Corrosion Resistance of Multilayer Coatings Deposited by PVD on Inconel 718 Using Electrochemical Impedance Spectroscopy Technique

AlCrN/TiSi, AlCrN/TiCrSiN and AlCrN/AlCrN + CrN coatings were deposited on Inconel 718 alloy by physical vapour deposition (PVD). The corrosion behaviour of uncoated and coated specimens was evaluated using electrochemical impedance spectroscopy (EIS) at open circuit potential in a 3.5 wt.% NaCl and 2 wt.% H2SO4 solutions. The EIS data acquired were curve fitted and analysed by equivalent circuit models to calculate the pore resistance, the charge transfer resistance and the capacitance. The Nyquist diagrams of all systems showed one part of the semicircle which could relate that reaction is a one step process, except for the AlCrN/TiCrSiN and AlCrN/AlCrN + CrN coatings in H2SO4 solution, for which two semicircles related to active corrosion in substrate alloy were found. However, from the Bode plots, it was possible to identify two the time constants for all systems exposed to NaCl and H2SO4 solutions. According to electrochemical results, the corrosion resistance of the AlCrN/TiSiN coating was better in the NaCl solution, whereas the AlCrN/AlCrN + CrN coating show better performance in the Sulphuric Acid solutions.


Introduction
The aviation industry demands improvements in the characteristics of the structural and functional materials components of aircraft based on scientific research conducted on new materials. The intrinsic resistance of alloys alone is insufficient to protect structural components from an aggressive environment. Coatings have been applied since the 1950´s as a method of surface protection that does not have enough resistance to corrosion at high temperatures [1][2][3]. Initially, aluminium diffusion coatings by packing cementation on Ni-based superalloys were used [4]. These coatings provided better corrosion protection to the substrate against corrosive agents. Subsequently, other forms of coating deposition such as plasma spraying and vapour deposition techniques for high-temperature applications such as thermal barrier coatings in aircraft gas turbines were used [5]. The basic physical vapour deposition nitride have a columnar structure with pinholes and pores and the electrolyte lead to the substrate. For iron and steels, Grabke [30] has shown the beneficial effect of N against localized corrosion i.e., the presence of nitrides ion may inhibit pit initiation by suppressing chlorides ion.
From the above, it can be seen that there are a good body of research about the corrosion behaviour of multilayer coatings by PVD techniques deposited on metallic substrates from mild steel, stainless steel (304SS and 17-4PH), tool steels and Ti alloys. However, none has involved AlCrN/TiSiN, AlCrN/TiCrSiN and AlCrN/AlCrN + CrN PVD multilayer coatings on Inconel 718 alloy corrosive behavior.
The advanced coating materials used for wear resistance, have proven themselves to be effective hard barriers for resisting wear of the substrate metal/metal alloys. These coating varies from cobalt boride, hard diamond-like carbide (DLC) coatings, boron nitride composites, chromium nitride, aluminium trioxide, aluminium-chromium nitride, nickel chromium-chromium trioxide, zirconium dioxide [31][32][33][34] Srinatah and Ganesha, comment that the literature have shown that both wear and corrosion resistance can be achieved through coatings and selective heat treatments of metals/metal alloys. The Studies of Wear and corrosion resistance of titanium carbo-nitride coated Al-7075 produced through PVD to characterize the bonding, using a potentiodynamic test and electrochemical impedance spectroscopy. The results of both the wear and the corrosion performance investigations were in good agreement with each other. Thus, the tribo-corrosive deterioration of titanium carbo-nitride coated Al-7075 had escalated with an increase in durations of the heat treatment [34].
The aim of this work was to study the corrosion behavior of multilayered coatings (AlCrN/TiSiN, AlCrN/TiCrSiN and AlCrN/AlCrN + CrN) deposited by PVD on Inconel 718. The electrochemical properties of the coatings have been studied using electrochemical impedance spectroscopy in two solutions: 3.5 wt.% NaCl and 2 wt.% H 2 SO 4 . This coating system on this superalloy might find potential applications in temperature sensors from the brake systems an aircraft.

