Corrosion behaviour of multilayer CrN coatings deposited by hybrid HIPIMS after oxidation treatment

2.1 Coating deposition

CrN coatings were fabricated by a semi-industrial PVD unit equipped with six cathodic arc ion evaporators (4G-CAE1028B) and Hüttinger magnetron sputtering sources: two bipolar power supplies, a high pulse power supply and a DC bias supply. Coating A was deposited by arc evaporation, and Coating B was prepared by Hüttinger magnetron sputtering sources. The deposition of Cr/CrN multilayer coatings onto 304 stainless steel substrates with a diameter of 40 mm was conducted in this system. The chromium targets used to deposit both groups of coatings has a purity of 99.9%.

Before deposition, the substrates were mirror polished, ultrasonically cleaned in an acetone solution and an industrial anhydrous ethanol reagent and dried. The coating process was carried out in a vertical cylinder vacuum chamber with an effective coating space of φ 600 mm × 600 mm.

The samples were clamped on a rotating carousel with a rotation speed of 2 min⁻¹. To remove oxygen, the deposition chamber was pumped down to a base pressure of 5 × 10⁻⁴ Pa and heated up to 400°C. Then, the pressure was increased to 2 Pa by argon supply and a sputter cleaning of the substrates with argon and chromium ions at a bias voltage of −1,000 V and a high pulse power of 10 kW for 15 min.

Deposition of coating A was performed at the substrate bias voltage of −40 V and an arc current of 120 A under a nitrogen–argon atmosphere at 7 Pa. The deposition time of each layer was controlled at 920 s. During deposition of coating B, the pressure was kept at 0.55 Pa (33% nitrogen and 67% argon), and the pulse width was set to 100 µs for HIPIMS and 2 µs for BIPPMS. Both the bipolar power and the high pulse power were adjusted to 8 kW. The frequency of the bipolar pulse unit is 40 kHz. Multilayer coatings were deposited by superimposing alternating layers of Cr and CrN. Each layer was deposited for 900 s, and a total of six layers are deposited with −60 V substrate bias voltages. The main experimental parameters are presented in Table 1.

2.2 Coating characterization

The thickness of the coatings and the different layers were evaluated by ball crater tests (SuPro Instruments). The sample and the ball (a ball bearing steel) with a diameter of 30 mm were ground for 90 s with a diamond suspension liquid. The coating thickness is calculated using a simple geometrical equation [13].

The adhesion of the coatings was characterized by the well-known Rockwell C indentation test prescribed by the VDI 3198 norm, as a destructive quality test for coated compounds. This simple low-cost test can qualitatively characterize the adhesion of the coating to the substrate [14]. The principle is that the tapered diamond indenter penetrates the surface of the coating, causing plastic deformation of the substrate and fracture of the coating. As with the indentation test, the indentation depth must follow the 1/10 rule, so the total thickness of the specimen must be at least ten times greater than the indentation depth. The coated sample is evaluated by scanning electron microscopy (SEM).

The samples, divided into four groups, were oxidized in a tubular heating device. The four groups of samples were raised from room temperature in air at a rate of 5 K/min to 500, 600, 700 and 800°C. After a holding time of 2 h, the samples were allowed to cool down again.

The surface morphology of the coatings before and after oxidation treatment was observed by field emission scanning electron microscopy (FESEM, JSM-7800F). Energy-dispersive X-ray spectroscopy (EDX) with FESEM was conducted to determine the elemental composition. The phases were examined using an X-ray diffractometer (Bruker-AXS D8 Advance; U = 40 kV, I = 40 A) with copper Cu Kα radiation (λ = 0.1541 nm), and a scanning speed of 4°/min.

The electrochemical properties of the CrN coatings were tested by means of a three-electrode system in a 3.5 mass% NaCl aqueous solution with a CHI660C Electrochemical Workstation (HuaChen ShangHai, China). The working electrode was the CrN-coated sample inlaid in epoxy resin, and the exposed area was about 1 cm². A graphite electrode served as an auxiliary electrode. A saturated calomel electrode (SCE) was used as a reference electrode. The potentiodynamic polarization tests were started after a 1,000 s stabilization period of the samples in the open-circuit conditions, so that a quasi-stable potential was reached. The potentiodynamic polarization curves were carried out at a scan rate of 0.001 V s⁻¹ over the potential range from −0.6 to 0 V.

