Effect of Molybdenum Content on Mechanical and Tribological Properties of Diamond-Like Carbon Coatings over Titanium β -21S Alloy

: Molybdenum-doped diamond-like carbon (Mo-DLC) coatings have been deposited on titanium β -21S alloy (Ti–15Mo–3Nb–3Al–0.2Si) using plasma-enhanced chemical vapor deposition (PECVD) equipped with pulsed-DC magnetron sputtering. Mo contents in the deposited coatings have been controlled with an applied sputtering pulse duty cycle. Chemical composition, structure, morphology, and topography have been studied using X-ray photoelectron spectroscopy (XPS), X-ray diffraction (XRD), ﬁeld emission scanning electron microscopy (FESEM), and atomic force microscopy (AFM), respectively. XRD pattern of the coating with highest Mo content shows the presence of carbide phase. Smooth features with low roughness values are observed in low Mo content coating, whereas high Mo content coating shows granular characteristics with a high roughness value. Raman spectra reveal the increased graphitic content for the highest metal concentration in the coatings, wherein the nanohardness is also determined to be highest, at about ~18–19 GPa. Furthermore, the ball-on-disk tribometry test on Mo-DLC coating with highest nanohardness shows a low coefﬁcient of friction (COF) of 0.2 with low wear loss compared to that of the substrate (0.62). the deposition of Mo-DLC coatings on titanium β -21S alloys using plasma-enhanced chemical vapor deposition (PECVD). Deposition of Mo-DLC coatings has been carried out using methane (CH 4 ) as the precursor gas for DLC coating with simultaneous Mo incorporation by magnetron sputtering, employing aMo target. The study includes a variation of Mo content through magnetron sputtering along with chemical vapor deposition of carbon matrix in the coatings. We have extensively studied the properties of the coatings toward tribological application. on the formation of sp 2 sites. As sp 2 carbon has lower binding energy compared to sp 3 carbon, it prefers to bond with Mo to form Mo carbide phase. It is clearly observed that molybdenum carbide was found to increase with the increase in molybdenum content in turn to the applied duty cycle. This is also found to inﬂuence the disordered nature where graphitic nature improves over the diamond-like behaviour of the coatings.

Titanium alloys are excellent materials for biomedical research and they have widely been considered as bioimplants for their lightweight, corrosion-resistant, high strength, low C 2021, 7, 1 2 of 20 modulus properties [33][34][35][36][37][38][39]. Among various phases of titanium alloys, β class-containing elements such as Mo, Sn, Ta, Nb, and Zr [40] are preferred for orthopedic implants because of their non-toxic nature and superior properties [41,42]. Particularly, the β class of Ti-Mo alloys have extensively been reported as biomaterials for their superior mechanical and corrosion properties [42][43][44][45][46][47]. However, titanium alloys have some limitations because of their poor tribological performance, high coefficients of friction, and the absence of mechanical stability due to formation of their native oxides. In this regard, surface modification of titanium alloys is a well-known process to improve their mechanical and tribological properties [48,49]. Surface modification with biocompatible DLC coating is one of the practicable routes for enhanced mechanical and tribological properties of titanium alloys [49].
In this research work, we have investigated the mechanical and tribological behaviors of surface-modified titanium β-21S alloys (Ti-15Mo-3Nb-3Al-0.2Si) using molybdenumdoped DLC coatings. In general, pure DLC and M-DLC coatings are commonly deposited using chemical vapor deposition (CVD) and physical vapour deposition (PVD) methods [50][51][52][53][54][55][56][57]. These coatings are also reported to be deposited on titanium and its alloys to improve their overall properties [58][59][60]. Molybdenum is a common transition metal found to exist in several oxidation states. It readily forms thermally stable alloys and hard carbides [61] and shows enhanced blood compatible [62] and biocompatible [63] properties. The present work has been motivated by these characteristics of Mo metal. Some of the earlier reports on Mo-doped DLC coatings have demonstrated their improved thermal, mechanical, tribological, and biocompatible properties [6,51,[64][65][66][67]. However, development of Mo-DLC coatings on the titanium β-21S alloy is limited in the literature [6]. Here, we present the deposition of Mo-DLC coatings on titanium β-21S alloys using plasmaenhanced chemical vapor deposition (PECVD). Deposition of Mo-DLC coatings has been carried out using methane (CH 4 ) as the precursor gas for DLC coating with simultaneous Mo incorporation by magnetron sputtering, employing aMo target. The study includes a variation of Mo content through magnetron sputtering along with chemical vapor deposition of carbon matrix in the coatings. We have extensively studied the properties of the coatings toward tribological application.

