Effect of oxygen pressure during HVOF spraying on the structural-phase state of Zr 2 CN coatings

This study investigates the effect of oxygen pressure on the structural-phase state, microstructure, mechanical properties and tribological behavior of Zr(C,N)-based coatings deposited on U8G tool steel by high-velocity oxygen fuel (HVOF) spraying. The coatings were produced at oxygen pressures of 4.0 × 10 5, 4.3 × 10 5 and 4.5 × 10 5 Pa, while the spraying distance, propane pressure and air pressure were kept constant. Variation in oxygen pressure was found to affect the relative XRD peak intensities, Rietveld-refined phase fractions, diffraction-line broadening parameters, coating thickness, porosity and tribological response. However, since the coatings exhibited differences in thickness, the interpretation of XRD intensity variations was performed with consideration of possible thickness-related effects and substrate contribution. X-ray diffraction and Rietveld refinement indicated the formation of a multiphase Zr(C,N)-based system containing carbonitride, carbide, nitride, oxide and α-Fe phases. The coating deposited at 4.3 × 10 5 Pa showed the most favorable combination of Zr(C,N)-type phase retention, reduced oxide fraction, low apparent cross-sectional porosity and relatively uniform microstructure. Williamson–Hall analysis suggested the largest coherent scattering domain size of approximately 82 nm and the lowest microstrain of approximately 0.30% for this condition, although these values should be considered as comparative estimates due to partial overlap of Zr 2 CN, ZrC and ZrN reflections. SEM/EDS analysis showed that the coating obtained at 4.3 × 10 5 Pa had the most homogeneous distribution of Zr, C and N and relatively low oxygen content. This coating also exhibited the highest microhardness (1826 ± 38.4 HV 0.5), the highest apparent pull-off adhesion strength (18.4 MPa) and the lowest friction coefficient (μ ≈ 0.54 ± 0.01). Quantitative wear assessment according to ASTM G99 showed that this coating had the lowest wear volume and specific wear rate among the investigated samples. SEM examination of wear tracks suggested a correlation between oxygen pressure, coating morphology and wear-track features; however, direct post-wear chemical analysis of the tribofilm would be required to confirm the detailed wear mechanisms. Overall, the results indicate that an oxygen pressure of 4.3 × 10 5 Pa provides the most balanced coating characteristics under the investigated HVOF conditions.

Carbon tool steel components, particularly U8G steel, are widely used in components operating under contact loads, abrasive and adhesive wear, and cyclic friction. However, their service life is significantly limited by intense surface wear, increased friction coefficient, localized overheating in the contact zone, and accelerated development of surface defects (Liu et al., 2022). Therefore, one of the most effective approaches to increasing the durability of such components is the formation of functional protective coatings on their surfaces that combine high hardness, wear resistance, and friction stability while maintaining the load-bearing capacity of the steel base (Wang et al., 2025).

Among the promising materials for tribological coatings, zirconium carbonitrides (Zr 2 CN), a type of nitride-carbide ceramic, are of particular interest. These materials are characterized by high hardness, wear resistance, thermal stability, and a reduced tendency to seize under friction due to their chemical inertness and ability to form stable surface films (Ul-Hamid, 2020). Unlike binary nitrides or carbides, carbonitride systems allow for the variation of the C/N ratio, providing targeted control over the structure and, consequently, the balance of mechanical and tribological properties (Yao et al., 2005).

However, the production of Zr 2 CN coatings on steel substrates using traditional vacuum methods (PVD/CVD) is associated with a number of technological limitations, including stringent surface preparation requirements, limitations on coating thickness, product geometry, and process productivity (Constantin and Miremad, 1999; Braic et al., 2011; Sun et al., 2009). For example, Hollstein et al. (2001) demonstrated the feasibility of producing Zr 2 CN coatings using a cathodic arc method and their subsequent use in medical instruments.

Studies devoted to thin-film ZrCN systems have made a significant contribution to the study of the structure and properties of coatings. In particular, it was shown that an increase in the N 2 flux during reactive magnetron sputtering leads to the formation of an FCC structure of ZrCN, whereas at high nitrogen concentrations an amorphous phase is observed, contributing to a decrease in the friction coefficient (Cotruț et al., 2013). The introduction of hafnium (Hf) provides an increase in hardness (∼5 GPa), adhesion, corrosion resistance, and biocompatibility of the coatings (Cotruț et al., 2013).