3.1 Thickness and adhesion

Many research groups tried to enhance the adhesion between the CrN coating and the substrate by applying a chromium interlayer with a thickness of 300–400 nm. The transition of the composition is beneficial to reduce the internal stress of the coating. Therefore, to ensure that the composition has a certain gradient, the first layer is always a metal chromium layer. The results of the ball crater test of two typical six-layer coatings by arc evaporation and hybrid HIPIMS are shown in Figure 1. The thickness of coating A and coating B is 3.6 and 3.5 µm, respectively. Thus, the deposition rates of these two coatings are 38 and 40 nm/min, respectively. It can be found that under the parameter settings in this article, the deposition rate of hybrid HIPIMS is equivalent to that of arc evaporation plating, and the deposition rate has been improved.

Figure 1

Crater grinding in (a) coating A and (b) coating B.

Figure 2 shows SEM images of coatings A and B of both the surface micromorphology and the Rockwell indentation. The indentation of coating A (Figure 2a) shows some delaminations and some cracks. Compared with coating A, the indentation on the inner wall of coating B (Figure 2b) is relatively uniform with fewer cracks on the edge. Quantification of the appearance according to VDI 3198 [15,16] results in HF.2 and HF.3 (Figure 3), indicating that both coatings meet the requirements. Adhesion between bodies is strongly influenced by surface roughness [17,18].

Figure 2

Rockwell indentation and surface microstructure of as-deposited coatings A (a and c) and B (b and d).

Figure 3

The principle of VDI 3198 indentation test [16].

Figure 2 shows that the surface quality of coating B is superior to that of coating A. The surface of coating A, shown in Figure 2c, reveals large numbers of defects, which include, among others, microdroplets of evaporated material of various sizes and shapes that are formed in the plasma during the coating process. In contrast, the surface of the sputtered coating B (Figure 2d) is dense and shows very few defects. This is typical for sputtering processes and may be connected with the high-power pulses applied here [19,20].

3.2 Composition and phases

The chemical composition of the cross-section of coatings was measured by EDX and is shown in Figure 4. The results reveal that in both cases, multilayer Cr/CrN coatings with a stoichiometric ratio (Cr/N) of nearly 3/1 were obtained. In the meantime, the boundary of the multilayer morphology of the cross-section of coating A is clearly visible, and the interlayer between the coating and the substrate can be observed. In comparison, the famous columnar crystal structure appears in the cross-section of coating B, but the boundary between the layers cannot be easily distinguished. The interlayer and the first layer are almost completely melted into one, which further shows its good adhesion.

Figure 4

Cross-section morphology and composition structure of coatings A (a and b) and B (c and d).

The XRD patterns of the as-deposited coatings A and B, shown in Figure 5, exhibit a NaCl-type face centre cubic (fcc) structure. Mainly three phases, CrN, Cr and Cr₂N, were found in both coatings. The CrN phase and the Cr phase were the main components of the experimentally designed multilayer coating. The diffraction peaks’ orientation of CrN were mainly on the (111), (200) and (311) planes. The preferred orientation of both CrN layers (coatings A and B) was the (200) plane with a strong, sharp and symmetrical diffraction spectrum. This proves the good crystallinity and a complete lattice. The preferred orientation of the Cr phase was the (200) plane. The appearance of the Cr₂N phase in the multilayer coatings may be attributed to the fluctuations in the flow rate of nitrogen [21,22]. In principle, under normal conditions, the nitrogen flow rate was sufficient to produce stoichiometric CrN, but in the actual coating process, it took a certain time to reach the nitrogen concentration in the gas phase. The main diffraction peaks of the Cr₂N phase were broad and diffuse, indicating that the crystal grains were very small, combined with defects such as dislocations. This may cause the secondary nucleation and the growth of crystal grains due to the increase of the bias voltage, thereby reducing the half width.

Figure 5

XRD patterns of the as-deposited coating.

