Growth of Anodic Layers on 304L Stainless Steel Using Fluoride Free Electrolytes and Their Electrochemical Behavior in Chloride Solution

Anodic layers have been grown on 304L stainless steel (304L SS) using two kinds of fluoride-free organic electrolytes. The replacement of NH4F for NaAlO2 or Na2SiO3 in the glycerol solution and the influence of the H2O concentration have been examined. The obtained anodic layers were characterized by scanning electron microscopy (SEM), energy-dispersive X-ray spectroscopy (EDX), X-ray photoelectron spectroscopy (XPS), Raman spectroscopy, and potentiodynamic polarization tests. Here, it was found that, although the anodic layers fabricated within the NaAlO2-electrolyte and high H2O concentrations presented limited adherence to the substrate, the anodizing in the Na2SiO3-electrolyte and low H2O concentrations allowed the growth oxide layers, and even a type of ordered morphology was observed. Furthermore, the electrochemical tests in chloride solution determined low chemical stability and active behavior of oxide layers grown in NaAlO2-electrolyte. In contrast, the corrosion resistance was improved approximately one order of magnitude compared to the non-anodized 304L SS substrate for the anodizing treatment in glycerol, 0.05 M Na2SiO3, and 1.7 vol% H2O at 20 mA/cm2 for 6 min. Thus, this anodizing condition offers insight into the sustainable growth of oxide layers with potential anti-corrosion properties.


Introduction
Significant progress has been reached in nanoscience and nanotechnology fields to improve the properties, performance, and durability of materials. Mainly, surface and interface engineering has developed surface properties in metallic materials [1].
One of the most common surface modification methods is the anodizing process, which provides a unique combination between functionality and surface morphology, different from bulk material [2]. During the anodizing, a target metal is connected to the positive terminal of a power supply unit (anode) and a platinum or very stable metal to the negative terminal of a power supply (cathode), and a direct current is applied to an electrolyte [3]. Due to the applied stimulus, the metal is oxidized. It generates metal cations that migrate outward up to the metal/electrolyte interface, which react with the anions (e.g., O 2 − , OH − ) coming from the electrolyte. As a result, an oxide layer begins to form on the surface of the metal. Concurrently, the migration of these ions continues through the oxide developed, enabling the growth of the anodic layer at both the metal/oxide and oxide/electrolyte interface [4]. Their structures and chemical properties depend on several Before anodizing, the samples were manually polished using successive grades of SiC paper to 2000 grit, rinsed and cleaned with ethanol and distilled water, then dried in the air stream.

Anodizing Conditions
The anodizing was carried out using a two-electrode cell with the 304L SS sample as anode and a platinum electrode as a cathode located at a distance of~15 mm from each other. A freshly prepared electrolyte was used for each treatment, maintaining a solution volume/sample area ratio of 38 mL/cm 2 . The electrolytes used were organic glycerol-based solutions containing NaAlO 2 or Na 2 SiO 3 and different H 2 O concentrations, as outlined in Table 1. In addition, the treatments were carried out at 25 • C under galvanostatic conditions with three different values of current density (i) and anodizing time ( Table 1). The voltage-time curves of the above anodizing treatments that follow a similar behavior to others carried out under the same conditions are annexed in the Supplementary Material. After anodizing treatment, the samples were rinsed with distilled water and ethanol to remove the remaining ions from the solution, then dried in an airstream.

Anodic Layers Characterization
The morphological characterization of the anodic layers was examined by field emission gun scanning electron microscopy (FEG-SEM) using a Hitachi S-4800 and equipped with energy-dispersive X-ray spectroscopy (EDX, Manufactured by Hitachi Ltd, Tokyo, Japan), which operated at 15 keV for EDX analysis and 7 keV for secondary electron imaging. Raman spectra were recorded in a Thermo Scientific DRX Raman (Manufactured by Thermo Fisher Scientific, Waltham, MA, USA) with a HeNe gas laser (λ = 633 nm) and power of 2 mW. Each spectrum was recorded by an accumulation of 10-50 scans in a range of 50-1500 cm −1 . Chemical surface composition was analyzed by X-ray photoelectron spectroscopy using a Thermo Scientific K-Alpha (Manufactured by Thermo Fisher Scientific) equipped with Al Kα radiation (hν = 1486.6 eV) and operated at 12 kV, and 40 W. XPS spectra were obtained using the small windows mode with 0.1 eV/step and pass energy of 50 eV. Binding energies of the photoelectrons were calibrated using a contaminant carbon peak (C1s at 284.8 eV), and the spectra were analyzed with the Thermo Scientific Advantage data system (Manufactured by Thermo Fisher Scientific) for surface analysis. The X-ray diffraction (XRD) pattern was obtained using a Panalytical Xpert Pro diffractometer (Manufactured by Malvern Panalytical a spectrics company, Malvern, UK) with Cu-Kα (1.5406 Å), acceleration voltage of 45 kV, and working current of 40 mA.
