Influence of Oxygen Contents on the Microstructure , High Temperature Oxidation and Corrosion Resistance Properties of Cr – Si – O – N Coatings

Cr–Si–O–N coatings with different oxygen contents were deposited by multi-arc ion plating, where various O2/(N2 + O2) reactive gas rates were adopted. The XRD and XPS results showed that the CrN crystals disappeared with the increasing of the oxygen flux ratio to 10 at.%. The microhardness of all the Cr–Si–O–N coatings was approximately 2000 Hv0.05, which were dramatically plummeted compared to that of the Cr–Si–N coatings (≈3300 Hv0.05). The Cr–Si–O–N coatings were annealed under 800 ◦C and 1200 ◦C in the air atmosphere for 2 h to study the high-temperature oxidation resistance of the coatings. Meanwhile, Cr–Si–O–N coatings with different O2/(N2 + O2) rates were also used to carry out the corrosion resistance testing using the electrochemical working station in 3.5% NaCl solution under free air condition at room temperature. The results indicated that the coatings containing oxygen were more vulnerable to the high-temperature destruction and more easily corroded in the NaCl electrolyte.


Introduction
CrN has been one of the most extensively used hard coatings for various forming and casting applications [1,2], due to its excellent high-temperature oxidation resistance and good corrosion resistance [3,4].In recent years, efforts have been continuously devoted in order to further improve the properties of CrN coatings.One of the most promising approaches is to synthesize the Cr-Si-N nanocomposite coatings, which follows the design concept of superhard nc-MeN/a-Si 3 N 4 proposed by Veprek et al. [5].
In the nanocomposites coatings, the oxygen has a great impact on the properties.On one hand, it has been reported that the oxygen impurity could be critically harmful to the superhard nanocomposite coatings [6,7].Veprek et al. [8] synthesized the Ti-Si-N superhard nanocomposites coatings with a hardness of ≥50 GPa, which was difficult for some other groups to achieve [9][10][11].Verprek et al. [8] figured that one of the important reasons for the lack of reproducibility was the oxygen impurity incorporated into the coatings during the deposition process.It was impossible to achieve the hardness of 40 GPa when the oxygen impurity was larger than 0.4 at.%. Ma et al. [12] also reported that oxygen impurity of 1-1.5 at.% could lead to a sharp decrease in hardness to 30 GPa, while they could obtain 45-55 GPa in the nc-TiN/a-Si 3 N 4 coatings by controlling the oxygen impurity below 0.2 at.%.On the other hand, however, Lee et al. [13] reported that the mechanical and tribological The coatings deposited on Si wafer were used for oxidation experiment, and the high-temperature oxidation resistance ability was evaluated by comparison of the change of coatings phase using XRD (X-ray diffraction, X'Pert-MPD System, PHILIPS, Amsterdam, The Netherlands) and observation of separation extent between Cr-Si-O-N coatings and substrates using SEM (scanning electron microscope, MERLIN, ZEISS, Jena, Germany).The experiments were carried out at a heating rate of 10 • C/min and the coatings were annealed at 800 • C and 1200 • C in the air for 2 h.The electrochemical corrosion test resistance of the deposited Cr-Si-O-N coatings and WC-Co cemented carbide substrate were tested by an electrochemical working station (CS350, Wuhan Kesite Instrument Co., Ltd., Wuhan, China) in 3.5% NaCl solution.The crevice corrosion was carried out under free air condition at room temperature, with scanning range from −0.5 to 1. 5