Fundamental aspects of the formation of the carbonitride phase the mechanism of Zr 2 CN synthesis by carbothermal reduction was investigated, showing that the process proceeds through successive stages of intermediate compound formation, with diffusion processes and thermodynamic phase stability playing a key role (Zhang et al., 2020). Recent studies have also demonstrated that multilayer TiCN-ZrCN coatings exhibit improved structural stability and enhanced mechanical performance (Kolchev et al., 2024; Yin et al., 2025).

Additional information on the structure and properties of carbonitride systems was obtained from studies of bulk materials. It was shown that ZrCN-ZrO 2 ceramics obtained by spark plasma sintering formed at approximately 1,562 °C, with maximum hardness of 17.63 GPa, fracture toughness of 4.22 MPa × m 1/2, and Young’s modulus of 426.49 GPa (Kuzmenko et al., 2025). It was found that an increase in the oxide phase content contributes to a decrease in porosity and an increase in the density of the material, which has a positive effect on the mechanical properties. The maximum values of the character-istics are: hardness - 17.63 GPa, fracture toughness - 4.22 MPa m1/2, Young’s modulus - 426.49 GPa and strength - 551.41 MPa, while the highest strength (623.12 MPa) is observed for the ZrN-ZrO 2 system (Kuzmenko et al., 2025). It has also been shown that increasing the nitrogen and oxide phase contents leads to increased oxidation resistance due to the formation of a protective ZrO 2 layer at temperatures above 900 °C.

Currently, thermal spraying methods are widely used to improve the performance of cutting tools, including atmospheric and vacuum plasma spraying, detonation, arc, and flame spraying, as well as plasma electrolytic hardening (Al-Mutairi et al., 2015; Najafizadeh et al., 2023; Torkashvand et al., 2021; Lyphout et al., 2019; Kurbanbekov et al., 2025; Williamson and Hall, 1953; Bao et al., 2018). Among these, HVOF occupies a special place due to its high flow velocity (∼1,500 m/s) and flame temperature (∼3,000 °C) (Kurbanbekov et al., 2024; Mittal et al., 2025; Krutskii et al., 2023). This ensures the formation of dense, low-porosity, minimally oxidized coatings with improved mechanical and tribological properties (Mouadji et al., 2022).

Thus, HVOF technology is an effective tool for producing high-quality wear-resistant coatings that meet modern industrial requirements and sustainable development goals (Younes et al., 2015; Younes et al., 2016).

Therefore, this study presents a comprehensive study of the formation of Zr 2 CN coatings using HVOF, focusing on the effect of oxygen pressure on the microstructure, phase composition, and functional characteristics. The primary objective is to establish quantitative relationships between spraying process parameters and coating performance properties, including microhardness, adhesion strength to U8G steel, friction coefficient, and wear mechanisms.

U8G steel (15 × 45 × 3 mm 3) was used as the substrate. Before coating, the substrate was sandblasted with white alumina grit, which improved coating adhesion (Table 1) (Kurbanbekov et al., 2025).

Zirconium carbide (ZrC) and zirconium nitride (ZrN) powders were purchased from Hebei Suoyi New Material Technology Co., Ltd. (Handan, China). Zirconium carbonitride (Zr 2 CN) powder was synthesized in accordance with the current technical specifications (TU). Figure 1 shows the particle size distribution of the original Zr 2 CN powder, obtained by laser diffraction. The distribution is expressed as a volume fraction (Q3(x), %) depending on the particle diameter (x, μm). The main peak is located in the range from ∼10 to ∼50 μm, with a modal value (maximum frequency) of about 25–30 μm. The C-shaped cumulative distribution curve confirms a narrow and symmetrical particle distribution, indicating good powder homogeneity. The absence of distinct fractions in the nano- or submicron region (<1 µm) indicates the absence of agglomerates or highly finely ground phase.

Figure 2 shows the X-ray diffraction analysis of the original Zr 2 CN powder selected for sputtering. The presence of the Zr 2 CN peak, along with reflections attributed to ZrC and ZrN, indicates partial retention of the original zirconium carbonitride with an FCC lattice. At the same time, the main diffraction lines of this phase coincide in position with the shifted reflections of the (ZrC)/(ZrN) FCC phases.