PVD Coating Deposition
The nitride AlCrN/TiSiN, AlCrN/TiCrSiN and AlCrN/AlCrN + CrN coatings were deposited by cathodic arc plasma assisted physical vapour deposition (PVD) using an Al 70%-Cr 30% target with nitrogen gas (N 2 ) ware used to obtain AlCrN coatings at 450 • C. Subsequently, TiSi, TiCrSi and Cr targets were used at 350 • C. The deposition were: (i) 60 min for AlCrN layer and (ii) 20 min for TiSiN, TiCrN and AlCrN + CrN layers. The Ar and N 2 flow were independently controlled using a mass flow controller [Fox Thermal, CA, USA], (approximately 20 and 10 sccm, standard cubic centimetres per minute). The Vacuum chamber [Oerlikon, Balzers, Leichtenstein] (10 −5 mbar), DC-substrate bias voltage was in the range of −40 to −150 (V). The cathode current (A) was 60 A and total gas pressure was 1.5 (Pa). Table 1 shows the classification of the different PVD coatings used.

Microstructural Characterization
Scanning Electron Microscopy (SEM, Jeol JSM 6510LV, Tokyo, Japan) was as well utilized for identifying the cross-sectional micrographs of PVD coatings a magnification of 2000X, operating at 20 kV, WD = 10 mm. The chemical composition of these multilayered structures was obtained by Energy Dispersive X-ray Spectroscopy (EDS, EDAX, Tokyo, Japan).

Electrochemical Technique
Electrochemical impedance spectroscopy (EIS) tests on uncoated and coated samples were carried out at 25 • C at the corrosion potential using a Gill-AC potentiostat/galvanostat/ZRA from ACM Instruments (Manchester, UK). Tests were performed in a conventional three-electrode cell where: the specimen (uncoated and coated, 1.0 cm 2 exposed area) was connected to a working electrode; reference and auxiliary electrodes were saturated calomel (SCE) and platinum wire, respectively. The two electrolytes were used: 3.5 wt.% NaCl and 2 wt.% H 2 SO 4. After the corrosion potential stabilized, a sinusoidal a. c. signal of amplitude 10 mV (root mean square, rms) was applied coupled with the corrosion potential over a frequency range from 10 kHz to 1.0 mHz obtaining 10 points per decade according to the ASTM G106-15 standard [35]. The experimental results were interpreted through the development of typical impedance models for the electrode surfaces, and curve fitting based on equivalent circuits (using the Zview impedance program).

Morphological Analysis
SEM/EDS cross-sectional analysis of the AlCrN/TiSiN (C1 coating) show strong peak signals for Ti, Al, Si and N at the top of the coating, whilst a strong signal for Al, Cr, and N near the coating/substrate interface was detected, indicating that the coating is mainly formed by TiAlN at the top and AlCrN near the metallic substrate see (Figure 1a) Coatings 2020, 10

Microstructural Characterization
Scanning Electron Microscopy (SEM, Jeol JSM 6510LV, Tokyo, Japan) was as well utilized for identifying the cross-sectional micrographs of PVD coatings a magnification of 2000X, operating at 20 kV, WD = 10 mm. The chemical composition of these multilayered structures was obtained by Energy Dispersive X-ray Spectroscopy (EDS, EDAX, Tokyo, Japan).

Electrochemical Technique
Electrochemical impedance spectroscopy (EIS) tests on uncoated and coated samples were carried out at 25 °C at the corrosion potential using a Gill-AC potentiostat/galvanostat/ZRA from ACM Instruments (Manchester, UK). Tests were performed in a conventional three-electrode cell where: the specimen (uncoated and coated, 1.0 cm 2 exposed area) was connected to a working electrode; reference and auxiliary electrodes were saturated calomel (SCE) and platinum wire, respectively. The two electrolytes were used: 3.5 wt.% NaCl and 2 wt.% H2SO4. After the corrosion potential stabilized, a sinusoidal a. c. signal of amplitude 10 mV (root mean square, rms) was applied coupled with the corrosion potential over a frequency range from 10 kHz to 1.0 mHz obtaining 10 points per decade according to the ASTM G106-15 standard [35]. The experimental results were interpreted through the development of typical impedance models for the electrode surfaces, and curve fitting based on equivalent circuits (using the Zview impedance program).

