Comparative study of corrosion and corrosion-wear behavior of TiN and CrN coatings on UNS S17400 stainless steel

2.1 Test materials and coating technique

UNS S17400 martensitic stainless steel bar (0.02% C, 16.46% Cr, 3.93% Ni, 0.94% Mn, 0.2% Nb, 3.24% Cu, and balanced Fe in wt%) was machined to 12 mm diameter and 4 mm length specimens as the substrate in this study. Two coatings were deposited by cathodic arc physical vapor deposition in an industrial chamber. After cleaning in an ultrasonic bath in acetone, the specimens were fixed in the chamber, evacuated and exposed to argon ions bombardment at 2 kV for 30 min for sputter cleaning of the surface. This treatment preheated the samples to above 200°C and ensured the surface cleanliness necessary for vapor deposition. The deposition process was started according to the parameters shown in Table 1.

On some specimens, at first, TiN with a thickness of approximately 0.6 μm was deposited as an intermediate layer, and then a TiAlN layer and another 0.6 μm TiN as a top coat was added to provide a total thickness of 2 μm. In other specimens, CrN with a thickness of approximately 0.3 μm was deposited at first, followed by the CrAlN layer and, finally 0.2 μm CrN top coats to provide a total thickness of 2 μm. Multilayer coating was performed by switching the cathodes of different targets on and off (Ti, Al, and Cr). After cooling in the chamber, the specimens were removed, cleaned by ethanol, and quickly dried. For metallography, the coated specimens were cut and mounted vertically, then ground progressively down to 2500 grade sand paper and ultrasonically cleaned in acetone.

2.2 Characterization

The phase structures were studied by an X-ray diffractometer using CuK_α radiation (λ=0.154 nm), with grazing incidence angle of 2°. The scanning angle ranged from 15° to 90° at a time step of 1 s with 0.1° step size. The PVD process commonly leads to a preferred growth of crystallography planes due to selective nucleation of crystallites, normally referred to as crystallographic texture (Matei et al., 2015). Quantitative estimation of the texture coefficient [TC(hkl)] can be expressed by Equation 1:

where I(hkl) is the measured relative intensity of the (hkl) line, I₀(hkl) is the intensity assigned to the same (hkl) planes of random orientation, or according to ICDD, and n is the number of peaks used to recognize the TC. A texture coefficient of greater than one indicates the abundance of crystallites oriented with (hkl) direction normal to the coating surface (Abadias, 2008).

Crystal lattice parameter described with the Nelson–Riley equation and the least squares method using all four fundamental diffraction data from each X-ray pattern as shown in Equation 2:

where θ is the diffraction angle and f(θ) is Nelson–Riley function. The lattice parameters obtained are plotted against cos², cos²θ/sinθ and the Nelson–Riley function, and the extrapolation of the curves to cos²=0 will provide the true value of the lattice parameter.

Lattice spacing is obtained by the Bragg relationship according to Equations 3 and 4:

For the cubic system,

where λ is the wavelength of incident radiation, d is the lattice spacing, θ is the diffraction angle, a is the lattice parameter and h, k, l are the Miller indices of the crystal plane (Yang et al., 2017). The coating surface, wear track morphologies, and specimen cross-section were observed by a field emission scanning electron microscope and the surface roughness was measured according to DIN EN ISO 4287-2010.

2.3 Corrosion and corrosion-wear test

Before the tests, the specimens were ultrasonically cleaned for 20 min, and 3.5% NaCl was used as an electrolyte in corrosion-wear and corrosion tests. Corrosion-wear tests were carried out in a ball-on-disk tribometer with a three-electrode system. Ag/AgCl electrodes as the reference electrode, Pt as the auxiliary electrode and specimen (coated and uncoated UNS S17400 stainless steel) as the working electrode were used in the tests. All specimens were soaking in 3.5% NaCl for 1 h to fix the open-circuit potential (OCP) before the test was begun. To start the test, the samples were placed inside the tribometer containing the electrolyte (Figure 1) and OCP was recorded for 2000 s before the sliding. A normal load of 5 N was applied on the SiC ball with 7 mm stroke length and the wear cycles of 0.5 Hz. The OCP recording was continued for 2000 s after the sliding was ended.

Figure 1:

Setup of corrosion-wear test.

In potentiodynamic polarization tests, the specimens were connected to a working electrode, and saturated calomel electrode (SCE) and pure graphite rods were used for reference and counter-electrodes, respectively. A total of 1 cm² of the specimen surface was in contact with the electrolyte and the other directions were covered by an insulating layer. The potentiodynamic polarization test was carried out using PARSTAT model 2273 potentiostat with potentials in the range –250 mV_OCP to 250 mV_SCE with a scan rate of 0.1 mV/s at room temperature (25°C). All tests were carried out at ambient temperature. To obtain the polarization resistance, the Stern–Geary relationship near the OCP potential was used, as expressed by Equation 5.

where R_p (cm² Ω) is the polarization resistance, β_a is the anodic Tafel slope, β_c is the cathodic Tafel slope, and i_corr (A/cm²) is the corrosion current density (Stern, 1958; Poursaee, 2010). On the assumption that the coating is electrochemically inert at low anodic overpotential, the porosity of the coating is estimated using Equation 6 (Creus et al., 2000).

where P is the total coating porosity rate, Rp_S and Rp are the polarization resistance of the substrate and coated steel, respectively. ΔE_corr is the difference potential between the free corrosion potentials of the coated steel and uncoated substrate, and β_a is the anodic Tafel slope for the substrate.

