was mainly composed of Cr oxide and the inner film was dominantly consisted of Fe-oxides. The variation in the chemical composition of the passive films would eventually lead to a change in corrosion resistance.

To date, several investigations have been conducted with regard to the passive and pitting behaviours of stainless steels under alkaline conditions [10, 11, 12, 13]. Agreements have been reached on the pH effect on the pitting behaviour of materials. It has been well accepted that with increasing pH, the pitting potential increased, while the oxygen evolution potential decreased. Therefore, it was possible that the conversion from pitting corrosion to OER occurred at a critical pH. However, in some literature [13], the OER was ignored even though the pH was significantly high.

In the present work, the corrosion behaviour of 316L stainless steel was investigated in alkaline solutions at various pH values. The study was conducted using the electrochemical techniques of cyclic polarization, electrochemical impedance spectroscopy (EIS), and galvanostatic measurements. The composition of the passive films formed on 316L at various pH values was studied by X-ray photoelectron spectroscopy (XPS) analysis. According to the electrochemical results, the critical pH for the conversion from pitting corrosion to OER was found. The pitting corrosion and OER were compared from the thermodynamic and kinetic aspects. The correlation between the passive film composition and its ability to protect from corrosion was discussed herein.

2 Experimental

2.1 Materials and Solutions

All test specimens were cut from a hot-rolled 316L stainless steel plate. The chemical composition of 316L stainless steel is listed in Table 1. All the specimens were cut to 10 mm × 10 mm × 3 mm in size. The electrochemical coupons were mounted by epoxy resin with a working area of 1 cm². Prior to the experiments, the specimens were successively polished with SiC paper up to 2000 grit in order to obtain a fine specimen surface and minimize the errors due to surface roughness. After polishing, the specimens were cleaned with distilled water and ethanol and consequently dried under cool air.

Test solutions contained 15 g/L NaCl were prepared by using deionized water and analytically pure NaCl. The pH of the solutions was adjusted by NaOH. The solutions were used immediately after preparation to avoid carbonation effects. All tests were conducted at ambient temperature (25 °C).

2.2 Electrochemical Measurements

All electrochemical measurements were carried out in a three-electrode cell. A platinum plate and a saturated calomel electrode (SCE) were utilized as the counter and reference electrodes, respectively. All the work electrodes were cathodically prepolarized at - 1 V (SCE) for 5 min to remove the easily reducible surface films and then stabilized at the open circuit potential (OCP) for 1 h. The pH of solution was measured and adjusted again to avoid pH variation. The cyclic polarization curves were recorded from - 0.1 V (vs. OCP) in the positive direction with a potential sweep rate of 0.5 mV/s. The scan reversed when the current densities reached 5 mA/cm². The surface morphologies of the work electrodes were examined by scanning electron microscopy (SEM, JSM-6510A). To investigate the kinetics of OER and pitting corrosion, the galvanostatic tests were conducted at a constant current density of 1 mA/cm² for 1600 s. The current density was selected since it was high enough for both reactions. The potentiostatic measurements were employed at a constant anodic potential of 0.1 V (SCE) for 3600 s just after the cathodic polarization. EIS measurements were then taken at 0.1 V (SCE) using an alternating current voltage amplitude of 10 mV. The frequency varied from 100,000 to 0.01 Hz. All the electrochemical tests were performed at least two times to ensure reproducibility.

2.3 XPS Analysis

After the anodic polarization at 0.1 V (SCE) for 3600 s under various pH conditions, the specimens were rinsed in deionized water and dried by highly purified nitrogen. The compositions of the passive films formed at various pH values were analysed by XPS with an AXIS-UltraDLD instrument [monochromatic AlK_α (hυ = 1486.6 eV)]. Additionally, the C1s peak at 284.8 eV was used as a reference to adjust the shifted charge. An analysis of the peaks on the measured curves was conducted via the commercial software XPSpeak version 4.1.