On the other hand, corrosion behavior was evaluated by potentiodynamic polarization using a Gamry Reference 600 potentiostat (Manufactured by GAMRY Instruments, Warminster, PA, USA). Electrochemical measurements were carried out in duplicate (less in aluminate solutions) using a conventional three-electrode cell with a 0.3 M NaCl solution at 25 • C. The working electrodes were the 304L SS samples, the reference electrode was a Ag/AgCl (3 M KCl) electrode, and the counter electrode was a platinum wire. Before the tests, the samples were masked with Red Stop-Off Lacquer (product code: 166054, MacDermid Española S.A, Barcelona, Spain), leaving exposed only the area to be evaluated (~0.5 cm 2 ). The potentiodynamic curves were conducted at a scan rate of 0.16 mV/s. The potential scan was performed in the anodic direction from a cathodic potential value of 300 mV with respect to the corrosion potential to 1 V with respect to the reference electrode or until the current density had reached a limit value of 0.25 mA/cm 2 .

Characterization of the Anodic Layers Obtained in NaAlO 2 Electrolytes
The voltage-time curves and the macroscopic images of the surface appearance after anodizing processes in NaAlO 2 electrolytes with different concentrations of H 2 O are shown in Figure S1 and Figure 1, respectively. It can clearly be seen that the H 2 O content has an important influence on the growth of the anodic layers since more compact and adherent oxide films were obtained when its percentage decreases. Indeed, treatments with percentages of H 2 O ≥ 50 vol% presented a loose layer on their surface, while the oxide layer formed in 10 vol% was more uniform and coarser. Additionally, no significant effect was observed on the layers for anodizing using different current densities (10-20 mA/cm 2 ). On the other hand, corrosion behavior was evaluated by potentiodynamic polarization using a Gamry Reference 600 potentiostat (Manufactured by GAMRY Instruments, Warminster, USA). Electrochemical measurements were carried out in duplicate (less in aluminate solutions) using a conventional three-electrode cell with a 0.3 M NaCl solution at 25 °C. The working electrodes were the 304L SS samples, the reference electrode was a Ag/AgCl (3 M KCl) electrode, and the counter electrode was a platinum wire. Before the tests, the samples were masked with Red Stop-Off Lacquer (product code: 166054, Mac-Dermid Española S.A, Barcelona, Spain), leaving exposed only the area to be evaluated (~0.5 cm 2 ). The potentiodynamic curves were conducted at a scan rate of 0.16 mV/s. The potential scan was performed in the anodic direction from a cathodic potential value of 300 mV with respect to the corrosion potential to 1 V with respect to the reference electrode or until the current density had reached a limit value of 0.25 mA/cm².

Characterization of the Anodic Layers Obtained in NaAlO2 Electrolytes
The voltage-time curves and the macroscopic images of the surface appearance after anodizing processes in NaAlO2 electrolytes with different concentrations of H2O are shown in Figure S1 and Figure 1, respectively. It can clearly be seen that the H2O content has an important influence on the growth of the anodic layers since more compact and adherent oxide films were obtained when its percentage decreases. Indeed, treatments with percentages of H2O ≥ 50 vol% presented a loose layer on their surface, while the oxide layer formed in 10 vol% was more uniform and coarser. Additionally, no significant effect was observed on the layers for anodizing using different current densities (10-20 mA/cm 2 ). On the other hand, a lumpy precipitate was generated under all anodizing conditions, perhaps due to the decomposition and agglomeration of the aluminate compounds within the electrolyte (Figure 2a). Such precipitate could contain relatively unstable aluminate ions that originate from the dissociation of NaAlO2 in the electrolyte. Previous studies by Carreira et al. [31] concluded through the Raman and IR techniques that there are two species of aluminate ions: 1) Al(OH)4 -, predominant in a pH range between 8-12, and 2) AlO2 -, abundant in solutions with a pH above 12.5. However, Yerokhin et al. [32] affirmed that the main aluminum species in concentrated NaAlO2 solutions is Al(OH)4 -. Considering this, Cheng et al. [33] associated the hydrolysis of NaAlO2 and its decomposition into aluminate ions during micro-arcs oxidation (MAO) treatments with Equation On the other hand, a lumpy precipitate was generated under all anodizing conditions, perhaps due