Deposition, microstructure and nanoindentation of multilayer Zr nitride and carbonitride nanostructured coatings

Nitrides, carbides, and carbonitrides of transition metal elements like Zr, W, Ti, etc. are generally employed to produce hard coatings. Zirconium-based hard coatings have shown useful applications in the areas of tribology, biomedicine and electrical due to their high thermal stability, hardness, biocompatibility, good erosion, wear, and corrosion resistance. In this study, we created homogeneous and tenacious nanostructured hard coatings based on Zr with good mechanical properties. The magnetron sputter deposition technique was utilized to coat stainless steel 316L substrates with multilayers of Zr/ZrN and ZrN/ZrCN with individual layer thicknesses of 250 and 500 nm for each coating composition. The deposition conditions were adjusted to create two different coating thicknesses of 2 and 3 µm. The thickness of the coating was confirmed using Calotest and the coatings’ morphology and elemental composition were determined utilizing the atomic force microscope and scanning electron microscope equipped with energy dispersive x-ray spectrometer. Coating thickness and adhesion were measured using cross-sectional samples and XRD was utilized to analyze the coatings structure. Nanoindenter was employed to determine the instrumental nanoindentation hardness and elastic modulus. The influence of coating thickness on tribological behavior was further investigated using the ratio of nanohardness-to-elastic modulus (H/E). No evidence of decohesion was observed at the substrate/coatings interface, and the grains of all the coatings were observed to show columnar growth which were homogeneous, compact and dense. The grains of the ZrN/ZrCN coatings were observed to be denser, finer and more compact compared to those of the Zr/ZrN coatings. Correspondingly, higher hardness, modulus and H/E values were exhibited by ZrN/ZrCN than Zr/ZrN coatings. This suggests that the ZrN/ZrCN coatings are capable of exhibiting better wear resistance and fracture toughness. The coatings developed in this investigation are anticipated to be suitable for applications in tribology due to their excellent hardness and H/E properties.

www.nature.com/scientificreports/ have also been reported [41][42][43][44][45][46] . The number of studies on ZrCN coatings in the literature are limited. Among the properties and applications explored for ZrCN are morphology, crystallinity & composition 47 , oxidation and corrosion resistance 48,49 , tribological 50 , synthetic parameters 51 , , biomedical [52][53][54][55][56][57][58][59] , characterization 60 and optical 61 . The hardness of a given coating is affected by the method of synthesis technique, processing, elemental, coating structure and elemental composition. Because of the influence of the underlying substrate, measuring the hardness of thin coatings is difficult. Nanoindentation is a technique for evaluating thin coatings with small indenters under low stress and penetration depths of tens to hundreds of nanometers. Without perceiving or quantifying the size of the indent, the mechanical properties of the coatings can be determined. The values of elastic modulus and nanoindentation hardness are extracted using analytical models. The microstructures, properties and characterization of single-layer Zr, ZrN and ZrCN deposited by magnetron sputtering have been reported, and ZrN was said to exhibit the highest elastic modulus and hardness while Zr exhibited the least 61 . Multilayer coatings, especially the ones composed of sub-layers of thicknesses in the nanometer range have been reported to show improved properties over the single layer ones [62][63][64][65][66][67][68][69] . The properties and microstructure of a multilayer Zr/CrN coatings grown by the multi-arc ion plating method have been reported 70 . Braic et al. 71 , developed a super lattice TiAlN/TiAlZrN multilayer coatings of different bilayer periods deposited by cathodic arc method and investigated their performances as wear resistant coatings. The tribological, mechanical, microstructural and microchemical properties of TiN/ZrN multilayer coatings with different bilayer periods and thicknesses have been reported 72 . Recently, many laser-assisted manufacturing techniques for producing multilayer coatings have been used with greater flexibility of control 73,74 . Multilayer nitride coatings of TiN/TiAlN and Ti/TiN/TiAlN have been investigated for biomedical applications 75 . Because of its cost effectiveness and reproducibility, the magnetron sputter deposition technology is widely used in industry 76 . It generates adherent and uniform coatings and uses low deposition temperature, which allows for the creation of coatings with improved properties.