Figure 3 shows SEM images of Zr 2 CN powders with a particle size distribution of 10–100 µm. It is evident that they have an irregular polyhedral fragmented shape, which is known to have lower flow properties than spherical particles.

Zr 2 CN coatings were produced using the HVOF method using a Termika-3 system. This system is equipped with a control system that regulates the pressure of the supplied gases (propane, oxygen, and air), enabling control of the spraying process parameters.

Figure 4 shows a schematic and general view of the HVOF system used (Kurbanbekov et al., 2024). During operation, the gas components (propane and oxygen) are fed into a combustion chamber, where they are intensively mixed at elevated pressure, forming a high-temperature gas flow. Zr 2 CN powder is fed into the burner at a predetermined velocity and transported by compressed air supplied by a compressor, after which it is accelerated and deposited onto the substrate surface.

The spraying process parameters are listed in Table 2. Spraying was carried out at a fixed distance of 0.3 m and constant propane (2 × 10 5 Pa) and air (3.2 × 10 5 Pa) pressures, while oxygen pressure varied from 4.0 × 10 5 to 4.5 × 10 5 Pa. Three samples (a, b, c) were obtained with different oxygen supply conditions, allowing us to study the effect of this parameter on the structure and properties of the coatings.

The powder particle size distribution was determined by laser diffraction using a FRITSCH ANALYSETTE 22 laser particle size analyzer. Measurements were performed in the 0.01–3,800 µm range with the suspension circulating under the following dispersion conditions: pump 60%, ultrasound 100%. Data were recorded using 60 channels; the obscuration was 19%–20%. Fraunhofer optical theory and the broad (100) distribution model (Fritsch GmbH, Germany, 2025) were used to calculate the particle size distribution.

Microstructural studies and analysis of the sprayed coatings were performed using scanning electron microscopy (SEM) on a Tescan Vega four system, which provides both a surface microstructure overview and cross-section analysis (Brno, Czech Republic, 2024).

The phase composition of the resulting Zr 2 CN-based coatings was studied using an X'Pert PRO X-ray diffractometer with Cu-Kα radiation (λ = 1.5406 Å), a voltage of 40 kV, and a current of 30 mA. The scanning angle range was from 20.01° to 89.99° with a step of 0.02° and a data accumulation time of 2 s. Diffraction patterns were processed using High Score Plus software with a PDF5+ (2026) database (Malvern Panalytical Empyrean, Almelo, Netherlands, 2009).

Tribological sliding wear tests were conducted using a TRB3 tribometer (Anton Paar, Buchs, Switzerland) in a ball-on-disc configuration under the following conditions: normal load N = 5 N, sliding speed υ = 2 cm/s, ball diameter = 6 mm (material: 100Cr6), sliding distance L = 30 m, and dry friction environment. The wear rate of the samples and counterbody was quantified according to the ASTM G99 standard (Younes et al., 2020).

Microhardness measurements were performed using the Vickers method in accordance with ASTM E384, on an HLV-1DT microhardness tester. A diamond tetrahedral pyramid indenter with an angle of 136° was used. The notation “HV0.5” refers to a Vickers hardness test performed under a load of 0.5 kgf (4.9 N) with a dwell time of 10 s. The load was applied to the sample surface, and the diagonals of the indentation traces (d1 and d2) were measured with high accuracy.

One of the important aspects of ensuring coating quality is the adhesion to the substrate. To improve the adhesive bond, the surface of the coating at the “dolly” attachment site was roughened with sandpaper and degreased with ethyl alcohol. Adhesion tests using the pull-off method were conducted in accordance with ASTM D4541-22 at a temperature of 20 °C ± 5 °C, no earlier than 3 days after coating deposition. The adhesive was applied following the manufacturer’s instructions. “Epoxy Adhesive 2,214” was evenly spread on the dolly surface, which was then pressed onto the coating and held until curing, ensuring proper alignment of the bonded surfaces. Excess adhesive was removed if necessary. Using a cutting tool (ring cutter), the coating was cut down to the metal around the dolly. Cuts were made through the entire coating thickness until the metal substrate was exposed, with a minimum cut width of 1 mm.

To evaluate the adhesion properties of the coatings, tests were performed in accordance with ASTM D4541-22 using a hydraulic adhesion tester Elcometer 510 (Elcometer Instruments, Manchester, United Kingdom). The adhesive strength of the coatings was measured under the following conditions: dwell time 0.50 s, target loading rate 1.00 MPa/s, and dolly diameter of 20 mm.