to the decomposition and agglomeration of the aluminate compounds within the electrolyte (Figure 2a). Such precipitate could contain relatively unstable aluminate ions that originate from the dissociation of NaAlO 2 in the electrolyte. Previous studies by Carreira et al. [31] concluded through the Raman and IR techniques that there are two species of aluminate ions: (1) Al(OH) 4 − , predominant in a pH range between 8-12, and (2) AlO 2 − , abundant in solutions with a pH above 12.5. However, Yerokhin et al. [32] affirmed that the main aluminum species in concentrated NaAlO 2 solutions is Al(OH) 4 − . Considering this, Cheng et al. [33] associated the hydrolysis of NaAlO 2 and its decomposition into aluminate ions during micro-arcs oxidation (MAO) treatments with Equations (1) and (2). These authors affirmed that if the electrolyte is decomposed, blocks of precipitated Al(OH) 3 can be easily found in the bottom of the reaction vessel [34], similar to those obtained in this study. Moreover, it is well known that Al(OH) 3 is transformed into AlOOH (Equation (3)) and then converted to Al 2 O 3 (Equation (4)) [35]. The XRD pattern confirmed that the anodic layer grown in NaAlO 2 contained a mixture of Al 2 O 3 , AlOOH, and FeOOH but exhibited limited adhesion to the 304L SS ( Figure 2b).

5
tern confirmed that the anodic layer grown in NaAlO2 contained a mixture of Al2O3, AlOOH, and FeOOH but exhibited limited adhesion to the 304L SS ( Figure 2b). Regarding the SEM analysis, the only adherent oxide layer fabricated in this type of electrolyte is presented in Figure 3a. The result corroborated the lack of surface uniformity mentioned above by observing the coated and devoid areas of the anodic layer. These zones can result from the evolution of fine bubbles formed as a result of the release of oxygen during the anodizing process, hindering contact of the surface with the electrolyte and, therefore, the growth of the layer in those areas.
The electrochemical stability of the anodic films was determined by potentiodynamic polarization curves shown in Figure 3b. The non-anodized 304L SS substrate evidences a corrosion potential (Ecorr) of −63.2 ± 3.1 mV vs. Ag/AgCl (3 M KCl), a pitting potential (Epitt) of 271.7 ± 36.3.6 mV vs. Ag/AgCl (3 M KCl) and a passive current density ( pass) of 7.75 × 10 −7 ± 8.00 × 10 −8 A/cm 2 . In contrast, the anodic branch of the film at 20 mA/cm 2 and 50 vol% H2O describes a more active behavior than the non-anodized 304L SS, while the other treatments show a similar electrochemical behavior than the substrate with pass in the same range of magnitude, but with a smaller passive branch length. In general, all polarization curves move towards negative potentials, both Ecorr and Epitt. The Ecorr for the anodized ones varies between ~−199.1 and ~−68.3 mV vs. Ag/AgCl (3 M KCl), while the Epitt were between ~51.3 and ~98.1 mV vs. Ag/AgCl (3 M KCl). Moreover, the pass of the anodic layers varies between ~1.86 × 10 −7 and ~9.86 × 10 −7 A/cm 2 . These results indicate that the anodic layers exhibit a higher susceptibility to localized corrosion than the non-anodized specimen but have a similar corrosion rate. Regarding the SEM analysis, the only adherent oxide layer fabricated in this type of electrolyte is presented in Figure 3a. The result corroborated the lack of surface uniformity mentioned above by observing the coated and devoid areas of the anodic layer. These zones can result from the evolution of fine bubbles formed as a result of the release of oxygen during the anodizing process, hindering contact of the surface with the electrolyte and, therefore, the growth of the layer in those areas. This behavior can be attributed to the cavities observed in the SEM images and the poor adhesion of the layers in most of the conditions evaluated with this electrolyte. According to Yang et al. [36], a hole that permits direct contact between the substrate and the electrolyte decreases the anticorrosive performance since it acts as a channel, allowing a pitting corrosion mechanism to develop. In addition, the electrochemical behavior of the anodic layers can also be associated with the absence of a stable barrier layer in the metal/oxide interface, as mentioned by Andrei et al. [28,37]. They advised the formation of a barrier layer of iron oxide without Ni and Cr before the advanced anodizing treatments (MAO) to obtain coatings on austenitic stainless steel using aqueous solutions of NaAlO2. Thus, it is likely to be required to get effective results in conventional anodizing as well.

