Corrosion Performance of Chemically Passivated and Ion Beam-Treated Austenitic–Martensitic Steel in the Marine Environment

Abstract

In the present work, chemical and ion beam surface treatments were performed in order to modify the electrochemical behavior of industrial austenitic–martensitic steel VNS-5 in 3.5 wt. % NaCl. Immersion for 140 h in a solution containing 0.05 M potassium dichromate and 10% phosphoric acid promotes formation of chromium hydroxides in the outer surface layer. By means of a new type of ion source, based on a high-current pulsed magnetron discharge with injection of electrons from vacuum arc plasma, ion implantation with Ar⁺ and Cr⁺ ions of the VNS-5 steel was performed. It has been found that the ion implantation leads to formation of an Fe- and Cr-bearing oxide layer with advanced passivation ability. Moreover, the ion beam-treated steel exhibits a lower corrosion rate (by ~7.8 times) and higher charge transfer resistance in comparison with an initial (mechanically polished) substrate. Comprehensive electrochemical and XPS analysis has shown that a Cr₂O₃-rich oxide film is able to provide an improved corrosion performance of the steel, while the chromium hydroxides may increase the specific conductivity of the surface layer. A scheme of a charge transfer between the microgalvanic elements was proposed.

1. Introduction

It is known that the characteristics of the corrosive environment, such as temperature, concentration of halogen anions, and pH level, are decisive for predicting and assessing the electrochemical behavior of stainless steels in the marine environment [1]. Based on the analysis of the scientific literature, austenitic steels, being the most important components of marine equipment and support tools, are very sensitive to ultraviolet radiation [2], mechanical stresses [3,4], and the presence of oxidizing agents [5]. The experimental examination of industrial steels exhibiting a high content of titanium, chromium, and nickel under natural climatic conditions (high humidity, direct contact with sea water, growth of bacteria) are performed to a very limited extent [6], and the main results on corrosion behavior and corrosion fatigue were obtained in laboratory conditions [7,8,9]. The corrosion durability of Cr-alloying steels is hard to predict reliably due to the fact that the chromium passivation in aqueous solutions is complex and consists of two main stages: water adsorption and anodic oxidation of chromium to oxides and hydroxides [10]. It was previously shown that the main reasons for the pitting resistance of Fe-Cr alloys were related to the crystallographic and electronic structure of the oxide film [11] and matrix phases [12].

Traditionally, the analysis of the accelerated corrosion destruction of austenitic steels is attributed to the number and size of pits [13], the rate of diffusion (transfer) of charges through the interfaces and ion transport between the corrosive environment and the substrate [14], bacterial activity [6], and the chemical state of the oxide layers [15]. To date, various methods of surface modification and passivation of stainless steels have been proposed. In particular, the deposition of thin (up to 200 nm) films consisting of chromium-based compounds (Cr₂O₃, (Fe,Cr)₂O₃, FeCrO₄) under constant external electrical voltage [16] is considered to be a promising duplex treatment. In order to synthesize corrosion-resistant layers onto the surface of austenitic steels, hot and cold quenching in air [17,18,19,20,21], chemical etching and electrochemical passivation [20,22,23,24], oxidation under alternating vacuum [25], physical vapor deposition [26,27], and ion implantation [28,29,30] are also used. However, durability of load-bearing parts and aircraft fasteners in the marine environment by implementation deposition of corrosion-resistant coatings cannot be achieved for a long-term period due to the limited adhesive strength of the “coating/substrate” system. Moreover, the protective coatings must be constantly renewed to prevent the development of either crevice or galvanic corrosion. Therefore, it is advisable to fabricate corrosion-resistant layers characterized by diffusion-type interfaces with the substrate material, i.e., having cohesive bonds with a substrate.

In our opinion, a modification of chromium–nickel steels by ion plasma and electrochemical methods [31,32,33] seems promising for increasing corrosion resistance without degradation of the grain structure that is responsible for the plasticity of the material. The adhesive strength of the “coating/substrate” system can be increased by ion-assisted deposition [34] or ion mixing [35]. The implementation of ion implantation for various substrates, including industrial steels and TiNi alloys, was examined [30,36,37,38,39,40]. In general, these studies demonstrate a positive effect of high-dose ion irradiation on the corrosion performance of the material. As a rule, the ion beam treatments promote a decrease in the corrosion current and shift the corrosion potential toward noble values. In particular, the ion implantation of steels with light (C, N) ions causes amorphization of the outer surface layer [36,37,41] due to the irradiation-induced phase transformations in the cascades of atomic collisions. In contrast, the Ti⁺-implantation of shape memory alloys [38] results in a decrease of the corrosion resistance despite the formation of a thicker TiO₂ oxide film. The reason for this negative result is due to the penetration of chlorine anions through the interface boundaries between the oxide particles into the Ni-rich sublayer. In turn, a defect-free glassy oxide film in the reference Nb⁺-implanted TiNi alloy shows the best corrosion resistance [39]. The in-depth redistribution of the elements is a common feature of the ion beam-treated alloys, exhibiting different self-passivation behavior and structure of the surface layer [30,42,43]. After irradiation, the new passive film is rich with implanted ions and could effectively hamper the ionic transport across the oxide in aqueous solutions. Also, the implantation of Al-bronze with Cr [44] inhibits a selective phase corrosion. However, the physical reasons for the improved corrosion performance of the irradiated alloys are not sufficiently studied. It is supposed that, in some cases, the ion beam treatment may promote the formation of a loose oxide layer with many defects, so the corrosion resistance may be reduced if the chosen irradiation dose is not optimal.

