Abstract
The enhancement of wear resistance by laser carburisation on TiZrN coatings was investigated in terms of bond state. Graphite paste was used to cover the TiZrN coatings and then pulsed laser ablation was used to solidify the carbon. Tribometer test showed that the friction coefficient was reduced from 0.64 to 0.17 after laser carburisation. The doped carbon was analysed as a mixture with sp2C and sp3C bonds using an X-ray photoelectron spectroscopy depth profile. The sp2/sp3 ratio was 2.29 at the surface and increased up to 2.91 in the depth direction. To verify the effect of the hybrid bonds on the atomic order, annealing was carried out to the carburised specimens. As the ratio increased to 3.28, graphitisation resulting in transition from sp3 to sp2 was confirmed, and variation in the lattice structure was demonstrated by a reduction in the lattice constant (4.45 Å to 4.40 Å) using Rietveld refinement.
Introduction
Titanium nitride (TiN) coating has been used in a variety of applications such as cutting tools and mechanical parts due to its good adhesion to metal substrates and mechanical properties [1-3]. Furthermore, Ti-based ternary nitride [Ti1−xMexN, (Me=Zr, Al, Cr, etc.)] coatings have been further developed and used, which complement the low oxidation resistance of TiN and exhibit enhanced mechanical properties [48]. Among these nitride coatings, TiZrN has the best thermal stability, so it is an appropriate candidate as a material for protective coating of parts in high-temperature environments [9,10].
However, Ti-based ternary nitride coatings have shown low wear resistance, which has drawn attention to amorphous carbon coatings with low friction coefficients and chemical inertness [11-13]. The excellent properties of amorphous carbon coatings appear in the form of a mixture of hybrid bonds (e.g. sp1, sp2 and sp3) [14]. Furthermore, the disorderly intermixed hybrid bonds play an important role in interfering with dislocation movement and increasing hardness by forming compressive stress inside the coatings [15]. sp2 and sp3, which are the main components of the hybrid bonds, are related to the tribological property and hardness, respectively, and so the chemical structure of the coating layer varies depending on the ratio of these bonds (sp2/sp3), which changes its mechanical properties [1618].
Various studies have been carried out on multi-layer design, modification of sputtering conditions, and doping method design in an attempt to apply the mechanical properties of amorphous carbon to hard coatings [19-24]. However, these studies have mainly focused on structural changes in the coating layer and performance enhancement. Thus, it is necessary to study the relationship between the chemical bonding structure of coexisting carbon in the coating layer and the enhancement of mechanical properties. In this study, we observed the variation in friction coefficient of TiZrN coatings after laser carburisation and investigated it in terms of the hybrid bonds of the carbon.
TiZrN specimens covered with graphite paste were carburised using pulsed laser ablation. After laser carburisation, variation in the friction coefficient was measured using ball-on-disc testing and changes in bond state of doped carbon were probed through X-ray photoelectron spectroscopy (XPS) depth profile analysis. In addition, the bonding environment in which thermal annealing had changed the degree of ordering of the carbon was analysed using XPS. The variation in lattice structure was observed through Rietveld refinement and transmission electron microscope (TEM) selected area electron diffraction (SAED) patterns.
Experimental
TiZrN coatings were deposited using an RF/DC magnetron sputtering system on stainless steel 304 as a substrate. Before deposition, ultrasonic wave cleaning was carried out in ethanol to remove contamination on the surface of the substrate. Coatings were deposited with a target with the composition of TiZr 50:50 (at.-%). Base and working pressure were respectively 1.33 × 10−3 Pa and 1.33 × 10−1 Pa. Deposition was carried out at 723 K for 6 h. The paste for laser carburisation was prepared by mixing graphite powder (diameter with <20 µm) and polyvinylidene fluoride (PVDF, binder) in a ratio of 90:10 (wt-%) and adding N-Methyl-2-pyrrolidone (NMP, solvent) to adjust viscosity. The graphite paste was covered on as-deposited TiZrN coating specimens to a thickness of about 150 µm using screen printing method, for a laser carburisation. The specimens covered the paste were dried at 80°C to improve adhesion. Pulsed laser irradiation was performed using a Nd:YAG pulsed laser ablation system (LSX-213, CERAC Technologies) with a spot diameter of 200 µm, energy output of 50%, frequency of 10 Hz and 10 shots. The same laser irradiation with 250 µm was repeated on the next area, after 10 shots with 10 Hz frequency on the first area.
After laser carburisation, the graphite paste remaining on the surface was removed by ultrasonic cleaning using ethanol and acetone. A ball-on-disc testing (Tribometer, J&L Tech) was employed to evaluate variation in wear resistance after laser carburisation. The testing was carried out in air under normal load 1N with a steel ball (6 mm diameter) as a counterpart material. The chemical characterisation was investigated using XPS (Theta Probe XPS system, Thermo Fisher Scientific) in KBSi Busan Center. Argon ion (Ar+) etching was performed to remove the surface contamination and make XPS depth profile analysis. The C 1s peaks were calibrated with adventitious or surface carbon located at 284.6 eV. The XPS peaks were decomposed by OriginPro 2015 (OriginLab Corporation) after background subtraction according to the Shirley method.
After removing the remained graphite paste, the carbon-doped specimens were annealed in air at 200°C, 300°C and 500°C for 30 min, in order to promote the crystallographic change. The structural characterisation was performed by using X-ray diffraction (XRD, Ultima IV, Rigaku) with Cu Kα and a field emission TEM (JEM-2100F, JEOL Ltd.) in KBSi Busan Center. Rietveld refinement of the diffraction patterns was undertaken using an X'pert high score (Panalytical).