Substrate Preparation
Titanium β-21S alloy (Ti-15Mo-3Nb-3Al-0.2Si) with a size of 2 cm × 2 cm, and Si(100) wafers with a size of 1 cm × 1 cm, were mainly used as the substrate materials for this study. The surface of each titanium alloy substrate, hereafter called Ti-Mo, was ground and polished to a mirror finish in order to remove any unwanted contamination and/or oxide layers. The Si wafer and polished Ti-Mo substrates were then separately ultrasonically cleaned for 15 min in acetone, dried, and loaded together into deposition chamber of PECVD system. Figure 1 shows the entire set up of the PECVD system used for this research. Before deposition of coatings, the chamber was evacuated to a base pressure of 5 × 10 −6 mbar through turbo-molecular pump backed with an oil rotary pump. At first, inductively coupled Ar plasma was created by 50 W RF power at 13.56 MHz frequency to clean the substrates for about 10 min at a bias of −200 V. The process was continued with deposition of coatings by introducing CH 4 gas with Ar in a mixing ratio of 3:7. At the same time, Mo was sputtered from a 99.9% pure Mo target with a 3 inch diameter through a magnetron gun powered by a symmetric bipolar-pulsed DC power supply from MAGPULS Stromversorgungs Systeme GmbH, Germany. During the deposition, the chamber pressure was maintained at 11 µbar. The substrates were negatively biased with −150 V. Keeping all the process conditions sputtering bias to 300 V and frequency at 10 kHz as the same, Mo content in the coatings were varied through the bipolar pulse on-time and off-time such that the duty cycles are fixed to 10%, 40%, and 85%, where duty cycle = positive pulse on-time/ (total pulse duration). Generally, all the experiments were conducted for a duration of 30 min. However, for wear testing, samples were prepared for 60 min with a pulse duty cycle of 40%.

Deposition
C 2021, 7, x FOR PEER REVIEW sure was maintained at 11 µbar. The substrates were negatively biased with −150 ing all the process conditions sputtering bias to 300 V and frequency at 10 kHz as Mo content in the coatings were varied through the bipolar pulse on-time and such that the duty cycles are fixed to 10%, 40%, and 85%, where duty cycle = posit on-time/ (total pulse duration). Generally, all the experiments were conducted fo tion of 30 min. However, for wear testing, samples were prepared for 60 min wit duty cycle of 40%.

Characterization
The chemical compositions and nature elemental species of Mo-DLC coati analyzed by X-ray photoelectron spectroscopy (XPS) using an SPECS spec (SPECS Surface Nano Analysis GmbH, Berlin, Germany). Spectra were record non-monochromatic AlKα radiation (1486.6 eV) as an X-ray source operated at 1 kV, 12.5 mA). The binding energies reported here were referenced with C 1s pea eV. Survey spectra were obtained with pass energy of 70 eV and a step increm eV. All the individual core level spectra of C 1s and Mo 3d were recorded wi energy and step increment of 25 and 0.05 eV, respectively. For XPS recording, sample was mounted on the sample holder and placed into a load-lock chambe ultrahigh vacuum (UHV) of 8 × 10 −8 mbar for 5 h in order to desorb any volatil present on its surface. After 5 h, the sample was transferred into the analyzing