Using the magnetron sputter deposition process, multilayer coatings based on Zr/ZrN and ZrN/ZrCN each of layer thicknesses 250 and 500 nm and coatings thicknesses of 2 and 3 µm were synthesized in this study. SEM equipped with energy dispersive x-ray spectroscopy, atomic force microscopy, and x-ray diffraction were employed to investigate the surface morphology and structure of the coatings. The nanoindentation method was used to measure the mechanical properties; modulus and nanoindentation hardness. The indentation hardness was determined as a function of normal load using the load displacement curve. The effects of the coating thickness and adding of carbon on the coatings properties were also investigated.

Materials and methods
Coating synthesis. Magnetron sputter deposition was used to develop multilayer Zr/ZrN and ZrN/ZrCN coatings (total thickness 2 and 3 µm) on (100) Si wafers and stainless steel SS316L substrates (of 16 mm diameter disc, thickness 3 mm). The coatings' cross-sectional fracture morphology was next studied from broken coated Si wafers. The stainless steel substrates were polished with emery paper (ranging from 240 to 2400 grit) before being mirror-polished with a diamond suspension prior to deposition. The samples were cleaned ultrasonically using distilled water, ethanol, and acetone for ten minutes. Following that, in-situ sputter etching was carried out for 900 s at Ar pressure of 1.32 Pa and a steady current of 480 mA. As described by Silva et al. 77 , unbalanced dual magnetron sputtering was used to create the coatings. Two highly pure (99.2%) 200 × 100 mm 2 Zr targets (rectangular in shape) mounted on unbalanced type-2 magnetrons were used to deposit the coatings in a reactive mode in an atmosphere of Ar/N 2 /C 2 H 2 . For compositional and morphological characterization, the films were deposited onto (100) silicon wafers, as well as onto 316L stainless steel discs with a diameter of 3 mm. As previously described 78 , the substrates were mounted in a matrix-like formation on a rotatable substrate holder. The parameters for creating multilayer Zr/ZrN and ZrN/ZrCN were chosen on the basis of each coating's deposition rates, adhesion properties, and composition. During the depositions, the bias voltage, chamber temperature and argon flux were respectively kept constant at −50 V, 100 °C and 60 sccm, while the current density of Zr was set between 1.8 and 2.0 A/cm 2 . Table 1 shows the values for the target's potential, working & base pressures with some other deposition parameters used to deposit the final coatings.

Materials characterization.
The surface morphology of coatings was investigated by undertaking high resolution secondary electron imaging of the coatings in their as-received state using field emission scanning electron microscope (FESEM, model TESCAN LYRA 3) equipped with an energy dispersive x-ray spectrometer from Oxford Inc. at an accelerating voltage of 20 kV. The backscattered electron images (BEI) and EDS spectra were collected using JEOL JSM-6610LV scanning electron microscope equipped with the energy dispersive x-ray spectrometer at an accelerating voltage of 10 kV. To avoid damaging the coating, the coated samples were cut with a slow-cutting diamond wheel and epoxy was used to mount them in cross-section. Later, the surface was ground with 600 grit silica carbide paper and polished using 1 µm diamond paste. In order to reduce charge build-up and improve the samples' surface conductivity during the SEM examination, a gold coater was utilized to sputter a thin layer of gold on the samples' surface in an argon atmosphere. Agilent 4500 series atomic force microscope (AFM) operated in a contact mode was also utilized to analyze the coatings' topography. The Rigaku Ultima IV x-ray diffractometer equipped with Ni-filtered Cu Kα radiation (λ = 0.154 nm) operated at 40 kV and 40 mA was utilized to assess the coatings phase composition from a range of 30 to 80° 2θ diffraction angle. The Scherrer equation was used to measure the average crystallite size.