Morphological Analysis.
SEM/EDS cross-sectional analysis of the AlCrN/TiSiN (C1 coating) show strong peak signals for Ti, Al, Si and N at the top of the coating, whilst a strong signal for Al, Cr, and N near the coating/substrate interface was detected, indicating that the coating is mainly formed by TiAlN at the top and AlCrN near the metallic substrate see (Figure 1a)   For the C3 coating, Figure 1c shows a homogeneous coating. Strong EDS signals for Cr, Al and N where detected at the top of the coating, whilst near the coating/substrate interface, Al, Cr, N and Ni were detected (Ni comes from the metal substrate). The coating thickness for C1 and C2 was about 2 µm on each case, while the C3 coating had a thickness of about 4 µm Figure 2 shows the Nyquist diagrams for Inconel 718 and AlCrN/TiSi (C1), AlCrN/TiCrSiN (C2) and AlCrN/AlCrN + CrN (C3) coatings exposed to 3.5 wt.% NaCl solution. For all systems, the Nyquist diagrams in Figure 2, show incomplete semi-circles, being this behaviour associated with high impedance values [36][37][38].  Figure 2. Nyquist diagram of substrate Inconel 718 and C1, C2 and C3 coatings exposed to 3.5 wt.% NaCl solution.

Corrosion Behaviour of PVD Coatings and Substrate
Only one semi-circle was shown for substrate SU alloy and AlCrN/TiSiN (C1) coating which are incomplete due to high impedance values when was tested in H2SO4 solution ( Figure 3). However, for the AlCrN/TiCrSiN (C2) and AlCrN/AlCrN + CrN (C3) coatings, the Nyquist plot show flattening of their semicircles under exposure to 2 wt.% H2SO4 solution. In this case, the smaller semi-circle (high frequency) can be related to the coatings capacitance whilst the lower frequency semi-circle is related to the double layer capacitance. This likely indicates initiation of corrosion of the substrate alloy, and subsequently a major coating damage is expected as the electrolyte penetrates the PVD coatings [39].  For the C3 coating, Figure 1c shows a homogeneous coating. Strong EDS signals for Cr, Al and N where detected at the top of the coating, whilst near the coating/substrate interface, Al, Cr, N and Ni were detected (Ni comes from the metal substrate). The coating thickness for C1 and C2 was about 2 µm on each case, while the C3 coating had a thickness of about 4 µm Figure 2 shows the Nyquist diagrams for Inconel 718 and AlCrN/TiSi (C1), AlCrN/TiCrSiN (C2) and AlCrN/AlCrN + CrN (C3) coatings exposed to 3.5 wt.% NaCl solution. For all systems, the Nyquist diagrams in Figure 2, show incomplete semi-circles, being this behaviour associated with high impedance values [36][37][38].  For the C3 coating, Figure 1c shows a homogeneous coating. Strong EDS signals for Cr, Al and N where detected at the top of the coating, whilst near the coating/substrate interface, Al, Cr, N and Ni were detected (Ni comes from the metal substrate). The coating thickness for C1 and C2 was about 2 µm on each case, while the C3 coating had a thickness of about 4 µ m Figure 2 shows the Nyquist diagrams for Inconel 718 and AlCrN/TiSi (C1), AlCrN/TiCrSiN (C2) and AlCrN/AlCrN + CrN (C3) coatings exposed to 3.5 wt.% NaCl solution. For all systems, the Nyquist diagrams in Figure 2, show incomplete semi-circles, being this behaviour associated with high impedance values [36][37][38].  Figure 2. Nyquist diagram of substrate Inconel 718 and C1, C2 and C3 coatings exposed to 3.5 wt.% NaCl solution.