The porosity of the coatings was quantitatively evaluated from cross-sectional SEM images using ImageJ software. For the analysis, only the coating region was selected, excluding the substrate. The images were converted to 8-bit format and subjected to threshold-based binarization to distinguish pore regions from the coating matrix. The porosity was calculated as the ratio of the total pore area to the total analyzed coating area: where P is the coating porosity, Apores is the total area of pores identified in the cross-section, and Acoating is the total analyzed coating area excluding the substrate.

X-ray diffraction patterns of Zr 2 CN coatings (samples a-c) obtained by the HVOF method at different oxygen pressures are shown in Figure 5. Analysis revealed the formation of a multiphase system including Zr 2 CN, ZrC, ZrN, ZrO 2, and α-Fe. The main diffraction peaks correspond to the FCC structure of the ZrC 2 N γ carbonitride solid solution (space group Fm-3m, lattice parameter a ≈4.636 Å), confirming the formation of the Zr 2 CN phase (Younes et al., 2016; Younes et al., 2020). The intermediate position of these peaks between the ZrC and ZrN reflections indicates the implementation of the mechanism of substitution of carbon and nitrogen atoms in the crystal lattice in accordance with Vegard’s law. Along with the main phase, secondary phases of ZrC and ZrN were identified, indicating partial decomposition of the original powder under high-temperature HVOF conditions (∼3,000 °C) and the nonequilibrium nature of phase formation during rapid cooling. The presence of the ZrO 2 oxide phase, detected by weak peaks, is due to the oxidation of particles during their transport in the flame. The intense α-Fe reflection indicates the contribution of the substrate, which is particularly pronounced for sample a, due to the lower coating thickness and density.

With an increase in oxygen pressure from 4.0 × 10 5 the to 4.5 × 10 5 Pa, a consistent evolution of the phase composition and structural state of the coatings is observed. Sample (a) is characterized by low intensity of the Zr 2 CN peaks and a high proportion of α-Fe, indicating insufficient coating formation. Sample b (4.3 × 10 5 Pa) exhibits the highest intensity of carbonitride phase reflections with minimal substrate contribution, indicating the formation of the densest and most homogeneous structure with a high degree of crystallinity. With a further increase in oxygen pressure sample (c), an increase in oxide reflections and a partial reduction in structural order are observed, which is associated with the intensification of oxidation processes (Figure 6).

Figure 7 shows the Williamson–Hall plots of β os θ versus 4sinθ for samples (a), (b), and (c), respectively. All Zr 2 CN coatings were obtained under the same process parameters (spraying distance 0.30 m, air pressure 3.2 × 10 5 Pa and propane pressure 2 × 10 5 Pa), with only oxygen pressure varying in the range of 4.0–4.5 × 10 5 Pa. Instrumental broadening was corrected using a standard reference material measured under identical diffractometer conditions, and six sample a, five sample b, four sample c diffraction peaks of the Zr 2 CN FCC phase were used for the linear fitting. The quality of the fits was evaluated using the coefficient of determination R 2, yielding values of 0.933, 0.952, and 0.906 for samples a, b, and c, respectively, confirming acceptable fit quality for all samples.

Analysis of microstructural parameters calculated by the Williamson–Hall method revealed a pronounced dependence on O 2 pressure. Sample b (4.3 × 10 5 Pa) is characterized by the lowest microstrain (∼0.30%) and the largest crystallite size (∼82 nm), suggesting more favourable coating formation conditions associated with higher Zr 2 CN phase retention and reduced internal stresses. With an increase in oxygen pressure to 4.5 × 10 5 Pa (sample c), an increase in microstrain (∼0.51%) and a decrease in crystallite size (∼16 nm) are observed, which is likely due to increased oxidation processes and partial degradation of the carbonitride phase. Sample a (4.0 × 10 5 Pa) exhibits the highest microstrain (∼0.59%) and the smallest coherent scattering domain size, consistent with incomplete coating formation and reduced structural order at lower oxygen pressure.