V versus open circuit potential, and the
Coatings 2018, 8, 19 3 of 11 mercury/mercury chloride combined with luggin probe as the reference electrode, and the platinum electrode (1 cm 2 ) as the counter electrode.To obtain accurate curve, Lujin capillary was as close as possible to the study electrode for decreasing liquid potential and the scan rate of 0.5 mV/s was used.The corrosion potential, corrosion current and polarization resistance obtained from the Tafel plots were used to evaluate the corrosion rate of deposited coatings.In addition, Electrochemical Impedance Spectroscopy (EIS, CS350, Wuhan Kesite Instrument Co., Ltd., Wuhan, China) measurements were recorded at open circuit potential in the frequency range of 10 −2 -10 −5 Hz with a signal amplitude of 10 mV in 3.5% NaCl solution.The experiment of Potentiodynamic Polarization Curves, Electrochemical Impedance Spectroscopy and high-temperature oxidation test were repeated for 3 times.
The structure and phases of the Cr-Si-O-N coatings deposited on Si wafer were analyzed by X-ray diffraction (XRD, X'Pert-MPD System, PHILIPS, Amsterdam, The Netherlands) with 0.04 • using Cu Kα radiation.The bonding status of Cr-Si-O-N coatings was studied by X-ray photoelectron spectra (XPS, PHI X-tool, ULVAC-PHI, Inc., Chigasaki, Japan) with Al Kα X-ray at 15 kV and 50 W.The chemical compositions of coatings were investigated by the electron probe microanalyzer (EPMA, EPMA 1600, Shimadzu, Kyoto, Japan).The field emission scanning electron microscope was used to obtain the surface morphology and the cross sectional images of the coatings before and after annealing.The microhardness of Cr-Si-O-N coatings deposited on cemented carbide substrate were measured by using an HX-1000 micro-hardness tester (ShangHai TaiMing Optical Instrument Co., Ltd., Shanghai, China) with a load of 50 gf (0.49 N) and the final values are taken from average values of 10 measurements.

Composition and Microstructure
Figure 1 shows the chemical composition of the Cr-Si-O-N coatings with different flux rates of O 2 /(N 2 + O 2 ).Since the current of the CrSi target was fixed at 70 A during all of the deposition processes, the Cr and Si contents of the as-deposited coatings were almost constant, being about 51 at.% and 9 at.%,respectively.The oxygen contents of the deposited coatings significantly increased from 0 to 40 at.% and the N contents sharply decreased from 31.5 at.% to 5 at.% when the O 2 /(N 2 + O 2 ) flux rate increased from 0% to 10%.The Gibbs free energy of Cr-O is lower than that of Cr-N [15], so the residual oxygen is more likely to remain in the coatings in the form of Cr x O y .However, both the O and N contents remained the same with the further increase of the O 2 /(N 2 + O 2 ) flux rate, which might result from the saturation of the O 2 in the chamber.The structure and phases of the Cr-Si-O-N coatings deposited on Si wafer were analyzed by Xray diffraction (XRD, X'Pert-MPD System, PHILIPS, Amsterdam, The Netherlands) with 0.04° using Cu Kα radiation.The bonding status of Cr-Si-O-N coatings was studied by X-ray photoelectron spectra (XPS, PHI X-tool, ULVAC-PHI, Inc., Chigasaki, Japan) with Al Kα X-ray at 15 kV and 50 W.The chemical compositions of coatings were investigated by the electron probe microanalyzer (EPMA, EPMA 1600, Shimadzu, Kyoto, Japan).The field emission scanning electron microscope was used to obtain the surface morphology and the cross sectional images of the coatings before and after annealing.The microhardness of Cr-Si-O-N coatings deposited on cemented carbide substrate were measured by using an HX-1000 micro-hardness tester (ShangHai TaiMing Optical Instrument Co., LTD, Shanghai, China) with a load of 50 gf (0.49 N) and the final values are taken from average values of 10 measurements.