All peaks associated with the Cr and CrN phases were shifted to lower 2-theta values compared to the standard ICCD card [10]. This indicates a state of compressive residual stress due to thermal stress and intrinsic stress [23]. The intrinsic stress is caused by the accumulation of defects, mostly point defects such as vacancies or interstitial atoms, as well as a high dislocation density [21,24]. Thermal stress is mainly due to the difference in the linear thermal expansion coefficient (CTE) between the substrate and the coating after deposition, resulting in a mismatch between the substrate and the coating during the postdeposition cooling. The CTE for CrN is 2.3 × 10⁻⁶ K⁻¹, while the CTE for the stainless steel is much larger (around 15 × 10⁻⁶ K⁻¹) [23,25].

To investigate the effect of the oxidation test on the corrosion resistance of the coating, both coatings were placed in a tube furnace, internally flowed by air, and oxidized for 2 h at 500, 600, 700 and 800°C. Heating rate was 5 K/min. The chemical compositions of the oxidized coatings’ surface at different temperatures are presented in Table 2. With increasing temperature, the content of nitrogen in both coatings decreased significantly, the content of oxygen increased gradually, whereas chromium shows only small changes.

Figure 5 shows the result of the EDX analysis of the cross-section composition of coatings A and B after different temperature treatments. Regardless of the oxidation temperature, it can be found that the boundaries between coatings layers are fused with each other, and the characteristics of columnar crystals disappear. This is because the increase in temperature accelerates the migration of O atoms to the inside and the diffusion of atoms inside the coating, so that the structure inside the coating is rearranged. During low-temperature oxidation (600°C, Figure 6a), the N in the inner layer of coating A diffuses into the outside of the coating and the external O diffuses into the inside of the coating, resulting in high N and O contents near the surface. For coating B, tight columnar crystal structure prevents the migration of the inner N atoms to the outside and the erosion of the outer O atoms to the inside to a certain extent. Finally, when oxidized at 800°C, the O atom diffusion depth of coating B (about 1.5 µm) is smaller than that of coating A (2.5 µm), achieving better oxidation resistance.

Figure 6

The diffusion depth of nitrogen and oxygen at different temperatures: (a) coating A at 600°C, (b) coating A at 800°C, (c) coating B at 600°C and (d) coating B at 800°C.

It can be seen from the XRD pattern (Figure 7) of coating A that after oxidation at 500°C, the coating contains the phases CrN, Cr₂N and Cr, and both Cr and CrN phases have strong diffraction intensity. At the same time, the diffraction peak appeared to be broadened compared to the as-deposited coating, which may be attributed to the incomplete crystal structure caused by the oxidation process. When oxidized at 600°C, the diffraction intensity of the CrN phase was greatly reduced, while the diffraction intensity of Cr₂N was obviously increased. This indicates that the phase structure was changed at the beginning, but no chromium oxide was formed. As the temperature continued to increase, the activity of Cr atoms increased [26].

Figure 7

XRD patterns of CrN coatings oxidized at different temperatures.

After oxidation at 700°C, chromium oxide Cr₂O₃, which is the thermodynamically stable compound, can be detected. Possible reactions are as follows [27,28]:

The Gibbs free energy changes corresponding to these reactions are as follows [28]:

In equations (3) and (4), pO₂ is the partial pressure of oxygen (21% in air), pN₂ is the partial pressure of nitrogen (78% in air), R is the gas constant 8.314 J/kmol, T is the absolute temperature (in K) of the reaction and ΔG ₀ is the standard Gibbs free energy for each substance. All the aforementioned data are derived from the NIST-JANAF thermodynamic database. It can be calculated that at 700°C, ΔG _R(1) is −513 kJ and ΔG _R(2) is −537 kJ. The XRD patterns of coatings A and B are substantially the same. A significant difference is the two crystal planes (101) and (211) of Cr₂N in coating B. Furthermore, coating B shows a higher intensity of other peaks together with a reduced broadening. Figure 8 shows the surface morphology of coating A, as deposited and oxidized at different temperatures. Figure 8a shows a relatively smooth surface with some microparticles or droplets, which are typical for arc evaporated coatings [29,30]. After oxidation at 500°C, the microparticles did not change much, but on the former smooth surface, a meshwork of lighter line becomes visible. It can be assumed that these areas consist of small oxide particles generated around grain boundaries. Here, the oxidation starts earlier because the energetic status of atoms in the disordered area is higher, and heterogeneous nucleation becomes possible earlier. After oxidation at 600°C, the surface is covered with small-sized (80–240 nm) oxide grains. The majority of the microparticles is still visible. A further increase to 700°C (Figure 8d) allowed the grain size of the oxide crystals to increase to 200–500 nm, and only a few of the microparticles remain visible. Finally, oxidation at 800°C (Figure 8e) allowed the oxide grains to grow up to 500–1,000 nm and then the microparticles disappeared completely.