where k is a constant (0.89), λ is x-ray wavelength (0.154 nm), β is peak broadening at full width half maximum (FWHM) in radians, and θ is the Bragg's diffraction angle. www.nature.com/scientificreports/ Nanoindentation. The tests were carried out using CSM nanoindentation tester having a load range 0.5 to 300 mN. In order to ensure for limited volume penetration, a sharp three-faced pyramid Berkovich diamond tip was used. The tip of the nanoindentation tester was driven into the material at a constant load and speed for each nanoindentation. At an unload-load speed of 20 mN/min, indentations were done at loads of 10 and 20 mN. The loading and unloading cycles were separated by 60 s. The curve of load-displacement for each nanoindentation cycle was constructed by continuously monitoring the applied normal load and the displacement of the indenter tip. Hardness was evaluated using the indent's area, which was obtained using defined geometry and tip size. For determining elastic modulus and hardness, various analytical models have been developed, including the Joslin-Oliver technique 79,80 , deformation energy technique 81,82 , the Oliver and Pharr technique 83,84 , the energy density technique 85 , and the force indentation function technique [86][87][88] . The Oliver and Pharr technique, used in this study, is the most popular of all the models. The resistance to permanent deformation under a perpendicular stress is known as the nanoindentation hardness (H), and it is calculated as follows: where P max is the maximum perpendicular load, A p is the projected contact area.
The elastic modulus (E) was computed from: where E i is the tip modulus, E τ is the indentation contact's reduced modulus, υ i is the tip Poisson's ratio and υ s is the test sample Poisson's ratio.

Results and discussion
Coatings synthesis. To vary the amount of N & C-atoms in the coatings and identify the rate of deposition, the optimum composition, and evaluate the cross section morphology, appropriate parameters for creating Zr, ZrN, and ZrCN coatings were selected using thick multilayer films with various N 2 and C 2 H 2 flows (Table 2). To determine the optimal synthesis parameters for achieving the best coating composition, thickness, and adhesion,  www.nature.com/scientificreports/ the depositions were conducted in three series. The settings for depositing the 2nd and 3rd series of samples were chosen on the basis of the 1st series' optimization, with minor changes due to the usage of a new zirconium target, an excess of zirconium was detected. As a result, the zirconium current density was reduced and, in certain cases, gas fluxes were raised to obtain a thickness and composition identical to the first sample series. In order to alter the deposition conditions, Fig. 1 depicts the composition and thickness of two films deposited previously to the multilayer system. The target potential increased as gas fluxes increased; this increase is explained by the target surface's high reactivity with N 2 and C 2 H 2 gases thereby poisoning surface of the target during reactive magnetron sputtering. The target potential of multilayer coatings varied, based on how much of gas is supplied to the reactor (Table 1). To obtain the requisite wear, the monolayers' thickness was measured utilizing the Calotest method, which involved spinning a 20 mm diameter sphere at 900 rpm for 120 s. Table 3 shows the results, whereas Fig. 1 shows a typical image for each composition. This test also provides a quantitative way to assess the films' adherence to the substrates. As, a result, coatings that failed the test were removed, and deposition conditions were optimized to increase film adherence. Figure 2 illustrates an instance of failed film during the test. To increase adhesion, a zirconium interlayer was placed between the functional layer and the substrate for each coating composition, with thicknesses ranging from 250 to 300 nm. The multilayer coating thicknesses were optimized making use of the monolayer thickness and SEM images in order to determine the deposition conditions of each compound. The coatings' cross-sectional FESEM views mostly exhibited columnar structures with a few small grains at the coating/alloy interface (Fig. 3). Such structure is distinctive of sputter-deposited coatings, which are formed at high super-saturation conditions, resulting in high rate of nucleation with tiny nuclei at the base 89,90 . The primary columnar shape found