Characterization of the Anodic Layers Obtained in Na2SiO3 Electrolytes
The voltage-time curves and the macroscopic images of the surface appearance of anodic layers grown at different current densities in a glycerol solution containing 0.1 M Na2SiO3 and 2.5 vol% H2O are shown in Figure S2 and Figure 4, respectively.  The electrochemical stability of the anodic films was determined by potentiodynamic polarization curves shown in Figure 3b. The non-anodized 304L SS substrate evidences a corrosion potential (E corr ) of −63.2 ± 3.1 mV vs. Ag/AgCl (3 M KCl), a pitting potential (E pitt ) of 271.7 ± 36.3.6 mV vs. Ag/AgCl (3 M KCl) and a passive current density (i pass ) of 7.75 × 10 −7 ± 8.00 × 10 −8 A/cm 2 . In contrast, the anodic branch of the film at 20 mA/cm 2 and 50 vol% H 2 O describes a more active behavior than the non-anodized 304L SS, while the other treatments show a similar electrochemical behavior than the substrate with i pass in the same range of magnitude, but with a smaller passive branch length. In general, all polarization curves move towards negative potentials, both E corr and E pitt . The E corr for the anodized ones varies between~−199.1 and~−68.3 mV vs. Ag/AgCl (3 M KCl), while the E pitt were between~51.3 and~98.1 mV vs. Ag/AgCl (3 M KCl). Moreover, the i pass of the anodic layers varies between~1.86 × 10 −7 and~9.86 × 10 −7 A/cm 2 . These results indicate that the anodic layers exhibit a higher susceptibility to localized corrosion than the non-anodized specimen but have a similar corrosion rate.
This behavior can be attributed to the cavities observed in the SEM images and the poor adhesion of the layers in most of the conditions evaluated with this electrolyte. According to Yang et al. [36], a hole that permits direct contact between the substrate and the electrolyte decreases the anticorrosive performance since it acts as a channel, allowing a pitting corrosion mechanism to develop. In addition, the electrochemical behavior of the anodic layers can also be associated with the absence of a stable barrier layer in the metal/oxide interface, as mentioned by Andrei et al. [28,37]. They advised the formation of a barrier layer of iron oxide without Ni and Cr before the advanced anodizing treatments (MAO) to obtain coatings on austenitic stainless steel using aqueous solutions of NaAlO 2 . Thus, it is likely to be required to get effective results in conventional anodizing as well.

Characterization of the Anodic Layers Obtained in Na 2 SiO 3 Electrolytes
The voltage-time curves and the macroscopic images of the surface appearance of anodic layers grown at different current densities in a glycerol solution containing 0.1 M Na 2 SiO 3 and 2.5 vol% H 2 O are shown in Figure S2 and  This behavior can be attributed to the cavities observed in the SEM images and the poor adhesion of the layers in most of the conditions evaluated with this electrolyte. According to Yang et al. [36], a hole that permits direct contact between the substrate and the electrolyte decreases the anticorrosive performance since it acts as a channel, allowing a pitting corrosion mechanism to develop. In addition, the electrochemical behavior of the anodic layers can also be associated with the absence of a stable barrier layer in the metal/oxide interface, as mentioned by Andrei et al. [28,37]. They advised the formation of a barrier layer of iron oxide without Ni and Cr before the advanced anodizing treatments (MAO) to obtain coatings on austenitic stainless steel using aqueous solutions of NaAlO2. Thus, it is likely to be required to get effective results in conventional anodizing as well.