At present, the technical problems associated with improving the corrosion resistance of high-strength austenitic–martensitic steels in marine environments have been solved only partially [26,45]. To our knowledge, a fundamental issue is attributed to the relationship between the chemical composition, the thickness of oxide layers, and the corrosion properties of austenitic–martensitic steels. In this paper, an idea based on the implementation of pulsed ion beams for surface alloying with Cr of the industrial steel VNS-5 is proposed. It is expected that the application of the ion plasma method of surface modification is able to promote a significant enhancement of the service life of steel fasteners operating in direct contact with seawater. The purpose of this work is to determine the chemical state of the constituent elements (Fe, Cr, Ni) within the oxide layers of the austenitic–martensitic steel VNS-5, modified by chemical and ion beam methods, and to assess the effect of surface treatments on its electrochemical properties.

2. Materials and Methods

The substrate was an austenitic–martensitic steel VNS-5 (manufactured by VIAM, Moscow, Russia) with the composition shown in

Table 1. The chemical composition of the steel substrate (at. %).

C	S	P	Mn	Cr	Si	Ni	N	B	Mo	Fe
0.11–0.16	≤0.02.	≤0.03	≤1	14–15	≤0.7	4–5	0.05–0.11	0.11–0.16	2.3–2.8	balance

. The given chemical composition is adjusted to the manufacturer certificate (Scientific Research Institute of Aviation Materials, Moscow, Russia) and has been evaluated by an energy dispersive spectroscopy in the previous paper of the authors [46]. The steel was subjected to heat treatment (hot and cold quenching), which promotes the formation of a two-phase (austenitic–martensitic) structure and increases the mechanical properties [46]. The test samples were disks of 3 mm thickness and 10 mm in diameter. The substrates were mechanically ground using an abrasive SiC paper until a mirror-like surface was obtained. This corresponds to the average roughness of Ra = 0.16 μm. The measurement was performed using a New View 6200 optical interference microscope (Zygo, Weiterstadt, Germany). A subsequent ultrasonic treatment of the substrates in solvent and ethyl alcohol was employed. Mechanically polished steel was considered to be the initial state.

One group of the initial samples was chemically passivated by immersion in a 10% H₃PO₄ solution with 0.05 M K₂Cr₂O₇ for 140 h to form a layer of chromium compounds onto the surface. For this, a hand-made sample holder was employed. The non-reactive support tool, made of graphite-filled caprolon, was utilized, and the test samples were glued by epoxy resin to the sample holder so that a uniform exposure in the solution was achieved.

The other group of samples was treated with ion beams in the high-dose ion implantation mode. The ion beam treatment was performed using a new type of ion source developed on the basis of a magnetron sputtering system exhibiting an additional injection of electrons from the plasma of a vacuum arc emitter located on the back side of the target [47]. The additional injection of electrons and their acceleration in the cathode layer of the magnetron discharge allows the stable operation of a high-current pulsed magnetron discharge at a significantly lower pressure compared to a conventional planar magnetron in HIPIMS mode [48,49]. By reducing the operating pressure, ion transport losses and the recharge effect can be significantly reduced, which in turn allows the extracted ion current in such a discharge system to be multiplied. Traditionally, MEVVA-type vacuum arc ion sources have been used for metal ion implantation [50,51]. In contrast to a vacuum arc, the plasma of a magnetron discharge is characterized by the absence of a microparticle fraction and a large noise in the ion current of the extracted ion beam. In addition, the vacuum arc emitter in this system operates at lower current values and, consequently, has an increased service life. On the other hand, the geometry of the electrodes of the discharge system of the planar magnetron provides protection against an insignificant micro-droplet fraction from the vacuum arc emitter. A distinctive feature of the magnetron discharge ion source for material research by high-energy ion irradiation is the monoenergetics of the ion beam.

A target of 120 mm diameter was made of pure chromium (99.95% purity) and used as the cathode of the magnetron discharge. Argon served as the working gas. The pressure was measured directly in the vacuum chamber using a Micro-Ion Plus sensor (Model 392402-0-YE-T, Granville Phillips, Andover, MA, USA). The vacuum in the chamber was provided by an oil-free pumping system (dry spiral and turbomolecular pumps) to a residual pressure of ~6 × 10⁻⁴ Pa. During the implantation, the mass–charge composition of the beam was in situ examined using a time-of-flight spectrometer according to the technique described in detail [51,52]. The duration of the magnetron discharge current pulse was 250 μs at a pulse repetition rate of 10 Hz. The accelerating voltage of the ion source was 30 kV. The maximum irradiation dose of the samples was 1 × 10¹⁷ cm⁻². The magnetron discharge and electron injection parameters were selected so the ratio of metal ions (Cr⁺) in the magnetron discharge plasma was predominant in relation to the ions of the working gas (Ar⁺). For this purpose, the amplitude of the magnetron discharge current pulse was increased to 40 A, and the working pressure was reduced to 0.05 Pa. The current of the injected electrons was maintained at the level of 1–2 A, which, on the one hand, ensured a stable operation of the magnetron discharge in the high-current form at low pressure and, on the other hand, ensured a high voltage of the magnetron discharge (600–700 V). High voltage of the magnetron discharge promotes an effective ion sputtering of the magnetron target and subsequent ionization of sputtered atoms at a low concentration of neutrals of the working gas. The amount of chromium ions in the beam was ~80–90% (Figure 1) with an average charge of 1.1. Thus, the maximum energy of the accelerated ions did not exceed ~33 keV. During implantation, the samples were placed on a sample holder made of 12X18N10T stainless steel. Before implantation, all the samples were subjected to a final ultrasonic cleaning (20 min) in distilled water. After implantation, the samples were kept in vacuum for 12 h.