Results and discussion
The results of the ball-on-disc testing are shown in Figure 1 to identify the effect of doped carbon on wear resistance. The friction coefficient was considerably reduced from 0.64 to 0.17 after laser carburisation. An investigation of the bonding configuration of carbon in the coating layer was required as the doped carbon played an important role in reducing the friction coefficient.
Variation of friction coefficient after laser carburisation.
Figure 2 shows XPS C 1s peaks, which reveal the bond state of carbon before and after laser carburisation. The peak at the lowest binding energy corresponds with that of carbide (TiC or ZrC) bonding, and represents that the portion of doped carbon substituted for nitrogen and involved in the bonding with transition metals (Ti or Zr) [25,26]. Peaks of hybrid bonds of carbon-carbon (sp2 C=C and sp3 C–C) were highly evident, and CN bonds (sp2 C=N and sp3 C–N) were also revealed [27-30]. This indicates that some carbon atoms substituted for nitrogen, forming carbide, and others occupied interstitial sites in the lattice, distributed in the form of hybrid bonds. Peaks of the CN bonds can be interpreted as the bonds between substituted nitrogen atoms and interstitial carbon or dangling bonds between nitrogen remaining on the surface and hybridised carbon. The sp2/sp3 ratio [(sp2 C=C + sp2 C=N)/(sp3 C–C + sp3 C–N)] was 2.32, indicating that sp2 composition was dominant in the mixture. It is speculated that sp2, which is thermodynamically stable and has a shorter bonding length (1.46 Å), found it easier to form bonds in the pre-existing TiZrN lattice. The peak of CF2 (F–C–F) bonding was due to the fluorine contained in the graphite paste, and the oxidation was accelerated during the laser carburisation process, resulting in an increase in the CO (C=O) bonding peak intensity [30,31].
XPS C 1s peaks before and after laser carburisation.
Figure 3 shows the XPS depth profile comparing the bonding structure of the doped carbon depending on depth, and the relative areas of the peaks are reported in Table 1. These results indicate that there was a carbonaceous phase not only on the surface but also at a certain depth and that hybrid bonds were present inside. The carbide peak increased in the depth direction and CN bonds decreased, which means that the carbon atoms were more likely to substitute for nitrogen atoms, unlike on the surface. The sp2/sp3 ratio gradually increased to 2.59 owing to the decrease of the CN bonds, and the sp2 predominance tended to intensify. The sp3 bonding of the CN bonds particularly decreased since the length of sp3 bonds (1.54 Å) is relatively longer than that of sp2 (1.46 Å). The intensified predominance of sp2 may represent differences in the local structure, such as the ordering of the sp2 phase [32]. The profile depth was not enough to confirm the difference in the local structure, thus supplementary information was required.
Comparison of XPS C 1s peaks by depth. Relative area of XPS C 1s peaks by depth.
As described earlier, thermal annealing was performed to change lattice ordering and variation in the bonding configuration was compared, as shown in Figure 4. Moreover, the relative area and full width at half maximum (FWHM) of each peak are exhibited in Table 2. The peak shift to lower binding energy of about 0.3 ± 0.1 eV suggests a change in the charge relaxation of excited carbon atoms [33]. The sp2/sp3 ratio increased to 2.47, 3.16, and 3.28 at 200°C, 300°C and 500°C, respectively, and the FWHM of sp2 C=C gradually decreased. The results provide unique evidence to describe thermally induced graphitisation or transition from sp3 to sp2 in the coating [34]. Thus, this implies that the lattice rearrangement in the coating layer can be attributed to an increase in degree of ordering by the graphitisation.
Comparison of XPS C 1s peaks by annealing temperature. Relative area and FWHM of XPS C 1s peaks by annealing temperature.
Figure 5 shows the XRD patterns of as-deposited, carbon-doped, and annealed specimens in terms of crystallography. TiN and ZrN phases were identified in all three cases, and lattice constants were calculated using Rietveld refinement with the XRD data, as reported in Table 3. The increase in lattice constants after laser carburisation represents lattice expansion caused by both interstitial and substitutional carbon atoms. Furthermore, lattice deformation was confirmed by the diffuse ring pattern observed in TEM SAED patterns (Figure 5).
XRD patterns of as-deposited, carbon-doped, and annealed specimens with SAED patterns. Lattice constants of as-deposited, carbon-doped, and annealed specimens.
The decrease in lattice constants after annealing was regarded as the influence of graphitisation. As the sp2/sp3 ratio was increased by graphitisation, the lattice constants decreased due to the shorter bonding length and an increase in the degree of ordering. The thermally induced graphitisation was considered to have affected the lattice rearrangement of the coating layer.
Conclusion
Enhancement of wear resistance of carbon-doped TiZrN coatings was investigated by analysing the bond state of doped carbon on the surface and inside. TiZrN coatings covered the graphite paste were carburised using pulsed laser ablation, after which a reduction in friction coefficient was observed. The doped carbon not only partially substituted with nitrogen, thereby forming carbides with transition metal, but also occupied interstitial sites in the lattice. The interstitial carbon existed in the form of a mixture of sp2 C and sp3 C bonds, and the sp2/sp3 ratio was 2.32 at the surface, and gradually increased to 2.59 inside. Heat treatment was performed to identify the effect of hybrid bonds on the lattice arrangement, and the thermally induced graphitisation was verified by an increase in the ratio up to 3.28 and a decrease in the FWHM of sp2 C=C bonding down to 1.31. It was confirmed that the increased degree of ordering by graphitisation led a reduced lattice constant (from 4.45 Å to 4.40 Å).
Footnotes
Acknowledgements
We are especially appreciative of Ms Sunyoung Park for her language editing of the manuscript.
Disclosure statement
No potential conflict of interest was reported by the authors.