3 Results

3.1 Cyclic Polarization

Figure 1 presents the cyclic polarization curves obtained for 316L stainless steel measured under various pH conditions. 316L displayed typical passive behaviour under these conditions. No active-passive transition region appeared on the curves, indicating that the material could spontaneously passivate in its natural state. The current oscillation could be observed when pH was less than 11, suggesting that metastable pitting occurred on the sample surface. Since all the metastable pits were likely to stably propagate, the pitting potentials at lower pH were more discrete. Besides, it was clearly observed that hysteresis loops were present at pH lower than 12. However, the hysteresis loops disappeared at pH higher than 13. The presence and absence of the hysteresis loops were related to the pitting corrosion and OER. In the author’s previous study [4], a sharp increase in current density at relatively positive potentials was related to transpassivity, oxygen evolution, or localized corrosion (commonly pitting corrosion). Since the OER on the sample surface would not damage the passive film, this process should be reversible in theory. This meant that the current densities measured from the forward and reversing scan at the same potentials should be nearly identical. Therefore, no hysteresis loop could be observed. However, when localized corrosion occurred at a certain anodic potential, repassivation in this location would be more difficult due to the local acidity in pits. Therefore, at the same potentials, the current densities during the reversing scan would be higher than that during the forward san and eventually resulted in the presence of a hysteresis loop in the cyclic polarization curves. Actually, the surface was checked after the polarization tests and the SEM images are presented in Fig. 2. Pits were clearly observed on the sample surface at pH ranging from 7 to 11, while no pit was found at pH 13 and 13.5. Actually, it was observed that at pH higher than 13, oxygen generated on the electrode surface at the potentials where the current densities increased sharply. Therefore, the corresponding potentials were taken as oxygen evolution potentials at pH higher than 13. The results indicated the conversion from pitting corrosion to OER of 316L stainless steel with increasing pH at higher potentials.

The cyclic polarization plots of 316L measured at pH 12 presented similar hysteresis loops, which implied that the critical pH for the conversion from pitting to OER was around 12.5. The cyclic polarization plots of 316L measured at pH 12.5 are shown in Fig. 3. The curves showed good reproducibility in the forward scan. However, during the reversing scan, the work electrodes suffered OER for Tests 1 and pitting corrosion for Test 2, respectively. For Test 2, the surface morphologies of the work electrode during the polarization test are shown in Fig. 4. At the potential point 1, a large amount of oxygen was generated on the surface (Fig. 4a), which was responsible for the dramatic increase in current densities. Once the current density reached 5 mA/cm², the scan reversed. At the potential point 2, the passive film locally broke down and pitting corrosion occurred (Fig. 4b). Meanwhile, the OER stopped and no new oxygen bubble was generated. This induced the higher current densities during the reversing scan than those in the forward scan region. It was noteworthy that the reversing and forward curves still overlapped at the beginning of reversing scan. This was somewhat different from the typical cyclic polarization curves when the work electrodes suffered pitting corrosion as shown in Fig. 1a-c, but was similar to Test 1. Therefore, combined with the surface morphologies, the potential where the current densities increased sharply was taken as the oxygen evolution potential and the potential where the forward and reversing plots separated was regarded as the pitting potential. This was theoretically reasonable since pitting corrosion could increase the reversing current densities and lead to the presence of a hysteresis loop. The oxygen evolution potential and pitting potential appeared in one plot at the critical pH. This was not currently reported in the literature. The pitting corrosion did not occur during the forward scan where the potentials were more positive than the pitting potential. This was due to the pitting process being postponed by OER. The detailed discussion is given in Sect. 4.2.

The passive current densities (i_p) extracted from the polarization curves are plotted in Fig. 5. As could be observed, i_p slightly increased from 1.87 to 2.61 μA/cm² as the pH increased from 7 to 11. At pH higher than 11, i_pdramatically increased to 4.07, 4.74, 6.11 and 6.17 μA/cm², respectively. Similar results were reported by Freire et al. [14, 15], who found a reduction of more than one order of magnitude in the passive current density when the pH dropped from 13 to 9. The increase in i_p was related to the decrease in Cr and Mo content with pH in the passive film, which would be illustrated in Sect. 4.3.

3.2 Potentiostatic Measurements

The current density versus time curves were measured in the solutions at various pH values. The potentials were set to be 0.1 V (SCE) at pH from 7 to 13. Figure 6 presents the potentiostatic polarization results of 316L measured in NaCl solutions at ambient temperature under various pH conditions. The current densities decreased dramatically at the beginning, which was attributed to the charging of the electric double layer at the specimen/solution interface and the thickening of the passive film [16]. The current densities finally stabilized at 0.206, 0.499, 0.764 and 1.45 μA/cm² at pH values of 7, 9, 11 and 13, respectively. The current densities increased with increasing pH.

3.3 EIS Results

Prior to the EIS tests, the samples were pre-passivated at 0.1 V (SCE) for 3600 s to form a stable passive film. Figure 7 shows the Nyquist plots of 316L measured at 0.1 V (SCE) under various pH conditions. The plots presented typical capacitance characteristics. The Nyquist plots were composed of one capacitive reactance arc, corresponding to the passive film. The capacitance corresponding to the double layer was not observed. This might be due to the fact that the film resistance was much higher than charge transfer resistance and that the double layer capacitance was much larger than the measured film capacitance and therefore the double layer capacitance was negligible [17].