The Rietveld refinement confirms that sample b contains the highest Zr 2 CN content (23.4 wt%) and the lowest ZrO 2 fraction (11.2 wt%), consistent with optimal phase stabilization at 4.3 × 10 5 Pa. Sample a exhibits the highest α-Fe contribution (37.8 wt%), confirming the significant substrate signal arising from its lower coating thickness. These quantitative phase fractions support the phase-composition trends discussed in the manuscript and provide additional evidence beyond qualitative comparison of peak intensities alone. However, due to partial overlap of Zr 2 CN, ZrC and ZrN reflections, the Rietveld results should be considered as semi-quantitative estimates (Table 3).

Figure 8 shows SEM images (SE and BSE mode) of the cross-section of coatings deposited by the HVOF method under various deposition conditions. The micrographs demonstrate the coating morphology, as well as a clearly defined coating-substrate interface, indicating good adhesion between the layers.

Analysis of the coating thicknesses showed that for sample (a) it was 8.75 ± 1.25 μm, for sample (b) it was 13.75 ± 2.00 μm, and for sample (c) it was 11.33 ± 1.50 μm. The average thickness for all samples was approximately 11.28 μm. The coating of sample (a) is characterized by pronounced porosity and a non-uniform structure near the interface with the substrate, which is likely due to particle overoxidation during flight and partial decomposition in the flame. Meanwhile, coating (c) exhibits isolated pores and localized inhomogeneities, which may indicate insufficient particle kinetic energy during deposition. In contrast, coating (b) exhibits a dense and uniform structure with minimal porosity and a clearly defined interface, indicating high interfacial adhesion and good adhesion of the layer to the substrate.

The quantitative porosity analysis showed that the porosity values of the coatings were 0.96%, 0.31%, and 0.39% for samples a, b, and c, respectively. The lowest porosity was observed for sample b, obtained at an oxygen pressure of 4.3 × 10 5 Pa. This result confirms that the intermediate oxygen pressure provides more favorable conditions for particle heating, acceleration, and lamellar consolidation during HVOF spraying. Sample a exhibited the highest porosity, which may be associated with insufficient consolidation of the lamellar structure and incomplete stabilization of the Zr 2 CN-based coating. In contrast, sample b showed the lowest porosity, the largest crystallite size of approximately 82 nm, and the lowest microstrain of about 0.30%, indicating a more structurally stable and dense coating. For sample c, the porosity slightly increased, while the crystallite size decreased to approximately 16 nm and the microstrain increased to 0.51%. This behavior can be related to enhanced oxidation and partial decomposition of Zr 2 CN at higher oxygen pressure.

It should be noted that the porosity values reported here were obtained from a single representative cross-sectional SEM image per sample using threshold-based binarization in ImageJ. Therefore, these values should be interpreted as indicative estimates reflecting relative trends between samples rather than absolute volumetric porosity measurements. Single-image SEM-based porosity assessment is widely employed in the thermal spray coating literature as a practical comparative screening method, which was the primary objective of the present study. Figure 5 shows the results of EDS elemental mapping for cross-sections of the Zr 2 CN coatings, illustrating the distribution of key elements throughout the coating thickness and near the coating–substrate interface. A comparative analysis of the elemental distribution in the coatings (samples a–c) revealed a significant influence of HVOF spraying parameters on their chemical homogeneity and oxidation state. Sample a shows a pronounced zirconium diffusion zone; however, the distribution of carbon and nitrogen is relatively non-uniform, particularly near the coating–substrate interface.

The elevated oxygen content in the upper part of the layer indicates intense oxidation processes and the formation of oxide phases. In contrast, sample (b) is characterized by the most uniform distribution of the main elements (Zr, C, and N) across the entire coating thickness with minimal oxygen content. A clearly defined Fe concentration gradient at the substrate interface indicates the formation of a dense, weakly oxidized layer, ensuring high adhesion and effective protective properties. Sample (c), on the other hand, is characterized by an elevated oxygen content in the upper coating zone, accompanied by a decrease in the intensity of the carbon and nitrogen signals. This indicates partial decomposition of the Zr 2 CN carbonitride phase and the formation of oxide compounds. Zirconium remains uniformly distributed throughout the coating, but its phase state likely changes due to oxidation.

Thus, sample b exhibits the most favorable combination of microstructural homogeneity, relatively low oxidation and high coating density, indicating improved mechanical and protective behavior under the investigated conditions.

