Corrosion Behavior of Passivated CUSTOM450 and AM350 Stainless Steels for Aeronautical Applications

: Custom 450 stainless steel and AM 350 stainless steel are both precipitation hardening stainless steels, which are widely used in a variety of aerospace applications. The former steel exhibits very good corrosion resistance with moderate strength, whereas the latter is used for applications requiring high strength along with corrosion resistance. In this study, the corrosion behavior of CUSTOM 450 and AM 350 stainless steels passivated in (a) citric acid and (b) nitric acid solutions for 50 and 75 min at 49 and 70 ◦ C, and subsequently exposed in 5 wt. % NaCl and 1 wt. % H 2 SO 4 solutions are investigated. Two electrochemical techniques were used: electrochemical noise (EN) and electrochemical impedance spectroscopy (EIS) according to ASTM G199-09 and ASTM G106-13, respectively. The results indicated that passivation in nitric acid made the surface prone to localized corrosion. Statistical and PSD values showed a tendency toward pitting corrosion. On the whole, passivated CUSTOM 450 stainless steel showed the best corrosion behavior in both, NaCl and H 2 SO 4 test solutions.


Introduction
Nowadays, stricter environmental regulations suggest the need to seek environmentallyfriendly corrosion protection processes in the aeronautical sector. Passivation is a chemical treatment typically for stainless steel that enhances the ability of the treated surfaces to resist corrosion [1][2][3][4][5][6].
Stainless steels (SS are widely used in aircraft components due to their very good mechanical properties, weldability, thermal expansion, impact resistance, and corrosion resistance, among others). According to their chemical composition and metallurgical phases, stainless steels are classified into five families, i.e. austenitic, ferritic, martensitic, duplex (consisting of a mixture of ferrite and austenite), and PH (precipitation hardened) [5,7]. Typically, austenitic, martensitic, and PH stainless steels are used in the aeronautical industry. Their corrosion, mechanical and high-temperature properties allow them to perform well in aggressive environments and makes them suitable for various aircraft parts, such as landing gear supports, actuators, and fasteners [6,8].
The aim of this work is to evaluate the corrosion behavior of the CUSTOM450 and AM350 stainless steels passivated in citric and nitric acid solutions for 50 and 75 min at 49 and 70 • C, and subsequently exposed in 5 wt. % NaCl and 1 wt. % H 2 SO 4 solutions. Two electrochemical techniques were used: electrochemical noise (EN) and electrochemical impedance spectroscopy (EIS). Martensitic stainless steel they are precipitation-hardenable alloy used in aeronautics and exposed to different atmospheres such as marine and industrial (acid rain).

Materials
The materials used in this work were CUSTOM450 (UNS S45000) and AM350 (UNS35000) stainless steel, tested in the as-received condition. The nominal chemical composition of these stainless steels [38,39] is shown in Table 1. Stainless steels samples were prepared according to ASTM E3-11 (2017) [40]. The various alloys were ground using 400, 600, and 800 grade SiC sandpaper followed by ultrasonic cleaning in ethanol (C 2 H 5 OH) and deionized water for 10 min for each sample.

Passivation Process
The passivation process was carried out under ASTM A967-17 [21]. Gaydos et al. [22] reported that extended passivation treatments gave better protection against corrosion for a series of stainless steels. The passivation treatment consisted of the following stages:

2.
The different combinations of citric and nitric acid baths available for testing were too many, so a 3-factor, 2-tier design of experiment 5 (DoE) was conducted to determine the optimal concentration of citric and nitric acid. The temperature of the solution and the passivation time.

3.
Passivation treatment: two passivation baths of nitric acid (20%v) and citric acid (55%v) solutions were used. A constant temperature of 49 and 70 • C was maintained through the passivation process. Specimens were immersed in the solutions for 50 and 75 min.

4.
Final treatment: Rinsed in distilled water. Table 2 shows the passivation exposure conditions for each type of HP stainless steel.