Composition and Microstructure
Figure 1 shows the chemical composition of the Cr-Si-O-N coatings with different flux rates of O2/(N2 + O2).Since the current of the CrSi target was fixed at 70 A during all of the deposition processes, the Cr and Si contents of the as-deposited coatings were almost constant, being about 51 at.% and 9 at.%,respectively.The oxygen contents of the deposited coatings significantly increased from 0 to 40 at.% and the N contents sharply decreased from 31.5 at.% to 5 at.% when the O2/(N2 + O2) flux rate increased from 0% to 10%.The Gibbs free energy of Cr-O is lower than that of Cr-N [15], so the residual oxygen is more likely to remain in the coatings in the form of CrxOy.However, both the O and N contents remained the same with the further increase of the O2/(N2 + O2) flux rate, which might result from the saturation of the O2 in the chamber.Cr-Si-N coatings.However, with the addition of oxygen, the diffraction peaks became weaker and the diffraction peaks of crystalline phase were no longer recognizable when the O 2 /(N 2 + O 2 ) flux rate above 10%.The more oxygen in the chamber, the more Cr from the CrN was despoiled to form the Cr x O y .Unlike the results from other researchers [13], the formed Cr x O y in our as-deposited coatings existed in the amorphous phase, which is because the deposition temperature in our system is not high enough to form the crystalline phase.In addition, no crystalline peaks of Si 3 N 4 and SiO 2 were detected in the XRD spectra, which can also be attributed to the low temperature deposition process.
Coatings 2018, 1, x FOR PEER REVIEW 4 of 11 coatings.However, with the addition of oxygen, the diffraction peaks became weaker and the diffraction peaks of crystalline phase were no longer recognizable when the O2/(N2 + O2) flux rate above 10%.The more oxygen in the chamber, the more Cr from the CrN was despoiled to form the CrxOy.Unlike the results from other researchers [13], the formed CrxOy in our as-deposited coatings existed in the amorphous phase, which is because the deposition temperature in our system is not high enough to form the crystalline phase.In addition, no crystalline peaks of Si3N4 and SiO2 were detected in the XRD spectra, which can also be attributed to the low temperature deposition process.3a), which is in good agreement with that of Si3N4 phase [14].In addition, little peaks of SiO2 phase appeared at 102.9 eV [13].In the N 1s spectra, we can find peaks at 396.7 and 397.9 eV, which correspond to CrN [16] and Si3N4 [17], respectively, in the coating with lower O2/(N2 + O2) flux rate.However, when the flux rate increased to 20%, the peak corresponding to CrN disappeared, indicating that there was no crystalline CrN phase existing in the coatings anymore, and the left N in the coatings was all in the amorphous form of Si3N4.This also implied that crystal phase of CrN was present at lower O2/(N2 + O2) flux rate in accordance with XRD results.3a), which is in good agreement with that of Si 3 N 4 phase [14].In addition, little peaks of SiO 2 phase appeared at 102.9 eV [13].In the N 1s spectra, we can find peaks at 396.7 and 397.9 eV, which correspond to CrN [16] and Si 3 N 4 [17], respectively, in the coating with lower O 2 /(N 2 + O 2 ) flux rate.However, when the flux rate increased to 20%, the peak corresponding to CrN disappeared, indicating that there was no crystalline CrN phase existing in the coatings anymore, and the left N in the coatings was all in the amorphous form of Si 3 N 4 .This also implied that crystal phase of CrN was present at lower O 2 /(N 2 + O 2 ) flux rate in accordance with XRD results.coatings.However, with the addition of oxygen, the diffraction peaks became weaker and the diffraction peaks of crystalline phase were no longer recognizable when the O2/(N2 + O2) flux rate above 10%.The more oxygen in the chamber, the more Cr from the CrN was despoiled to form the CrxOy.Unlike the results from other researchers [13], the formed CrxOy in our as-deposited coatings existed in the amorphous phase, which is because the deposition temperature in our system is not high enough to form the crystalline phase.In addition, no crystalline peaks of Si3N4 and SiO2 were detected in the XRD spectra, which can also be attributed to the low temperature deposition process.3a), which is in good agreement with that of Si3N4 phase [14].In addition, little peaks of SiO2 phase appeared at 102.9 eV [13].In the N 1s spectra, we can find peaks at 396.7 and 397.9 eV, which correspond to CrN [16] and Si3N4 [17], respectively, in the coating with lower O2/(N2 + O2) flux rate.However, when the flux rate increased to 20%, the peak corresponding to CrN disappeared, indicating that there was no crystalline CrN phase existing in the coatings anymore, and the left N in the coatings was all in the amorphous form of Si3N4.This also implied that crystal phase of CrN was present at lower O2/(N2 + O2) flux rate in accordance with XRD results.The surface and cross-sections SEM images of Cr-Si-O-N coatings deposited under different O2/(N2 + O2) flux rates are shown in Figure 4.Many micro-particles can be seen on the surface of both coatings, which were introduced by the multi-arc ion plating during experiments [18].