The EIS data were fitted with the electrochemical equivalent circuit, which is shown in Fig. 8. In this circuit, R_s is the solution resistance, R_f is the resistance of passive film, and Q_f represents the passive film capacitance. In the circuit, Q is the constant phase element (CPE). CPE was commonly used in the case of uneven current distribution at the surface or in the event of increased surface roughness. The impedance of CPE is given by:

$Z_{\text{CPE}} = \left( {1/Y_{0} } \right)\left( {j\omega } \right)^{ - n} ,$ (1)

where Y₀ is a proportionality constant, j² = - 1, and n is an empirical exponent between 0 and 1. When n = 1, the CPE represents a pure capacitive element. When n = 0, the CPE represents a resistor, and when n = 0.5, the CPE represents the Warburg impedance.

The values of the equivalent circuit elements and the error of the impedance data calculated at various pH are listed in Table 2. It could be seen that the max SE for the fitting parameters was below 10%, which was acceptable for the EIS fitting. The good fit of the simulated curves to the measured data indicated the reliability of the proposed equivalent circuit and the fitting results. R_f decreased from 320 to 79 kΩ as the pH increased from 7 to 13, indicating that the corrosion resistance of 316L was weakened at the higher pH. This was consistent with the polarization results.

3.4 Passive Film Characterization

After the potentiostatic tests, the chemical composition of the passive films formed at various pH values was investigated by XPS measurements. The main cation elements Cr, Fe, Mo, and Ni and the anion element O were analysed. However, Ni was barely detected in the passive films and thus the XPS result of Ni is not presented in the following section. Figure 9 shows the high-resolution XPS spectra of Cr 2p_3/2 obtained on the specimens after the potentiostatic tests. Three constituent peaks were identified in the Cr 2p_3/2 signal, representing Cr⁰ in metallic state (574.1 eV), Cr³⁺ in oxide (Cr₂O₃, 576.4 eV), and hydroxide (Cr(OH)₃, 577.1 eV), respectively. At pH 7, the outermost layer was composed of Cr and Cr(OH)₃. In order to characterize the inner part of the passive film, the XPS measurement was also conducted at 1 nm depth by XPS sputtering. It should be emphasized that this depth was estimated by sputtering on the Si/SiO₂subtract. Hence, all the depth data reported in this paper were referred to Si/SiO₂, if not specially pointed out. At 1 nm depth, the peak for Cr₂O₃ appeared, indicating that chromium oxides lay underneath the hydroxides [18]. The presence of metallic Cr at pH 7 suggested that the passive film was relatively thin under this condition. When pH > 7, the Cr⁰ signal disappeared. It was noted that the peaks corresponding to Cr(OH)₃ were significantly stronger than those of Cr and Cr₂O₃, which suggested that more Cr(OH)₃ was generated in the films. At pH 13, the passive film was only composed of Cr(OH)₃ at 0 and 1 nm, which demonstrated the strong hydroxylation of the passive film at higher pH. The similar result was reported by Lu et al. [19]. They found that Cr existed mainly in hydroxides and this was enhanced with increasing pH.

Figure 10 shows the high-resolution XPS spectra of Fe 2p_3/2 obtained on the specimens after the potentiostatic tests. The Fe spectra were decomposed into four peaks contributed to Fe⁰ (706.8 ± 0.1 eV), Fe₃O₄ (709.5 ± 0.2 eV), Fe₂O₃(710.7 ± 0.1 eV), and FeOOH (711.5 ± 0.2 eV). At pH 7, the passive film was composed of Fe, Fe₃O₄, Fe₂O₃, and FeOOH. As the sputtering depth increased from 0 to 1 nm, the main composition of the film changed from iron hydroxides to oxides. Besides, the peak corresponding to Fe⁰ increased significantly. This result implied that the film formed at pH 7 was thin, which was consistent with the XPS results of Cr. As the pH increased to 9, a larger peak of Fe₂O₃ was detected and the Fe⁰ peak was not as obvious as that at pH 7. The Fe⁰ signal finally disappeared at pH 11 and 13, which indicated that the passive film thickened with pH. In the outermost layer, the passive film formed at pH 11 was composed of Fe₂O₃ and FeOOH, while at pH 13, only FeOOH was detected. The passive film tended to hydroxylate at higher pH. It was noted that the main peaks of the Fe signal shifted towards the left side as the sputtering depth increased. In general, the peak located in the high binding energy state always corresponded to the high oxidation state or the hydroxide form of the elements. Therefore, the left movement of the peak with the sputtering depth implied that the main composition of the film changed from iron hydroxides in the outer layer to oxides in the inner layer.