Figure 8

SEM surface micrograph of CrN coating A: (a) as deposited, (b) oxidized at 500°C, (c) oxidized at 600°C, (d) oxidized at 700°C and (e) oxidized at 800°C.

The surface morphologies of coating B, as deposited and oxidized, are shown in Figure 9. The surface structure of coating B at 10,000 times magnification, as shown in Figure 9(a), was a agglomerated structure, and the surface was smooth without the appearance of microparticles (Figure 2(d)). At the same time, there was no droplet splash during the sputtering process, and the coating structure was uniform.

Figure 9

SEM surface micrograph of CrN coating B: (a) as deposited, (b) oxidized at 500°C, (c) oxidized at 600°C, (d) oxidized at 700°C and (e) oxidized at 800°C.

After heat treatment at 500°C, only a small amount of oxidation becomes visible (Figure 9b). At higher magnification, fine oxide crystals can be detected. Oxidation at 600°C (Figure 9c) results in the generation of hexagonal chromium oxide Cr₂O₃ crystals with a size of 40–80 nm. Compared to coating A (Figure 9c), the oxide grains are finer and more uniform. Oxidation at 700°C (Figure 9d) results in a grain size of 100–500 nm and finally reached about 1,000 nm at 800°C (Figure 9e).

Summarizing it shows that in both cases oxide crystal grains are formed on the surfaces of both coatings at 500°C, but the oxide grains on the surface of coating B are not obvious. Higher oxidation temperatures let the oxide grains grow up, but grains formed on coating B at the same oxidation temperature are finer.

3.3 Corrosion resistance

The metal nitride film can provide a good corrosion protection for the substrate in most cases, but defects in the surface of the coating can decrease the corrosion resistance significantly. Different corrosion mechanisms are well known and described in the literature [31].

To investigate the effect of oxidation on the corrosion resistance of the coatings, electrochemical analysis before and after oxidation was carried out in an aqueous solution of 3.5 mass% NaCl at different temperatures. The dynamic potential polarization curves under the open circuit potential are shown in Figure 10. As the substrates are connected as anode, atoms of the coating’s surface – preferentially from coating defects – will discharge ions and diffuse into the solution, thus generating a self-corrosion current [32]. The polarization curves were evaluated by Tafel linear extrapolation. The relevant corrosion data are presented in Table 3, where E _corr is the self-corrosion potential at the open circuit potential, J ₀ is the self-corrosion current density and R _p is its corresponding polarization resistance.

Figure 10

The polarization curves of coating A and coating B oxidized at different temperatures in 3.5% NaCl solution.

With the aim to better reflect the positive effect of the coating on improving the corrosion resistance of the substrate, the same electrochemical corrosion test was performed on 304 stainless steel substrates. The corrosion resistance of the substrate is much lower than that of the coating. At the same time, the polarization curve of the substrate is passivated at potentials of −0.69 and −0.31 V (Figure 10).

Table 3 states the following: coating A, as deposited, shows a corrosion potential (E _corr) of −0.278 V, a self-corrosion current density (J ₀) of 3.467 × 10⁻⁷ A cm⁻² and a polarization resistance (R _p) of 300.2 kΩ. After oxidation at 500°C, E _corr is decreased to −0.413 V, J ₀ is increased to 5.402 × 10⁻⁷ A cm⁻² and R _p is decreased to 265.2 kΩ. Due to the heterogeneity of the chemical composition and the heterogeneity of the structure, the meshwork of lighter line that composed of the chromium oxide (Figure 8b) on the surface of the coating will form a potential difference with the surrounding chromium nitride, thereby causing galvanic corrosion [32,33,34]. This increases the activation energy of the surface and causes a negative shift in the corrosion potential.