Characterization of the Anodic Layers Obtained in Na2SiO3 Electrolytes
The voltage-time curves and the macroscopic images of the surface appearance of anodic layers grown at different current densities in a glycerol solution containing 0.1 M Na2SiO3 and 2.5 vol% H2O are shown in Figure S2 and The development of a compact layer was favored just as the current density increased. For the anodizing at 20 mA/cm 2 , a uniform and adherent layer was observed on the surface of the 304L SS. In this case, SEM images ( Figure 5a) also reveal the formation of small craters in the oxide layer, possibly originated by eruptions during the process. However, unlike the cavities obtained in NaAlO2, these craters showed an oxide film inside ( Figure 5b). This outcome indicated that the formation of a barrier layer exists, which prevents direct contact of the metal with the surrounding medium (Figure 5b).  The development of a compact layer was favored just as the current density increased. For the anodizing at 20 mA/cm 2 , a uniform and adherent layer was observed on the surface of the 304L SS. In this case, SEM images (Figure 5a) also reveal the formation of small craters in the oxide layer, possibly originated by eruptions during the process. However, unlike the cavities obtained in NaAlO 2 , these craters showed an oxide film inside (Figure 5b). This outcome indicated that the formation of a barrier layer exists, which prevents direct contact of the metal with the surrounding medium (Figure 5b). The EDS results ( Figure S3) revealed that the anodic layer was composed of 4.28 wt.% C, 3.05 wt.% O, 0.65 wt.% Si, 17.84 wt.% Cr, 63.79 wt.% Fe, 8.23 wt.% Ni, and 2.16 wt.% Mn. The presence of C in the anodic layers is usually detected in anodizing with organic electrolytes due to its incorporation from the solution [15,20,21]. Moreover, the increase in Si content concerning the concentration of Si in 304L SS substrate can be attributed to the formation of a silica-based gel at the oxide/electrolyte interface that favors the mobility of ions such as SiO 3 2− , HSiO 3 − , Si 4+ , and retains them in its structure according to Mato et al. [38].
The potentiodynamic polarization curves of each anodic layer are presented in Figure 6. The anodic layers showed similar behavior among them and disclosed similar values of E corr around −198.4 ± 23.3 mV vs. Ag/AgCl (3 M KCl), E pitt about 37.0 ± 21.9 mV vs. Ag/AgCl (3 M KCl), and i pass around 3.12 × 10 −7 ± 3.25 × 10 −8 A/cm 2 . This implies that, regardless of the current density applied, the dissolution rate of the anodic layers in NaCl solution was similar to the non-anodized 304L SS. In addition, the corrosion behavior of the Na 2 SiO 3 anodic layers resembled the NaAlO 2 ones. Nevertheless, the decrease of the H 2 O concentration in the anodizing bath seems to produce a uniform anodic layer, possibly caused by the attenuation of the chemical dissolution mechanism in organic electrolytes as mentioned in the studies carried out by LaTempa et al. [39]. Likewise, Klimas et al. [21] pointed out that the growth of anodic layers on SS in organic electrolytes with fluorides is not possible when the H 2 O concentration exceeds 3 vol%; therefore, we decided to study the effect of the H 2 O concentration in the electrolyte containing Na 2 SiO 3 for this work. The EDS results ( Figure S3) revealed that the anodic layer was composed of 4.28 wt.% C, 3.05 wt.% O, 0.65 wt.% Si, 17.84 wt.% Cr, 63.79 wt.% Fe, 8.23 wt.% Ni, and 2.16 wt.% Mn. The presence of C in the anodic layers is usually detected in anodizing with organic electrolytes due to its incorporation from the solution [15,20,21]. Moreover, the increase in Si content concerning the concentration of Si in 304L SS substrate can be attributed to the formation of a silica-based gel at the oxide/electrolyte interface that favors the mobility of ions such as SiO3 2-, HSiO3 -, Si 4+ , and retains them in its structure according to Mato et al. [38].
The potentiodynamic polarization curves of each anodic layer are presented in Figure 6. The anodic layers showed similar behavior among them and disclosed similar values of Ecorr around −198.4 ± 23.3 mV vs. Ag/AgCl (3 M KCl), Epitt about 37.0 ± 21.9 mV vs. Ag/AgCl (3 M KCl), and pass around 3.12 × 10 −7 ± 3.25 × 10 −8 A/cm 2 . This implies that, regardless of the current density applied, the dissolution rate of the anodic layers in NaCl solution was similar to the non-anodized 304L SS. In addition, the corrosion behavior of the Na2SiO3 anodic layers resembled the NaAlO2 ones. Nevertheless, the decrease of the H2O concentration in the anodizing bath seems to produce a uniform anodic layer, possibly caused by the attenuation of the chemical dissolution mechanism in organic electrolytes as mentioned in the studies carried out by LaTempa et al. [39]. Likewise, Klimas et al. [21] pointed out that the growth of anodic layers on SS in organic electrolytes with fluorides is not possible when the H2O concentration exceeds 3 vol%; therefore, we decided to study the effect of the H2O concentration in the electrolyte containing Na2SiO3 for this work.  