Microhardness, High Temperature Test and Corrosion Test
Figure 5 shows the hardness value of Cr-Si-O-N coating as a function of O2/(N2 + O2) flux rate.It can be found clearly that the hardness sharply decreased with the addition of the oxygen from 0 to 2 at.% and slightly decreased with the further increase of the oxygen above 5 at.%.The hardness value of cemented carbide substrate is about 1260 Hv0.05, and the hardness of all the Cr-Si-O-N coatings is ranging around 2000 Hv0.05 which is far less than the Cr-Si-N coatings.This phenomenon might be attributed to the formation of a large proportion of the CrxOy amorphous phase in the as-deposited coatings which disrupted the formation of the nanocomposite structure of amorphous matrix and crystalline [19].
The Cr-Si-O-N coatings deposited under O2/(N2 + O2) flux rate of 2% and 10% were chosen to do the high-temperature oxidation treatments.The XRD spectra of the coatings before and after oxidation under 800 °C and 1200 °C in the air were shown in Figure 6.After the coatings were annealed under 800 °C in the air for 2 h, some diffraction patterns of Cr2O3 and CrO2 were detected in both coatings, which suggested that small extent of oxidation behavior might occur on the surface.The number of oxide diffraction patterns in the coatings with lower oxygen was much less than that in the higher one, indicating that the high-temperature oxidation ability of the former might be better.As the oxidation temperature rose to 1200 °C, the diffraction peaks, corresponding to Cr2O3 and CrO2, became stronger, and their number increased, indicating that the oxidation progressed.The crystalline Cr2O3 and CrO2 appear due to oxidation of the coatings and crystallisation of amorphous CrxOy phase during annealing.Owing to SiO2 presents in terms of amorphous [20], it was noted that no diffraction pattern of SiO2 was detected in all the XRD spectra.Meanwhile, Kim et al. proposed that the crystal CrN phase of Cr-Si-N coatings did not disappear thoroughly and maintained good  It can be found clearly that the hardness sharply decreased with the addition of the oxygen from 0 to 2 at.% and slightly decreased with the further increase of the oxygen above 5 at.%.The hardness value of cemented carbide substrate is about 1260 Hv 0.05 , and the hardness of all the Cr-Si-O-N coatings is ranging around 2000 Hv 0.05 which is far less than the Cr-Si-N coatings.This phenomenon might be attributed to the formation of a large proportion of the Cr x O y amorphous phase in the as-deposited coatings which disrupted the formation of the nanocomposite structure of amorphous matrix and crystalline [19].
The Cr-Si-O-N coatings deposited under O 2 /(N 2 + O 2 ) flux rate of 2% and 10% were chosen to do the high-temperature oxidation treatments.The XRD spectra of the coatings before and after oxidation under 800 • C and 1200 • C in the air were shown in Figure 6.After the coatings were annealed under 800 • C in the air for 2 h, some diffraction patterns of Cr 2 O 3 and CrO 2 were detected in both coatings, which suggested that small extent of oxidation behavior might occur on the surface.The number of oxide diffraction patterns in the coatings with lower oxygen was much less than that in the higher one, indicating that the high-temperature oxidation ability of the former might be better.As the oxidation temperature rose to 1200 • C, the diffraction peaks, corresponding to Cr 2 O 3 and CrO 2 , became stronger, and their number increased, indicating that the oxidation progressed.The crystalline Cr 2 O 3 and CrO 2 appear due to oxidation of the coatings and crystallisation of amorphous Cr x O y phase during annealing.Owing to SiO 2 presents in terms of amorphous [20], it was noted that no diffraction pattern of SiO 2 was detected in all the XRD spectra.Meanwhile, Kim et al. proposed that the crystal CrN phase of Cr-Si-N coatings did not disappear thoroughly and maintained good oxidation resistance even the annealing temperature up to 1000 • C [21], which implied that CrN phase included in the coatings was reasonable, as shown in Figure 6a.The SEM images of the coatings after oxidation are illustrated in Figure 7.There are some chromium oxide crystallites (confirmed by EDS test) covering the surface of the coatings and those oxide crystallites may derive from the outward diffusion of the chromium.The higher the temperature is, the more oxide crystallites there are.The cross-sectional images in Figure 8 show that both of the coatings deposited under O2/(N2 + O2) flux rate of 2% and 10% keep intact after the 800 °C oxidation.However, after the higher temperature oxidation, the destruction occurred.It can be observed clearly that the extent of oxidation destruction are severer in the coatings with higher oxygen content and the coatings tend to be peel off from the substrate.The strong detriment of the high-temperature oxidation resistance of the as-deposited coatings may attribute to the lack of the nanocomposite structure [10,22,23].Figure 8 shows the SEM images and EDS spectra of Cr-Si-O-N coatings with the oxygen flow ratio of 2% and 10% after annealing.Although a large nubmer of fluffy macroparticles that were chromium oxide detected by EDS spectra were present on the surface of the samples, Figure 8c,f still showed that the coatings have not been damaged completely.Figure 8f also confirmed that the concentration of Si was not large.In other words, the coatings can resist to destruction at 1100 °C.The cross-sectional chemical compositions of the coatings deposited under both the oxygen contents after annealing at 800, 1100 and 1200 °C have been recited in Table 2.It was obvious that the Cr contents reduced, Si and O contents increased when annealing temperature ranged from 800 to 1200 °C, and N contents calculated by subtracting the sum of Cr, Si and O contents from 100% had a decreasing trendency.This revealed that the CrN phase degraded as the annealing temperature rose, which can be in agreement with XRD results.The SEM images of the coatings after oxidation are illustrated in Figure 7.There are some chromium oxide crystallites (confirmed by EDS test) covering the surface of the coatings and those oxide crystallites may derive from the outward diffusion of the chromium.The higher the temperature is, the more oxide crystallites there are.The cross-sectional images in Figure 8 show that both of the coatings deposited under O2/(N2 + O2) flux rate of 2% and 10% keep intact after the 800 °C oxidation.However, after the higher temperature oxidation, the destruction occurred.It can be observed clearly that the extent of oxidation destruction are severer in the coatings with higher oxygen content and the coatings