Figure 11 shows the high-resolution XPS spectra of Mo 3d obtained on the specimens after the potentiostatic tests. The Mo spectra exhibited a doublet due to the spin-orbit coupling of Mo 3d_5/2 and Mo 3d_3/2. The spectra of Mo 3d could be split into three components, i.e. MoO₃ (3d_3/2 235.6 eV and 3d_5/2 232.5 eV), MoO₂ (3d_3/2 232.6 eV and 3d_5/2 230.3 eV), and metallic Mo (3d_3/2 231.1 eV). It could be observed that MoO₃ was the primary molybdenum species. At pH 7, the passive film consisted of metallic Mo, MoO₂, and MoO₃ and the relative content of MoO₂ increased with the sputtering depth. As the pH increased to 9, the outermost layer was only composed of MoO₃. However, MoO₂ appeared at the depth of 1 nm. The results suggested that MoO₂ might be more favoured in the inner layer. At pH 11, MoO₂ signal was not detected and MoO₃ was the only molybdenum species in the passive film. At pH 13, the signal-to-noise ratio was low, which suggested the low content of Mo.

Figure 12 shows the high-resolution XPS spectra of O 1s obtained on the specimens after the potentiostatic tests. The O 1s spectra were characterized by two components, corresponding to O^2- (530.2 ± 0.1 eV) and OH^- (531.5 ± 0.1 eV). The high content of OH^- was attributed to the formation of Cr(OH)₃ and FeOOH, while O^2- was the constituent of the passive film, which was related to the formation of Cr₂O₃, iron oxides, and molybdenum oxides. As could be seen, the area of the peaks corresponding to OH^- was larger than that to O^2- in the outermost layer, indicating that hydroxides were the main composition of the outer passive films. As the sputtering depth increased to 1 nm, the area ratios of O^2-to OH^- increased, which suggested that the relative content of oxides increased in the inner film.

4 Discussion

4.1 Pitting Corrosion and OER: Thermodynamics and Kinetics Study

Figure 13 shows the characteristic potentials extracted from the cyclic polarization curves, including corrosion potentials (E_corr), pitting potentials (E_p), and oxygen evolution potentials (E_ox). In general, the corrosion potentials decreased from - 0.247 to - 0.435 V (SCE) as the pH increased from 7 to 13.5. The decrease in corrosion potential was related to the negative movement of the equilibrium potential of cathodic reaction [4, 14]. Conversely, the pitting potentials increased with the pH, as previously reported [20, 21]. This phenomenon was probably related to the following aspects: (1) the local acidity was relatively more difficult at higher pH, and (2) more OH^- anions would compete with Cl^- in adsorption on the sample surface at higher pH. Moreover, the oxygen evolution potentials decreased from 0.561 to 0.550 and then dropped to 0.458 V (SCE) as the pH increased from 12.5 to 13 and then 13.5. This trend was in accordance with that of the theoretical E_ox, which was indicated by the dashed line. The large overpotential for oxygen evolution was due to the stepwise four-electron oxidation process [22, 23]. It was noted that the oxygen evolution potentials were more negative than the pitting potentials when the pH value was higher than 12.5. Therefore, oxygen evolution was more favoured at higher pH values and this resulted in the absence of the hysteresis loop at pH 13 and 13.5.

The pitting corrosion and OER were compared by the cyclic polarization results from the thermodynamics aspect. The galvanostatic tests were conducted under various pH conditions to investigate the kinetics of the two reactions. The applied current densities were set to be 1 mA/cm². This current density was selected since it was high enough for both the pitting corrosion and the OER. Figure 14 shows the E-t curves obtained for 316L stainless steel measured under various pH conditions. At the very beginning, the potentials increased dramatically, which was related to charging of the electric double layer at the specimen/solution interface. At pH 7, 9 and 11, the E-t plots presented three regions, i.e. the high potential region (Stage 1), the transition region (Stage 2), and the low potential region (Stage 3). In Stage 1, the potentials stabilized at approximately 1.4 V (SCE). At this period, oxygen was generated on the sample surface. However, the potentials sharply dropped to around 0.5 V (SCE) at 506, 607 and 825 s at pH 7, 9 and 11, respectively. The decrease in potentials was attributed to the conversion from oxygen evolution to pitting corrosion. Pitting corrosion was somewhat postponed due to the occurrence of OER. Since the pitting initiation was more difficult at the higher pH, the width of Stage 1 broadened with pH. Thereafter, the sample suffered pitting corrosion at Stage 3. As the pits propagated, the potentials decreased with time and stabilized at approximately 0.3 V (SCE). At pH 13, once the current density was applied, the potential increased dramatically and then stabilized at approximately 0.6 V (SCE), which was much more negative than the potentials measured at pH 7, 9, and 11. Oxygen evolution was observed during the testing, and no pit was found after the test.