Electrochemical Tests
Electrochemical measurements were conducted at room temperature using an electrochemical interface mod. 1287A and Impedance Analyzer mod. Then 1260 Solartron (Bognor Regis, UK) two 5 wt. % NaCl and 1 wt. % H 2 SO 4 solutions were used with the latter reagent used to simulate acid rain conditions [5]. A conventional three-electrode cell configuration was used for the electrochemical corrosion studies, which consisted of a working electrode, WE (passivated CUSTOM450 and AM350 steels), a reference electrode, RE (calomel electrode (SCE)), and an auxiliary electrode, CE (platinum mesh) [6,42,43]. Before the test, the working electrode (passivated stainless steel sample) was held for about 1 h at open circuit potential (OCP). Corrosion tests were realized in triplicate.
Electrochemical noise (EN) measurements of current and potential values at onesecond intervals on the passivated stainless steel tested were taken. The time records consisted of 1024 data. The electrochemical cell for EN measurements consisted of an array of three electrodes: PHSS samples (working electrode WE1), a platinum electrode (WE2), and a saturated calomel electrode as reference electrode (RE). The EN data were detrended using a degree 9 polynomial and applying the fast Fourier transform (FFT) with a Hann window [49][50][51][52][53][54]. Data processing was performed with a program that can be used in MATLAB 2018a software (Math Works, Natick, MA, USA) [55][56][57].