The following designations are assigned for different groups of the studied samples: (i) samples after mechanical polishing—VNS-5^MP; (ii) chemically passivated samples—VNS-5^CP; (iii) samples subjected to the ion beam treatment—VNS-5^Cr+.

Electrochemical experiments were performed using a P-40X potentsostat (Electrochemical instruments, Moscow, Russia) by means of a three-electrode cell consisting of a graphite (counter) electrode, a Ag/AgCl reference electrode filled with 4.2 M KCl solution, and a working electrode. Tafel curves were recorded in a linear sweep potential mode with a rate of 1 mV/s. The electrolyte was 3.5 (wt. %) solution of the chemically pure sodium chloride. Impedance spectra in Bode and Nyquist coordinates were obtained at an input sinusoidal signal of 0.01 mV amplitude and within the range of frequencies of 10⁵–0.05 Hz. The elements of the equivalent electrical circuit were analyzed using the EIS Spectrum Analyser 1.0 program. The Pearson criterion (χ²), used for fitting of the impedance data, did not exceed 10^–3. The capacitances of the electrical double layer (EDL) and the oxide layer were described using a constant phase element (CPE). The corrosion rate was estimated using the polarization resistance (Rp) according to the Stern–Geary equation [53]. It should be emphasized that there are many limitations of laboratory corrosion studies, associated with unaccounted for dynamic factors of the marine environment such as humidity, fluctuations of halogen anions, biofouling, etc. [15]. Also, the issues of residual mechanical stresses and fatigue-induced corrosion must be taken into account in order to predict the corrosion rate of the load-bearing materials [54]. Taking into account these considerations, the effect of ion implantation on the electrochemical response of the steel is mainly related to the repassivation ability of the Cr-rich outer surface layer. Nevertheless, the measured corrosion currents and EIS results could be employed to find possible synergistic effects that are related to the formation of the Fr- and Cr-bearing oxide films. In order to reliably determine the resistance of stainless steels to pitting or crevice corrosion, the development of accelerated testing methods based on the chemical and electrochemical methods, as well as full-scale native testing, is required [55,56].

The surface morphology of the steel before and after corrosion tests was examined by Apreo 2S scanning electron microscope (SEM) (Thermo Fisher Scientific, Waltham, MA, USA) using a secondary electron detector at an accelerating voltage of 20 KV.

The chemical state of the elements {Fe, Ni, Cr, C, O} before and after surface treatments was studied by X-ray photoelectron spectroscopy (XPS) on a K-Alpha Nexsa spectrometer (Thermo Scientific, Waltham, MA, USA) using a monochromatic Al Kα X-ray source (λ = 1486.6 eV). Before collection of the spectra, the surface of the samples was subjected to ion bombardment for 100 s at an accelerating voltage of 3 kV in order to remove contaminations. The survey and high-resolution spectra were processed (fitting, peak indexing, background subtraction) using single Gaussian peaks implemented in the Avantage Data System software (Avantage^TM V5, Thermo Scientific, USA). The “chemistry fitting” protocol was employed in order to define the center of gravity of the XPS peaks belonging to various oxidation states. Spectral lines were identified by means of the National Institute of Standards and Technology database [57], so the complex multiplet structure of the Fe 2p_3/2 and Cr 2p_3/2 spectra [58,59] were not taken into account. Under such considerations, the performed XPS analysis is a semi-quantitative one.

3. Results

3.1. Electrochemical Study

Typical potentiodynamic curves (in 3.5 wt. % NaCl) of the steel VNS-5 with different surface treatments are shown in Figure 2a. The corrosion parameters (J_cor, E_cor) were estimated using the Tafel extrapolation method (

Table 2. Corrosion parameters, evaluated through Tafel extrapolation, and the results of the EIS fitting of the steel substrate exhibiting various surface treatments.

	Mechanical Polishing	Chemical Passivation	Ion Implantation
Corrosion current density, μA/cm2	0.583	0.513	0.045
Corrosion potential, mV	–294	–222	–227
Polarization resistance, MΩ·cm2	0.038	0.055	0.296
Solution resistance, Ω·cm2	7	15	16
Oxide resistance, Ω·cm2	–	3150	39,180
Charge transfer resistance, Ω·cm2	9.66 × 106	9939	2.18 × 107
Constant phase element of oxide, Ω−1cm2s−n	–	2.60 × 10−5	6.13 × 10−6
Constant phase element of the double layer, Ω−1cm2s−n	1.34 × 10−5	4.46 × 10−5	2.96 × 10−6

). It is important to note that the studied steel, regardless of the surface treatment method, is characterized by a passive state region (Figure 2a). The corrosion potential (E_cor), determined at the point of equality of the rates of cathodic and anodic processes, shifts toward higher values from −294 mV to −133 mV (Table 2). This is due to a change in the chemical composition of the outer (oxide) layer of the steel, which hinders the oxidation reactions (in particular, Fe and Ni). A significant (by an order of magnitude) decrease in the corrosion current density (J_cor) is found for the VNS-5^Cr+ samples as compared to the initial substrate (VNS-5^MP). As a result, the polarization resistance (Rp), which is inversely proportional to the corrosion rate, increases by ~7.8 times for the sample modified with chromium and argon ion beams. In turn, for the VNS-5^CP sample, the corrosion rate decreases by ~1.4 times. Non-monotonic changes in current densities on the anodic branch at electrical potentials E > –50 mV indirectly indicate the onset of corrosion pitting. These regions are marked with dotted lines in Figure 2a and have often been recorded in the potentiodynamic curves of the stainless steels (VNS-5, UNS N08367, AISI 316L, 17–4 PH SS) under the similar corrosion conditions [46,60,61,62]. The breakdown potential and the transpassivation region cannot be determined in the polarization curves (Figure 2a). When an electrical potential is above +350 mV, a steady growth of the anodic currents is observed due to a continuous dissolution of the surface layer of the samples. To analyze the predominant type of corrosion damage, the potential sweep was stopped when the anodic current was above 0.1 A/cm².