The corrosion behaviour of 316L stainless steel during the galvanostatic tests is illustrated in Fig. 15. The anodic polarization curves of 316L stainless steel under these conditions were plotted as the black lines. At pH 7, 9, and 11, the oxygen evolution potentials were more positive than pitting potentials and therefore the dashed line indicated the polarization curve if no pitting corrosion occurred. Once the current density was applied on the work electrode, pitting corrosion should have been more favoured from the perspective of thermodynamics since the pitting corrosion potential was more negative. However, OER preferentially occurred in all cases, indicating that OER was kinetically faster than pitting corrosion. This resulted in the high potentials at pH 7, 9 and 11 in Stage 1. Since the potentials at this period were more positive than pitting potentials, pitting corrosion occurred afterwards, leading to the significant decrease in potentials from E_ox′ to E_p′ (transition region). As the pits propagated, the potentials decreased with time (Stage 3). At pH 13, the pitting potential was more positive than oxygen evolution potential. Once the current density was applied on the work electrode, OER would preferentially occur due to the thermodynamic and kinetic advantage. Therefore, the potential stabilized at 0.6 V (SCE) and pit was not observed on the sample surface after the test.

4.2 Discussion on the Cyclic Polarization Curves Measured at the Critical pH

As discussed in Sect. 4.1, OER would preferentially occur when it was thermodynamically possible for both reactions (OER and pitting corrosion). Meanwhile, the occurrence of OER would postpone pitting corrosion, presenting a broader Stage 1 at higher pH. According to the potential analysis, the pitting potential was slightly more positive than oxygen evolution potential at pH 12.5, and accordingly in Fig. 16 the dash line indicated the polarization curve if no oxygen evolution occurred. The black and blue lines show the forward and reversing plots, respectively. During the forward scan, E_ox and E_p were reached successively. At E_ox, only OER could occur. As the scan continued, the pitting potential was reached. However, pitting corrosion would not immediately occur due to the hindering effect caused by OER. If it took long enough for the pitting initiation, no hysteresis loop would be observed. On the contrary, if pitting corrosion occurred during this period, a hysteresis loop would appear. The time it took for pitting initiation at the critical pH might also be significantly influenced by the surface condition, for example, inclusions. Liu et al. [24, 25] have investigated the negative effect of inclusions on the pitting corrosion of 316L. The various surface conditions resulted in the different reversing behaviours in the actual cyclic polarization plots at the critical pH.

4.3 Enrichment of the Cation Elements in the Passive Film

According to the XPS analysis, the passive films formed at various pH were mainly composed of Cr₂O₃, Cr(OH)₃, Fe₃O₄, Fe₂O₃, FeOOH, MoO₃, and MoO₂. The atomic concentrations of Fe, Cr, Mo, and Ni in the passive film are shown in Fig. 17. The Ni content in the passive film was very low, and thus the corresponding high-resolution XPS spectra were not present. Similar results were reported by Hamada et al. [26] and Oh et al. [3]. Hamada et al. [26] demonstrated that the oxygen affinity of Ni was lower compared to that of Fe and Cr. Therefore, Ni remained in the metal phase and hardly participated in the passive film formation. Moreover, the total at.% of Fe and Cr was more than 80%, implying that Fe and Cr species were the primary constituents of the passive films, regardless of the pH values. The at.% of Cr and Mo decreased as the sputtering depth increased, which indicated that Cr and Mo were relatively enriched in the outer layer. Conversely, Fe content increased with the sputtering depth, which was in accordance with the previous result [16]. Besides, the at.% of Fe increased with the pH, which was due to the fact that Fe was more stable at the higher pH. However, the at.% of Mo and Cr in the passive film decreased as the pH increased, revealing the decreased stability of these species at more alkaline pH [14, 27]. The presence of Mo in the passive film could decrease the activity of the metallic iron, leading to the enrichment of chromium in the passive film [28]. Thus, the higher concentration of Mo at lower pH was somewhat responsible for the enrichment of Cr.