Electrochemical Noise Analysis
The electrochemical noise (EN) technique is non-perturbative and describes the spontaneous low-level potential and current fluctuations during an electrochemical process. During the corrosion process, predominantly electrochemical anodic and cathodic reactions can cause small transients in the electrical charges of the material under study. Figure 1 shows the time series obtained for the passivated steels after exposure to the NaCl electrolyte. Figure 1a shows the electrochemical potential noise (EPN), where all samples show a decrease in amplitude over time. This behavior can be related to a decrease in ionic exchange due to thermodynamic stabilization. Sample anodized in nitric acid AM350-NA-50 min-70 • C showed a higher potential amplitude (3 × 10 −2 V). Figure 1b shows the electrochemical current noise (ECN), where the behavior of CUSTOM450-CA-50 min-49 • C, which in these conditions shows the higher amplitude fluctuations This behavior can be related to the higher corrosion kinetic of the passive layer in this condition, and the maximum value is 2.5 × 10 −5 A/cm 2 . On the other hand, AM350-CA-50min-49 • C and AM350-NA-50min-70 • C presented the lowest amplitude value of about ×10 −7 A/cm 2 . This result can be related to its corrosion resistance in NaCl. The passivated AM350 steel is more corrosion resistant than passivated CUSTOM 450 steel when these steels are exposed to NaCl. fluctuations This behavior can be related to the higher corrosion kinetic of the passive layer in this condition, and the maximum value is 2.5 × 10 -5 A/cm 2 . On the other hand, AM350-CA-50min-49 °C and AM350-NA-50min-70 °C presented the lowest amplitude value of about ×10 -7 A/cm 2 . This result can be related to its corrosion resistance in NaCl. The passivated AM350 steel is more corrosion resistant than passivated CUSTOM 450 steel when these steels are exposed to NaCl.  Figure 2a shows the EPN time series when passivated steels are exposed to H2SO4. The ionic exchange is of low order (×10 -5 V). The low fluctuation amplitude is related to a generation of the passive layers created by the H2SO4 electrolyte. As can be seen, all samples presented the same behavior. Figure 2b shows the EPN time series where the sample anodized in citric acid AM350-CA-75min-49°C presents the higher current transients of 1.5 × 10 -6 A/cm 2 . Meanwhile, CUSTOM450-NA-75min-70 °C presented transients of 8 × 10 -7 A/cm 2 , related to localized processes, but they occur due to the passive layer's breakdown and regeneration. Furthermore, a transitory reaction is induced by a non-uniform passive layer, and OH-reactions occur easier. The results show that passivated AM350-CA-75min-49 °C and passivated CUSTOM450-CA-50min-49 °C are more susceptible to pitting corrosion.  Figure 2a shows the EPN time series when passivated steels are exposed to H 2 SO 4 . The ionic exchange is of low order (×10 −5 V). The low fluctuation amplitude is related to a generation of the passive layers created by the H 2 SO 4 electrolyte. As can be seen, all samples presented the same behavior. Figure 2b shows the EPN time series where the sample anodized in citric acid AM350-CA-75min-49 • C presents the higher current transients of 1.5 × 10 −6 A/cm 2 . Meanwhile, CUSTOM450-NA-75min-70 • C presented transients of 8 × 10 −7 A/cm 2 , related to localized processes, but they occur due to the passive layer's breakdown and regeneration. Furthermore, a transitory reaction is induced by a non-uniform passive layer, and OH-reactions occur easier. The results show that passivated AM350-CA-75min-49 • C and passivated CUSTOM450-CA-50min-49 • C are more susceptible to pitting corrosion. Statistical Analysis EN can be studied by various methods such as time domain, where analysis is made by visual analysis, skewness, localization index (LI), and noise resistance (R n ). Another method to study EN signals is by power spectral density (PSD); this method evaluates frequency signals. Nevertheless, since the EN shows different signals (see Equation (1)), it is necessary to separate components that do not provide any information about the corrosion process. The DC (m t ) component can be separated by different methods, such as the polynomial filter or wavelets. The random (S t ) and stationary (Y t ) components are considered because the corrosion process occurs in those signals [58][59][60][61][62].
To determine noise resistance (R n ) it is necessary to obtain the standard deviation from time series values; these statistical values give corrosion kinetics and mechanistic information. Cottis and Turgoose [53] found a relationship between an increase in variance and standard deviation with increased corrosion rate. For standard deviation, (σ) evaluation applying Equation (1) is required, for which Rn can be obtained (Equation (2)) using the relationship between the standard deviations from EN potential (σ E ), and current signals (σ I ): The statistical analysis (Table 3) showed that mixed corrosion is predominant in both electrolytes. Only AM350-NA-50min-70 • C and CUSTOM450-NA-50min-49 • C systems showed localized corrosion when exposed to sulfuric acid, which can be related to a non-uniform layer. However, when steels under passivation conditions are exposed to NaCl, only CUSTOM450-NA-75min-70 • C showed mixed corrosion, corrosion, and the other samples presented uniform corrosion. Skewness shows that localized processes are predominant in corrosion systems due to passive layer generation. Furthermore, the samples passivated in citric acid presented a lower corrosion resistance. However, samples passivated in nitric acid presented a non-uniform passive film due to localized corrosion. Power Spectral Density Analysis and Noise Impedance (Z n ) The separation of the DC signal is necessary because DC creates false frequencies and interferes in visual, statistical, and PSD analysis. The polynomial method is governed by Equation (3), where x n is the EN signal with all the components, a polynomial of "n" grade (p o ) at n-th term (a i ) in "n" time to obtain a signal without trend (y n ) [53,54]. In this work, a 9th-degree polynomial was applied. For power spectral density (PSD) analysis, it is necessary to apply a polynomial filter to the original signal, followed by the application of a fast Fourier transform (FFT) to transform from time to frequency domain. Equations (4) and (5) are employed to obtain spectral densities [47][48][49][50][51][52][53][54][55][56][57][58][59][60].
The interpretation of PSD is based on the slope and frequency zero limits (Ψ 0 ). The slope is related to the type of corrosion present in the system [35,36]. The slope is mathematically defined by β x and is represented by Equation (6): To obtain information about the material dissolution, it is required to analyze the frequency zero limits (Ψ 0 ) in current PSD [24]. The following table helps to determine the corrosion process in a given system [60][61][62][63][64]. Results of Table 4 show the parameters obtained from the first test because in general experiments did not show any significant variations. Table 4. β intervals to indicate the type of corrosion [47].
The spectral noise resistance is the noise impedance (Zn), which is expressed by the following equations [6,47,62,65] Noise impedance is calculated as the square root of the ratio of the PSD of EPN divided by the PSD of ECN. The electrochemical noise impedance zero (Z n 0) is related to corrosion resistance [59,66]. Figure 3a shows the PSD of the ECN. Sample CUSTOM450-CA-50min-49 • C has the highest corrosion dissolution with −99.05 dBi, and a lower noise impedance of 856 Ω·cm 2 (see Table 5); these results may be related to high corrosion kinetics. On the other hand, Figure 3b sample AM350-NA-50min-70 • C showed a higher noise impedance value (75,382 Ω·cm 2 ) and lower material dissolution with −129.55 dBi (see Table 5). All the slope values are related to pitting corrosion when exposed to NaCl. Figure 4 a shows the current PSD, and sample AM350-CA-75min-49 • C showed the higher corrosion kinetic even in Figure 4a with −128.66 dBi. The passivated steel that showed the lower noise impedance value was CUSTOM450-CA-50min-49 • C (1972 Ω·cm 2 ). (Figure 4b) Compared to samples passivated in nitric acid and exposed to sulfuric acid, those passivated in citric acid presented lower corrosion resistance. Samples passivated with nitric acid presented a better performance in both NaCl and H 2 SO 4 solutions. Table 5 shows that all samples exposed to NaCl presented pitting corrosion. The results obtained by statistical analysis showed a uniform system. This behavior may be related to a uniform distribution of pitting. When samples were exposed to H 2 SO 4, uniform corrosion occurred on the sample surface. This process could be attributed to the uniform breakdown and regeneration of the passive layer.   Table 5 shows that all samples exposed to NaCl presented pitting corrosion. The results obtained by statistical analysis showed a uniform system. This behavior may be related to a uniform distribution of pitting. When samples were exposed to H2SO4, uniform corrosion occurred on the sample surface. This process could be attributed to the uniform breakdown and regeneration of the passive layer.      Table 5 shows that all samples exposed to NaCl presented pitting corrosion. The results obtained by statistical analysis showed a uniform system. This behavior may be related to a uniform distribution of pitting. When samples were exposed to H2SO4, uniform corrosion occurred on the sample surface. This process could be attributed to the uniform

