The effect of chloride, sulfate, and ammonium ions on the semiconducting behavior and corrosion resistance of AISI 304 stainless steel passive film

ABSTRACT Water cooling systems usually receive additional chemical treatment and different features to prevent pitting corrosion on stainless steel which is dependent on medium factors such as species concentration. Chloride is an aggressive ion with which sulfate can act as an inhibitor. By applying cyclic voltammetry, electrochemical impedance spectroscopy (EIS), and Mott-Schottky plots, the behavior of 304 stainless steel (304 SS) was evaluated in solutions containing chloride, sulfate and ammonium ions. The polarization curves indicated the inhibiting effect of sulfate ions against pitting corrosion and the increase of the repassivation potential with increasing ammonium concentration. The most pronounced inhibiting effect occurred with a smaller ratio between chloride and sulfate ions (1:2). By EIS, the values obtained for charge transfer resistance of 304 SS sample corroborated the results of the cyclic voltammetry, ranging between 0.4 and 0.7 MΩcm2. The passive film showed semiconductor behavior, in agreement with the model of the chromium oxide inner layer and iron oxide outer layer. The passive films formed were highly doped, with doping density values on the order of 1020-21 cm–1. The difference between the corrosion potential and the flat band potential was found in the solutions where the material was less susceptible to corrosion.


INTRODUCTION
The electricity generation in Brazil is dominated by hydropower plants, but thermoelectrics represent 13% of the energy power sector which reached 621.2 TWh in 2020 [1].Due to the need for cooling to remove waste heat, thermoelectric power plants require large amounts of water and usually draw water from rivers or lakes to cool down their units [2,3].Water reuse practices consist of the use of alternative sources and the reduction of volumes captured [3,4].
In a recirculating cooling system, water is kept in a closed loop with a cooling tower and uses ambient air to cool its temperature.This kind of system draws much less water from the environment, compared to the open cycle, but consumes more than the closed cycle due to water evaporation [2][3][4].Blowing down is removing part of the circulating water and replacing it with fresh water to decrease the concentration of chemical substances.This process can help to slow down system corrosion.In industrial terms, the main requirements imposed on the quality of cooling water are low temperature, less formation of mineral deposits and biofouling, and prevention of corrosion processes in the equipment and pipelines [2,3,5].The cooling water from rivers or alternative sources usually receives additional treatment inside the plant such as biocide applications or acid addition for pH control [6,7].Thereafter, the ions resulting from the treatment as chloride and sulfate or ions from the source as ammonium still feature in the water [3,5].
Austenitic stainless steels (SS) are useful materials with many industrial applications as they are extensively used in cooling water systems [8,9] especially when ammonium is present, and the utilization of copper is not advisable [10].Their corrosion behavior is dependent on several medium factors such as ion concentration, temperature, and water velocity.Thus, the composition changes in the cooling water promoted by evaporation and blow down process can result in different corrosivity.
The influence of sulfate and chloride on the passive film properties and corrosion resistance of SS is associated with the ion concentration in the solution.Chloride is an aggressive ion since it has a critical level to promote pitting, which is very difficult to prevent [9,11].On the other hand, the corrosion risk decreases with increasing sulfate concentration which acts as a pitting inhibitor [12,13].The effect of ammonium still is not totally known, and fewer experimental studies are conducted in an environment containing ammonium, chloride, and sulfate.YANG et al. [14] reported that the presence of ammonium and chloride ions in supercritical water provided a severe corrosion environment for 316 SS promoting pitting and intergranular corrosion.According to TIAN et al. [15] the passive film formed on 316 SS in ammonium chloride (NH 4 Cl) solution is more compact and has a better barrier effect on the aggressive ions, compared with the sodium chloride (NaCl) solution with the same concentration.Conversely, the corrosion resistance of 316 SS is minor with the increase in the concentration of chloride and ammonium ions.
The corrosion resistance of SS is also dependent on the composition and structure of the passive film formed on the metal surface.SS passive films are often reported as having a duplex structure, consisting of an inner barrier of the oxide film and outer oxide or hydroxide film, acting as p-n heterojunction [16,17] When these oxides have perfect crystalline structure and stoichiometry, they act as insulators, but the presence of defects in the structure provides characteristics of extrinsic semiconductors when exposed to an aqueous solution [18].In this work, the Mott-Schottky equation (Equation 1) was applied to estimate the reciprocal of the square of the capacitance of the oxides formed on SS in different solutions.
Where: Nq is the doping density, ε is the dielectric constant, ε 0 denotes the vacuum permittivity, q is the elementary charge, kB is the Boltzmann constant, T denotes temperature, E fb is the flat band potential, and E is the applied potential.
It was possible to study the properties and fundamental parameters of the semiconductor behavior of the SS passive films and relate them to the corrosion resistance of SS [19][20][21][22][23][24].In the present work, measurements of Mott-Schottky plots, electrochemical impedance spectroscopy (EIS), and polarization curves were used to investigate the influence of chloride, sulfate, and ammonium ions on the semiconducting properties of passive films and the corrosion resistance of AISI 304 SS.
The solutions were prepared to simulate the cooling water conditions in a sodium borate buffer at pH 8.0 and varying concentrations of the reagents: NaCl, Na 2 SO 4 , and NH 4 OH, whose concentrations were reported in Table 1.
The electrochemical measurements were performed in a setup employing a three-electrode cell using a Gamry 3000 potentiostat.AISI 304 SS in a cylindrical shape with a 0.6 cm diameter, cold mounted in epoxy resin was featured as the working electrode, a platinum foil as the counter electrode, and a saturated calomel electrode (SCE) as the reference electrode.