Characterization
The chemical compositions and nature elemental species of Mo-DLC coatings were analyzed by X-ray photoelectron spectroscopy (XPS) using an SPECS spectrometer (SPECS Surface Nano Analysis GmbH, Berlin, Germany). Spectra were recorded using nonmonochromatic AlKα radiation (1486.6 eV) as an X-ray source operated at 150 W (12 kV, 12.5 mA). The binding energies reported here were referenced with C 1s peak at 284.6 eV. Survey spectra were obtained with pass energy of 70 eV and a step increment of 0.5 eV. All the individual core level spectra of C 1s and Mo 3d were recorded with a pass energy and step increment of 25 and 0.05 eV, respectively. For XPS recording, a coating sample was mounted on the sample holder and placed into a load-lock chamber with an ultrahigh C 2021, 7, 1 4 of 20 vacuum (UHV) of 8 × 10 −8 mbar for 5 h in order to desorb any volatile species present on its surface. After 5 h, the sample was transferred into the analyzing chamber with UHV of 5 × 10 −10 mbar and spectra were recorded. The atomic percentages of Mo in the Mo-DLC coatings were estimated using the areas under the Mo 3d and C 1s core level peaks and the respective XPS relative sensitivity factors (RSF). C 1s and Mo 3d core level spectra were also curve-fitted with Gaussian-Lorentzian peaks to identify various carbon and molybdenum species after Shirley background subtraction. Relative concentrations of different carbon and molybdenum species were evaluated from the area under the respective peak in the curve-fitted spectrum.
XRD patterns of Si substrate and Mo-DLC coatings on the substrate with high and low Mo contents were recorded, employing an X-ray diffractometer (Bruker D8 advance) operated with CuKα radiation (λ = 1.5406 Å) at 35 kV and 10 mA. The samples were scanned over 2θ range of 30-55 • .
Surface morphologies of the coatings on Si substrate were obtained by FESEM with Carl Zeiss SUPRA 40VP.
Surface topography was investigated by AFM with Surface Imaging Systems on a 10 µm × 10 µm area on the coatings on Si substrate (CSEM Instruments (Model SSI)) operated in contact mode.
Raman spectra were collected using confocal micro Raman spectrometer with DILOR-JOBIN-YVON SPEX (LABRAM 010A) and a He-Ne laser source at a wavelength of 632.8 nm. Raman spectra were also obtained on the wear tracks of Mo-DLC coatings to understand the change in materials during the friction process.
Optical micrographs of wear tracks of Mo-DLC coatings and substrate were collected using an Olympus microscope to find out the nature of wear that occurred on the coatings and the substrate.

Properties
Nanohardnesses of the Mo-DLC coatings were tested using a Nano Hardness Tester (CSM Instruments, Peseux, Switzerland) with a Berkovich diamond indenter at a 5 mN load. The values were determined from the indentation curves using the Oliver and Pharr method [68].
Wear studies of bare substrate and Mo-DLC-coated substrate were carried out in a reciprocating wear tester (model CM 9084 DuCom) as per ASTM G133-02 standard [37]. However, these tests are not in full compliance with the provisions of Test Method G 133, Procedure A, because the normal forces in these experiments were 3 and 5 N instead of 25 N as prescribed by the standard. Moreover, the stroke length was 10 mm and an alumina ball of 6 mm diameter was used as the counter surface. Furthermore, the experiments were carried out at a frequency of 100 Hz min −1 for 20 min. Relative humidity during tribology measurements was around 65%. After the experiments, wear profiles of the samples were examined by a 3D-profilometer with Nanomap 500 LS (AEP Technology, Santa Clara, CA, USA). Considering the ASTM G133-02 method, the wear losses were calculated from the cross-sectional areas of the respective wear profiles.

