Electrochemical Corrosion Behavior of TiAlN/CrN Nanoscale Multilayer Coatings by Multi-Arc Ion Plating in 3.5% NaCl Solution

Stainless steel with a minimum of 10.5% chromium content by mass is applied extensively to the medical field and industrial equipment owing to its excellent corrosion resistance and its simple maintenance in aggressive environments.^1-4  The superior anticorrosion property of stainless steel is derived from the protective surface Cr-rich oxide film, which can insulate the substrate against corrosive attack. However, film failures, such as pitting, crevice, intergranular, and galvanic corrosion, occur because of a destruction of the passive film, especially in certain critical environments.^5-9  In recent years, various surface modification techniques, such as plasma nitriding,^10-12  chemical vapor deposition,^9,13  and physical vapor deposition^14-19  have been used to improve the abrasion and corrosion resistance of stainless steel.

Transition-metal-nitride coatings, such as TiN,^16,19-22  CrN,^23-24  and TiAlN²⁵  that have been prepared by physical vapor deposition, can improve the resistance of a variety of steel substrates against corrosion damage when they are exposed to working conditions with a wide range of chemical corrosive gaseous or liquid media. The existence of growth defects in coatings (e.g., microcracks, pores, pinholes, and grain boundaries) is deleterious as they allow corrosive electrolytes to reach the coating/substrate interface, resulting in localized galvanic corrosion, despite the fact that the coating materials themselves are highly corrosion resistant.²⁶  This process results in a breakdown of monolayer coatings because of their columnar crystal structure. Some attempts have been made to obtain desirable coatings with a better corrosion resistance. Compared with monolithic films, a multilayer architecture can inhibit defect growth owing to its high interfacial energy density and improve the coatings’ ability to withstand corrosive media.^22,27 

Of the nitride-based hard coatings, TiAlN was developed to possess a high corrosion-protection ability, which is attributed to the existence of elemental Al that formed a passive film against corrosive damage. Despite the formation of the film, the anticorrosion capacity of the TiAlN coating does not reach requirements for poor working conditions because of these intrinsic defects.^28-31  To improve the corrosion resistance, because CrN is a prospective anticorrosion film owing to its dense passive surface Cr-rich oxide film, the TiAlN/CrN multilayer coating is attractive. In a previous study, a TiAlN/CrN multilayer coating from magnetron sputtering showed a satisfying oxidation resistance and hardness at a high temperature. Therefore, previous studies^32-34  have concentrated mainly on physical and mechanical properties, with little attention given to investigate the electrochemical behavior of TiAlN/CrN multilayer coatings in saline corrosive environments.

A layered design method can improve the fracture toughness and adhesion strength of the TiAlN/CrN multilayer coating considerably,^35-36  and is associated with a high interfacial energy density owing to the small modulation wavelength (also termed the bilayer thickness, i.e., the thickness of each successive pair). It is assumed that the layered structure may increase the interfacial energy density and strengthen its corrosion resistance.

To prove this assumption, three TiAlN/CrN multilayer coating types were prepared on stainless steel by cathodic arc physical vapor deposition (CA-PVD) at varying bilayer periods (2 bilayers, 4 bilayers, and 20 bilayers). Microstructure characterization and composition analysis of the as-deposited coatings were obtained by x-ray diffraction (XRD), scanning electron microscopy (SEM), and transmission electron microscopy (TEM). The electrochemical behavior of the coatings was investigated in 3.5% NaCl aqueous solution at ambient temperature using the potentiodynamic polarization technique and electrochemical impedance spectroscopy (EIS). The corrosive morphology of the substrate and coatings after 15 d of immersion was investigated to establish the corrosive mechanism. The uncoated stainless steel substrate was analyzed for comparison. The relationship between the layer design mode and the corrosion mechanisms of the TiAlN/CrN multilayer coatings is discussed.

TiAlN/CrN multilayer coatings were prepared on substrates of Type 316L (UNS S31603⁽¹⁾) stainless steel and commercial monocrystal Si wafers with a (100) plane by CA-PVD (Oerlikon Balzers^† mini). The Type 316L stainless steels with a chemical composition of C (0.04), Si (0.65), Mn (1.61), Cr (15.57), Mo (1.65), Ni (11.76), and Fe (balance) were cut into 10 mm × 10 mm × 6 mm pieces. Prior to deposition, the substrates were given a mirror-like surface by grinding with a series of silicon-carbide papers and polishing with 1.5 μm diamond paste. The polished substrates and Si wafers were cleaned with ultrasound in ethyl alcohol and dried in cold air. The Si wafers were used for the microstructural features and composition analysis of the coatings, and the Type 316L stainless steels were used in the electrochemical experiments and immersion test.