tend to be peel off from the substrate.The strong detriment of the high-temperature oxidation resistance of the as-deposited coatings may attribute to the lack of the nanocomposite structure [10,22,23].Figure 8 shows the SEM images and EDS spectra of Cr-Si-O-N coatings with the oxygen flow ratio of 2% and 10% after annealing.Although a large nubmer of fluffy macroparticles that were chromium oxide detected by EDS spectra were present on the surface of the samples, Figure 8c,f still showed that the coatings have not been damaged completely.Figure 8f also confirmed that the concentration of Si was not large.In other words, the coatings can resist to destruction at 1100 °C.The cross-sectional chemical compositions of the coatings deposited under both the oxygen contents after annealing at 800, 1100 and 1200 °C have been recited in Table 2.It was obvious that the Cr contents reduced, Si and O contents increased when annealing temperature ranged from 800 to 1200 °C, and N contents calculated by subtracting the sum of Cr, Si and O contents from 100% had a decreasing trendency.This revealed that the CrN phase degraded as the annealing The SEM images of the coatings after oxidation are illustrated in Figure 7.There are some chromium oxide crystallites (confirmed by EDS test) covering the surface of the coatings and those oxide crystallites may derive from the outward diffusion of the chromium.The higher the temperature is, the more oxide crystallites there are.The cross-sectional images in Figure 8 show that both of the coatings deposited under O 2 /(N 2 + O 2 ) flux rate of 2% and 10% keep intact after the 800 • C oxidation.However, after the higher temperature oxidation, the destruction occurred.It can be observed clearly that the extent of oxidation destruction are severer in the coatings with higher oxygen content and the coatings tend to be peel off from the substrate.The strong detriment of the high-temperature oxidation resistance of the as-deposited coatings may attribute to the lack of the nanocomposite structure [10,22,23].Figure 8 shows the SEM images and EDS spectra of Cr-Si-O-N coatings with the oxygen flow ratio of 2% and 10% after annealing.Although a large nubmer of fluffy macroparticles that were chromium oxide detected by EDS spectra were present on the surface of the samples, Figure 8c,f still showed that the coatings have not been damaged completely.Figure 8f also confirmed that the concentration of Si was not large.In other words, the coatings can resist to destruction at 1100 • C. The cross-sectional chemical compositions of the coatings deposited under both the oxygen contents after annealing at 800, 1100 and 1200 • C have been recited in Table 2.It was obvious that the Cr contents reduced, Si and O contents increased when annealing temperature ranged from 800 to 1200 • C, and N contents calculated by subtracting the sum of Cr, Si and O contents from 100% had a decreasing trendency.This revealed that the CrN phase degraded as the annealing temperature rose, which can be in agreement with XRD results.The electrochemical behavior of corrosion resistance of the deposited Cr-Si-O-N films and cemented carbide substrate was tested by Electrochemical Working Station (CS350) in 3.5% NaCl solution under free air condition at room temperature.The Tafel plots of the base cemented carbide substrate and Cr-Si-O-N coatings with oxygen contents are shown in Figure 9.The results such as corrosion potential, corrosion current, and polarization resistance fitted from the Tafel plots are listed in Table 3.It can be found that the corrosion potential (E corr ) of uncoated cemented carbide substrate is about −0.4650 V and the E corr of the coated samples shows a shift towards cathodic side compared to the substrate.Moreover, the corrosion current (I corr ) is decreased while the polarization resistance (R p ) is increased with the oxygen content increasing, which indicates the better corrosion resistance of the coated samples.However, the Cr-Si-N coatings have better corrosion rate than that of the Cr-Si-O-N coatings.The Nyquist plots can also be used to evaluate the corrosion resistance performance of the coatings, and the results are shown in Figure 10.  3. It can be found that the corrosion potential (Ecorr) of uncoated cemented carbide substrate is about −0.4650 V and the Ecorr of the coated samples shows a shift towards cathodic side compared to the substrate.Moreover, the corrosion current (Icorr) is decreased while the polarization resistance (Rp) is increased with the oxygen content increasing, which indicates the better corrosion resistance of the coated samples.However, the Cr-Si-N coatings have better corrosion rate than that of the Cr-Si-O-N coatings.The Nyquist plots can also be used to evaluate the corrosion resistance performance of the coatings, and the results are shown in Figure 10.The equivalent circuit for substrate is shown in Figure 11A, and the proposed equivalent circuit for Cr-Si-O-N coatings on the WC-Co cemented carbide substrate with defects (large particles and pinholes) is shown in Figure 11B [24].Here, R s is the solution resistance, CPE c represents the constant phase element related to the coatings, CPE dl represents the constant phase element of the electrical double layer, WE is the working electrode, and R c is the coating resistance [25].Table 4 shows the fitting results of the equivalent electrical circuit of Figure 11.The fit errors of the values presented in the table are less than 10%.From the table, it can be seen that the R ct which is the charge transfer resistance decreased from 50.15 to 23.48 kΩ•cm 2 with the increasing of oxygen contents, which also provides the evidence that the Cr-Si-N coatings have better corrosion resistance than that of Cr-Si-O-N coatings.The equivalent circuit for substrate is shown in Figure 11A, and the proposed equivalent circuit for Cr-Si-O-N coatings on the WC-Co cemented carbide substrate with defects (large particles and pinholes) is shown in Figure 11B [24].Here, Rs is the solution resistance, CPEc represents the constant phase element related to the coatings, CPEdl represents the constant phase element of the electrical double layer, WE is the working electrode, and Rc is the coating resistance [25].Table 4 shows the fitting results of the equivalent electrical circuit of Figure 11.The fit errors of the values presented in the table are less than 10%.From the table, it can be seen that the Rct which is the charge transfer resistance decreased from 50.15 to 23.48 kΩ•cm 2 with the increasing of oxygen contents, which also provides the evidence that the Cr-Si-N coatings have better corrosion resistance than that of Cr-Si-O-N coatings.