in coatings explored here is a consequence of motion of grain boundary in the course of the epitaxial and coalescence developments. The Thornton model 91 , which proposes that faults from atomic shadowing & voided growths contribute to limited sample's surface diffusion, supports this type of growth. The columnar morphology for zirconium-based coatings have been documented extensively 14 . ZrN is a dazzling gold, whereas ZrCN is a dull gold and pure Zr coating is silver grey in color. This is consistent with previous findings of silver-colored metallic coatings containing high Zr and low nitrogen 92,93 . The creation of stoichiometric ZrN, which has a gold color, occurs when the nitrogen content in the coatings rises 44,92,93 . Additionally, Calotest was used to characterize the thickness of the multilayer where possible and the results are presented in Table 3. It must be stressed that Calotest allows the observation of the multilayer, as observed in Fig. 4, however the thickness of each individual layer was not calculated since the error of the technique may exceed the thickness of the individual layers. Figure 4 presents representative images  www.nature.com/scientificreports/   www.nature.com/scientificreports/ of the multilayer for 250 of thickness of each individual layer. When nitrogen is added to the process, the film deposition rate decreases, as expected due to Zr targets' poisoning; however, when acetylene is used instead, the deposition rate increases, this is explained by carbon's ability to deposit directly into the films. Multilayer cross-sectional SEM images are presented in Fig. 5 for two representative samples PN10 and PN14, corresponding to Zr/ZrN and ZrN/ZrCN multilayer, respectively. Figure 5a reveals a clear differentiation between the Zr and ZrN layer, with individual layer thicknesses close to 250 nm. This individual layer is not clearly noticed for the ZrN/ZrCN films, since structural and morphological changes for those materials are expected to be very subtle; nonetheless, the layer differentiation was detected by optical images after Calotest, as previously discussed. In contrast, Zr and ZrN may present different crystalline phases and the large compositional/morphology variation, as observed for surface SEM images and hence, each layer is clearly identified. Materials characterization. SEM/EDS. The morphology of the coatings' top surfaces was studied using FESEM. The FESEM secondary electron micrographs of the coatings are shown in Fig. 6. The morphology of the coatings was essentially homogeneous, compact and dense with fine spherical grains. This is consistent with the previously observed columnar grain growth (Fig. 1). Aside from usual grain growth, localized grain growth near the thickening film's surface was seen, which was aided by nearby grain coalescence. Figure 6 demonstrates that for all coating compositions, coatings with a 3 µm thickness had larger grain sizes than those with a 2 µm thickness. The larger grain growth in the 3 µm coatings is attributed to the prolonged time of deposition, which may have allowed adjacent atoms to reorganize via surface diffusion, resulting in bigger grains. Generally, the grains of the ZrN/ZrCN coatings were more compact and smaller compared to those observed in the Zr/ZrN coatings. It has been reported that the carbon and nitrogen atoms in coatings that contain these atoms tend to limit grain growth by reducing neighboring grains' mobility, resulting in fine grain size 78 , thus, resulting in the smaller grain size of the ZrN/ZrCN as compared to the Zr/ZrN coatings observed in this study. This is because carbon and nitrogen limit the mobility of adjacent grains and restrict them to the fine grains. Furthermore, densification of coatings was observed in the present study due to introduction of carbon and nitrogen. This leads to filling of interstitial positions due to diffusion of carbon and nitrogen which reduces the vacant cites, which in turn leads to more dense and smooth morphology of the coatings 93 .