SEM images showed that the anodic layer grown at the lowest concentration of H 2 O presented a smooth surface with fewer defects (Figure 8a,b). However, rupture of the layers was observed when the concentration increased from 1.0 vol% to 1.7 vol% H 2 O, revealing the presence of apparent nanopores with small craters inside them (Figure 8c,d). Instead, micrometer-sized craters were formed in the anodic films fabricated at the higher H 2 O concentration (Figure 8e,f). This oxide film is similar to those described by Asoh et al. [12] in 304 SS using a H 2 SO 4 and H 2 O 2 containing electrolyte. In addition, the EDS analysis summarized in Figure S5 and Table 2 confirmed the presence of silicon in the oxide layer, probably coming from both the substrate and the electrolyte. The voltage-time curves and the surface appearance of the 304L SS samples after anodizing in the Na2SiO3 solution with different H2O concentrations and a constant current density of 20 mA/cm 2 are shown in Figure S4 and Figure 7, respectively. It can be seen that the H2O concentration plays an important role in the uniformity of the layers. The layer turned out to be more compact when H2O concentration was 1.7 vol%, while the surface coverage decreased for H2O concentrations of 1.0. vol% and 2.5 vol%. SEM images showed that the anodic layer grown at the lowest concentration of H2O presented a smooth surface with fewer defects (Figure 8a,b). However, rupture of the layers was observed when the concentration increased from 1.0 vol% to 1.7 vol% H2O, revealing the presence of apparent nanopores with small craters inside them (Figure 8c,d). Instead, micrometer-sized craters were formed in the anodic films fabricated at the higher H2O concentration (Figure 8e,f). This oxide film is similar to those described by Asoh et al. [12] in 304 SS using a H2SO4 and H2O2 containing electrolyte. In addition, the EDS analysis summarized in Figure S5 and Table 2 confirmed the presence of silicon in the oxide layer, probably coming from both the substrate and the electrolyte.  SEM images showed that the anodic layer grown at the lowest concentration of H2O presented a smooth surface with fewer defects (Figure 8a,b). However, rupture of the layers was observed when the concentration increased from 1.0 vol% to 1.7 vol% H2O, revealing the presence of apparent nanopores with small craters inside them (Figure 8c,d). Instead, micrometer-sized craters were formed in the anodic films fabricated at the higher H2O concentration (Figure 8e,f). This oxide film is similar to those described by Asoh et al. [12] in 304 SS using a H2SO4 and H2O2 containing electrolyte. In addition, the EDS analysis summarized in Figure S5 and Table 2 confirmed the presence of silicon in the oxide layer, probably coming from both the substrate and the electrolyte.    Therefore, the H 2 O concentration in the electrolyte resulted in a key factor for obtaining uniform films in Na 2 SiO 3 -electrolyte. Considering this, the results in Figures 7 and 8 confirmed that the H 2 O concentration of 1.7 vol% was the ideal condition to balance the mechanisms of formation and dissolution of the anodic layer in this type of electrolyte.

H2O Concentration
Regarding the Raman analysis, all the spectra of the anodic layers grown at 20 mA/cm 2 in the Na 2 SiO 3 -electrolyte with different H 2 O concentrations are presented in Figure 9a. It is observed that the anodizing treatments did not show important changes between them, so the spectrum of the oxide layer obtained in 1.7 vol% H 2 O is analyzed as an example in   Therefore, the H2O concentration in the electrolyte resulted in a key factor for obtaining uniform films in Na2SiO3-electrolyte. Considering this, the results in Figures 7 and 8 confirmed that the H2O concentration of 1.7 vol% was the ideal condition to balance the mechanisms of formation and dissolution of the anodic layer in this type of electrolyte.
Regarding the Raman analysis, all the spectra of the anodic layers grown at 20 mA/cm 2 in the Na2SiO3-electrolyte with different H2O concentrations are presented in Figure 9a. It is observed that the anodizing treatments did not show important changes between them, so the spectrum of the oxide layer obtained in 1.7 vol% H2O is analyzed as an example in Figure 9b. The higher intensity bands identified at 725 cm −1 (Fe 3+ ) and 550 cm −1 (Fe 2+ ) were associated with the Fe-O stretching mode. This bond occurs in compounds such as magnetite (Fe3O4), maghemite (γ-Fe2O3), and goethite (α-FeOOH) [40]. The low intensity bands at 1357 and 1446 cm −1 corresponded only to the α-FeOOH and γ-Fe2O3 species, respectively, The higher intensity bands identified at 725 cm −1 (Fe 3+ ) and 550 cm −1 (Fe 2+ ) were associated with the Fe-O stretching mode. This bond occurs in compounds such as magnetite (Fe 3 O 4 ), maghemite (γ-Fe 2 O 3 ), and goethite (α-FeOOH) [40]. The low intensity bands at 1357 and 1446 cm −1 corresponded only to the α-FeOOH and γ-Fe 2 O 3 species, respectively, indicating their presence in the anodic layer [41]. On the other hand, the presence of an intense band at 550 cm −1 could be assigned to Ni-O stretching modes of Ni(OH) 2 [42], which overlaps with iron oxides peaks. In addition, the bands at 575 and 497 cm −1 indicated that the formation of mixed oxides of Fe and Ni is possible [43]. The band at 939.6 cm −1 represented the Cr=O stretching mode [44], and it is related to the presence of chromium oxide in the anodic layers. Raman weak band around 1070 cm −1 should be assigned to the vibrational mode of the Si-O group, suggesting the formation of SiO 2 in the anodic layers [45].