The electrochemical response was assessed by the impedance spectroscopy in the same electrolyte (3.5 wt. % NaCl). The Nyquist diagrams (Figure 2b) of the studied samples can be divided into two types: (i) with a clear linear part in the low-frequency region, which is characteristic for the VNS-5^MP and VNS-5^Cr+; (ii) with an incomplete semicircle, found in the case of the VNS-5^CP sample. This indicates that the samples VNS-5^MP and VNS-5^Cr+ are less susceptible to polarization and, as a result, do not act as the donors of additional electrons for realization of the electrochemical (redox) reactions with oxidation agents. Additionally, for the relevant selection of the equivalent circuits, the Bode diagrams were plotted, in which the phase angle maxima (time constants) in the frequency range from 10,000 to 1 Hz are marked with arrows (Figure 2c). The dependence of the phase angle indicates that, in addition to the formation of a double electric layer, an extra interface between the electrolyte and the substrate is formed. This interface is considered to be an oxide layer that exhibits its own resistance and capacitance and contributes to the total impedance of the system. The proposed equivalent electrical circuits, which satisfactorily describe the Nyquist and Bode diagrams in the entire frequency range, are shown in Figure 2d. Here, in addition to the solution resistance (R_s), the constant phase element of the double electric layer (CPE_dl) and the charge transfer resistance (Rct), the resistance of the oxide film (R_oxide), and its capacitance (CPE_oxide) are present. It should be noted that the deviation of the phase angle from 90° (Figure 2c) in the Bode diagrams at frequencies of 10 < f < 1000 Hz is a result, most likely, of two factors: (i) the presence of microroughness (e.g., traces of mechanical grinding) on the surface of the samples; (ii) a non-monotonic change in the volume charge density in the double electric layer. The interface between the outer oxide layer and underneath substrate could be considered as an ionically blocking interface, through which only electrons may penetrate. In this regard, the charging and discharging of the pseudocapacitors are the predominant mechanisms for the charge storage. If the real volume charge density accumulates nonuniformly and depends on the electron current through the oxide layer, then the competition between the resistive and capacitor contributions takes place. The frequency dependence of the CPE impedance includes the exponent component (n), which, as will be discussed below, is lying in the range of ~0.70–0.85, so the implementation of the CPE element instead of capacitor (n = 1) is justified.

The fitting of the impedance curves by the proposed models (Figure 2d) allows us to conclude that the ohmic and capacitive components (Table 2) vary over a wide range of values depending on the applied surface treatments. Some elements (namely, Rct and CPE_oxide) of the electrical circuit seem to be important with respect to the corrosion properties, since they are related to the magnitude of the corrosion currents and dielectric properties of the oxide layer, respectively. As can be seen from Table 2, the charge transfer, characterized by the highest Rct values, is hindered in the case of steels with mechanically polished and ion beam-modified surfaces. In turn, the most electrochemically active is the steel after chemical passivation in a solution of potassium dichromate and orthophosphoric acid. At the same time, the capacity of the oxide layer is the lowest in VNS-5^Cr+, which is, most likely, due to the poor dielectric properties of the oxide layer. It is interesting that the chemically passivated steel demonstrates the lowest Rct values of ~ 10 kΩ·cm² and a relatively low polarization resistance of Rp = 0.055 MΩ·cm², estimated using the Stern–Geary equation (Table 2). This feature may be associated with the redistribution of the elements over the surface layer due to the contribution of the conductive Cr-bearing compounds, which increase the specific conductivity of the surface (oxide) layer. Summarizing the results of the polarization and impedance curves (Figure 2, Table 2), the VNS-5^Cr+ steel has the best corrosion resistance, while the VNS-5^CP exhibits the poorest corrosion properties.

3.2. Surface Morphology

SEM images of the surface of the steel VNS-5 after the applied treatments are shown in Figure 3. It can be seen that, after mechanical grinding, the residual traces of the abrasive wear remain on the surface (Figure 3a). With a decrease in the grain size of the SiC abrasive, the surface acquires a mirror shine due to a smaller depth of the scratches (Figure 3b). Chemical passivation (Figure 3c) in the 10% solution of orthophosphoric acid leads to etching of the surface layer and, as a consequence, to an increase in the size of the scratches inherited from the mechanical grinding (Figure 3a). Ion beam treatment (Figure 3d), on the contrary, changes the surface roughness due to the partial sputtering of the surface layer with accelerated ions. It can be seen that the ion bombardment with a relatively low energy (~33 keV) effectively smoothed out abrasive traces. Thus, the largest number of surface defects that can serve as corrosion concentrators are found in the VNS-5^MP and VNS-5^CP.