Characterization of the SS
The SS samples were procured in a cylindrical format, possessing a diameter of 0.6 cm, and analyzed as received.
Characterization of the SS microstructure was carried out through metallography.The samples were wet ground with silica papers having granulometry of #180 -#240 -#400 -#600 -#1500.Afterward, the samples were polished with alumina 1.0 µm and finely polished with alumina 0.3 µm.Electrolytic etching was performed with a solution of 10.0% (w/w) oxalic acid with an applied potential of 6 V for 60 s.A Zeiss Axio Vert.A1 inverted optical microscope was used to obtain the micrographs.
Phase identification through X-ray diffractometry was also employed using a Bruker D8 Advance Eco diffractometer with Bragg Brentano geometry.The scanning was done in the 2θ range of 40° to 100°, at a scan rate of 6° min -1 using a Cu-Kα radiation source (λ = 1.5406Å).
The potentiodynamic polarization curves were obtained at a potential scan rate of 5 mV.s -1 with starting potential at -2.00 V, and when it reached +1.60 V, the direction was reversed until it reached the repassivation potential.For the EIS and capacitance measurements by Mott-Schottky plots, the passive film was formed by applying a constant potential of -0.10 V for 2 h.The EIS measurements were performed at the same film formation potential, applying an excitation voltage of 10 mV (amplitude) with a frequency ranging from 100 kHz to 10 -2 Hz.Mott-Shottky plots were obtained by starting the potential scan at -0.10 V and concluding the experiment at -1.50 V, with a 50 mV step at 2000 Hz.
All the experiments were carried out at room temperature and to ensure the repeatability of the experiment, the curves were performed leastwise 3 times.

Characterization
Figure 1 exhibits the optical micrography of the SS.Austenite grains containing shear bands (arrows with designation A) and twins (arrows with designation B) mainly constitute the microstructure of the AISI 304 SS [25][26][27].The presence of shear bands indicates that this material bar was strengthened through cold work.As is widely known, this can induce the formation of martensite and defects such as deformation bands through plastic deformation of the austenite grain, which can decrease its corrosion resistance [27,28].Moreover, the microstructure was also composed of rounded particles uniformly distributed, probably inclusions resulting from the SS bar manufacturing process [29,30].Nevertheless, the material exhibited a step structure in the austenite grain boundaries, indicating that the material was not sensitized.
The XRD analysis shown in Figure 2 indicated the material was mainly composed of austenite (γ) since only its main peaks were identified in the diffractogram.This confirms that the cold deformed state of the sample did not produce any martensite that could affect the corrosion resistance of the SS and that no significant deleterious phase content, such as intermetallic (σ phase) and carbide content, was identified.Therefore, the particles exhibited in the micrograph (Figure 1) were probably inclusions resulting from the manufacturing process.Moreover, the microstructure and phases of the studied material were in the as-received state, and hence faithfully represented the structure of the commercial SS widely applied in many industrial sectors.