The formation of the multiphase Zr(C,N)-based coating should be interpreted as the result of both thermodynamic driving forces and kinetic limitations during HVOF spraying. The Zr(C,N) carbonitride phase can be considered as a substitutional solid solution between ZrC and ZrN, since both compounds crystallize in a NaCl-type FCC structure. The measured lattice parameter for the carbonitride phase in sample b was approximately a = 4.636 Å, which lies between the reference lattice parameters of ZrC (a = 4.698 Å) and ZrN (a = 4.5675 Å). Using Vegard’s law for the pseudobinary ZrC–ZrN system, the lattice parameter was calculated according to Equation 1. the calculated value x ≈ 0.525 corresponds to a C/N ratio close to unity. This supports the assignment of the main FCC phase to a Zr(C,N)-type carbonitride. However, due to the strong overlap between ZrC, ZrN and Zr(C,N) reflections, this interpretation should be considered together with the Rietveld refinement results and EDS elemental mapping rather than as an independent proof of phase purity.

From a thermodynamic point of view, zirconium-containing carbide and nitride phases have a strong tendency to oxidize in the presence of oxygen, and the formation of ZrO 2 is energetically favorable. This explains the appearance of oxide-related reflections in the XRD patterns, particularly at higher oxygen pressure. However, HVOF spraying is a highly non-equilibrium process. Therefore, the final phase composition cannot be explained by thermodynamic arguments alone. The residence time of particles in the flame is very short, and the particles undergo rapid heating, acceleration, impact on the substrate and fast cooling. Under such conditions, complete equilibrium transformation is unlikely. Instead, partial decomposition of the initial carbonitride phase, limited surface oxidation and rapid solidification of molten or semi-molten particles may occur simultaneously.

At 4.0 × 10 5 Pa, insufficient particle heating and incomplete lamellar consolidation may contribute to higher porosity, lower coating thickness and a stronger substrate contribution in the XRD pattern. At 4.3 × 10 5 Pa, the balance between particle heating, velocity and limited oxidation appears to promote the formation of the densest and most uniform coating. At 4.5 × 10 5 Pa, higher oxygen availability may enhance surface oxidation of particles, which is consistent with the increased oxygen signal in EDS mapping and the presence of oxide-related reflections. Thus, the observed phase and microstructural trends should be considered as the combined result of oxygen pressure, particle thermal history, oxidation tendency, deposition efficiency, coating thickness and rapid solidification.

Figure 9 shows the dependence of the friction coefficient (μ) on the friction path for the original sample and Zr 2 CN coatings (samples a, b, c) produced by the HVOF method. All coatings exhibit a characteristic running-in phase, after which a quasi-steady-state friction regime is established (Mouadji et al., 2024).

The original sample exhibits the highest friction coefficient, reaching an average value of μ ≈ 0.86 ± 0.03, indicating unstable tribological behavior and intense wear. Sample (a) has a slightly lower friction coefficient of μ ≈ 0.80 ± 0.03, but retains significant fluctuations, indicating a heterogeneous coating structure. Sample (b) exhibits the lowest friction coefficient, μ ≈ 0.54 ± 0.01, and the curve is highly stable and exhibits minimal fluctuations.

This indicates the formation of a dense, uniform coating with good antifriction properties and wear resistance. Sample c exhibits intermediate friction coefficient values of μ ≈ 0.74 ± 0.02, with relatively stable behavior, but a higher friction level compared to sample b, which is likely due to the increased content of oxide phases.

Thus, the best tribological properties are observed for sample b, which is consistent with the results of X-ray phase and microstructural analysis and confirms the optimality of this spraying mode.

To complement the tribological analysis, SEM examination of the wear tracks was carried out to identify the dominant wear mechanisms for each coating condition (Figure 10). The morphology of the worn surfaces revealed distinct differences depending on the oxygen pressure applied during HVOF spraying.

Sample (a), deposited at the lowest oxygen pressure (4.0 × 10 5 Pa), exhibited pronounced deep grooves and macro-cracks oriented along the sliding direction, along with scattered debris pits - features characteristic of abrasive wear. The relatively high porosity of this coating, as confirmed by SEM cross-section analysis, led to accelerated ploughing by the hard 100Cr6 counterball, and the porous microstructure provided insufficient mechanical support to resist crack propagation under cyclic contact loading. These observations are consistent with the elevated friction coefficient (μ ≈ 0.80) measured for this sample. This interpretation is further supported by the EDS elemental mapping of the coating cross-section (Figure 5a), which revealed non-uniform distribution of C and N throughout the coating and an elevated O content - confirming the chemically heterogeneous, partially oxidized state of sample (a) prior to tribological testing.