After corrosion tests in the potentiodynamic mode (in 3.5 wt. % NaCl), numerous corrosion pits reaching ~200 μm in diameter are observed on the surface of the samples (Figure 4). This is evidenced by enlarged images of the microrelief of the individual corrosion pits (Figure 4 b,d,f), the inner surface of which consists of many cells. This result is consistent with the previous work of the authors [46], in which the pitting corrosion mechanism was the main one for this steel. As a result of the selective dissolution of metals (Fe and Ni) in a marine environment containing Cl^– ions, the corrosion pits were localized, probably in the grains of the martensitic phase α-(Fe,Cr) (symmetry group Im-3m, Pearson symbol cI2), depleted in nickel, rather than in the grains of the austenitic γ-phase (symmetry group Fm-3m, Pearson symbol cF4). Because of the similar type of pitting damage in all the studied samples, the main corrosion mechanism of the VNS-5 steel, as shown earlier in [46], includes the following stages:

(1)

O₂ + 4e⁻ + 2H₂O → 4OH⁻ (cathodic reaction);

(2)

Fe⁰ → 2e⁻ + Fe²⁺; Ni⁰ → 2e⁻ + Ni²⁺; (primary anodic processes);

(3)

Ni²⁺ + 2Cl⁻ →NiCl₂; Fe²⁺ + 2OH⁻ → Fe(OH)₂;

(4)

Fe⁰ − 2e⁻ + 2H₂O → Fe(OH)₂ + 2H⁺; Ni⁰ − 2e⁻ + 2H₂O→Ni(OH)₂ + 2H⁺ (anodic reactions);

(5)

The formation of corrosion products (Fe(OH)₂, Fe(OH)₃, FeOOH, Fe₂O₃, FeCl₂, NiCl₂, NiO) upon interaction with hydroxide and chlorine ions.

3.3. XPS Study

In order to perform a reliable XPS analysis, the positions of the observed spectral lines [17,18,20,37,63,64] were compared with published data and adjusted with the “envelope fitting” results of M.C. Biesinger et al. [59,65]. The XPS spectra of the studied steel, subjected to the mechanical polishing, chemical passivation, and ion beam treatment, are shown in Figure 5 and Figure 6. Representative survey XPS spectra contain the spectral lines of carbon, oxygen as well as the traces of Ni, Cr, and Fe, as marked in Figure 5. Also, some Auger signals corresponding to Cr-LMM and Fe-LMM are seen. Carbon is found to be a predominant contamination that cannot be completely removed from the surface upon sputtering of the surface layer prior to the high-resolution spectra collection. The high-resolution XPS spectra (Figure 6) clearly reveal oxygen, carbon, and metallic species (Fe, Cr, Ni) belonging to the valence bonds. The semi-quantitative results of the XPS peak fitting are listed in

Table 3. Chemical state of the constituent elements {C, O, Ni, Fe, Cr} identified via XPS analysis of the VNS-5 steel subjected to mechanical polishing, chemical passivation, and ion implantation.

Surface Treatment	Constituent	Peak Energy, eV	Chemical State
Mechanical polishing	C 1s O 1s Ni 2p3/2 Ni 2p1/2 Fe 2p3/2 Cr 2p3/2	283.0 284.5 530.4 531.5 853.2 870.4 706.9 707.7 709.8 574.3 575.7 577.7	C-(Fe, Cr) in γ-(Fe,Cr,Ni) phase C-C (graphite) O2– in oxide phase OH− Ni0 (metallic) Ni0 (metallic) Fe0 (metallic) Fe2+ and Fe3+ in Fe3O4 Fe2+ and Fe3+ in Fe2O3 Cr0 (metallic) Cr3+ in Cr2O3 Cr3+ in CrO(OH), Cr(OH)3
Chemical passivation	C 1s O 1s Ni 2p3/2 Ni 2p1/2 Fe 2p3/2 Cr 2p3/2	283.2 284.9 530.6 531.9 853.2 870.4 706.9 707.8 709.9 574.3 577.0	C-(Fe, Cr) in γ-(Fe,Cr,Ni) phase C-C (graphite) O2– in oxide phase OH− Ni0 (metallic) Ni0 (metallic) Fe0 (metallic) Fe2+ and Fe3+ in Fe3O4 Fe2+ and Fe3+ in Fe2O3 Cr0 (metallic) Cr3+ in CrO(OH), Cr(OH)3
Ion implantation	C 1s O 1s Ni 2p3/2 Ni 2p1/2 Fe 2p3/2 Cr 2p3/2	283.3 285.2 530.6 531.9 853.2 870.5 707.1 708.1 710.2 574.3 575.5 577.3	C-(Fe, Cr) in γ-(Fe,Cr,Ni) phase C-C (graphite) O2– in oxide phases OH− Ni0 (metallic) Ni0 (metallic) Fe0 (metallic) Fe2+ and Fe3+ in Fe3O4 Fe2+ and Fe3+ in Fe2O3 Cr0 (metallic) Cr3+ in Cr2O3 Cr3+ in CrO(OH), Cr(OH)3

. It has been found that the C 1s line (Figure 6a) is a multiplet corresponding to C-(Fe, Cr) and C-C (graphite) components. For the VNS-5^CP sample, the C-C bonds of a graphite-type have a larger ratio than other ones (C-Fe, C-Cr). This could be attributed to the enhanced conductivity of the chemically passivated steel and, as consequence, a lower value of a charge transfer resistance according to the EIS analysis (Table 2). It is noteworthy that the oxygen peak (Figure 6b) in all cases is a multiplet corresponding to the oxidation state O^2– (530.4–530.6 eV) and hydroxide OH^– (531.5–531.9 eV). The component of the O 1s peak (Figure 6b), having a higher binding energy, may be assigned to the formation of hydroxides during passivation in open air or by applied surface treatments. For example, the bonding of chromium with water molecules and OH^– groups may occur during chemical passivation, while Cr-OH bonds, most likely, are formed during ion implantation because of a certain amount of the residual water vapor preserved in the vacuum chamber.