After oxidation at 600°C, E _corr of the coating is −0.363 V, which is positively shifted compared to that at 500°C, but still lower than E _corr of the as-deposited coating. The corresponding J ₀ is reduced to 1.117 × 10⁻⁸ A cm⁻² and R _p is increased to 2244.3 kΩ. From XRD and surface topography (Figures 6c and 7), it is known that the coating surface is covered with Cr₂O₃ in the shape of polyhedral particles. This increases the total areas of the surface, preventing penetration of the fluid into the interior of the coating, thus improving the corrosion resistance [35].

During oxidation at 700°C, the oxide particles on the surface of coating A formed a dense oxide film, which hinders the corrosion of the coating by Cl⁻ in the corrosion solution and makes it difficult to form corrosion holes, thus improving the corrosion resistance. This phenomenon is also reflected in the table data. The corrosion current density J ₀ decreases and the polarization impedance R _p increases. Finally, oxidation at 800°C shifts E _corr to −0.256 V, lower than that of the as-deposited state. With that, J ₀ decreased to 5.494 × 10⁻⁹ A cm⁻² and the corresponding R _p increased to 5366.2 kΩ, leading to the conclusion that oxidation at 800°C results in the best corrosion resistance.

It is remarkable that – compared to the as-deposited state – oxidation at 500°C led to a decreased corrosion resistance with E _corr shifted negatively, increased J ₀ and decreased R _p. But higher oxidation temperatures gradually increased corrosion resistance from 600 to 700°C and finally 800°C.

Coating B in the as-deposited state shows E _corr of −0.204 V, J ₀ of 1.004 × 10⁻⁷ A cm⁻² and R _p of 445.1 kΩ. This means that coating B – compared to coating A – is less corrosive. This can be explained by the rough surface of coating A with a high particle density and a small amount of pinholes, compared to the relatively smooth surface of coating B [33,36,37,38,39]. After oxidation at 500°C, E _corr decreased to −0.361 V, the polarization curve also shifted negatively, but J ₀ decreased to 9.5 × 10⁻⁹ A cm⁻², and R _p increased to 4448.9 kΩ. This can be explained by the fact that there are no droplet particles on the surface of the coating, and the oxidation can take place uniformly (Figure 9b).

When the oxidation temperature increased to 600°C, E _corr resulted in −0.457 V, and a negative shift of the polarization curve occurred again. The results of XRD analysis (Figure 7), and the surface micromorphology (Figure 9c) shows that CrN, Cr₂N and Cr₂O₃ are present on the surface of the coating, resulting in heterogeneity of chemical composition and the structure of the coating. Therefore, at 500°C, galvanic corrosion is caused by the potential difference. While at 600°C, the number of these galvanic corrosion systems increases, thereby enhancing the surface activity of the coating. Thus, the corrosion potential is negatively shifted.

As the oxidation temperature increased to 700°C, the oxide particles on the surface of the coating B increased, the oxide layer became thicker. As a result, the surface composition becomes uniform and the potential difference of the coating’s surface gradually disappears, causing the corrosion potential shift positively.

After oxidation at 800°C, the surface of the coating was coated with a very corrosion-resistant Cr₂O₃ phase (Figure 9e), E _corr was changed to −0.181 V, lower than the as- deposited state, and J ₀ was reduced to 2.414 × 10⁻⁹ A cm⁻². R _p was increased to 12443.1 kΩ. In summary, during oxidation of coating B, as the temperature increases, J ₀ decreases, R _p increases and the corrosion resistance of the coating is improved. Starting from 700°C, the corrosion potential will gradually move to the positive axis [40,41].

It can be seen that with increasing oxidation temperature, the surface composition and the structure of the coating will change, leading to an obvious change in corrosion potential. However, the direct reason for the increase in corrosion resistance of the coating is the decrease in the self-corrosion current density. Therefore, the self-corrosion current density is more convincing in evaluating the corrosion resistance of the coating. Comparing the corrosion patterns of the two groups after oxidation, it was found that the corrosion resistance of coating B after oxidation at 800°C was the best.