3.1 Depositions, chemical composition and structure
Ti-C-N coatings with the thickness of about 8 mm were deposited by HiPIMS at various target power from 4 to 8 kW and negative substrate bias from -60 to -200 V in closed field unbalanced magnetron sputtering system in the mixture gas of acetylene, argon and nitrogen. The typical HiPIMS pulses are shown in Fig.1a. At constant width of micro-pulse, the increased HiPIMS power indicates the increased number of micro-pulse, ie., the frequency was increased. Thus, the increased unit of power density produce much more ionized target species to form high density of plasma. With an increase of negative substrate bias, the ion bombardment into substrate surface was enhanced by the increased impact energy of energetic ions in plasma. In addition, the ion bombardment has a sensitive correlation with films growth in terms of the migration of adatoms and growth of metastable phases in a nonequilibrium condition. A strong impact into surface always results in rough surface morphology along with the increased sputtering yield and residual stress [35-37]. Hence, the optimized reactive HiPIMS depositions with the appropriate target sputtering power and negative substrate bias are beneficial for high-performance coatings.
Fig.1b shows XRD patterns of Ti-C-N coatings deposited by HiPIMS at various power and substrate bias. As shown in Fig.1b, (111) and (222) peaks with broadening width overlapped by TiN and Ti-C-N phases are detected in XRD patterns of Ti-C-N coatings except the diffraction peaks labeled as “SS” from 316L stainless steel substrates and Ti adhesion layers. The diffraction peaks from TiN and Ti-C-N are hard to be completely distinguished to the same face-centered structure and similar atomic radius. In addition, the broadening width and shifting towards low angle of (111) and (222) peaks indicate significant refinement of grains. At 4 kW and -60 V, the coatings are dominated by cubic Ti-C-N phase even at a lower density of Ti plasma due to a wide range of chemical composition of TiCxN1-x (0<x<1) phase [38,39]. When the negative substrate bias is fixed at -60 V, the (111)-textured coatings are gradually obtained with increasing target power from 4 to 8 kW. It is noted that the coatings deposited at 4 kW and -60 V show random orientations of (111) and (220). According to our previous researches, the disappearance of the preferred orientation is attributed to the formation of ultrafine grains [40,41].
Fig.1c-e presents TEM images of cross-sectional morphologies of Ti-C-N coatings deposited by HiPIMS at 5 kW and -60 V. As shown in Fig.1c, Ti-C-N coatings are composed of Ti adhesion layer, TiN transition layer and Ti-C-N coatings. Both Ti and TiN layer possess dense columnar structure, but Ti-C-N coatings have a highly dense and discontinuous nano-columnar. Moreover, there are also highly dense interfaces, so that no obvious interfaces between layers are observed. Hence, according to structure zone diagram (SZD) modified by Anders , the fine-nanograins with preferred orientation of coatings belongs to Zone T. The inserted selected area electron diffraction (SAED) patterns reveal distinct diffraction rings assigned to (111), (200) and (220) reflections from f.c.c crystal structure of Ti-C-N coatings including Ti(C,N) and TiN, due to their little difference in lattice constant. These agree well with XRD results of Fig.1b.
As shown in Fig.1d, high resolution TEM image of selected red frame in Fig.1c presents dense microstructure without distinct grain boundaries. (111) and (220) reflections for TiN and Ti(C,N) are detected in selected area electron diffraction (SAED) patterns inserted in Fig.1c. Fig.1e shows uniform element distributions of Ti, C and N in the selected red frame of Fig.1c detected by EDS lateral scanning. It is clearly observed that much more Ti than N and C elements, indicating Ti is enough to react with N and C to form TiN and Ti(C,N) in TiCN coatings. Based on XRD and XPS results, it is also noted that there are no halos and less content of C than Ti and N, revealing few amorphous CNx phase in the coatings. The increased HiPIMS target sputtering power and appropriate substarte bias are efficient to generate high ionization rate of target species and the increased metal ion bombardment [40,43]. Thus, a mass of momentum transfer between the growing film and the incoming metal ions results in the refinement of microstructure with a less coarse columnar. Hence, high surface integrity including high density, free of defects and uniformity is obtained for Ti-C-N coatings deposited by HiPIMS at appropriate conditions (5 kW and -60 V).