XPS Studies and Compositions
The nature of chemical bonds present in the coatings prepared with different duty cycles and their compositions have been identified using XPS. Survey spectra of Mo-DLC coatings in Figure 2 reveal the peaks related to C 1s, Mo 3d, and O 1s. These peaks confirm the presence of molybdenum and carbon in the coatings. The presence of oxygen could arise because of exposure of coatings to atmosphere prior to XPS measurements. The atomic percentages of Mo in Mo-DLC coatings evaluated from their core level spectra are 3.9%, 6.9%, and 16.2%, respectively. As high-resolution C 1s core level spectra are found to be asymmetric in nature, an in-depth chemical analysis has been carried out to decompose the spectra into various carbon component species using non-linear curve-fitting.  presents curve-fitted C 1s core level spectra of Mo-DLC coatings. The component peaks at 284.6 and 285.0 eV are identified to be sp 2 hybridized carbon (C=C) and sp 3 hybridized carbon (C-C) [69], whereas the peak at lower binding energy of 283.5 eV is related to Mo-C bonds [51]. The peaks at 286.6 and 287.7 eV are attributed to C-O and C=O bonds, respectively [2,4]. The difference in binding energies between sp 3 and sp 2 carbons is associated with their coordination, hydridization, and core-hole relaxation [70]. Broad Mo 3d core level spectral envelopes indicate the presence of different Mo species in the coatings and are resolved into component peaks by curve-fitting as displayed in Figure 4. Mo 3d 5/2,3/2 doublet at 228.4 and 231.5 eV is identified as a molybdenum carbide species containing Mo-C bond. Observed doublet peaks in the curve-fitted spectra at 229.9 and 233.2 eV are assigned for Mo 4+ species. Doublet binding energies at 232.4 and 235.3 eV correspond to Mo 6+ species in the coatings [6,51]. However, relative concentrations of molybdenum carbide, Mo 4+ and Mo 6+ species are estimated from areas under the curves in respective Mo 3d core level spectra and are found to be different in the coatings. Binding energies and relative concentrations of carbon and molybdenum species are summarized in Table 1. Overall Mo contents are found to increase with the increase in the pulse duty cycles according to the table. Further, the effect of metal content is also understood through its carbide content and disordered nature. Figure 5 shows the estimated molybdenum carbide contents and sp 2 to sp 3 carbon ratios for various amounts of molybdenum inclusion in the DLC coatings. It has been found that sp 2 carbon concentration gets increased with increase in Mo concentration. Nevertheless, the increase in sp 2 /sp 3 ratio is because of the Mo doping that can be attributed to the catalyst effect of Mo atoms on the formation of sp 2 sites. As sp 2 carbon has lower binding energy compared to sp 3 carbon, it prefers to bond with Mo to form Mo carbide phase. It is clearly observed that molybdenum carbide was found to increase with the increase in molybdenum content in turn to the applied duty cycle. This is also found to influence the disordered nature where graphitic nature improves over the diamond-like behaviour of the coatings.
C 2021, 7, x FOR PEER REVIEW peaks at 284.6 and 285.0 eV are identified to be sp 2 hybridized carbon (C=C) and bridized carbon (C-C) [69], whereas the peak at lower binding energy of 283.5 e lated to Mo-C bonds [51]. The peaks at 286.6 and 287.7 eV are attributed to C-O a bonds, respectively [2,4]. The difference in binding energies between sp 3 and sp 2 is associated with their coordination, hydridization, and core-hole relaxation [70 Mo 3d core level spectral envelopes indicate the presence of different Mo specie coatings and are resolved into component peaks by curve-fitting as displayed in F Mo 3d5/2,3/2 doublet at 228.4 and 231.5 eV is identified as a molybdenum carbide containing Mo-C bond. Observed doublet peaks in the curve-fitted spectra at 22 233.2 eV are assigned for Mo 4+ species. Doublet binding energies at 232.4 and 2 correspond to Mo 6+ species in the coatings [6,51]. However, relative concentration lybdenum carbide, Mo 4+ and Mo 6+ species are estimated from areas under the c respective Mo 3d core level spectra and are found to be different in the coatings. energies and relative concentrations of carbon and molybdenum species are sum in Table 1. Overall Mo contents are found to increase with the increase in the pu cycles according to the table. Further, the effect of metal content is also und through its carbide content and disordered nature. Figure 5 shows the estimated denum carbide contents and sp 2 to sp 3 carbon ratios for various amounts of moly inclusion in the DLC coatings. It has been found that sp 2 carbon concentration creased with increase in Mo concentration. Nevertheless, the increase in sp 2 /sp 3 because of the Mo doping that can be attributed to the catalyst effect of Mo atom formation of sp 2 sites. As sp 2 carbon has lower binding energy compared to sp 3 c prefers to bond with Mo to form Mo carbide phase. It is clearly observed that denum carbide was found to increase with the increase in molybdenum content in the applied duty cycle. This is also found to influence the disordered nature wh phitic nature improves over the diamond-like behaviour of the coatings.               [72]. Mo 2 C is known to be an extremely hard ceramic material which can help in improving thermal stability. Thus, the diffraction peaks confirm the presence of molybdenum carbide in the Mo-DLC coating with high Mo content. Nevertheless, this finding is also corroborated with XPS results, wherein the presence of molybdenum carbide is clearly observed with respect to C 1s and Mo 3d core level spectra.

FESEM and AFM Studies
Surface morphology and topography, including surface roughness of Mo-DLC coatings for highest and lowest molybdenum content, are shown in Figures 7 and 8, respectively. At 3.9 at.% Mo content, coating is found to be smooth, with corresponding surface roughness values of around 0.73 nm and appears as a highly smooth coating, as seen in Figure 7a. However, at higher Mo content of about 16.2 at.%, coating exhibits more granular features, possibly from the formation of nanoaggregate features with an increased

FESEM and AFM Studies
Surface morphology and topography, including surface roughness of Mo-DLC coatings for highest and lowest molybdenum content, are shown in Figures 7 and 8, respectively. At 3.9 at.% Mo content, coating is found to be smooth, with corresponding surface roughness values of around 0.73 nm and appears as a highly smooth coating, as seen in Figure 7a. However, at higher Mo content of about 16.2 at.%, coating exhibits more granular features, possibly from the formation of nanoaggregate features with an increased roughness value of about 1.2 nm. This could be due to the presence of a higher fraction of molybdenum carbide phase and increased metallicity of the coating. Thus, scanning electron micrographs of the coatings corroborate the observation from atomic force microscopy as shown in Figure 8.