Pitting corrosion
To understand the effects of the ions on the SS corrosion resistance, the results were separated into two different ammonium concentrations and ratios between the concentration of chloride and sulfate.Figure 3a shows the comparison between solutions C and E (Table 1), with the same concentrations of chloride and sulfate but different ammonium concentrations.The electrochemical processes were not significantly different and both ammonium concentrations showed similar corrosion potential (E corr ) and passive potential range.However, the repassivation potential (E repass ) values were different.The highest concentration had more difficulty in stopping the pitting processes and returning to the passivated state.
The substantial difference brought about by the variation in ammonia concentration was evident in the highlighted anodic peak in the figure.In the polarization curves an anodic peak was detected at 1.10 V and was attributed to a passive film change, resulting from the oxidation reaction from Cr (III) to Cr (VI), which occurred close to the pitting potential [31].The increase in the concentration of ammonium promoted a localized increase in the pH value, favoring the oxidation reaction of chromium and enhancing the peak current.According to TIAN et al. [15] initially, the ammonium ion contributes to the formation of the passive film on the SS surface, however, these cations hydrolyze and produce excess H + , resulting in the value of pH reduction which will promote the growth and propagation process of pitting.As depicted in Figure 3b, following this peak, a subsequent rise in current was observed as the scan reaches 1.37 V, coinciding with the initiation of pit corrosion.
The parameters obtained from polarization curves for SS in all solutions investigated were summarized in Table 2.The pitting potential (E pit ) is the potential value that stable pits start to grow, and repassivation (E repass )   1).
which is reached after reversal of the potential sweep direction is the potential value below which the already growing pits are repassivated and the growth is stopped.The maximum current is linked to the breakdown of the passive film and the highest level of corrosion.For Solution B, this value was attained at the scan's limit potential.In solutions with a higher concentration of sulfate ions, the breakdown of the passive film is hindered, resulting in less pitting corrosion on the material.Consequently, the measured current is also lower.
With constant ammonium concentration, the corrosion resistance of SS was dependent on the proportion of chloride and sulfate ions (Cl -:SO 4 2-).The passive range was smaller for solution B, with a higher ratio between chloride and sulfate ions (2:1).The pitting potential (E pit in red -Figure 3b) value was lower, about 0.7 V and the current density reached 9 mAcm -2 , indicating intense corrosion pitting.
Previous studies have reported that chloride and sulfate ions participate in the pitting corrosion process since chloride is an aggressive ion and sulfate is an inhibitor.For the sulfate inhibitor effect to occur, a lower chloride/sulfate ionic ratio is needed [13,16] The inhibitor action of the sulfate ions can be observed by increasing the value of the pitting potential to more positive values [13,32,33].
According to EL-EGAMY and BADAWAY [33], the inhibitor effect of sulfate does not include a stage of propagation of pitting; it only hinders the initiation process.This inhibitor effect on the pitting initiation could be observed by the E pit , changing the ratio Cl -:SO 4 2-2:1 to 1:1 or 1:2.However, in the solutions where the ratios between chloride and sulfate ions were 1:1 and 1:2, respectively, the pitting potential (E pit in black -Figure 3b) was the same as for 1.30 V.At this potential, the electrolyte reaction also took place and the local acidification from H + formation promoted the breakdown of passivation, so pitting corrosion occurred at the electrode.
Variations in the maximum current density and repassivation potential could be observed, indicating different intensities of pitting corrosion.At the 1:2 ratio in solutions C and E, the highest sulfate concentration promoted lower current density and more positive repassivation potential values.In the case of the 1:1 ratio, the corrosion increased with the total ion concentration, thus, in the more concentrated solution D the current reached a bigger value.
The pits on the surfaces of 304 SS generated at different chloride and sulfate ion concentrations with greater (solution B) and lower corrosion (solution C) were shown in Figure 4.The results were consistent with the potentiodynamic polarization curves.Solution B with ratio Cl -:SO 4 2-2:1 showed a greater number of pits as shown in Figure 4a and the width of the pinholes reached 68 µm (Figure 4b).On the other hand, solution C with ratio Cl -:SO 4 2-1:2 showed less and thinner (28 µm) pits as shown in Figures 4c and 4d.
The pitting number density difference was not obvious with the increase in ammonium concentration, indicating that the addition of NH 4 + had little influence on the transition of metastable to stable pitting corrosion [34].The growth rate of pits was controlled by diffusion, independent of the ions in the solution, thus in solution B the pitting corrosion starts before, with a smaller E pit resulting in greater pinholes.Pitting corrosion extended along the depth direction and along the circular direction.The presented morphology showed a single initiation pit surrounded by a few other small holes that assemble to form a larger hole [35].