3.1 Structure and analysis of the coatings

X-ray diffraction patterns of typical coatings, produced by grazing angle technique, are shown in Figure 2 and the parameters are provided in Table 2. Both TiN and CrN coatings have fcc crystal lattice with a strong (200) preferred orientation. Generally, by increasing the thickness of the coating, preferred orientation changes from (200) to (111); (200) texture is stabilized by a high ion to neutral ratio during vapor deposition. This orientation can improve tribological properties and wear resistance. Because the active sliding plates in a cubic crystal are (111), the change of preferential orientation to (200) improved tribological and wear properties (Paulitsch et al., 2008). Four crystalline planes were indexed on diffraction patterns including (111), (200), (220), and (311) peaks of CrN and TiN phases. Titanium atomic radius is greater than that of chromium, thus, it has a larger lattice parameter with all the diffraction peaks are located toward smaller angles as compared with chromium nitride. The lower lattice parameter of the chromium nitride coating indicates that the coating is more compact than titanium nitride. This can improve the corrosion resistance of the coating.

Figure 2:

XRD patterns of the TiN and CrN top coatings.

SEM micrographs of the coated and uncoated specimens are shown in Figure 3. The substrate was composed of martensite and a small amount of ferrite as in Figure 3A; a typical inclusion is also observed in the micrograph, probably composed of MnS. Figure 3B and C show the surface topography of CrN and TiN top-coated specimens, respectively. There are some microparticles and pinholes on the surface of coatings that are common features of cathodic arc vapor deposition. Average surface roughness (R_a) of the coatings was estimated at 115 and 75 nm for TiN and CrN top coatings, respectively. SEM micrographs of the cross-sections of multilayer coatings are shown in Figure 3D and E. These coatings exhibit a typical compact structure without cracking and delamination that indicate, conforming a good adhesion between layers and the substrate. The total thickness of the coating was estimated at approximately 2±0.2 μm for TiN and CrN multilayers.

Figure 3:

SEM micrographs of surface and cross-sections of PVD-coated UNS S17400 stainless steel: (A) substrate, (B, D) topography and cross-section images of CrN top coat, (C, E) topography and cross-section images of TiN top coat.

3.2 Evaluation of corrosion and corrosion-wear

Figure 4 shows the polarization curves of the substrate and multilayer coatings. As shown in Table 3, the corrosion potential of the uncoated UNS S17400 stainless steel was –456 mV. TiN coating exhibited a higher corrosion resistance with a corrosion potential of –462 mV, and CrN still had a better corrosion behavior with a corrosion potential of –485 mV. By increasing the corrosion resistance, the corrosion current density of the specimen was decreased; these indications confirm that transition metal nitrides are not affected by chemical attacks, thereby increasing the corrosion resistance (Grips et al., 2006). The passive film on the nitride coating is enhanced by the presence of nitrogen and has an effective role in increasing the corrosion resistance. Nitrogen in the film reinforced the formation of NH₃ and further NH₄⁺ compounds in addition to metal oxides on the surface, increasing the pH and decreasing the material oxidation driving force (Ningshen et al., 2007; Lavigne et al., 2011). Grabke (1996) has demonstrated that nitrogen in a negatively charged state N^δ− at the metal surface under the passive film will become rich. Some nitrogen may traverse the film as NH_x and will be dissolved as NH₄⁺. The presence of N^δ− may inhibit pit initiation by suppressing Cl⁻, therefore, nitride coating on the surface can increase the corrosion resistance.

Figure 4:

Potentiodynamic polarization curves of the specimens.

SEM micrographs of the specimens after corrosion test are shown in Figure 5. Figure 5A reveals bare martensitic stainless steel surface containing a 127 μm diameter cavity. Unlike the uncoated steel, the coated specimens (Figure 5B and C) showed few pores as a result of pitting corrosion. Further examination of the SEM micrographs confirmed that CrN coating exhibited excellent corrosion resistance with a relatively smooth morphology.

Figure 5:

Corrosion morphologies of specimens: (A) UNS S17400, (B) CrN top coating, (C) TiN top coating.