Sample (b), obtained at the optimal oxygen pressure (4.3 × 10 5 Pa), showed the mildest wear morphology: only shallow, fine abrasion grooves were visible along the sliding direction, accompanied by small oxidized wear particles distributed across the surface. The absence of macro-cracks and large spalling regions indicates that adhesive wear was the dominant mechanism, with the formation of a thin tribofilm from the oxidized particles acting as a lubricating layer and contributing to the low friction coefficient (μ ≈ 0.54). This behaviour is attributed to the dense, well-crystallized Zr 2 CN phase with minimal porosity and internal stress, which provided a high load-bearing capacity under the applied contact conditions. Chemical support for this mechanism is provided by the EDS mapping of sample (b) (Figure 5b), which confirmed the lowest oxygen content and the most uniform distribution of Zr, C, and N of all three coatings—consistent with a coating that resists oxidative decomposition under frictional contact and sustains a continuous, low-oxide tribofilm rather than generating brittle debris.

Sample (c), sprayed at the highest oxygen pressure (4.5 × 10 5 Pa), displayed a combination of surface cracks, large spalling pits, and localized delamination zones. These features are characteristic of oxidative and delamination wear, consistent with the elevated ZrO 2 content identified by XRD analysis. The brittle oxide-rich regions are prone to fracture under contact loading, leading to subsurface crack initiation and subsequent spalling. Consequently, the intermediate friction coefficient (μ ≈ 0.74) for sample (c) reflects a mixed regime of oxidative and mild abrasive wear. Indirect chemical support for this interpretation is provided by the EDS elemental mapping of sample (c) (Figure 5c), which shows a pronounced oxygen enrichment in the upper coating zone accompanied by a reduction in C and N signal intensity - confirming partial decomposition of Zr 2 CN to ZrO 2 according to reaction (8) prior to tribological testing. This pre-existing oxide-rich surface layer is the proximate cause of the brittle fracture morphology observed in the wear tracks.

Overall, the SEM analysis of wear tracks suggests that oxygen pressure influences the operative wear mechanisms. The transition from predominantly abrasive wear (sample a) through adhesive/mild-oxidative wear (sample b) to delamination/oxidative wear (sample c) is fully consistent with the measured friction coefficients and the phase-structural evolution established by XRD and Williamson–Hall analysis. The proposed mechanism assignments are underpinned not only by SEM morphology of the wear tracks but also by the pre-wear chemical characterization of each coating: EDS elemental mapping (Figure 5) documenting the oxygen content and elemental homogeneity of each coating, and XRD phase analysis (Figure 6) quantifying the ZrO 2 fraction -together providing indirect support for a correlation between coating composition to tribological response. It should be noted, however, that direct post-wear chemical characterization of the wear tracks - such as EDS point analysis or Raman spectroscopy of the tribofilm and debris - was not performed in the present study. Such analysis would allow unambiguous identification of transfer-layer composition (e.g., Fe-rich debris in sample a or a ZrO 2 tribofilm in sample b) and is recommended as a direction for future work to further validate the proposed wear mechanism map.

To quantify the wear resistance of the coatings, the wear volume (V) was calculated from the cross-sectional area of the wear track (A) measured from SEM micrographs using ImageJ software, in accordance with ASTM G99: V = A × L, where L = 30 m is the sliding distance. The specific wear rate was determined as k = V/(F × L), where F = 5 N is the applied normal load.

The results are summarised in Table 4. Sample (b), deposited at an oxygen pressure of 4.3 × 10 5 Pa, exhibits the lowest wear volume (V = 12.33 × 10 -3 mm 3) and the minimum specific wear rate (k = 0.82 × 10 −4 mm 3 /N·m), which is 30% lower than sample (a) (k = 1.18 × 10 −4 mm 3 /N·m) and 21% lower than sample (c) (k = 1.04 × 10 −4 mm 3 /N·m). These quantitative results are fully consistent with the friction coefficient data, XRD phase analysis, and microstructural characterisation, all of which identify sample (b) as the optimal coating in terms of phase stability, density, and minimum porosity (0.31%).