Corrosion Behaviour of PVD Coatings and Substrate
Only one semi-circle was shown for substrate SU alloy and AlCrN/TiSiN (C1) coating which are incomplete due to high impedance values when was tested in H2SO4 solution ( Figure 3). However, for the AlCrN/TiCrSiN (C2) and AlCrN/AlCrN + CrN (C3) coatings, the Nyquist plot show flattening of their semicircles under exposure to 2 wt.% H2SO4 solution. In this case, the smaller semi-circle (high frequency) can be related to the coatings capacitance whilst the lower frequency semi-circle is related to the double layer capacitance. This likely indicates initiation of corrosion of the substrate alloy, and subsequently a major coating damage is expected as the electrolyte penetrates the PVD coatings [39]. Only one semi-circle was shown for substrate SU alloy and AlCrN/TiSiN (C1) coating which are incomplete due to high impedance values when was tested in H 2 SO 4 solution ( Figure 3). However, for the AlCrN/TiCrSiN (C2) and AlCrN/AlCrN + CrN (C3) coatings, the Nyquist plot show flattening of their semicircles under exposure to 2 wt.% H 2 SO 4 solution. In this case, the smaller semi-circle (high frequency) can be related to the coatings capacitance whilst the lower frequency semi-circle is related to the double layer capacitance. This likely indicates initiation of corrosion of the substrate alloy, and subsequently a major coating damage is expected as the electrolyte penetrates the PVD coatings [39].

Discussion
The electrochemical results obtained show a variation between impedance measurements and the AC polarization from the different coatings and metallic substrate.A quantitative analysis of EIS was performed using two different equivalent circuit models for the different systems in the present work, Figure 4.  The capacitive behavior exhibited for the SU substrate in NaCl solution suggests that formation of a highly stable film at pores is also capacitive. It has been reported that a dense and stable Cr2O3 oxide film layer is responsible of the passive behavior observed for the 718 alloy exposed in the 3.5 wt.% NaCl solution [24]. The EIS behaviour for the PVD coatings, can be related to two electrochemical process, i.e., two kinetic phenomena in each sub-interface-electrolyte/coating and electrolyte/substrate [40]. This corrosion mechanism may be related to a decrease in defect resistance as the electrolyte solution penetrates through the C1, C2 and C3 coatings and creates a diffusion path towards the base metal. It has been mentioned that galvanic corrosion may take place due to the fact that hard coatings are generally more noble than the metal [41]. The Nyquist plot of mono and multilayered CrAlN and CrAlN/SiNx coatings on 420 SS exposed in 3.5 wt.% NaCl solution, revealed a single semicircle for all samples, which was associated with insufficient exposure time in order to start the corrosion in the substrate alloy. The CrAlN coating show pitting corrosion resulting from the coating defects existing within the coating. The multilayered coatings had better corrosion protection than the single-layered one [42,43].
These figures also show the fitting data of the EEC model (see Figures 3 and 4 upper part), which matches well with the experimental data shown as individual points. The agreement of the data and the low χ 2 values corroborate the accuracy of the proposed EEC model.
The distributed electrical parameter, the constant phase element (CPE) is defined by Equation (1):

Discussion
The electrochemical results obtained show a variation between impedance measurements and the AC polarization from the different coatings and metallic substrate.A quantitative analysis of EIS was performed using two different equivalent circuit models for the different systems in the present work, Figure 4.  Figure 3. Nyquist diagram of substrate Inconel 718 and C1, C2 and C3 coatings exposed to 2 wt.% H2SO4 solution.