The relative intensities and positions of the Ni 2p peaks (Figure 6c) do not change after surface modification. Indeed, the metallic state Ni⁰ is found to be a major chemical state, while no oxidation species (for example, Ni²⁺) have been revealed via a deconvolution of the Ni 2p_3/2 and Ni 2p_1/2 peaks. The strongest series of Fe 2p_1/2 and Fe 2p_3/2 lines are located near the ~719–722 eV and ~707–710 eV, respectively (Figure 6d). Besides a metallic state of iron, two different oxidation species, Fe²⁺ and Fe³⁺, appear in the high-resolution spectra after fitting of the Fe 2p_3/2 multiplet. Surprisingly, iron oxides in the Cr-alloying steel remain on the surface notwithstanding the chromium passivation using chemical or ion beam methods. However, the relative ratio of iron and chromium-bearing oxides (and hydroxides) changes due to the alloying of the surface layer, as will be discussed later. It has been revealed (Table 3) that the chemical states of Fe correspond to the oxide mixture (Fe₃O₄ and Fe₂O₃). The shifting of the Fe 2p_3/2 peaks by 0.2–0.4 eV relative to the positions in the VNS-5^MP (Table 3) is related to a deviation of the chemical stoichiometry in the Fe_xO_y oxides. The XPS spectra were collected from the sample’s surface after Ar⁺ sputtering, so the most intense peaks (near ~707 eV) of iron in the metallic state (Figure 6d) originated from the Fe-Fe or Fe-(Ni, Cr) bonds. Therefore, the results of the XPS analysis are attributed to both the oxide layer and underneath sublayer. Finally, the spectrum of Cr 2p_3/2 is complex (Figure 6e) and consists of multiplets located at ~574, ~576, and ~577–578 eV. The peak at the low-energy side (Figure 6e) originates from the metallic state (Cr⁰). Deconvolution of the XPS peaks reveals a Cr³⁺ oxidation state that could belong to either Cr₂O₃ or Cr-hydroxides (Table 3). The complex shape of the Cr 2p_3/2 (Figure 6e) peak is far from the pure Cr₂O₃ and Cr(OH)₃ [65]. According to the results of [65], the O 1s peak may have several multiplets: the hydroxyl peak is assigned to the intense peak (~531.8 eV); the peak at the lower binding energy results from Cr₂O₃ oxides; and the peak at the higher energy is assigned to water from hydration. Our results confirm a formation of the hydroxyl group OH^– (~531.4 eV), while the peak of the lower energy (~530.4 eV) corresponds to the metallic oxides (Figure 6b). The high-energy component (~577–578 eV) of the Cr 2p_3/2 peak is presumably related with CrO(OH), but, in some cases, this component may be assigned to Cr(OH)₃. Thus, the surface layer in the VNS-5^MP and VNS-5^Cr+ consists of Fe₂O₃, Fe₃O₄, and Cr₂O₃ oxides and CrO(OH) hydroxide, while the Fe₂O₃, Fe₃O₄ oxides and chromium hydroxides are found in the outer layer of the chemically passivated steel.

Deconvolution of the complex XPS peaks is in good agreement with the existing research results. In general, the Fe 2p_3/2 peak energy is consistent with the results of the XPS studies of the AM355 [17], AISI 304 [18] stainless steels, as well as the reported data of M.C. Biesinger (et al.) [58,59]. Moreover, the Cr 2p_3/2 multiplets could be analyzed in such manner (Figure 6e) according to [59,62,65]. However, some shifts of the Fe and Cr peaks are detected (Table 3) in comparison with XPS database [57], most likely due to the vacancy-type defects and nonstoichiometric composition of the oxide phases.

4. Discussion

The results of the electrochemical (Table 2) and XPS experiments (Table 3) allow us to consider that the corrosion resistance of the steel VNS-5 does depend on the chemical state and type of oxides as well as the passivation ability of the outer electroactive layer. Several issues attract the most interest: (i) an increase in the conductivity of the oxide layer by ~970 times (Table 2) (as a value inversely proportional to Rct) in the steel after the chemical passivation; (ii) a decrease in the corrosion rate of the ion-modified sample (by ~7.8 times) while maintaining a similar pattern of the observed corrosion (pitting) damage (Figure 4). In our opinion, the changes in the electrochemical response of the samples are associated with a change in the thickness and dielectric properties of the electroactive Fe- and Cr-rich surface layer that exhibits various ratios of the oxide and hydroxide compounds (Figure 6). In particular, the Cr-bearing hydroxides, formed onto the VNS-5^CP according to the XPS data, promote a charge transfer, most likely due to its specific conductivity and low values of Rct ~10 kΩ·cm² (Table 2). Notwithstanding the absence of the relevant data on the dialectic constant of the CrO(OH) or Cr(OH)₃ compounds, our findings indicate that passivation of steel in a 10% H₃PO₄ solution with the addition of 0.05 M K₂Cr₂O₇ is able to manage the charge transfer ability and conductance of the oxide film. In turn, the ion implantation leads to an impeded charge transfer through the oxide film. The origin of the observed phenomenon could be revealed by implementation of the potentiodynamic impedance experiments and will be discussed in a separate paper.

In order to estimate the thickness of the outer electroactive layer (δ), an approach, described in [66], was used. Previously, the EIS method was employed to evaluate the impedance response of passivated stainless steels [18,21]. For this, the following fitting results of the impedance spectroscopy (Table 2) have been considered: the impedance of the constant phase element (CPE_oxide), the resistance (R_oxide) of the oxide layer, and the frequency index (n). The EIS measurements, performed in the low-frequency range (0.05–1 Hz), correspond to the primarily capacitive response, while the ohmic resistance has the minor effect on the total impedance. Here, the simplified resistor–capacitor circuit for the oxide layer (Figure 2d) was utilized. The capacitance has a frequency-dependent response and is related to the stored charge at the electrode at a given frequency when the AC perturbation oscillates around the open circuit potential. The quantitative value of capacitance depends on the thickness of the electroactive layer and its dielectric properties [66]. The resistivity component is inversely proportional to the electrical conductivity of the surface layer. Thus, considering the quantitative values of the CPE_oxide and R_oxide, some conclusions of the electrochemical response of the electroactive surface layer could be made.