XPS core level spectra of Ti 2p, C 1s and N 1s were determined to analyze chemical bonding state of the elements in Ti-C-N coatings deposited by HiPIMS at 4-8 kW and -60 V, as shown in Fig.2. At 4 kW, the asymmetric Ti 2p spectrum is composed of two main peaks originated from Ti 2p3/2 and Ti 2p1/2 peaks at Binding energies (BE) of 457.1/463.1 eV, 455.9/461.8 eV and 458.4/464.4 eV (Fig.2a), indicating 3 possible species for Ti-(C,N), Ti-N and Ti=O bonds, respectively [38,40,44]. There are no Ti-C bonds detected in the coatings. Fig.2b shows a broaden peak overlapped by four peaks in C 1s spectra at 282.9 eV for C-(Ti,N), 284.5 eV for sp2C-C/C-N, 285.4 eV for sp3C-C/C-N and C-O bonds for 286.7 eV. As for N 1s spectra (Fig.2c), a main asymmetric peak with two shoulder peaks is obtained indicating N-(C, Ti) at 396.5 eV, TiN at 397.3 eV, sp2C-N at 398.4 eV and sp3C-N at 399.9 eV. Hence, it is inferred that TiCN and TiN phase in the coatings, while TiC phase is hard to form at plasma temperature conditions (lower sputtering power) due to it’s a higher formation energy as compared with that for Ti(C,N) and TiN .
Small C-O signal results from the surface contamination. With an increase of target sputtering power, Ti-C bond appears and its intensity increases, which reveals the formation of TiC phase (Fig.2d-i). Thus, Ti-C-N coatings are mainly composed of Ti(C,N) and TiN phases, agreed with XRD results in Fig.1b. Based on XPS results above, Ti ions are easy to react with N to form TiN, while C atoms exit in the TiN crystal structure as interstitial atoms to form Ti(C,N). The small amount of sp2 C would react with N to form CNx phase. The increased sputtering power produce more Ti ions to promote the formation of TiN, which would reduce the amount of sp2 C amorphous and CNx phases in the coatings. Therefore, Ti-C-N coatings mainly contain Ti(C,N) and TiN, as well as a small amount of CNx phases. As is reported that Ti-C-N coatings are easy to be reactively deposited to form Ti(C,N) and TiN , while the amorphous CNx phases would be generated at a higher C content conditions .
3.2 Surface morphologies and wettability
As is reported, surface integrity including smooth surface, highly dense microstructure, free of defects, low residual stress and high uniformity is vital for high-performance coatings [4,45]. Surface morphologies of Ti-C-N coatings deposited by HiPIMS at various target power and negative substrate bias were investigated using AFM. As shown in Fig.3a, rough surface morphologies with large package of grains and distinct gap between grain boundaries are observed for the coatings deposited at 4 kW and -60 V. Smooth surface morphology with smaller grains is formed at 5 kW and -60 V due to the increased number of energetic ions and ion bombardment leading to the migration of adatoms to promote films growth (Fig.3b). At 8 kW and -60 V, the increased ion bombardment results in the decrease of grain size, but also rough surface with craters is obtained on coating surface (Fig.3c). Meanwhile, residual stress in the coatings sharply increases from -1.3 to -4.1 GPa. When HiPIMS sputtering power was fixed at 5 kW, the evolution of surface morphology changes from dune shape to foothills feature with increasing negative substrate bias from -60 to -200 V, as shown in (Fig.3d-f). Thus, strong ion bombardment results in a rougher surface with craters and foothills. It is inferred that the appropriate energetic ions flux and ion bombardment closely related with target power and substrate bias promoted the high surface integrity and low residual stress by controlling the migration of surface adatoms. The combined improvement is attributed to the precise control of micro pulse of HiPIMS and reactive deposition conditions of the coatings [42,46].