Discussion
The electrochemical results obtained show a variation between impedance measurements and the AC polarization from the different coatings and metallic substrate.A quantitative analysis of EIS was performed using two different equivalent circuit models for the different systems in the present work, Figure 4.  The capacitive behavior exhibited for the SU substrate in NaCl solution suggests that formation of a highly stable film at pores is also capacitive. It has been reported that a dense and stable Cr2O3 oxide film layer is responsible of the passive behavior observed for the 718 alloy exposed in the 3.5 wt.% NaCl solution [24]. The EIS behaviour for the PVD coatings, can be related to two electrochemical process, i.e., two kinetic phenomena in each sub-interface-electrolyte/coating and electrolyte/substrate [40]. This corrosion mechanism may be related to a decrease in defect resistance as the electrolyte solution penetrates through the C1, C2 and C3 coatings and creates a diffusion path towards the base metal. It has been mentioned that galvanic corrosion may take place due to the fact that hard coatings are generally more noble than the metal [41]. The Nyquist plot of mono and multilayered CrAlN and CrAlN/SiNx coatings on 420 SS exposed in 3.5 wt.% NaCl solution, revealed a single semicircle for all samples, which was associated with insufficient exposure time in order to start the corrosion in the substrate alloy. The CrAlN coating show pitting corrosion resulting from the coating defects existing within the coating. The multilayered coatings had better corrosion protection than the single-layered one [42,43].
These figures also show the fitting data of the EEC model (see Figures 3 and 4 upper part), which matches well with the experimental data shown as individual points. The agreement of the data and the low χ 2 values corroborate the accuracy of the proposed EEC model.
The distributed electrical parameter, the constant phase element (CPE) is defined by Equation (1): The capacitive behavior exhibited for the SU substrate in NaCl solution suggests that formation of a highly stable film at pores is also capacitive. It has been reported that a dense and stable Cr 2 O 3 oxide film layer is responsible of the passive behavior observed for the 718 alloy exposed in the 3.5 wt.% NaCl solution [24]. The EIS behaviour for the PVD coatings, can be related to two electrochemical process, i.e., two kinetic phenomena in each sub-interface-electrolyte/coating and electrolyte/substrate [40]. This corrosion mechanism may be related to a decrease in defect resistance as the electrolyte solution penetrates through the C1, C2 and C3 coatings and creates a diffusion path towards the base metal. It has been mentioned that galvanic corrosion may take place due to the fact that hard coatings are generally more noble than the metal [41]. The Nyquist plot of mono and multilayered CrAlN and CrAlN/SiN x coatings on 420 SS exposed in 3.5 wt.% NaCl solution, revealed a single semicircle for all samples, which was associated with insufficient exposure time in order to start the corrosion in the substrate alloy. The CrAlN coating show pitting corrosion resulting from the coating defects existing within the coating. The multilayered coatings had better corrosion protection than the single-layered one [42,43].
These figures also show the fitting data of the EEC model (see Figures 3 and 4 upper part), which matches well with the experimental data shown as individual points. The agreement of the data and the low χ 2 values corroborate the accuracy of the proposed EEC model. The distributed electrical parameter, the constant phase element (CPE) is defined by Equation (1): where Y p is the admittance, j 2 = (−1), ω is the angular frequency (rad/s), and the dimensionless n (−1< n <1) is the power of the CPE [44]. The value of Y p is of the order of the double-layer capacitance (10−100 µF/cm 2 ) typical of a charge transfer process [45].
The electrical equivalent circuit model proposed in Figure 4b, the constant phase elements CPE c and CPE s which represent the pseudo-capacitances at the solution/coating and coating/substrate interfaces and their empirical exponent n 1 and n dl respectively, are listed in Table 2. Also, Table 2 shows the resistance R e which represents the ionic solution resistance between the working electrode and reference electrode, pore resistance R pore which relate the resistance of ion-conducting paths the develop the coating and transfer resistance at the coating/substrate interface R s for uncoated alloy and PVD coatings. The CPE element represents a non-ideal capacitor according to the n value and its capacity is determined by faradaic and adsorption charging at the double layer. The n value is related to surface homogeneities (n = 1 for a pure capacitor, n = 0 for pure a resistor and n = 0.5 for a Warburg impedance (diffusion)) [22]. Given the fact that the n 1 value for the C3 coating was higher than that of the C1 and C2 coatings, the C3 coating had a less active behaviour in the 3.5 wt.% NaCl solution. However, once the electrolyte reaches the substrate, the system shows a resistive behavior. For this reason, two inflection points appeared in the frequency range studied in the phase angle vs frequency plot, suggesting the involvement of mass transfer through pores. Similar case occurs for the C2 coating which had high susceptibly to corrosion the coating/substrate interface. On the other hand, the n 1 value for the C1 coating was lower than its corresponding n dl value. Table 2. EIS data obtained by equivalent circuit simulation of coated and uncoated specimens exposed in 3.5 wt.% NaCl solution. Electrical equivalent circuits for PVD coatings has been reported widely to fit the electrochemical impedance data [29,[46][47][48][49][50][51][52]. An EIS study on CrAlSi x N nitride coatings on 420 SS substrate revealed important improvements in charge-transferred resistance (R s ) and n values with increasing Si content. These improvements were associated to a dense microstructure of the CrAlSi x N coatings [53]. Another EIS study of TiN and Ti-Si-N coatings on 304 SS reported the corrosion improvement when Si was added to the TiN. For TiN alone, the presence of holes provides a diffusion path for the electrolyte eventually reaching the coating/substrate interface. For the Ti-Si-N coatings, their microstructure provides longer diffusion paths. When localized corrosion occurs, the corrosion products block the diffusion path and hinder the penetration of the electrolyte, which will improve the R ct value in the Ti-Si-N coating [48]. According to the above mentioned, this could relate to why the PVD AlCrN/TiSiN C1 coating obtained values of n 2 of 0.880 and R pore of 11323 Ω cm 2 .