Figure 1 shows a map of the deposition system inside the chamber. The substrates were fixed on rotating equipment with a two-fold rotation at a rotary speed of 2 r/min. The deposition device was equipped with two targets at the different chamber heights. Alternating layers of CrN and TiAlN in the TiAlN/CrN multilayer coatings were deposited with a pure Cr target and a Ti_0.5Al_0.5 alloy target (99.99%, 100 mm diameter). Before coating deposition, the substrates were cleaned in an argon-ion-etching process for 30 min in an argon atmosphere of 1 Pa and with a substrate bias of −200 V DC. The deposition parameters of the TiAlN/CrN multilayer coatings with different bilayer periods were a N₂ pressure of ∼3 Pa, a deposition temperature of 300°C, a substrate bias of −100 V DC, a target current of 130 A, and a deposition time of 2 h. During the deposition, a pure Cr buffer layer was deposited sequentially at a 1 Pa argon pressure for 10 min to enhance the bond strength. The CrN and TiAlN bilayers were deposited alternately, and the bilayer thickness was controlled by the deposition time of each layer. As shown in Table 1, multilayer coatings with different bilayer thicknesses were obtained and termed C1, C2, and C3, respectively.

The crystalline structure and preferred orientation of the as-prepared multilayer coatings were determined by grazing incidence x-ray diffraction (GIXRD, PANalytical X-Pert^†) with Cu Kα radiation (λ = 0.154 nm) in 2θ scan mode from 20° to 90° at a 10°/min scan speed. The cross-sectional microstructural characterizations of the as-prepared coatings were studied by SEM (Hitachi SU8010^†, Japan). The detailed microstructures and interfacial characterization of the as-prepared multilayer were examined by TEM by using a JEOL JEM-2200FS^† at 200 kV. TEM specimens were prepared using a dual-beam focused ion beam (FIB) system (FEI Helios Nanolab 600i^†) following a standard lift-out procedure. In the final preparation step, surface cleaning was applied at 5 kV and 41 pA to reduce the damage that was induced by FIB.

Electrochemical potentiodynamic polarization and EIS were conducted in 3.5% NaCl diluted aqueous solution at 25°C using an electrochemical workstation (CS2350^†, Wuhan, China). The standard three-electrode cell included a platinum counter electrode, a reference electrode of a saturated calomel electrode (SCE), and a working electrode of a specimen with a 1.0 cm² exposed area for both electrochemical measurements. Samples were loaded in the Teflon^† sample fixture with the same short distance to the Luggin capillary of the SCE for all tests after cleaning in distilled water. The open-circuit potential (OCP) was established by keeping the sample in solution for 60 min prior to the electrochemical tests. For each experiment, three samples were used to make sure the repeatability of the experiments.

The potentiodynamic polarization test was implemented at a 1 mV/s scan speed from −1 V to 1 V to measure the sample corrosion potential (E_corr) and corrosion current density (i_corr).

The EIS measurements were performed to obtain the difference in impedance behavior between the stainless steel and coated samples from 10⁵ Hz to 10⁻² Hz with an AC amplitude of 10 mV at the respective OCP. The EIS experimental data were displayed as Nyquist and Bode plots, and were fitted to an appropriate electrical equivalent circuit (EEC) using ZView^† software.

The crystalline structure of the TiAlN/CrN multilayer coatings with different bilayer periods are presented in GIXRD patterns in Figure 2(a). The diffraction peaks for all multilayer coatings with a B1-NaCl cubic structure were identified as the (111), (200), (220), and (311) planes. The texture coefficient can be used to quantify the change in preferred orientation for all multilayer films from Figure 2(b), and calculated from:³⁷   where I_m(hkl) and I₀(hkl) are the relative intensities of the reflection from the (hkl) plane of the coating and the standard powder specimen, respectively, and n represents the total number of reflection peaks from the coatings. Figure 2(b) showed that the intensity of the (111) preferential orientation for TiAlN and CrN weakened with the decrease in bilayer thickness and increased the intensity of the (200) peaks and slightly increased the intensity of the (220).