Using the literature data (

Table 4. Dielectric constants (ε) of carbon, some Fe- and Cr-bearing oxides, stainless steel, and hydroxides *.

Compound/Material	ε	Ref.
Cr2O3	~13	[67]
Fe2O3	~12	[68]
Fe3O4	~20
FeO(OH)	~11	[69]
Stainless steel	~12	[70]

* The relevant data on dielectric properties of chromium hydroxides are absent in the published literature.

) on the dielectric constants (ε), it is possible to estimate the electroactive layer thickness exhibiting Fe- and Cr-rich compounds:

$$\delta ={\displaystyle \frac{{\left(\overline{\epsilon }{\epsilon }_0\right)}^n}{\left(1+2.88{\left(1-n\right)}^{2.375}\right)\times {CPE}_{oxide}\times R_{oxide}^{1-n}}}$$

(1)

where ε₀ = 8.85 × 10⁻¹⁴ F/cm is the permittivity of vacuum; the values of CPE_oxide and R_oxide were taken from the EIS data (Table 2); n is the exponent that determines the nature of the frequency dependence of the CPE impedance:

$$Z\left(CPE\right)=Q^{-1}{\left(j\omega \right)}^{-n}.$$

(2)

Here, $j$ is the imaginary unit; $\omega$ is the frequency; $Q$ is the capacitance. The typical values of n, used in Formula (2), are close to ~0.85, ~0.70, and ~0.80 for the mechanically polished, chemically passivated, and ion beam-treated samples, respectively. These values are far from the ideal capacitor (n = 1), so the implementation of the CPE element instead of capacitor is justified.

The average dielectric constant of the surface layer could be evaluated as

$$\overline{\epsilon }={\sum }_i{C}_i⨯{\epsilon }_i.$$

(3)

Here, C_i is the relative ratio of the metallic oxides; ε_i is the dielectric constant of the chemical compound. For example, for the initial steel having iron and chromium oxides, the average dielectric constant is estimated to be as follows:

$$\overline{\epsilon }=C_{Fe3O4}·{\epsilon }_{Fe3O4}+C_{Fe2O3}·{\epsilon }_{Fe2O3}+C_{Cr2O3}·{\epsilon }_{Cr2O3}=18.7.$$

As one can see in

Table 5. The ratios (C_i) of the metallic and oxidized form of metals (Fe, Cr, Ni) in the initial and modified steels according to the XPS analysis. The evaluated thickness of the outer electroactive layer (δ) was calculated using Formula (1).

Surface Treatment	Fe0, %	Fe2 and Fe3+, %	Cr0, %	Cr3+, %	Ni0, %	δ, nm
Mechanical polishing	27	57	5	7	4	N/A
Chemical passivation	24	45	5	23	3	~8.9
Ion implantation	21	43	11	21	4	~13.7

, the ratio of the metallic and oxidized forms of metals in the initial and modified steels vary significantly. Using the Formulas (1)–(3), the evaluated thickness (δ) of the outer electroactive layer is ~8.9 nm to ~13.7 nm after chemical passivation and ion implantation, respectively (Table 5). It is supposed that the passivation ability of the steel could be improved if the thickness of the Cr-rich layer increases. Consequently, the larger amount of Cr₂O₃ oxides and thicker electroactive layer in the VNS-5^Cr+ are able to provide an improved corrosion performance rather than a formation of the CrO(OH)-rich surface layer in the VNS-5^CP. It should be added that a certain fraction of Fe₂O₃ and Fe₃O₄ oxides is still observed on the surface of the steel after the specified surface treatments. It is supposed that, in order to prevent the formation of iron oxides, it is required (i) to completely remove a native oxide film prior to the surface alloying and (ii) to promote an outward diffusion of iron species from the outer surface to the bulk. Ion implantation, as a rule, leads to the surface oxidation of the ion beam-treated materials and promotes the formation of a thicker and complex oxide film [71]. When the surface is enriched with implanted atoms, the in-depth variation in the concentration of elements occurs due to the spraying of adsorbed and chemisorbed compounds (in particular, carbon-bearing film, oxides), which are always present on the surface of steels. In addition, the diffusion of chromium atoms toward the surface is also possible due to heat flows arising from the impact interaction of accelerated ions with the surface of the sample.

The relationship between the influence of microalloying of the alloy on nucleation rate, electrochemical properties, thickness, and defect density of a passive film was previously described [72]. In the marine environment, the higher concentration of dissolved anions, hydroxide groups, and oxygen species enhance the corrosion rate [73]. According to the paper [74], the process of the Fe-Cr alloy passivation accelerates with a decrease in the acidity of the solution, as well as an increase in the concentration of chromium. In turn, the pitting resistance of austenitic steels improves abruptly with an increase in the concentration of chromium to the limiting values in the region of the active–passive transition [75]. The protective ability of the Cr-bearing passive films is also related to an average breakdown potential.