XPS analysis was performed to provide information about the chemical composition of the layers grown in the Na 2 SiO 3 electrolytes with different H 2 O concentrations (Figure 10a). In all samples, the characteristic peaks of Fe, Cr, Si, O, and C were identified in the survey spectra. Only the layers that grew using H 2 O concentrations ≤1.7 vol% showed the Ni peak, which became more noticeable when H 2 O concentration was 1.7 vol% (Figure 10b,g).
The high-resolution window of Fe2p, Figure 10c, allowed identification of the presence of the Fe II -O bond at 710.7 eV, suggesting the formation of FeO or Fe3O4. Furthermore, the Fe III -O and Fe III -OH bonds at binding energies of 712.11 and 713.8 eV, respectively, are related to the possible presence of Fe2O3 and FeO(OH) in the layer. These results were in agreement with results published by Pawlik et al. [22] in anodizing on pure iron using organic electrolytes with fluorides.  The high-resolution window of Cr2p, Figure 10d, evidenced two peaks at binding energies of 576.6 and 578.8 eV, indicating the presence of both Cr III (78%) and Cr VI (22%) in the oxide, as hydroxide or oxyhydroxide form [20]. Furthermore, the incorporation of silicate species in the anodic layer was observed in the deconvolution of the Si2p peak in   [22] in anodizing on pure iron using organic electrolytes with fluorides.
The high-resolution window of Cr2p, Figure 10d, evidenced two peaks at binding energies of 576.6 and 578.8 eV, indicating the presence of both Cr III (78%) and Cr VI (22%) in the oxide, as hydroxide or oxyhydroxide form [20]. Furthermore, the incorporation of silicate species in the anodic layer was observed in the deconvolution of the Si2p peak in 101.4 eV, Figure 10e. This indicated the presence of Si (23%) associated at a binding energy of 98.3 eV and silicate (77%), which peak appears at 101.7 eV. These results are analogous to those reported by Hsiao et al. [46] in anodizing treatments on Mg alloys using electrolytes containing Na 2 SiO 3 . Finally, the high-resolution window of O1s, Figure 10f, revealed the presence of three peaks: the first one at 529.8 eV (22%) ascribed to the O-Metal bond that suggests the formation of Fe and Cr oxides; the second one at 531.7 eV (48%), associated to hydroxides of Fe and Cr; and the third one at 532.4 eV (30%) assigned to O-C bond [21]. Therefore, the above results indicated that the anodic layer was mainly composed of oxides, hydroxides, and oxyhydroxides of Fe, Cr, Si, and carbonates of Fe. Furthermore, it may also contain Ni species that, due to the detection limit of the equipment (0.2 at.%), was not possible to carry out the deconvolution of the Ni peak in the high-resolution window and determine its concentration in the anodic layers (Figure 10g).