During deposition, the requirement for lowering the free energy, with strain and surface energies, governs the tendency of the preferred orientation for the thin film. Grains of the B1-NaCl structure could grow along the (100) plane with the lowest surface energy, whereas the (111) plane has the highest surface energy.³⁷  As shown in Figure 2(b), the TiAlN/CrN-1 multilayer film exhibits a preferential growth orientation of (111) and (311) and the TiAlN/CrN-2 and TiAlN/CrN-3 multilayer coatings changed to (200). For the TiAlN/CrN multilayer coating with a 200 nm bilayer thickness (C1), the preferential orientation is (111) and not (200) because the strain energy of the lattice distortion in the TiAlN from Ti substitution by Al cannot be balanced with the interfacial energy. The preferential orientation changed to (111) to reduce the strain energy for TiAlN coating. The growth orientation of the TiAlN/CrN multilayer films may change from the (111) to the (200) plane with the increase in bilayer period, which could be attributed to an increase in the density of interfacial energy that is associated with multilayer growth; growth of the TiAlN film was interrupted constantly by the growing CrN coating. This interruption suppressed the growth of (111) grains and promoted (200) preferential grain growth by providing a sufficient driving force, which resulted from the lower strain energy density of the (200) grains compared with that of the (111) grains.^38-39 

The fracture cross-sectional field emission SEM (FESEM) morphologies of the TiAlN/CrN multilayer films with different bilayer thicknesses are shown in Figure 3. The thicknesses of the three multilayer films with 2, 4, and 20 bilayers were 1.7 μm, 1.8 μm, and 1 μm, respectively. As shown in Figure 3(a), the dark and light layers are TiAlN and CrN coatings, respectively. The CrN layer thickness is approximately equal to that of the TiAlN layer in all cases, which is attributed to the same ratio in deposition time of both bilayers. Figure 3 shows that the bilayer thicknesses of the TiAlN/CrN multilayer coatings are ∼200 nm, 100 nm, and 20 nm. All three coatings exhibit a columnar morphology that expands from the initial substrate growth to the thickness of the entire coating epitaxially in Figure 3.

As shown in Figure 4(a), a cross-sectional scanning TEM (STEM) bright-field (BF) image of the C3 coating with the alternating bright TiAlN and dark CrN sublayers shows columnar grains elongated along the growth direction from the Cr buffer layer.³⁴  The cubic B1 structure for both layers was shown from the selective area electron diffraction (SAED) pattern (marked by the yellow region in Figure 4[a]) in accordance with the XRD results. The high-angle annular dark field (HAADF) image taken in STEM mode, as indicated in Figure 4(b), reveals an alternate bright and dark contrast of a nanoscale layered structure that corresponds to high (CrN sublayers) and low (TiAlN sublayers) average atomic numbers, respectively. The relative contrast of different sublayers is opposite for the HAADF image and TEM BF image. The high-resolution TEM (HR-TEM) micrographs (in Figures 4[c] and [d]) confirmed the coherent interfaces between the CrN and TiAlN layers in individual columnar grains of the C3 coating, where Moire stripes verify the polycrystalline structure.

The average bilayer thickness and modulation ratio of C3 are 20 nm and 1:1, which agree with the SEM results. The B1 cubic structure with (200) and (111) lattice planes was affirmed by two fast Fourier transformation (FFT) spectra of the square region, which agrees with the XRD results. The columnar grain morphology in the SEM fracture cross sections and single-phase structure in the FFT spectra may affirm the coherent multilayer interfaces, which is demonstrated by the inverse fast Fourier transformation (IFFT) images. IFFT shows a perfect lattice match between TiAlN and CrN sublayers with respective (111) and (200) lattice plane in the square region outlined by the box in Figures 4(c) and (d). No dislocation was observed, which verifies that in the structure of the coherent interface exists a slight lattice strain, which results from the different lattice parameters of TiAlN and CrN. This coherent interfacial bilayer state in multilayered coatings is driven by the minimum free energy, which depends on the strain energy and interfacial energy.³⁴ 