Passive film properties
The passive film was potentiostatically formed at -0.10 V, which was below the repassivation for all solutions.The EIS measurements were performed at the passive film formation potential and the Nyquist plots are shown in Figure 5a.The equivalent circuit used to adjust the EIS data was shown in Figure 5b.This circuit was also simulated in previous works for SS in different medium [19,23,31].
The series resistance (R1) accounts for the resistance of the electrolyte, cables, and current collectors, while the charge transfer resistance (R2) refers to the charge transfer at the electrode/electrolyte interface, and the constant phase element (CPE) relates to the double-layer capacitance at the electrode/electrolyte interface (Q DL ).The model assumes that the passive film is composed of a non-homogeneous layer with the presence of defects.Together with the electrode's rough surface and some adsorbed species, the behavior can deviate from an ideal capacitor.The parameter α can range from 0 to 1, where unity represents a perfectly flat electrode.The parameters obtained through fitting EIS data are shown in Table 3.
R1 variations were related to solution concentration.R2 values, which were related to the resistance to 304 SS oxidation and its dissolution through the passive film, were about 0.4-0.7 MΩ.cm 2 , indicating greater corrosion resistance at the potential applied.The results for R2 in the different solutions agreed with the polarization curves.The charge transfer resistance was lower for solution B with Cl -:SO 4 2-2:1 and greater for Cl -:SO 4

2-
1:2 ratio.The increase in the ammonium concentration produced a less resistant passive film.The capacitances were calculated with Brug's Equation [36] and the obtained values remained virtually unchanged.
The Mott-Schottky plots (Figure 6) showed the variation of capacitance values with the potential applied, which was similar to the behavior reported in the literature for SS [17,18,37,38].Region I, from -0.10 V to -0.40 V, presented a positive slope, indicating an n-type semiconductor behavior of the iron oxide (γ-Fe 2 O 3 ).In Region (III), between -0.70 V and -1.05 V, a negative slope was noted, consistent with the p-type semiconductor behavior of the chromium oxide (Cr 2 O 3 ) inner layer.Region II was characterized by the flat band potential region and the capacitance was related to the Helmholtz layer.
The densities of dopants and the flat band potential were shown in Table 4.The dielectric constant (ε) used was 15.6 which was like the reported [15,20,39].The measured capacitance was associated with the space charge layer formed in the passive layer of the 304 SS on contact with the different solutions [40,41].The values of the donor densities were lower than those of the acceptor densities due to the difference in space charge layer thickness developed by the iron oxide and the chromium oxide layers [38].The higher concentration of species   1).
in the solution favored the formation of defects in the oxide structure and increased the number of dopants, aligning with the findings of [42].The small increase in the ammonium concentration almost did not change the dopant densities.The change in the number of dopants was more pronounced in the p-type chromium oxides as can be visualized by the slope of the linear region.Even though the densities of dopants did not exhibit an evident correlation with the corrosion resistance [15].Since the protection performance of SS passive film was closely related to chromium oxides, these changes could affect the corrosion.Solutions containing higher concentrations of sulfate had more negative flat-band potential values.Anions adsorbed on the passive film cause an increase in negative surface charge, which results in a decrease in the potential drop in the Helmholtz layer, resulting in a negative change in the flat band potential [43].
Solutions C and E showed less pitting corrosion, and the flat band potentials from p-type were more negative.When the corrosion potential is more positive than the flat band potential, the p-type semiconductor      will be in an accumulation situation and the n-type semiconductor will form a depletion layer that will inhibit electron transfer, which agrees with the EIS measurements [44].The corrosion resistance was more subject to this effect than the dopant densities for the evaluated concentrations.