Tribological hard coatings have been wildly used in the field of marine engineering as protective coatings [47,48]. Hence, the surface and interface features have a key influence on the performance due to surface hydrophily and hydrophobicity . The surface wettability reveals unique properties of gas-liquid-solid interfaces, which strongly depends on surface nanostructures and roughness attributed to the varied surface free energy [50,51]. Surface wettability is thus of interest in the field of tribocorrosion behaviors working on water solution environment. The wettability can be characterized by surface free energy using the measurement of contact angles. Thus, Fig.3g-m illustrates contact angles of droplets of distilled water on the surface of Ti-C-N coatings deposited using HiPIMS at various target power and negative substrate bias. Fig.3g presents the schematic diagram of contact angles. As shown in Fig.3h-j, the increase of target power results in the decrease of contact angle from 64.6o to 42.4o, followed by increase to 55.2o. The increased negative substrate bias results in the increased contact angle from 56.7o to 81.5o (Fig.3k-m). The evolution of surface morphology of Ti-C-N coatings is responsible for changes of contact angle. It is clearly seen from surface morphology that the variation of contact angle is induced by surface roughness. The Ti-C-N coatings deposited at 5 kW and -60 V with smooth surface morphology exhibits good wettability with low contact angle of 42.4o.
3.3 Mechanical properties and adhesion
Fig.4 exhibits hardness (H), effective Young’s modulus (E*), H/E* and H3/E*2 ratios of Ti-C-N coatings deposited using HiPIMS at various target power and negative substrate bias. As shown in Fig.4a-d, at 4 kW and -60 V, the values of H, E*, H/E* and H3/E*2 are 30.5 GPa, 332.4 GPa, 0.092 and 0.257, respectively. Whereafter, H and E* gradually increase to 42.8 GPa and 437.4 GPa, respectively, with increasing target power to 8 kW. However, H/E*, and H3/E*2 reach the highest value of 0.104 and 0.433 at 5 kW, respectively. Hence, when target power was fixed at 5 kW, substrate bias was increased from -60 to -200 V, H and E* gradually increase to 44.2 GPa and 476.3 GPa, respectively. However, H/E* and H3/E*2 show a decrease to 0.093, and 0.381, respectively. In addition, the micro-indentation inserted in Fig.4b,d under the load of 0.1 N is well declared that the coatings possess excellent fracture toughness. It is clearly seen that no radial cracks of Vickers indentations were observed with increasing HiPIMS power and negative substrate bias. At 5 kW and -60 V, the coatings have smooth and smaller indentations, indicating a higher fracture tougness[12-14].
Although hardness and toughness are always considered as the contradiction because they are hard to achieve simultaneously, heterogeneous interfaces are gradually introduced into nanomaterials to enhance them both as interfacial engineering rapidly develops. Hence, hard yet tough and flexible hard multilayer or nanocomposite coatings formed by nanocrystalline, amorphous materials and their interfaces exhibit excellent friction and wear properties [5,13,40]. Moreover, superhard yet tough coatings are always expected by the precise control of chemical composition and microstructure to achieve the combined improvement of hardness and toughness 12. Thus, it is demonstrated that Ti-C-N coatings deposited at 5 kW and -60 V exhibit superhard yet tough feature. Moreover, the related mechanism of superhard yet tough is attributed to the block of dislocation motion across interfaces at boundaries between dense and refined grains, as well as the contribution of the solution strengthening of Ti(C,N) with C interstitial atoms in TiN. In addition, the fully dense composite structure with refined Ti(C,N) and TiN grains promotes the enhanced cohesion strength.
The combined improvements in hardness, toughness and cohesion/adhesion are essential for hard coatings in wear and corrosion applications. Thus, Rockwell C indentations under a standard load of 150 kg were used to determine the cohesion/adhesion between the coatings and substrates by the large deformation of the coatings along with substrates. In addition, as for superhard yet tough coatings deposited on soft substrates (316L stainless steels), the enhanced hardness and toughness also account for the improved adhesion due to their close correlations with H, H/E* and H3/E*2, [14,43].
Fig.4e-j reveal Rockwell C indentations for the coatings Ti-C-N coatings deposited using HiPIMS at various target power and negative substrate bias. With an increase of target power from 4 to 8 kW, the adhesion shows an initial increase from HF4 to HF1, followed by a decrease to HF3. While the adhesion level presents a gradually decrease from HF1 to HF3 with increasing substrate bias from -60 to -200 V. Besides, residual compressive stress shows an increase from -1.3 to -4.1 GPa with increasing target power and from -2.4 to -3.8 GPa with increasing substrate bias, respectively. It is noted that the coatings with high H, H/E* and H3/E*2 exhibits the reduction of cracking at the edge and inside of indentations corresponding to HF adhesion level and cohesion, respectively, due to the enhanced cracking resistance. The similar results are also reported that CrN/TiN superlattice coatings and TiAlSiN nanocomposite coatings deposited by deep oscillation magnetron sputtering, known as an alterative HiPIMS technique. The improvement in cohesion/adhesion of coatings is ascribed to the enhanced H, H/E* and H3/E*2 due to smooth surface morphology, highly dense microstructure and low residual stress.