The respective potentiodynamic polarization of the three coatings (i.e., C1, C2, and C3) and the Type 316L stainless steel (SS) were used in natural 3.5% NaCl solution at room temperature, as shown in Figure 5. The values of the related electrochemical parameters extracted from the polarization curves, including the corrosion potential (E_corr), corrosion current density (i_corr), pitting potential (E_pit), and passivation potential interval (PPI), are summarized in Table 2. Figure 5 shows that all coating specimens in solution are automatically in passivation with no activated observable peaks. The pitting potential of the three multilayer coatings increased from −0.1 to 0.2 and 0.4 with the decrease in modulation periods. The passivation potential interval is almost 0.7 V for the C3 coating, and decreases to 0.6 V for the C2 coating and to 0.4 V for the C1 coating, respectively. For the stainless steel substrate, it dissolved in anodic initially and a uniform dissolution process was taking place through an unstable layer from 0 V. Corrosion potentials (E_corr) and corrosion current density (i_corr) were obtained using ZView^† software. Compared with the stainless steel, the Tafel curves of the three coatings shifted simultaneously to lower current densities and to positive corrosion potentials; this indicates that the corrosion reaction in anodic and cathodic branches of the three coatings occurs with difficulty. The corrosion potentials (E_corr) and corrosion current density (i_corr) of TiAlN/CrN multilayer coatings imply that the decrease in modulation period can improve the anticorrosion performance of the TiAlN/CrN multilayer coating.

Detailed information on the corrosion reaction, mass transfer, and electrical charge transfer characteristics of the multilayer coatings in saline solution can be investigated by important parameters that are deduced from EIS.^40-42 

Figure 6 shows the Nyquist and Bode plots of the three coated and uncoated samples at the OCP in natural 3.5% NaCl solution. The response for all investigated samples in the Nyquist plots (Figure 6[a]) exhibits a single semicircle that corresponds to only one time constant, and which originates from a passive film on the substrate and an exposure time that is too short to degrade the substrate.^43-44  Much larger semicircle diameters for the three coated samples display flattened capacitive arcs over the frequency range, which implies that the coated samples have a higher corrosion resistance compared with the uncoated sample. The corrosion resistance of the multilayer coated samples is enhanced further with a decrease in modulation period, as evidenced by the observation that C3 with the smallest modulation period is almost a straight line, which suggests the superior dielectric properties of the protective coating. The potentiodynamic polarization test yielded the same trend of corrosion resistance.

As shown in Figure 6(b), the electrochemical response of Type 316L stainless steel in the Bode plot was hardly distinguished from that of the multilayer coated samples, mainly because of the formation of a dense protective film on the stainless steel substrate and insufficient immersion time to corrode the substrate.²⁶  The Bode plots of all samples showed a phase angle at nearly −90° and a slope that approaches −1 in a broad frequency range, which suggests an obvious capacitive behavior for all samples. In general, the phase angle is associated with solution and coating resistance. For the same solution resistance, a more homogeneous and denser coating yield a higher phase angle, which implies a higher corrosion resistance.^45-46  Figure 6(b) shows that the substrate presented the highest phase angle, which is attributed mainly to a dense and homogeneous passive film on the surface.

A further interpretation of the impedance response can be obtained by analyzing the “equivalent circuit,” which models the physical and electrical characteristics of the electrode/solution interface by an assembly of electrical circuit elements. To describe the coated systems, an equivalent circuit model is used to adopt two hierarchically distributed time constants: dielectric behaviors of the coating and coating/steel interface at coating defects.^41,44,47-49  In this case, one time constant was observed, which suggests that the electrochemical behavior of the electrolyte/substrate interface does not have to be considered for such a short immersion time.²⁶  Based on this analysis and considering these three aspects: having a rational fitting for the experimental values, minimizing the number of circuit elements, and connecting these elements with the corrosion reactions at the electrode, the proposed equivalent circuit for such a system was fitted using software, shown in Figure 7. The physical meaning of the symbols in the circuit is: R_p is the polarization resistance, which evaluates the porosity of the coatings, and CPE and R_s are the constant-phase element and solution resistance, respectively. If the distributed relaxation feature of the coatings is considered, and to obtain a better fitting, a CPE was used to replace the double-layer capacitance (CPE_dl), which relates to the coating-state characteristic. The circuit description code for the equivalent circuit for all samples can be written as R(QR).