FESEM and AFM Studies
Surface morphology and topography, including surface roughness of Mo-DLC coatings for highest and lowest molybdenum content, are shown in Figures 7 and 8, respectively. At 3.9 at.% Mo content, coating is found to be smooth, with corresponding surface roughness values of around 0.73 nm and appears as a highly smooth coating, as seen in Figure 7a. However, at higher Mo content of about 16.2 at.%, coating exhibits more granular features, possibly from the formation of nanoaggregate features with an increased roughness value of about 1.2 nm. This could be due to the presence of a higher fraction of molybdenum carbide phase and increased metallicity of the coating. Thus, scanning electron micrographs of the coatings corroborate the observation from atomic force microscopy as shown in Figure 8.

Raman Spectroscopy Studies
Structural and phase disorder of the coatings have been studied using micro Raman spectroscopy. Figure 9a shows typical Raman spectra of Mo-DLC coatings on Si substrate recorded in the range of 1000 to 1800 cm −1 for various Mo concentrations. The spectra provides information about the microstructural changes due to doping of Mo into the carbon matrix. As spectra are asymmetric, they are curve-fitted into several component peaks using a Gaussian function [73]. A typical curve-fitted Raman spectrum of 3.9 at.% Mo-DLC coating is presented in Figure 9b. The choice of using this function is because Gaussian line profile closely accounts to the nature and chemical disorder of bonds, whereas other fitting functions, such as Lorentzian functions, relate to the photoionization process [73]. Ferrari and Robertson have reported that Gaussian line function is resulted from random distribution of phonon lifetimes in disordered materials [74]. Therefore, the curve fitting using Gaussian function can provide a better insight into the coating. In the present study, all the spectra are curve-fitted into four components. G peak located around 1560-1580 cm −1 occurring due to E2g stretching of sp 2 bonds in the ring as well as in chains and D

Raman Spectroscopy Studies
Structural and phase disorder of the coatings have been studied using micro Raman spectroscopy. Figure 9a shows typical Raman spectra of Mo-DLC coatings on Si substrate recorded in the range of 1000 to 1800 cm −1 for various Mo concentrations. The spectra provides information about the microstructural changes due to doping of Mo into the carbon matrix. As spectra are asymmetric, they are curve-fitted into several component peaks using a Gaussian function [73]. A typical curve-fitted Raman spectrum of 3.9 at.% Mo-DLC coating is presented in Figure 9b. The choice of using this function is because Gaussian line profile closely accounts to the nature and chemical disorder of bonds, whereas other fitting functions, such as Lorentzian functions, relate to the photoionization process [73]. Ferrari and Robertson have reported that Gaussian line function is resulted from random distribution of phonon lifetimes in disordered materials [74]. Therefore, the curve fitting using Gaussian function can provide a better insight into the coating. In the present study, all the spectra are curve-fitted into four components. G peak located around 1560-1580 cm −1 occurring due to E 2g stretching of sp 2 bonds in the ring as well as in chains and D peak at 1360 cm −1 is usually because of the disordered A 1g breathing modes of sp 2 carbon in the ring [75,76]. Along with these two peaks, two more peaks are found around 1200 cm −1 and 1480 cm −1 related to modes of trans polyacetylene, which are the sum and difference of C=C chain stretching and C-H wagging modes respectively [77]. It is to be noted that Raman spectrum indicate the ordering of sp 2 sites and thereby measures the sp 3 fraction in the coating. The observation on the sp 2 clusters formation and ID/IG ratio can be substantiated with XPS analysis, wherein sp 2 to sp 3 C ratios are noted to increase with Mo contents. In short, as the sp 2 to sp 3 C decreases, the G peak intensity and ID/IG also decreases [84]. Many earlier works on M-DLC composite films have similarly been accounted for this type of linear behaviour [85]. Ji et al., in their Mo-DLC nanocomposite coating, have observed not only increased sp 2 content with higher Mo doping, but also that Mo undergoes preferential bonding to sp 2 carbon to form MoC [51]. A similar trend is also observed in the XPS analysis ( Figure 5). Thus, these sp 2 clusters in the Mo-DLC coatings mostly form chains and dimers undergo larger grain boundaries that lie around MoC [84,86].