The cross-sectional backscattered electron image (BEI) of all coatings are shown in Fig. 7. It's clear that the deposition parameters used were successful in achieving the intended number of layers and thickness for each coating. The individual layers for each coating are of equal thickness. The Zr interlayer between the metal substrate and the functional layers is clearly visible (the first bright color) layer in all the coatings. The thickness of the Zr interlayer was 250 nm to 300 nm. The reason to insert Zr interlayer between substrate and functional layer is to improve their adhesion. The Zr interlayer is marked in Fig. 7. The coated samples exhibit well-defined multilayer structure with good planarity of individual layers and distinctive interfaces. For the Zr/ZrN multilayer coating, the alternating bright and grey color regions represent the Zr and ZrN layers respectively (Fig. 7a-d). While for the ZrN/ZrCN multilayer coatings, the alternating grey and dark color regions represent the ZrN and ZrCN layers respectively (Fig. 7e-h). Figure 8 depicts representative cross-sectional EDS spectra for PN8 (Fig. 8a) & PN12 (Fig. 8b) and x-ray mapping for PN8 (Fig. 8c,d). EDS spectrum of PN8, whose result is representative of all the Zr/ZrN coatings, reveals a higher atomic mass % for Zr as compared to N with N to Zr ratio 0.835 (Fig. 8a). This is expected due to the presence of single Zr layers with the coatings. The error in the Zr and N was observed as 54.5 ± 1.85 and 45.5 ± 1.85, respectively. The ZrN/ZrCN coatings, represented by the result for PN12 coating, exhibited an almost stoichiometric N to Zr ratio of 0.989 with the carbon atomic % almost twice the atomic % for either Zr or N (Fig. 8b). The error in the C, Zr and N was noted as 45.9 ± 2.38, 27.2 ± 2.68, and 26.9 ± 1.99, respectively. www.nature.com/scientificreports/ The representative x-ray mapping also shows the alternating multilayer structure of the coatings, the individual Zr and ZrN coatings for coating PN8 are clearly depicted (Fig. 8c,d).  www.nature.com/scientificreports/ interstitial sites 79 . Figure 9 shows the x-ray diffraction spectra obtained from all the multilayer coatings. The spectra match that of an analogous coating described in a recent study 94,95 , with obvious (fcc) ZrC, Zr 2 CN and ZrN, peaks that are almost homogeneous. The standard peak positions of (fcc)-ZrC (1CCD: 00-035-0784), (fcc)-Zr 2 CN (ICSD: 01-071-6065) and (fcc)-ZrN (ICCD: 00-035-0753) are plotted as dashed lines. All the coatings peaks were in agreement with the standard ZrN and Zr 2 CN. All the coatings exhibited preference for (111) plane. The (220) planes also exhibited a high peak. The (111) plane have been reported as the preferred orientations for ZrN and Zr 2 CN coatings 17,21,38,78 . Other planes that showed low-intensity peaks were planes (200) and (311). The ZrN/ZrCN coatings had smaller grain size than the Zr/ZrN coatings, this corresponds to the surface morphology of the coatings as observed in the FESEM micrographs (Fig. 6). This is because the carbon atoms in coatings limit grain growth, resulting in tiny grain size. Also in general, the coatings of 3 µm thickness had larger grain size than the ones in 2 µm thickness, this is a consequence of the prolonged deposition times for the coatings of 3 µm thickness as against the 2 µm ones (Table 3).
AFM analysis. The 2D AFM images of all the coatings are shown in Fig. 10. Just as it was observed in the SEM, all the coatings showed compact, uniform and dense morphology of fine spherical grains with the 3 µm coatings appearing more compact, coarse and having grains of larger size than the 2 µm coatings. Also, the 3 µm coatings have larger average roughness (Ra) value than the 2 µm coatings. The larger average roughness value of the 3 µm coatings as against the 2 µm coatings is a consequence of increased clustering and packing which leads to increase vertical size of grains 96 . Nanoindentation. All nanoindentation tests of all coatings were performed at two different loads (10 & 20 mN). After which, the load (change in applied normal force) vs displacement (depth of penetration) of the indenter over time was plotted. For the same indent, the normal force was plotted vs penetration depth. For all coating compositions, depth of penetration increases with load, ranging from 145 nm (for ZrN/ZrCN) to 435 nm (for Zr/ZrN). The indenter penetration depth for ZrN/ZrCN coatings was within the ideal range (10% of coating thickness). Whereas the Zr/ZrN coatings have penetration depths ranging from 11 to 20% of the coating thickness. This is owing to the fact that Zr/ZrN coatings have lower hardness. The Oliver & Pharr (O & P) method was used to calculate the elastic modulus and nanoindentation hardness from the plots 62,63 . Figure 11 shows bar charts of nanohardness and elastic modulus plotted for each coating mixture at loads of 10 and 20 mN.