Values of the circuit elements in Table 3 that were obtained from the fitting model showed a good fit (χ² < 0.01%). The simulated data agreed satisfactorily with the experimental data over the frequency range. Table 3 shows that R_p increases from 1.5 × 10⁵ Ω·cm² for the Type 316L stainless steel substrate to 5.8 × 10⁵ Ω·cm² and 7.4 × 10⁵ Ω·cm² for C1 and C2 coatings, respectively, and to 1.5 × 10⁶ Ω·cm² for the specimen that was protected by the C3 multilayer coating. The decrease of R_p and increase of CPE reflect the deterioration of coating barrier properties and a weakness of the protective ability. The capacitance values were calculated to consider whether the resistance values were related to the solution, whereas the capacitance is independent and becomes a more reliable approach to evaluate the electrochemical properties of the electrode.⁵⁰  Based on Brug’s approach,⁵¹  the capacitance values (C) can be calculated from Equation (2).

C for the Type 316L stainless steel is 32.5 μF/cm², and for TiAlN/CrN multilayer coatings with three modulation periods, it decreases to 11.6 μF/cm², 15.4 μF/cm², and 13.3 μF/cm², respectively. Compared with passive film grown on the uncoated substrate, the values of C for the three coated samples are lower, which reflects the higher anticorrosion performance. In terms of the time constant (τ), which is used to describe the rate of a relevant parallel RC circuit appearing in the EC,^26,52  this is expressed as Equation (3).

As shown in Table 3, the τ values of the TiAlN/CrN multilayer coated samples increased from 6.7 s to 11.5 s and 20 s with a decrease of bilayer thickness, and for the substrate, it was only 4.8 s. Higher τ data mean that the corrosion behavior is slower. For the coatings, the C3 multilayer coating with the smallest thickness of 1 μm has the largest time constant and displays the strongest resistivity. The C1 and C2 multilayer coatings with closely equal thicknesses have different time constants, 6.7 s and 11.5 s, respectively, which shows that the C2 coating has a stronger resistivity than the C1. Therefore, the TiAlN/CrN multilayer structure can suppress electrochemical corrosion compared with the uncoated Type 316L stainless steel substrate, although the dense passive films existed naturally on the substrate.

The role of multilayer construction in the corrosion resistance of coatings in 3.5% NaCl solution was determined, and EIS tests and corroded surface examinations after immersion for an extended time were performed. Figure 8 shows Nyquist and Bode plots of the three multilayer coatings and Type 316L stainless steel after immersion for 15 d in 3.5% NaCl solution. Compared with the EIS in Figure 6, an obvious decrease of the capacitive semicircle diameter and the maximum phase angle was visible in Figure 8. For all samples, Figure 8(b) presents two time constants that are distributed in high- and low-frequency regions, which indicates that interfacial corrosion occurred at the coating/substrate,^53-54  and corresponds to the dielectric feature of the coatings or oxide layer and the charge transfer reaction that is induced by electrochemical corrosion activity on the stainless steel surface through pinholes and current concentrated at the pores.^26,55  The decrease in phase angle in Figure 8(b) after immersion suggests that the impedance response of all specimens becomes less capacitive, and localized corrosion occurs at the coating and substrate interface because of the pinholes. Figure 8(b) shows that at a high frequency, the peak height decreases from C3 to C2 and to C1, which indicates that the corrosion resistance of the multilayer coating weakens with an increasing bilayer period. At a low frequency, the peak height of C1 is almost the same as that of C3, and the lowest is C2, which may be attributed to the fact that corrosion products blocked the pinhole to retard electrolyte arrival at the substrate interface.

An equivalent circuit model that corresponds to the EIS of samples that were immersed for 15 d is shown in Figure 9 to explain the two different electrochemical behaviors. Parameters in the first section of the equivalent circuit R_c and CPE_c relate to the characteristics of the coating or oxide layer; R_ct and C_dl in the second circuit section correlated with the charge transfer reaction at the coating/substrate interface.

Values for the electrical elements in the EEC (Figure 9) that were obtained using software are compiled in Table 4. Compared with the C1, C2 coatings, and the substrate, the C3 coating had a lower CPE but a higher phase angle of CPE (n), which suggests a higher capacitance and indicates that the C3 coating achieved the lowest penetration of corrosive medium to the substrate for all samples. Conversely, the capacity behavior of the C1 coating with the highest CPE and lowest CPE (n) leaked more, which implies that some corrosion pinholes formed during immersion. Parameter R_c can evaluate the consistency of the coating and is inversely proportional to the porosity amount in the coating;⁵⁶  therefore, the porosity level of the coatings increased from C3 to C2, and to C1 during the 15 d immersion in 3.5% NaCl solution. The C3 coating had the lowest CPE_dl and highest R_ct, which shows a satisfactory corrosion resistance at the substrate interface, whereas the two parameters of the C1 and C2 coating indicated that significant galvanic corrosion occurred at the substrate interface. Compared with the R_p of the as-deposited coatings, the relative decay of the R_p for the C3 coating was the smallest, followed by the C2 coating, and the largest was the C1, which proves that the corrosion resistance of the coating improved with a decrease in bilayer thickness.