CONCLUSIONS
This study was conducted in solutions containing ammonium, chloride, and sulfate to evaluate the effect of these ions on the semiconducting behavior and corrosion resistance of 304 stainless steel passive film.From the polarization curves, it was possible to observe the inhibiting effect of sulfate ions on SS.This effect was more pronounced in the solution where the ratio between the molar concentration of the aggressive chloride ions and the sulfate inhibitor ions was 1:2.On the other hand when the proportion was 2:1, there was a virtual absence of inhibiting effect.The corrosion behavior was not significantly different with the increase in the ammonium concentration, however, the repassivation potential (E repass ) presented a less positive value indicating more difficulty in stopping the pitting processes and returning to the passivated state.The EIS results agreed with the polarization curves, and the charge transfer resistance values were lower in the solutions with a higher ratio between chloride and sulfate ions (2:1).
The Mott-Schottky plots showed different semiconductor properties: an n-type behavior of the iron oxide (γ-Fe 2 O 3 ) outer layer and a p-type behavior of the chromium oxide (Cr 2 O 3 ) inner layer.The presence of sulfate ions in greater molar proportion concerning chloride ions promoted the formation of a more resistive passive film, regardless of the defect density in the crystal structure.The higher concentration of sulfate promotes a pitting inhibitory effect, due to preferential sulfate adsorption, but it also made the passive film more resistant due to the more negative flat band potentials.
Therefore, the control of species in recirculating water needs to be done in terms of concentration and in proportion between them.

Figure 1 :
Figure 1: Optical micrographs of the SS, electrolytically etched with 10% oxalic acid at 6 V for 60 s.Magnification of 200×.Austenite grains with shear bands (A) and twins (B).

Figure 2 :
Figure 2: X-Ray diffractogram of the SS in the 2θ range of 40° to 100°.The γ symbol represents the austenite phase.

Figure 3 :
Figure 3: Polarization curves (5 mV•s -1 ) of SS at (a) different ammonium ion concentrations and (b) different chloride and sulfate ion concentrations, where C, E, B, D, and A were the concentrations of the solutions used (Table1).

Figure 5 :
Figure 5: (a) Nyquist plots and (b) equivalent circuit for adjustment of the EIS measurements performed in different ammonium, chloride, and sulfate concentrations, where A, B, C, D, and E were the concentration of the solutions used (Table1).

Figure 6 :
Figure 6: Mott-Schottky plots for 304 SS in different ammonium, chloride, and sulfate concentrations, where A, B, C, D, and E were the concentrations of the solutions used (Table1).

Table 1 :
Solution compositions to study the effect of chloride, sulfate, and ammonium ions on the semiconducting behavior of SS passive film.

Table 2 :
Corrosion parameters obtained from polarization curves for SS in solutions A-E with different chloride, sulfate, and ammonium concentrations.

Table 3 :
Parameters obtained through fitting EIS data for 304 SS in solutions A-E with different chloride, sulfate, and ammonium concentrations.

Table 4 :
Semiconducting parameters from Mott-Schottky plots for 304 SS in solutions A-E with different chloride, sulfate, and ammonium concentrations.