The presence of inclusions in the stainless steel substrate could be the initial point of corrosion leading to large cavities due to different potentials of the passive stainless steel and the inclusion in chloride-containing solutions. When the uncoated steel is susceptible to free corrosion potential, the inclusion will be polarized and begins to dissolve. The inclusion has low electrical conductivity and caused localized dissolution at the interface of matrix-inclusion. The inclusion dissolution creates a gap between the matrix and inclusion, and thus, an increasing dissolution process (Eklund, 1974; Vuillemin et al., 2003; Nakhaie & Moayed, 2014). Baker and Castle (1993) have shown that the inclusions can either be dissolved partially or completely. When the inclusion is dissolved, the concentration of Cl⁻ increases in the pores leading to the formation of a cavity. Conversely, in the coated specimens, the inclusions have no contact with the electrolyte and corrosion must be initiated from the nitride surface. Multilayer coating has high corrosion resistance that depends on the contact area with the electrolyte and nonhomogeneities on the surface. As shown in Table 3 and Figure 5, the specimen with the CrN top coating represented better corrosion resistance than the TiN top coat. This is because of the ability of CrN to create a passive layer on the surface (Liu et al., 2001). Moreover, TiN coatings normally have columnar structure with some defects such as pinholes and pores that affect the corrosion resistance of the coating; porosity content of TiN layer was more than that of CrN. Lower coating density, columnar structure, and higher porosity content of the TiN coating caused penetration of electrolyte into the intercolumnar paths and premature corrosion attack. On the contrary, the microstructure of CrN is dense and provides no direct diffusion path for the corrosive solution (Grips, 2006; Beibei et al., 2016).

The Nyquist and the electrochemical impedance spectroscopy plots of coated and uncoated samples are shown in Figures 6 and 7. The Nyquist diagram for coated and uncoated specimens show a semicircular graph with a large radius for multilayer coated samples. The largest semicircle is the chromium nitride coating and the smallest is the substrate. Also, the maximum capacitive resistance is related to the CrN top coating and the lowest value is for the substrate. In general, the larger diameter of the circles in the diagram represents the higher corrosion resistance of the coating. As shown in Figure 7, the highest impedance at low frequency is also related to the sample with chromium nitride top coating. This reveals that the multilayer PVD coatings improved the corrosion resistance of UNS S17400 martensitic stainless steel. An equivalent circuit model was used to describe the mechanism of electrochemical reactions as shown in Figure 8. In this model, The electrode characterization includes the determination of the polarization resistance (R_s), C_coat, and R_coat are ascribed the coating properties and the interaction between the solution-coating; C_dl is related to double-layer capacitance, the R_ct that couples with C_dl corresponds to the charge transfer resistance due to the formation of a double layer of charge at the substrate/electrolyte interface. The values of fitted parameters of the equivalent circuit for all specimens, using the Zview software, are presented in Table 4.

Figure 6:

Nyquist plot of coated and uncoated specimens.

Figure 7:

Electrochemical impedance spectroscopy plot of coated and uncoated specimens.

Figure 8:

The equivalent circuit model.

OCP of uncoated stainless steel and coated specimens as a function of time during corrosion-wear test is shown in Figure 9. In the early stages of soaking, OCP was constant. The potential was suddenly reduced to negative values when the sliding motion was started; the reduction was due to the rupture or removal of the passive film that caused the surface became exposed to corrosive environment. An almost constant state of OCP, by establishing a balance between the recovery rate of electrochemical and mechanical damage to the passive layer, was experienced (Kok et al., 2005; Diomidis et al., 2009; Beibei, 2016). When the motion was stopped, the OCP gradually increased with time. Potential fluctuations were observed in all three specimens, uncoated stainless steel, and the coated specimens.

Figure 9:

E _corr for coated and uncoated stainless steel as a function of testing time.

Figure 10 presents SEM micrographs and energy-dispersive X-ray spectroscopy (EDS) microanalyses of the wear tracks after corrosion-wear tests in 3.5% NaCl solution. Figure 10A and C represent the delamination of layers in the track. TiN top coating was damaged more than CrN coating with the distribution of numerous cracks on the failure zone edges. Results of EDS analyses at zones A and B are seen in Figure 10B and D, respectively. Ti, Al, N, O, and small amounts of Cl elements were identified in zone A, whereas Cr, Al, N, and O elements were seen in zone B. Aluminum appeared in the EDS analysis from the intermediate layer because of low thickness of the top layer. The presence of oxygen confirmed the formation of oxide films on the surface; this oxidation caused the surface to become passive with increased corrosion resistance at the beginning of the test. As sliding continued, the oxide film broke and cracks appeared on the surface. Apparently, the CrN coating exhibited a better corrosion-wear performance than TiN top coating. The roughness and porosity of CrN top coat are less than those in TiN coating, leading to a more compact coating, thus, resisting against the formation of microcracks on the surface. It is clearly observed that microcracks on the CrN surface are less than those on the TiN surface. Such cracks act as an effective path for accelerated diffusion of NaCl solution inside the coating. Chlorine ions are considered an important factor in corrosion failure of materials (Bhandari et al., 2015; Wang et al., 2016). More cracks and porosity as well as the influence of Cl⁻ ions greatly deteriorated the interfacial bonding strength within the coating material; as a result, the top surface of TiN coating was destroyed before CrN coating.

Figure 10:

SEM images and EDS analyses of the wear track after corrosion-wear tests; (A, B) CrN top coating, (C, D) TiN top coating.