Electrochemical Impedance Spectroscopy Analysis
Electrochemical impedance spectroscopy (EIS) has been widely used to study the electrochemical mechanisms at electrodes. The characteristics of the EIS spectra of the samples investigated in the present study are presented in Figures 5 and 6 which show Nyquist and Bode plots for AM350 y CUSTOM 450 passivated stainless steels and immersed in 5 wt. % NaCl and 1 wt. % H 2 SO 4 solution. The EIS analysis was employed to analyze the corrosion susceptibility and protection properties of passivated HPSS and electrochemical noise. The electrochemical behavior of passivated HPSS may be due to an oxide layer with a constant phase element and a combination of resistive elements in the equivalent circuit. The aforementioned behavior can be seen in the Bode plots through the phase angle distribution. This is the case with untreated stainless steel, as they naturally form a barrier layer when exposed to oxygen from the environment. In the NaCl solution, the passivated alloy in citric acid may have a higher charge transfer resistance (AM350-CA-50min-49 • C and CUSTOM 450-NA-50min-49 • C). The CUSTOM 450-CA-75min-49 • C sample immersed in H 2 SO 4 had the highest charge transfer resistance i.e., 4.252 × 10 6 Ω·cm 2 passivated in a citric acid bath, and the lower value corresponding to sample AM350 (NA 20 wt. %; 50 min; 70 • C) with a value of 3.915 × 10 4 Ω·cm 2 , respectively. All impedance spectra have the same EIS measurement frequency range characteristics. Figures 5b,c and 6b,c show the Bode plots for all impedance data, one peak can be observed in the phase angle (θ) vs. frequency plot, indicating that HPSS passivated presents one-time constants.