Conclusions
Cr-Si-O-N coatings with different oxygen contents were deposited by multi-arc ion plating.The oxygen contents of the deposited coatings significantly increased from 0 to 40 at.% and the N contents sharply decreased from 31.5 to 5 at.% when the O2/(N2 + O2) flux rate increased from 0% to 10%.With the addition of oxygen, the diffraction peaks of CrN (111) and CrN (200) became weaker and the diffraction peaks of CrN phases were no longer recognizable when the O2/(N2 + O2) flux rate above 10%.The Cr-Si-O-N coatings remained stable after annealed at 800 °C in the air for 2 h but break appeared

Conclusions
Cr-Si-O-N coatings with different oxygen contents were deposited by multi-arc ion plating.The oxygen contents of the deposited coatings significantly increased from 0 to 40 at.% and the N contents sharply decreased from 31.5 to 5 at.% when the O 2 /(N 2 + O 2 ) flux rate increased from 0% to 10%.With the addition of oxygen, the diffraction peaks of CrN (111) and CrN (200) became weaker and the diffraction peaks of CrN phases were no longer recognizable when the O 2 /(N 2 + O 2 ) flux rate above 10%.The Cr-Si-O-N coatings remained stable after annealed at 800 • C in the air for 2 h but break appeared in the inner structure after annealing at 1200 • C. The XRD, SEM and electrochemical corrosion resistance results showed that the Cr-Si-N coatings have better oxidation and corrosion resistance properties than that of Cr-Si-O-N coatings with various oxygen contents.