Hardness is controlled by many factors such as preferred orientation, residual stress, grain size, defect density, stoichiometry and structure of the coatings which is turn depend largely on film growth conditions and process   (Fig. 6). It has been reported the denser and more compact a coating the harder it is 1,98,99 . As explained above the introduction of carbon leads to dense, firm, smooth and small grain size morphology. High hardness values are achieved for highly dense and nanocrystalline coatings. This is because a highly dense nanostructured coating inhibits sliding or rotation of crystallites due to the lack of free space, thus giving rise to high hardness values 97 . Also for both multilayer coating types, the hardness decreases with decrease in bilayer thickness (Fig. 11a). The influence of the substrate on hardness measurements is reduced by keeping the indent displacement to 10% of the coating thickness. In contrast, a small indent could be comparable in size to the sample surface roughness, reducing measurement accuracy. As a result, there will be a discrepancy in the measurement results. To avoid this, the penetration depth must be at least 20 times the roughness of the surface. The majority of the time, a compromise between these two competing conditions is reached. The change from an elastic to an elastic/plastic state in the indented region makes measurement difficult at low stresses 62 . Soft films on hard surfaces typically increase in hardness as the indenter penetration depth increases, whereas hard films on soft substrates frequently decrease in hardness as displacement increases. For all the coatings, the hardness value was observed to decrease with increase in load from 10 to 20 mN. Stainless steel substrate influence on coating hardness at both load 10 and 20 mN has been  www.nature.com/scientificreports/ discussed in a previous study 78 . It was reported there that the hardness of the substrate was measured as 2.10 GPa. The substrate was reported to have little effect on the coatings hardness at 10 mN while at the 20 mN load, the substrate showed a more significant effect on the coating hardness, it reduced the coating hardness at this www.nature.com/scientificreports/ load. Also the effect of substrate on coating hardness was also reported to reduce with coating thickness 78 . All these suggest that the underlying stainless steel substrate possessed a much lesser hardness than the coatings. Similar trends just as in the case of hardness was observed for elastic modulus. The Zr/ZrN multilayer coatings generally have lower values of elastic modulus as compared to the ZrN/ZrCN coatings where sample PN12 (3 µm, ZrN/ZrCN multilayer coating having individual layer each 500 nm in thickness) exhibited the highest elastic modulus of 335.3 and 325.2 GPa at load 10 and 20 mN respectively. Whereas, the 2 µm Zr/ZrN multilayer coating with individual layer thickness, 250 nm (sample PN9) showed the lowest elastic modulus of 190. 8 and 184.0 GPa at load 10 and 20 mN respectively (Fig. 11b).
It is well understood that hardness is not the only criterion for predicting coating wear, the ratio of hardness to modulus (H/E) which is related to elastic strain to failure and fracture toughness has been proposed as one of the key parameters controlling wear [100][101][102] . In fact, the H/E ratio has been shown by some authors to be a more appropriate parameter for predicting wear resistance compared to hardness, which can be regarded as a vital material property that defines wear resistance 101,102 . Similar to hardness and modulus, the H/E ratio of ZrN/ZrCN multilayer coatings was higher than that of Zr/ZrN coatings, with PN12 coatings showing the highest H/E value and PN9 coatings having the lowest H/E value (Fig. 11c). A high value of H/E suggests a higher fracture toughness. This implies that the ZrN/ZrCN multilayer coatings generally showed higher fracture toughness than the Zr/ZrN multilayer coatings. The higher hardness and H/E ratio of the ZrN/ZrCN coatings as compared to the Zr/ZrN coatings is due to the presence of carbon in the ZrN/ZrCN coatings as explained earlier.

Conclusions
Using a dc magnetron sputtering method and varied N 2 and C 2 H 2 flowrates so as to manipulate the coatings nitrogen and carbon contents, Zr/ZrN and ZrN/ZrCN multilayer coatings, each with individual layer thicknesses of 250 and 500 nm and coatings thicknesses of 2 and 3 µm were deposited on 316L stainless steel. The coatings morphology and microstructure were investigated by high resolution FESEM, BEI, EDS, XRD and AFM techniques. Nanoindentation method was employed to measure the hardness and modulus of the coatings. The ratio of hardness to modulus (H/E) which is related to elastic strain to failure was also employed in the study. The cross-section FESEM images of the coatings reveal the coatings to show mainly growth that is columnar in nature with some tiny grains at the substrate/coating interface. The top surface FESEM images show a homogeneous, compact and dense morphology of fine spherical grains for all the coatings with the ZrN/ZrCN coatings showing more compact and smaller grain size than the Zr/ZrN coatings. The BEIs show alternating bilayers of equal sizes for each coatings. The denser and more compact nature of the ZrN/ZrCN as compared to the Zr/ ZrN coatings due to the carbon presence in ZrN/ZrCN resulted in higher hardness, modulus and H/E exhibited by the ZrN/ZrCN coatings. This suggests the ZrN/ZrCN coatings exhibit a higher wear resistance and fracture toughness than the Zr/ZrN coatings. The effective synthesis and excellent characteristics shown by the studied multilayer coatings make them viable candidates for tribological applications. www.nature.com/scientificreports/