Electrochemical Impedance Spectroscopy Analysis
Electrochemical impedance spectroscopy (EIS) has been widely used to study electrochemical mechanisms at electrodes. The characteristics of the EIS spectra of samples investigated in the present study are presented in Figures 5 and 6 which sh Nyquist and Bode plots for AM350 y CUSTOM 450 passivated stainless steels and i mersed in 5 wt. % NaCl and 1 wt. % H2SO4 solution. The EIS analysis was employed analyze the corrosion susceptibility and protection properties of passivated HPSS a electrochemical noise. The electrochemical behavior of passivated HPSS may be due to oxide layer with a constant phase element and a combination of resistive elements in equivalent circuit. The aforementioned behavior can be seen in the Bode plots through phase angle distribution. This is the case with untreated stainless steel, as they natura form a barrier layer when exposed to oxygen from the environment. In the NaCl soluti the passivated alloy in citric acid may have a higher charge transfer resistance (AM3 CA-50min-49 °C and CUSTOM 450-NA-50min-49 °C). The CUSTOM 450-CA-75min-49 sample immersed in H2SO4 had the highest charge transfer resistance i.e., 4.252 × 10 6 Ω·c passivated in a citric acid bath, and the lower value corresponding to sample AM350 (N 20 wt. %; 50 min; 70 °C) with a value of 3.915 × 10 4 Ω·cm 2 , respectively. All impedan spectra have the same EIS measurement frequency range characteristics. Figures 5b,c a  6b,c show the Bode plots for all impedance data, one peak can be observed in the ph angle (θ) vs. frequency plot, indicating that HPSS passivated presents one-time constan   The physical and mathematical model of the metal/passive film/solution system is presented in Figure 7. The interface resistance is the resistance R1, and the passive electrode system is the capacitance C in parallel in the electrical equivalent circuit (EEC) [67][68][69][70][71]. The capacitance obtained can sometimes deviate from the "pure capacitance" when the actual system impedance is measured. A constant phase element (CPE) is introduced in data fitting to allow for depressed semicircles. The capacitor C is replaced with a CPE for better fit quality. Figure 7 proposes the following equivalent electrical circuit, where RS is the uncompensated solution resistance; R1 and CPE are the interface resistance and the constant phase element, respectively. The mathematical equation for the system impedance of HPSS steels in this model is [67].
The constant phase element CPE impedance is presented by Equation (9) [72].

Discussion
Electrochemical noise uses two data analysis methods: the time domain (