Table 1 .
The parameters for the deposition of Cr-Si-O-N coatings.

Figure 2 Figure 1 .
Figure 2 shows the XRD spectra of Cr-Si-O-N coatings as a function of O2/(N2 + O2) flux rate.It can be seen that only the diffraction peaks of CrN (111) and CrN (200) were detected for the Cr-Si-N Figure 1.Chemical composition (obtained by EPMA) of Cr-Si-O-N coatings as a function of O 2 /(N 2 + O 2 ) flux rate.

Figure 2
Figure 2 shows the XRD spectra of Cr-Si-O-N coatings as a function of O 2 /(N 2 + O 2 ) flux rate.It can be seen that only the diffraction peaks of CrN (111) and CrN (200) were detected for the

Figure 3
Figure3presents the Si 2p and N 1s spectra of Cr-Si-O-N coatings deposited under the O2/(N2 + O2) flux rate of 2% and 20%.The binding energy of Si for both coatings appeared at 101.5 eV (Figure3a), which is in good agreement with that of Si3N4 phase[14].In addition, little peaks of SiO2 phase appeared at 102.9 eV[13].In the N 1s spectra, we can find peaks at 396.7 and 397.9 eV, which correspond to CrN[16] and Si3N4[17], respectively, in the coating with lower O2/(N2 + O2) flux rate.However, when the flux rate increased to 20%, the peak corresponding to CrN disappeared, indicating that there was no crystalline CrN phase existing in the coatings anymore, and the left N in the coatings was all in the amorphous form of Si3N4.This also implied that crystal phase of CrN was present at lower O2/(N2 + O2) flux rate in accordance with XRD results.

Figure 2 .
Figure 2. The XRD spectra of Cr-Si-O-N coatings as a function of O 2 /(N 2 + O 2 ) flux rate.

Figure 3
Figure 3 presents the Si 2p and N 1s spectra of Cr-Si-O-N coatings deposited under the O 2 /(N 2 + O 2 ) flux rate of 2% and 20%.The binding energy of Si for both coatings appeared at 101.5 eV (Figure3a), which is in good agreement with that of Si 3 N 4 phase[14].In addition, little peaks of SiO 2 phase appeared at 102.9 eV[13].In the N 1s spectra, we can find peaks at 396.7 and 397.9 eV, which correspond to CrN[16] and Si 3 N 4[17], respectively, in the coating with lower O 2 /(N 2 + O 2 ) flux rate.However, when the flux rate increased to 20%, the peak corresponding to CrN disappeared, indicating that there was no crystalline CrN phase existing in the coatings anymore, and the left N in the coatings was all in the amorphous form of Si 3 N 4 .This also implied that crystal phase of CrN was present at lower O 2 /(N 2 + O 2 ) flux rate in accordance with XRD results.

Figure 3
Figure3presents the Si 2p and N 1s spectra of Cr-Si-O-N coatings deposited under the O2/(N2 + O2) flux rate of 2% and 20%.The binding energy of Si for both coatings appeared at 101.5 eV (Figure3a), which is in good agreement with that of Si3N4 phase[14].In addition, little peaks of SiO2 phase appeared at 102.9 eV[13].In the N 1s spectra, we can find peaks at 396.7 and 397.9 eV, which correspond to CrN[16] and Si3N4[17], respectively, in the coating with lower O2/(N2 + O2) flux rate.However, when the flux rate increased to 20%, the peak corresponding to CrN disappeared, indicating that there was no crystalline CrN phase existing in the coatings anymore, and the left N in the coatings was all in the amorphous form of Si3N4.This also implied that crystal phase of CrN was present at lower O2/(N2 + O2) flux rate in accordance with XRD results.