An enrichment factor F_Me was employed to characterize the tendency of cations to remain in the oxide film [29]. The factor was defined as:

$F_{\text{Me}} = \frac{{{\text{Me }}\,{\text{in}}\,{\text{oxide}}\,{\text{film}},\, {\text{wt\% }}}}{{{\text{Me }}\,{\text{in }}\,{\text{alloy}}, \,{\text{wt\% }}}},$ (2)

where F_Me is the enrichment factor for a metal element (Me). Me% is the weight percentage of Me to all the concerned cation elements. Me is enriched in the oxides, if F_Me > 1. Me is depleted in the oxides when F_Me < 1. If F_Me = 1, the concentration of Me in the oxide remains the same level as the bulk alloy.

The enrichment factors of Fe, Cr, Mo and Ni in the passive film are shown in Fig. 18. In general, Cr and Mo were enriched, while Fe and Ni were depleted in the passive films, regardless of the pH. It was commonly believed that the corrosion resistance of stainless steels was largely related to the composition of the passive films. Since Ni was depleted in the passive film (F_Ni < 0.2), it was reasonable to decrease the Ni content in materials and invent some nickel-saving stainless steels. The attempts to invent nickel-saving stainless steels were reported in the literature [30, 31, 32]. For Mo, the enrichment factor decreased dramatically from approximately 7 to less than 1, indicating that Mo might be unstable under strong alkaline conditions. It was well known that Mo could ameliorate the defect structure of chromium species and therefore the addition of Mo could increase the corrosion resistance of the materials [28]. However, under strong alkaline conditions Mo selectively dissolved and was depleted in the passive film. Accordingly, the addition of Mo in materials might not be beneficial to the corrosion resistance improvement in strong alkaline solutions. Actually, some investigations even demonstrated the negative effect of Mo addition at strong alkaline pH. Mesquita et al. [33] found that under alkaline conditions, Mo-containing austenitic stainless steels were even less corrosion resistant than Mo-free ones. Besides, the pitting potential of Mo was significantly more negative than those of pure Fe, Cr, and Ni under strong alkaline and S-free conditions [34] and accordingly in literature [20], the pitting potential for Mo-containing stainless steels was more negative.

The protectiveness of a passive film is largely owing to its composition and structure. Since chromium oxides were more stable than iron oxides, the atomic ratio of Cr to Fe in the passive film was always used as the assessment criteria for the film protectiveness [14, 26, 35, 36, 37]. Accordingly, the decrease in Cr content and increase in Fe content induced the degradation of the passive film, resulting in the higher passive current densities at more alkaline pH. Besides, the strong hydroxylation trend of the passive film at higher pH was also responsible for the passive film degradation, since hydroxides were less protective than the corresponding oxides [19].

5 Conclusions

The effect of pH on the corrosion behaviour of 316L stainless steel in alkaline solutions was studied using cyclic polarization, potentiostatic, EIS and galvanostatic measurements. The passive films formed at various pH values were characterized by XPS. The results were summarized as follows:

1. As the pH increased, the pitting potentials increased while the oxygen evolution potential decreased. The conversion from pitting corrosion to OER occurred at pH 12.5. The pH value had a significant effect on the electrochemical behaviour of 316L stainless steel.

2. The galvanostatic tests demonstrated that OER was kinetically faster than pitting corrosion, regardless of the pH. Moreover, OER would postpone the occurrence of pitting corrosion, which resulted in that the pitting initiation began during the reversing scan in the cyclic polarization.

3. The passive films formed at various pH were mainly composed of Cr₂O₃, Cr(OH)₃, Fe₃O₄, Fe₂O₃, FeOOH, MoO₃, and MoO₂. Hydroxides were more favoured at higher pH. The area ratios of O^2- to OH^- increased as the sputtering depth increased to 1 nm, indicating that the relative content of oxides increased in the inner film.

4. The atomic concentrations of Cr and Mo in the passive film decreased with pH and the sputtering depth, while the Fe content increased. Since chromium oxides were more stable than iron oxides, the lower Cr and Mo content at higher pH was responsible for the degradation of the passive film.

Acknowledgements This work was supported by the technology projects of State Grid Corporation (No. 52110417000N) and the National Science and Technology Major Project (No. 2016ZX05028-004).

The authors have declared that no competing interests exist.

---