3.4 Tribocorrosion resistance
Fig.5 presents tribocorrosion behaviors in 3.5 wt.% NaCl aqueous solution of Ti-C-N coatings deposited using HiPIMS at various target power and negative substrate bias, compared with those for hard coatings. Tribocorrosion resistance of the coatings are evaluated by instantaneous coefficient of friction (COF), in-situ response of open circuit potential (OCP) and worn surface morphologies. As shown in Fig.5a, fluctuant COF and OCP at 4 kW and -60 V are observed during tribocorrosion tests, and the average values are 0.21 and -0.036 V, respectively. That is because the coatings suffer from severe abrasive wear leading to distinct long grooves parallel to the friction direction resulted from hard wear debris polishing the sliding contact surface. Besides, low surface wettability is not useful for the decrease of wear and friction. Severe plastic deformation on worn surface is attributed to low hardness of the coatings. Noted that no macro cracks are observed on worn surface during sliding wear. Hence, the diffusion distance of Cl-1 in tribocorrosion is limited between sliding contact surface and sublayer. Thus, the severe polishing and plastic deformation result in micro cracks in sublayer of the coatings owing to dense microstructure and high cohesion/adhesion. The calculated specific wear rate is 3.52×10-6mm-3 N-1m-1.
When the target power is increased to 5 kW, a sharp decrease to 0.04 of COF and a slight decrease to -0.055 V of OCP are obtained, as shown in Fig.5b. At the beginning of 10 min, OCP is higher -0.05 V due to the formation of passive films. And then, the OCP gradually decrease and achieve the stable value of about -0.055 V during sliding wear, revealing durable passive films formed on contact surface. The worn surface presents smooth and clean morphologies along with shallow grooves, indicating a mild abrasive wear. The calculated specific wear rate is 1.3×10-7mm-3 N-1m-1.
Superhard yet tough Ti-C-N coatings exhibit high resistance of elastic and plastic deformation to reduce the formation of micro cracks and polishing. In addition, superhard yet tough coatings with the high cohesion/adhesion possess high cooperative deformability to reduce cracking and adhesion failure, which build a strong barrier for Cl ions diffusion to accelerate tribocorrosion failure, due to high surface integrity. Moreover, the high surface wettability promotes the uniform water films and passive films formed at sliding contact surface to reduce the friction and wear. As we reported in our previous works, the enhanced surface integrity and interface structure contribute for the combined improvement of hardness and toughness, leading to high wear, corrosion and tribocorrosion resistance [4,14,26]. At 8 kW, although COF is still about 0.05, it appears a quite unstable tendency. In addition, oscillating changes in corrosion potential is detected, and the average value is about -0.23 V (Fig.5c). The calculated specific wear rate is 3.34×10-6mm-3 N-1m-1. With an increase in substrate bias from -60 to -200 V (Fig.5d-f), the COF and OCP exhibits a sharp fluctuation and increases to 0.34 and -0.05 V due to severe abrasive wear resulting in cracking and adhesive failure during tribocorrosion tests. The calculated specific wear rate increases from 0.29 to 6.87×10-6mm-3 N-1m-1. Meanwhile, the deterioration of Ti-C-N coatings would be accelerated owing to high residual stress and surface defects caused by ion bombardment with increasing substrate bias. Fig.5i summarized that hardness against coefficient of friction in seawater for typical metal and ceramic coatings. It is clearly seen that COF of ceramic coatings presents an increase with increasing hardness. For traditional lubricating materials, DLC based coatings exhibit good tribocorrosion resistance with low coefficient of friction of 0.036 due to the combination of hardness and toughness [52,53]. However, in this work, superhard yet tough Ti-C-N coatings possess excellent seawater lubrication with ultra-low COF of 0.03.