3. 2 .
Figure 5 shows the hardness value of Cr-Si-O-N coating as a function of O 2 /(N 2 + O 2 ) flux rate.It can be found clearly that the hardness sharply decreased with the addition of the oxygen from 0 to 2 at.% and slightly decreased with the further increase of the oxygen above 5 at.%.The hardness value of cemented carbide substrate is about 1260 Hv 0.05 , and the hardness of all the Cr-Si-O-N coatings is ranging around 2000 Hv 0.05 which is far less than the Cr-Si-N coatings.This phenomenon might be attributed to the formation of a large proportion of the Cr x O y amorphous phase in the as-deposited coatings which disrupted the formation of the nanocomposite structure of amorphous matrix and crystalline[19].The Cr-Si-O-N coatings deposited under O 2 /(N 2 + O 2 ) flux rate of 2% and 10% were chosen to do the high-temperature oxidation treatments.The XRD spectra of the coatings before and after oxidation under 800 • C and 1200 • C in the air were shown in Figure6.After the coatings were annealed under 800 • C in the air for 2 h, some diffraction patterns of Cr 2 O 3 and CrO 2 were detected in both coatings, which suggested that small extent of oxidation behavior might occur on the surface.The number of oxide diffraction patterns in the coatings with lower oxygen was much less than that in the higher one, indicating that the high-temperature oxidation ability of the former might be better.As the oxidation temperature rose to 1200 • C, the diffraction peaks, corresponding to Cr 2 O 3 and CrO 2 , became stronger, and their number increased, indicating that the oxidation progressed.The crystalline Cr 2 O 3 and CrO 2 appear due to oxidation of the coatings and crystallisation of amorphous

Coatings 2018, 1 ,
x FOR PEER REVIEW 8 of 11 The electrochemical behavior of corrosion resistance of the deposited Cr-Si-O-N films and cemented carbide substrate was tested by Electrochemical Working Station (CS350) in 3.5% NaCl solution under free air condition at room temperature.The Tafel plots of the base cemented carbide substrate and Cr-Si-O-N coatings with different oxygen contents are shown in Figure 9.The results such as corrosion potential, corrosion current, and polarization resistance fitted from the Tafel plots are listed in Table

Figure 11 .
Figure 11.(A) Equivalent circuit to fit the electrochemical impedance data of the WC-Co cemented carbide substrate; (B) General Equivalent circuit to fit the electrochemical impedance data of the coated samples.

Figure 11 .
Figure 11.(A) Equivalent circuit to fit the electrochemical impedance data of the WC-Co cemented carbide substrate; (B) General Equivalent circuit to fit the electrochemical impedance data of the coated samples.

Table 4 .
Fitting results of EIS plots in Figure 11.Samples R s (Ω•cm 2 ) CPE c -Y 0 (µF•cm 2 ) R c (Ω•cm 2 ) CPE dl -Y 0 (µF•cm 2 ) R ct (kΩ•cm Coatings 2018, 1, x FOR PEER REVIEW 3 of 11 circuit potential, and the mercury/mercury chloride combined with luggin probe as the reference electrode, and the platinum electrode (1 cm 2 ) as the counter electrode.To obtain accurate curve, Lujin capillary was as close as possible to the study electrode for decreasing liquid potential and the scan rate of 0.5 mV/s was used.The corrosion potential, corrosion current and polarization resistance obtained from the Tafel plots were used to evaluate the corrosion rate of deposited coatings.

Table 2 .
The cross-sectional chemical composition of the coatings after annealing at 800 °C, 1100 °C, 1200 °C.

Table 2 .
The cross-sectional chemical composition of the coatings after annealing at 800 • C, 1100 • C, 1200 • C.

Table 3 .
Corrosion potential, corrosion current, polarization resistance and corrosion rate properties got from the Tafel plots.

Table 3 .
Corrosion potential, corrosion current, polarization resistance and corrosion rate properties got from the Tafel plots.

Table 4 .
Fitting results of EIS plots in Figure10.