The sample corrosion morphology was studied to investigate the corrosion behavior of the coatings. Figure 10 shows that differences in the corrosion resistance of the three coatings and the substrate are manifested by the surface morphology after immersion for 15 d in 3.5% NaCl solution. Figures 10(a) and (b) show that numerous dimpled pits with different dimensions appeared on the stainless steel surface, which is ascribed to the destroyed dense passive film in a corrosion medium.

In terms of the multilayer coatings, different levels of corrosion are visible on the surface. Figure 10(i) was an example for the surface characteristic of three as-deposited coatings as they were almost the same. For the C1 coating (in Figures 10[c] and [d]), a large degree of localized corrosion spread on the surface, where numerous droplets during deposition (as shown in Figure 10[i]) were dissolved and became smaller. The droplet pits increased in size and gradually became shallow corrosion pits in the corrosion medium. Some pinholes were not visible owing to the corrosive products as mentioned previously. Instead, some shallow corrosion pits with regular circular edges occurred on the surface. Figures 10(e) and (f) show that the corrosion surface of the C2 coating with a few flat droplets was very smooth compared with the C1 coating, but some corrosion pits and delamination still occurred. On the corrosion surface of the C3 coating, shown in Figures 10(g) and (h), localized corrosion was absent, small droplets and their pits still existed, and the morphology was close to the original coating surface morphology (in Figure 10[i]). Thus, the C3 coating shows superior anticorrosion properties, compared with the C2 and C1 coatings that experienced more severe corrosion damage, which agrees with the electrochemical test results.

It is believed that the PVD coatings contain cracks, pinholes, and micropores, which allow the corrosive media to reach the substrate and degrade the coating/substrate system. The electrochemical tests and immersion corrosion morphology showed that the C3 coating had a superior corrosion resistance compared with the other modulation ratio coatings. The smallest modulation period of a multilayer coating may contribute to block the pinholes, halt electrolyte arrival at the substrate and interrupt galvanic corrosion propagation. A higher interface density may provide a more homogeneous, uniform, and denser coating to improve the corrosion resistance. The corrosion propagation mechanism can be modified by the individual layers in coatings. Consequently, the existence of a large number of interfaces between individual layers in a multilayer structure inhibits pitting corrosion propagation significantly.

Three TiAlN/CrN multilayer coatings with different modulation periods were deposited on the stainless steel. The functional roles of the multilayer construction in improving the corrosion resistance of the coating were explored. The electrochemical properties of the multilayer coatings were investigated in naturally aerated 3.5% NaCl solution using electrochemical measurement and surface characterization techniques. The following conclusions can be drawn:

• •

  With a decrease in modulation thickness and an increase in interface number, the preferential growth orientation of the TiAlN/CrN multilayer films changes from (111) to (200).

• •

  The SEM fracture cross-sectional features of all TiAlN/CrN coatings exhibit significant columnar structures that are elongated in the growth direction with obvious laminate structures. The TEM data show a perfect lattice match at the interface of the CrN and TiAlN layers in individual columnar grains.

• •

  Specific electrochemical characteristics for all as-deposited coatings suggest that the multilayer structure can decrease the passive current density and shift the corrosion potential to a positive value with a decrease in bilayer thickness. The EIS shows an increase of R_p and a decrease in CPE for the multilayer coatings with a decrease in the bilayer thickness. The C3 coating, with 20 bilayers, had the lowest time constant.

• •

  After 15 d of immersion, the EIS plot shows that the C3 coating had the lowest CPE_dl and highest R_ct, which indicates a satisfactory corrosion resistance. Compared with the as-deposited coatings, the relative decay of the R_p of the C3 coating was the smallest in all samples.

• •

  The surface corrosion morphology of the multilayer coatings verified that the decrease in modulation thickness can enhance the corrosion resistance.

This work was supported financially by the National Key Research and Development Program of China (2017YFB0305900).