Mechanical Properties
Nanohardness and elastic modulus values of Mo-DLC coatings for various Mo contents are presented in Figure 10. The coatings show significant variation in nanohardness and elastic modulus values for various Mo contents. A high nanohardness of about 19.2 GPa was observed for the coating with 6.9 at.% Mo content, whereas the higher Mo content (16.2 at.%) coating shows a nanohardness value of 18.7 GPa. This slight variation in nanohardness with significant change in Mo content may possibly be due to variation in thickness of the coatings or distribution of MoC clusters in the carbon matrix. However, both the coatings (6.9 and 16.2 at.%) have same and lowest value of elastic modulus that can be corroborated with a saturated amount of MoC in the coatings as the Mo concentration increases. Thus, the consistency of nanohardness values across different places in the coating can be inferred from Figure 11, which shows series of nanoindentation curves obtained for the highest Mo content (16.2 at.%) coating.   [78]. Tuinstra and Koenig have proposed that I D /I G is inversely related to the size of graphite cluster [79]. However, in the work of Cho and coworkers [80], when I D /I G value crosses over 1.1, this inverse relationship may no longer hold, and instead the I D /I G values behave proportional to graphitic content in the film. Therefore, in the case of Mo doping, the observed increase in I D /I G ratios from 0.94 to 1.3 can be associated to increased graphitization of the coating. In addition, full-width-at-half-maxima (FWHM) represent the measure of cluster size and structure disorder in the coatings. This occurs due to distortion in bond length and bond angle, particularly in sp 2 sites [81,82]. Figure 9c reveals that FWHM of G peaks (FWHM G ) decrease with Mo content in the coating. Thus, Mo incorporation induces increased sp 2 clusters around the carbon matrix. As reported, increase in metal incorporation substantially reduces the residual stress, and in turn shows a decrease in FWHM of G peak and enhances its tribomechanical properties [83].
The observation on the sp 2 clusters formation and I D /I G ratio can be substantiated with XPS analysis, wherein sp 2 to sp 3 C ratios are noted to increase with Mo contents. In short, as the sp 2 to sp 3 C decreases, the G peak intensity and I D /I G also decreases [84]. Many earlier works on M-DLC composite films have similarly been accounted for this type of linear behaviour [85]. Ji et al., in their Mo-DLC nanocomposite coating, have observed not only increased sp 2 content with higher Mo doping, but also that Mo undergoes preferential bonding to sp 2 carbon to form MoC [51]. A similar trend is also observed in the XPS analysis ( Figure 5). Thus, these sp 2 clusters in the Mo-DLC coatings mostly form chains and dimers undergo larger grain boundaries that lie around MoC [84,86].

Mechanical Properties
Nanohardness and elastic modulus values of Mo-DLC coatings for various Mo contents are presented in Figure 10. The coatings show significant variation in nanohardness and elastic modulus values for various Mo contents. A high nanohardness of about 19.2 GPa was observed for the coating with 6.9 at.% Mo content, whereas the higher Mo content (16.2 at.%) coating shows a nanohardness value of 18.7 GPa. This slight variation in nanohardness with significant change in Mo content may possibly be due to variation in C 2021, 7, 1 12 of 20 thickness of the coatings or distribution of MoC clusters in the carbon matrix. However, both the coatings (6.9 and 16.2 at.%) have same and lowest value of elastic modulus that can be corroborated with a saturated amount of MoC in the coatings as the Mo concentration increases. Thus, the consistency of nanohardness values across different places in the coating can be inferred from Figure 11, which shows series of nanoindentation curves obtained for the highest Mo content (16.2 at.%) coating.

Mechanical Properties
Nanohardness and elastic modulus values of Mo-DLC coatings for various Mo contents are presented in Figure 10. The coatings show significant variation in nanohardness and elastic modulus values for various Mo contents. A high nanohardness of about 19.2 GPa was observed for the coating with 6.9 at.% Mo content, whereas the higher Mo content (16.2 at.%) coating shows a nanohardness value of 18.7 GPa. This slight variation in nanohardness with significant change in Mo content may possibly be due to variation in thickness of the coatings or distribution of MoC clusters in the carbon matrix. However, both the coatings (6.9 and 16.2 at.%) have same and lowest value of elastic modulus that can be corroborated with a saturated amount of MoC in the coatings as the Mo concentration increases. Thus, the consistency of nanohardness values across different places in the coating can be inferred from Figure 11, which shows series of nanoindentation curves obtained for the highest Mo content (16.2 at.%) coating.