In the present study, the formation of a mixed structure of chromium and iron oxides also contributes to the surface passivation. Considering the stainless steels, the beneficial effect of various oxide phases preventing a formation of microgalvanic corrosion currents and pitting damage was established [17,22,26,76,77]. Taking into account that the real distribution of oxides and microstructure of the ion beam-treated surface layer could be revealed only by cross-section TEM studies [37], it is suggested (for simplicity) that the studied steel exhibits a single layer oxide film consisting of oxide particles based on Fe₂O₃, Fe₃O₄, Cr₂O₃, and CrO(OH) phases (Figure 7). In this case, the interaction of H₂O molecules and oxygen dissolved in the 3.5 wt. % NaCl solution with matrix α- or γ-phases leads to formation of corrosion products via the following possible electrochemical reactions [10,19,20]:

(1)

Fe₃O₄ + H₂O → 3Fe₂O₃ + 2H⁺

(2)

Fe₂O₃ + FeCl₃ → 3FeOCl

(3)

Fe²⁺ + 2Cr³⁺ + 4OH⁻ → FeCr₂O₄ + 4H⁺

(4)

2Fe₃O₄ + 2OH⁻ + 2H₂O → 6FeOOH + 2e⁻

(5)

2Fe₃O₄ + 2OH⁻ → 3Fe₂O₃ + H₂O + 2e⁻

(6)

Cr → Cr³⁺ + 3e⁻

(7)

2Cr + 6H₂O → Cr₂O₃ + 3H₂O + 6H⁺ + 6e⁻

(8)

Cr + 3H₂O → CrOOH + H₂O + 3H⁺ + e⁻

(9)

CrO + H₂O → CrOOH + H⁺ + e⁻.

For a mixed oxide film consisting of both Fe- and Cr-based compounds (Table 3), the passive layer of the steel VNS-5 may be proposed (Figure 7). Here, the primary cathodic and anodic processes occurred on either the surface of the oxides or interface boundaries between the oxide phases. The formation of Cr-bearing hydroxides in aqueous solutions is thermodynamically favorable and accompanied by adsorption of lower-valence hydroxides that tend to transform into higher-valence ones. The observation of CrOOH and Cr₂O₃ is also expected in open air. Because of the electrochemical reactions, the additional electrons and protons appear and may transfer between the surface areas exhibiting different electric potentials (Figure 7). This will cause galvanic corrosion and facilitate a penetration of Cl^– anions into the bulk of the material that also accelerates the development of corrosion pits. The protective effect of the Cr-rich outer surface layer being in contact with an aggressive corrosive environment is that the processes of pitting corrosion of steels are induced preferentially in the oxide film of the chromium oxides [18,19].

The largest values of the charge transfer resistance (Table 2) were observed in the VNS-5^Cr+ sample. This may indicate that the enrichment of the surface layer with Cr enhances the dielectric strength of the native oxide film. This assumption is based on the fact that the breakdown potentials of the Cr-rich stainless steels are higher than those in the low-carbon steels [78,79]. During the corrosion process in the applied electrical field, other oxidation reactions occur, leading to the appearance of the less stable FeOCl, Cr(OH)₃, and CrO compounds. Prior to the onset of the potentiodynamic tests, the passive film serves as an electric barrier that suppresses the oxidation of the metals (Fe²⁺, Cr³⁺) and formation of the corrosion products (Figure 7). Moreover, the Fe-bearing oxides are transformed into soluble compounds (FeCl₂, etc.). If the surface is currently in the active region, then the anodic dissolution dominates, so the anodic current significantly increases (Figure 2a). Generally speaking, the appearance of many microgalvanic elements, consisting of corrosion products (spinel-type oxides and hydroxides), on the corroded surface results in the severe pitting damage observed by SEM (Figure 4) for all the studied samples. Thus, the surface oxides may inhibit the corrosion processes if the electric potential required for the anodic current flows is higher than an external electric potential. In practice, this condition is never fulfilled because of the diffusion of Fe, Ni, and Cr cations through the various defects (cracks, boundaries) within the oxide film, as revealed by TEM [73].

5. Conclusions

The implementation of the ion beam surface treatment, as well as the immersion (for 140 h) using potassium dichromate and 10% phosphoric acid, was performed in order to manage the passivation behavior, surface chemistry, and electrochemical response of the industrial austenitic–martensitic steel VNS-5. For the first time, a new type of ion source based on a high-current pulsed magnetron discharge with injection of electrons from vacuum arc plasma was employed for ion implantation of the steel substrate by Cr⁺ and Ar⁺ ions. The following conclusions were made:

(1).

Based on the EIS results, the mechanically polished and ion beam-treated steels are characterized by the largest values of a charge transfer resistance. In turn, the chemically passivated steel shows a significantly lower (by ~1000-times) resistance and promotes a transfer of charges in seawater (3.5 wt. % NaCl) due to the specific conductivity of the oxide layer.

(2).

After the ion beam treatment, the polarization resistance of the steel VNS-5 reaches the highest value, and the estimated corrosion rate via the Stern-Geary equation is ~7.8 times lower than that of the initial (mechanically polished) steel.

(3).

After corrosion tests in potentiodynamic mode (in 3.5 wt. % NaCl), the numerous corrosion pits of ~200 μm in diameter are observed on the surface of the steel, notwithstanding the applied surface treatments.

(4).

Passivation at various conditions leads to a noticeable change in the ratio of the oxidation species (Cr³⁺, Fe²⁺, and Fe³⁺) in the oxide layers, as confirmed by XPS analysis. It has been found that the oxide layer in the mechanically polished and ion beam-treated steels mainly consists of Fe₂O₃, Fe₃O₄, and Cr₂O₃ oxides and a low amount of CrO(OH) hydroxide, while the chemically passivated steel exhibits Fe₂O₃ and Fe₃O₄ oxides and chromium hydroxides.

(5).

Using the electrochemical impedance data, the thickness of the electroactive surface layer was evaluated to be ~8.9 nm and ~13.7 nm after chemical passivation and ion implantation, respectively. An improved corrosion performance of the ion beam-treated steel is associated with a better passivation ability of the Cr₂O₃-rich surface layer.