EIS
The corrosion behaviors of TiN/TiAlN multilayer coatings in 3.5 wt.% NaCl solution was reported [50]. In this study, the plots show two time-constants are representing the capacitive response of double layer and coatings. Values between 10,300 and 61,800 Ω cm 2 for R s were measured according the number of layers as of coating. In the present work for PVD C1, C2 and C3 coatings, the R s was 7,940,000, 763,000 and 9,220,000 Ω cm 2 , respectively. It is possible to relate that increased layered interfaces and high structural density can improve corrosion resistance as reported [51].
The total polarization resistance defined by R p = R pore + R s can be considered as an indicator of the corrosion resistance of the material. The R p represents the dielectric properties of coatings and passive substrate in pores which is inversely proportional to the corrosion rate [27,28,[51][52][53][54]. R p values around 3000 and 800 kΩ for multilayer Cr/CrN/CrAlN coatings immersed in 3.0 wt.% NaCl has been reported, indicating high corrosion resistance [54]. In the present work, R p values for the C1, C2 and C3 coating systems were 7950, 766 and 9230 kΩ cm 2 , respectively.
From the Nyquist plots obtained for the substrate alloy (SU) exposed in the 2 wt.% H 2 SO 4 solution, an equivalent circuit is proposed in Figure 4a, indicating the presence of a single layer (capacitance behavior) were R e is the solution resistance, R is the passive film resistance and CPE 1 the constant-phase element characteristic for a passive film [55,56]. For the coated specimens the EIS data were fitted by a two-time constants equivalent circuit, Figure 4b, and the parameters for this equivalent circuit are shown in Table 3. Table 3. EIS data obtained by equivalent circuit obtained simulation of coated and uncoated specimens exposed in 2 wt.% H 2 SO 4 solution. It is important to mention that R s and double layer capacitance (CPE -s ) are dependent on the affected area of the substrate. As the area increases, R s decreases, and the reverse is true for CPE -s . Due to this, the H 2 SO 4 electrolyte resistance is the lowest compared to the NaCl solution, thus, greater corrosion is often uncoated specimens could be expected under exposure to the H 2 SO 4 solution at the metal surface.

Conclusions
The PVD AlCrN/TiSiN and AlCrN/TiCrSiN coatings did not present a homogeneous and compact bonding with the bottom layer and substrate. It could have been due that atoms of the top layer did not diffuse into the bottom layer and substrate, respectively.
The EIS measurements of uncoated substrate showed the presence of a single time constant. However, the corrosion mechanism for all PVD coatings was determined by an equivalent circuit containing two-time constants, which described two electrochemical process in the electrolyte/coating and coating/substrate interfaces.
The corrosion resistance of the AlCrN/TiSiN coating was better in the sodium chloride solution, whereas the AlCrN/AlCrN + CrN coating show better performance in the Sulphuric Acid solution. In both cases they had, the lowest corrosion current density, highest protective efficiency and largest charge transfer resistance than that of other bilayer.
PVD technology is becoming an emerging alternative for the aeronautical industry to produce hard coatings with applications in temperature sensors from the brake system of an aircraft. Funding: This research was funded by the Conacyt, Proyect "Estudio de las propiedades electroquímicas y mecanismos de crecimiento de la película de pasivación de aceros inoxidables endurecibles por precipitación en ambientes ácidos", con clave No. A1-S-8882 and UANL (Dirección de Investigación).