The potentiodynamic polarization curves of the oxide layers fabricated in a Na 2 SiO 3 electrolyte with different H 2 O concentrations are presented in Figure 11. Changes in the polarization curves of the layers were shown regarding the non-anodized 304L SS. When the H 2 O concentration increased from 1.0 vol% to 1.7 vol% the i pass values varied from 3.85 × 10 −8 ± 1.17 × 10 −8 A/cm 2 to 9.45 × 10 −8 ± 2.55 × 10 −8 the E corr shifted from -162.35 ± 1.75 to −18.92 ± 5.41 mV vs. Ag/AgCl (3M KCl) and the E pitt changed from 42.0 ± 2.0 to 243.3 ± 6.5 mV vs. Ag/AgCl (3 M KCl). These findings indicated a better electrochemical response against corrosion for the anodic layer grown with 1.7 vol% H 2 O concentration. Moreover, it could be seen that the electrochemical behavior of the anodic layers worsened when H 2 O percentage increased from 1.7 vol% to 2.5 vol%, reaching values of 2.98 × 10 −7 ± 2.70 × 10 −8 A/cm 2 for i corr , −117.6 ± 1.1 mV vs. Ag/AgCl (3 M KCl) for E corr , and −57.5 ± 18.8 mV vs. Ag/AgCl (3 M KCl) for E pitt . These results imply that the anodic layers with better protective properties were obtained for 1.7 vol% H 2 O concentration in the electrolyte, improving the corrosion resistance of the non-anodized 304L SS substrate by approximately one order of magnitude.  This better corrosion behavior for anodic layers grown in the bath containing 1.7 vol% H2O may be associated with the presence of Ni in the oxide as it was found in the XPS spectra. It is commonly accepted that, in addition to Cr, Ni plays an important role in This better corrosion behavior for anodic layers grown in the bath containing 1.7 vol% H 2 O may be associated with the presence of Ni in the oxide as it was found in the XPS spectra. It is commonly accepted that, in addition to Cr, Ni plays an important role in austenitic SS to improve the corrosion resistance [47], while for steels with a high Ni content, this element is found in the native passive layer [48]. Figure 12 shows a schematic representation of Ni incorporation in the anodic layer during anodizing process. Figure 11. Potentiodynamic polarization curves for anodic layers grown in glycerol electrolyte con taining 0.05 M Na2SiO3 and different H2O concentrations. This better corrosion behavior for anodic layers grown in the bath containing 1.7 vol% H2O may be associated with the presence of Ni in the oxide as it was found in the XPS spectra. It is commonly accepted that, in addition to Cr, Ni plays an important role in austenitic SS to improve the corrosion resistance [47], while for steels with a high Ni content, this element is found in the native passive layer [48]. Figure 12 shows a schematic representation of Ni incorporation in the anodic layer during anodizing process. Firstly, there is an anion migration (O2 − , OH − , SiO3 2− ) from the electrolyte towards the SS interface as well as an ejection of cations from the material into the electrolyte triggered by the action of the electric field of the anodizing process ( Figure 12a). Secondly, such ion migration outwards could distribute Ni into the anodic film as follows: (1) Ni accumulation in the oxide/metal interface in metallic form and (2) Ni presence in the anodic layer as Ni oxide/hydroxide compounds (Figure 12b). The first possibility is reported in the literature by Olsson et al. [49], who explained that the less active potential of Ni favors 10 -13 10 -12 10 -11 10 -10 10 -9 10 -8 10 -7 10 -6 10 -5 10 -4 10 -3 10 -2  Firstly, there is an anion migration (O 2 − , OH − , SiO 3 2− ) from the electrolyte towards the SS interface as well as an ejection of cations from the material into the electrolyte triggered by the action of the electric field of the anodizing process ( Figure 12a). Secondly, such ion migration outwards could distribute Ni into the anodic film as follows: (1) Ni accumulation in the oxide/metal interface in metallic form and (2) Ni presence in the anodic layer as Ni oxide/hydroxide compounds (Figure 12b). The first possibility is reported in the literature by Olsson et al. [49], who explained that the less active potential of Ni favors preferential oxidation of Fe and Cr. Therefore, the anodic layer is mainly composed of compounds of Fe and Cr while a metallic Ni enrichment is generated in the oxide/metal interface that delays the dissolution of the material. The second possibility is a consequence of the oxidation of Ni cations that are eluted outside the surface during ionic migration, which allows the formation of Ni compounds such as NiO and Ni(OH) 2 [21].

Conclusions
Anodic layers were formed by a galvanostatic conventional anodizing process in fluoride-free electrolytes on 304L SS. The samples anodized in a based glycerol electrolyte with 0.3 M NaAlO 2 , and high H 2 O concentrations did not form uniform and protective oxide layers. In contrast, it was possible to obtain compact anodic layers using a glycerol electrolyte containing low concentrations of Na 2 SiO 3 and H 2 O.
The characterization of the anodic layers obtained in Na 2 SiO 3 solutions showed that they were mainly composed of oxides, hydroxides, and oxyhydroxides of Fe, Cr, and Si. However, XPS results showed the presence of Ni in the samples anodized at concentrations ≤1.7 vol% H 2 O. The presence of Ni influences the chemical stability, suggesting that its distribution within the anodic layer provides better corrosion behavior than 304 L SS.
Thus, we concluded that the most optimal condition of those evaluated in this study to enhance the behavior against corrosion was the anodizing in glycerol with 0.05 M Na 2 SiO 3 , and 1.7 vol% H 2 O for 6 min at 20 mA/cm 2 . This condition improved the electrochemical response one order of magnitude compared to the non-anodized 304L SS.