Tribological Properties
As the Mo-DLC coating deposited at 30 min (Mo content of 6.9 at.%) shows the highest nanohardness, this coating has been tested for tribological behaviour on β-21S Ti-Mo alloy substrate using a ball-on-disk tribometer. Besides, the test has been extended to Mo-DLC coatings deposited for 60 min. Furthermore, friction behaviours of the coatings deposited with 30 and 60 min have been studied at two different stroke loads of 3 and 5 N. Figure 12 shows the evolution of the coefficients of friction of Mo-DLC coatings over time for two different loads. It is observed that the substrate coated for a longer duration of 60 min shows a relatively stable friction-time profile with average coefficients of friction of 0.17 and 0.15, which are lower than 0.62 and 0.43 for bare β-21S Ti-Mo alloy substrate for 3 and 5 N loads, respectively as summarized in Table 2. On the other hand, in the case of coating with a shorter duration of 30 min, the coefficient of friction abruptly increases from a steady state of 0.2 and 0.17 to substrate values of 0.64 and 0.63. The sudden increase in friction coefficient is due to loss and transfer of the coating materials on the counter part (ball) during wear stroke [87]. It is also known that tribochemical reaction at the point of contact between ball and coating can deteriorate the Mo-DLC layer [67]. This chemical reaction can create a third superficial layer called a tribolayer that acts as a wear resistant. Graphitization can also take place that reduces the friction coefficient, which is observed in the case of Ti-doped DLC [88] and Zr-doped DLC coatings [89]. In the present study, Mo-DLC coating obtained with 60 min duration shows a constant friction coefficient, which is probably due to graphitization from tribochemical reactions occurred by transformation of sp 3 C to sp 2 carbon in the coatings, possibly by thermal process during wear stroke [90].