Discussion
Electrochemical noise uses two data analysis methods: the time domain (statistical analysis) and the frequency domain (power spectral density, PSD). Those methods were employed to obtain information on the corrosion kinetics of passivated CUSTOM 450 and AM 350 precipitation hardened stainless steels immersed in NaCl and H 2 SO 4 solutions.
In the time series of passivated stainless steels, the amplitude transients are associated with the corrosion kinetics, which decreased with time for steels immersed in the NaCl solution. However, the PHSS immersed in H 2 SO 4 solution showed transients associated with rupture and regeneration of the passive layer. This behavior is attributed to the anodic reactions of the OH -. The R n values are higher for AM 350 stainless steel in both passivation solutions, indicating a high corrosion resistance for this steel. On the other hand, the samples passivated in citric acid presented a lower corrosion resistance.
However, the samples passivated in nitric acid showed non-uniform passivation according to the LI values obtained, which corresponds to localized corrosion [5,29,[47][48][49]. The type of corrosion that occurred for the passivated PHSS in both citric acid and nitric acid solutions indicates that the LI and asymmetry parameters were evaluated from the electrochemical noise data and found to be 0.003 to 0.2, see Table 1. The LI and skewness values obtained for the passivating solutions, the type of corrosion indicates localized corrosion due to the breakdown of the passive film. Various research groups have used the LI parameter [54][55][56][57][58][59], such as Cottis [53], as criteria to determine the possible type of corrosion occurring. In the present study, the LI and skewness parameters were estimated based on an analysis in the time domain and supported by an analysis in the frequency or time-frequency domain.
According to some authors [67,72,74], the impedance data obtained for passivated PHSS and exposed in sodium chloride and sulfuric acid solution were adequately represented by an equivalent electrical circuit model, comprising two elements: the interface resistance and the constant phase element in parallel and uncompensated solution resistance.
Macdonalds D. [81,82] indicates that EIS measurements allow obtaining information about the corrosion mechanism, establishing a theoretical transfer function, and developing the passive film growth model.
The passivation of stainless steel involves the formation of chromium and iron oxide films [83][84][85][86][87]. Dissolution on the steel surface generates a superficial enrichment of Cr 3+, giving rise to the formation of chromium trihydroxide Cr(OH) 3 , as shown in Equation (10) and Figure 7. Further dissolution of the hydroxide leads to a continuous layer of Cr 2 O 3 , according to chemical reaction (5) [5,6,[86][87][88]. Figure 8 shows a schematic diagram showing the corrosion mechanism for CUSTOM 450 and AM 350 stainless steels after passivation process in C 6 H 8 O 7 and HNO 3 baths. The samples passivated in nitric acid showed a higher trend of pitting corrosion. Noh et al. [91] conclude that nitric acid increases the chromium presence of the passive layer, removing MnS inclusions from the surface; also, the probability of individual pitting increases. In this research, the samples passivated in nitric acid presented more pitting than those passivated within citric acid due to the presence of MnS. To reduce the presence of MnS, changes in the acid concentration or the use of other solutions similar to citric acid, According to the literature, the anodic reactions during the film growth period come mainly from the oxidation of iron and chromium. The following chemical reactions indicate the oxidation of iron [5,6,86,89,90]: The samples passivated in nitric acid showed a higher trend of pitting corrosion. Noh et al. [91] conclude that nitric acid increases the chromium presence of the passive layer, removing MnS inclusions from the surface; also, the probability of individual pitting increases. In this research, the samples passivated in nitric acid presented more pitting than those passivated within citric acid due to the presence of MnS. To reduce the presence of MnS, changes in the acid concentration or the use of other solutions similar to citric acid, where the pitting process was more controlled, should be used.
From the above reactions, MnS can be removed, and the passivation stability could be related to the acid concentration [92].
Research by Lara et al. and Gaydos et al. [5,22] conclude that citric acid can be used as an alternative to nitric acid since citric acid passivation showed better results compared to the nitric acid solution specified in AMS-QQ-P-35. Furthermore, citric acid helps to remove iron and some particle contamination; for this reason, samples passivated within citric acid presented a lower trend to localized process. Since the passivation process was not generated in preferential zones, the oxide layer formed will be more uniform, and because of this, it could be expected that the corrosion process will be one of uniform nature.

Conclusions
After studying the corrosion behavior of CUSTOM450 and AM350 stainless steels passivated in citric and nitric acid solutions for 50 and 75 min at 49 and 70 • C and exposed to 5 wt. % NaCl and 1 wt. % H 2 SO 4 solutions, the following conclusions can be drawn:

•
After evaluation by EN and irrespective of the method used, i.e., time domain (statistical analysis) and the frequency domain (power spectral density, PSD), PHSS samples passivated in citric acid showed lower corrosion resistance. • AM 350 stainless steel passivated samples showed higher corrosion resistance than passivates of Custom 450 passivated stainless steel samples. • Passivation in nitric acid baths made the surface prone to the localized corrosion process. In this condition, statistical and PSD values showed a tendency toward pitting corrosion. • Passivated samples immersed in H 2 SO 4 solution showed an increase in corrosion kinetics due to the breakdown and passivation of the passive layer.

•
The samples passivated in citric acid showed lower corrosion resistance. However, samples passivated in nitric acid showed non-uniform passivation due to localized corrosion.

•
The samples passivated in the citric acid bath and immersed in NaCl solution are the ones that presented the best corrosion behavior. Likewise, and from the EIS results the AM 350 passivated stainless steel samples immersed in H 2 SO 4 solution presented the highest corrosion rates.

•
The EIS results obtained for passivated PHSS have been fitted using an equivalent electrical circuit model where RS is the uncompensated solution resistance; R1 and CPE are the interface resistance and the constant phase element, respectively.
• On the whole, the results indicate the effectiveness of citric acid in the passivation of PH stainless steel, and its potential as an alternative to replacing nitric acid passivation solutions currently used in the aeronautical and aerospace industries. Institutional Review Board Statement: Not applicable.
Informed Consent Statement: Not applicable.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.