Tribological Properties
As the Mo-DLC coating deposited at 30 min (Mo content of 6.9 at.%) shows the highest nanohardness, this coating has been tested for tribological behaviour on β-21S Ti-Mo alloy substrate using a ball-on-disk tribometer. Besides, the test has been extended to Mo-DLC coatings deposited for 60 min. Furthermore, friction behaviours of the coatings deposited with 30 and 60 min have been studied at two different stroke loads of 3 and 5 N. Figure 12 shows the evolution of the coefficients of friction of Mo-DLC coatings over time for two different loads. It is observed that the substrate coated for a longer duration of 60 min shows a relatively stable friction-time profile with average coefficients of friction of 0.17 and 0.15, which are lower than 0.62 and 0.43 for bare β-21S Ti-Mo alloy substrate for 3 and 5 N loads, respectively as summarized in Table 2. On the other hand, in the case of coating with a shorter duration of 30 min, the coefficient of friction abruptly increases from a steady state of 0.2 and 0.17 to substrate values of 0.64 and 0.63. The sudden increase in friction coefficient is due to loss and transfer of the coating materials on the counter part (ball) during wear stroke [87]. It is also known that tribochemical reaction at the point of contact between ball and coating can deteriorate the Mo-DLC layer [67]. This chemical reaction can create a third superficial layer called a tribolayer that acts as a wear resistant. Graphitization can also take place that reduces the friction coefficient, which is observed in the case of Ti-doped DLC [88] and Zr-doped DLC coatings [89]. In the present study, Mo-DLC coating obtained with 60 min duration shows a constant friction coefficient, which is probably due to graphitization from tribochemical reactions occurred by transformation of sp 3 C to sp 2 carbon in the coatings, possibly by thermal process during wear stroke [90].   Figure 13 presents typical 3D micrographs of wear tracks and their corresponding 2D cross-section line profiles of bare β-21S, Mo-DLC-coated substrates at 30 and 60 min duration after friction test. Wear loss has been estimated as per ASTM G 133 standard. Table 3 lists the wear loss values of bare substrate and two Mo-DLC-coated substrates.   Figure 13 presents typical 3D micrographs of wear tracks and their corresponding 2D cross-section line profiles of bare β-21S, Mo-DLC-coated substrates at 30 and 60 min duration after friction test. Wear loss has been estimated as per ASTM G 133 standard. Table 3 lists the wear loss values of bare substrate and two Mo-DLC-coated substrates. For the coating with 30 min deposition at 3 N load, it is observed that coated substrate shows a wear loss of about 0.0025 mm 3 compared to bare substrate wear loss of 0.00475 mm 3 , whereas with 5 N load, the wear loss is 0.0095 mm 3 that is comparatively less than 0.015 mm 3 of the bare substrate. For the coating with 30 min deposition at 3 N load, it is observed that coated substrate shows a wear loss of about 0.0025 mm 3 compared to bare substrate wear loss of 0.00475 mm 3 , whereas with 5 N load, the wear loss is 0.0095 mm 3 that is comparatively less than 0.015 mm 3 of the bare substrate.   Figure 14 shows the optical micrographs of wear tracks of Mo-DLC-coated β-21S Ti-Mo alloy substrate and uncoated substrate. The wear track on the bare substrate shows abrasive wear with a highly rough and scratchy surface. On the other hand, wear tracks of coated substrates are found to be smoother than bare substrate, which may be due to adhesive wear. Thus, it can be inferred from the optical images that the wear tracks are an abrasive kind for substrate and an adhesive for coated substrate.  Figure 14 shows the optical micrographs of wear tracks of Mo-DLC-coated β-21S Ti-Mo alloy substrate and uncoated substrate. The wear track on the bare substrate shows abrasive wear with a highly rough and scratchy surface. On the other hand, wear tracks of coated substrates are found to be smoother than bare substrate, which may be due to adhesive wear. Thus, it can be inferred from the optical images that the wear tracks are an abrasive kind for substrate and an adhesive for coated substrate. Raman spectroscopic analyses have been carried out on the wear tracks to throw the light on the changes in the characteristics of material during the friction process. Figure  15 displays the Raman spectra of as-deposited coatings and the wear track of the same coatings. The wear track of the substrate with Mo-DLC coating for 30 min deposition shows no carbon signal, which clearly indicates that coating is completely removed dur- Raman spectroscopic analyses have been carried out on the wear tracks to throw the light on the changes in the characteristics of material during the friction process. Figure 15 displays the Raman spectra of as-deposited coatings and the wear track of the same coatings. The wear track of the substrate with Mo-DLC coating for 30 min deposition shows no carbon signal, which clearly indicates that coating is completely removed during wear stroke. However, substrate coated for 60 min deposition shows low intense carbon peaks, indicating the presence of carbon bonds after frictional testing formed during wear stroke. This is because of graphitization induced from the wear process as observed by Müller et al., in their coating [90]. Thus, Mo-DLC coatings deposited for 60 min duration shows superior tribological performance in terms of wear loss and the lowest coefficient of friction on a titanium β-21S alloy.

Conclusions
Molybdenum-doped diamond-like carbon coatings have successfully been deposited using PECVD with pulsed-DC magnetron sputtering of molybdenum target particularly on titanium β-21S alloy (Ti-15Mo-3Nb-3Al-0.2Si). Three different Mo contents (3.9, 6.9, and 16.2 at.%) have been incorporated by varying the pulse duration of sputtering bipolar pulse as determined from chemical analysis through XPS. Mo is present in carbidic form, Mo 5+ , and Mo 6+ oxidation states as observed in XPS studies. The deposited coatings are

Conclusions
Molybdenum-doped diamond-like carbon coatings have successfully been deposited using PECVD with pulsed-DC magnetron sputtering of molybdenum target particularly on titanium β-21S alloy (Ti-15Mo-3Nb-3Al-0.2Si). Three different Mo contents (3.9, 6.9, and 16.2 at.%) have been incorporated by varying the pulse duration of sputtering bipolar pulse as determined from chemical analysis through XPS. Mo is present in carbidic form, Mo 5+ , and Mo 6+ oxidation states as observed in XPS studies. The deposited coatings are also found to be more graphitic at a higher Mo content along with the presence of its carbide phase as evidenced through XRD studies. Moreover, surface morphology and topography measurements reveal that the coatings are, in general, smooth, with roughness ranging from 0.7-1.2 nm. Results obtained from Raman spectroscopy studies on Mo-DLC coatings with three Mo contents are complimentary to XPS results revealing that the coatings behave with a more graphite-like nature with increase in Mo content. Furthermore, nanoindentation measurements show a hardness ranging around~18-19 GPa with no significant increase with higher Mo content. Finally, the ball-on-disk wear testing shows low coefficient of friction of about~0.15 than β-21S Ti-Mo alloy and coating behaves like adhesive wear on the counter surface with negligible wear loss.