The microhardness of the coatings was determined using the Vickers method. The results showed a significant increase in hardness after deposition of the Zr 2 CN coatings compared to the original substrate. The microhardness of the original material was 304.9 HV, while the coatings exhibited substantially higher values: 1725 ± 61.3 HV for sample a, 1826 ± 38.4 HV for sample b, and 1776 ± 42.9 HV for sample c. The average microhardness of the coatings was 1775.7 ± 41.3 HV, where the standard deviation reflects the variation between the different deposition modes (Figure 11).

The highest microhardness value was observed for sample (b), indicating the formation of the densest and most uniform coating structure. Samples (a and c) exhibit slightly lower values, which may be due to increased porosity (sample a) and enhanced oxidation (sample c). Overall, the increase in microhardness by more than 5 times compared to the substrate confirms the high efficiency of HVOF spraying for surface hardening.

The results of the adhesion strength of the Zr 2 CN coatings (Figure 12) showed that the 413 adhesion strength significantly depends on the HVOF spraying parameters. The adhesion strength 414 values were: 14.7 ± 3.6 MPa for sample (a), 18.4 ± 2.1 MPa for sample (b), and 16.4 ± 2.4 MPa for sample (c). The 415 average adhesion strength of the coatings is 16.5 ± 1.5 MPa, where the standard deviation reflects the 416 variation in properties under different spraying conditions.

The highest adhesion strength was observed for sample (b), indicating the formation of the densest and most uniform coating with a high degree of adhesion to the substrate. The lower value for sample a may be due to insufficient density and the presence of structural defects, while for sample (c), a slight decrease in strength is due to the increased proportion of oxide phases. In general, the obtained results are consistent with the data of microstructural and phase analysis and confirm the optimality of the sputtering mode for sample (b).

Although detailed fractographic analysis was not performed, visual inspection after pull-off testing suggested a mixed adhesive/cohesive failure character, especially for sample b. Therefore, the adhesion values are discussed as an apparent pull-off strength of the coating–adhesive–substrate system rather than as a purely interfacial adhesion value.

The highest adhesion strength was observed for sample (b), indicating the formation of the densest and most uniform coating with a high degree of adhesion to the substrate. The lower value for sample a may be due to insufficient density and the presence of structural defects, while for sample (c), a slight decrease in strength is due to the increased proportion of oxide phases (Younes et al., 2025; Qiao et al., 2025). In general, the obtained results are consistent with the data of microstructural and phase analysis and confirm the optimality of the sputtering mode for sample (b).

This study evaluated the influence of oxygen pressure on the structural-phase state, microstructure, mechanical properties and tribological behavior of HVOF-sprayed Zr(C,N)-based coatings deposited on U8G tool steel. The oxygen pressure was varied from 4.0 × 10 5 to 4.5 × 10 5 Pa, while the other main spraying parameters were kept constant. The results show that oxygen pressure affects coating formation, phase composition, microstructure, porosity and friction behavior; however, the observed trends should be interpreted together with the accompanying differences in coating thickness, density and substrate contribution.

XRD analysis combined with Rietveld refinement indicated the formation of a multiphase Zr(C,N)-based system containing Zr 2 CN/Zr(C,N), ZrC, ZrN, ZrO 2 and α-Fe. The coating deposited at 4.3 × 10 5 Pa showed the highest refined Zr 2 CN fraction (23.4 wt.%) and the lowest ZrO 2 fraction (11.2 wt.%), while sample a exhibited the highest α-Fe contribution due to its lower coating thickness. Because of the strong overlap of Zr 2 CN, ZrC and ZrN reflections, the Rietveld results should be considered semi-quantitative and were interpreted together with peak-position analysis, Vegard’s law calculations and EDS elemental mapping.

Williamson–Hall analysis showed that sample b had the largest coherent scattering domain size of approximately 82 nm and the lowest microstrain of approximately 0.30%. In contrast, samples a and c exhibited higher microstrain values, which may be associated with incomplete coating formation at lower oxygen pressure and enhanced oxidation at higher oxygen pressure. Due to partial peak overlap and limitations of the Williamson–Hall approach, these values should be regarded as comparative estimates rather than absolute crystallite-size and microstrain values.

SEM cross-sectional analysis showed that the coating deposited at 4.3 × 10 5 Pa had the most uniform morphology, the highest average thickness and the lowest apparent cross-sectional porosity of 0.31%. Samples a and c showed higher porosity and/or more pronounced oxygen enrichment. Since porosity was estimated from thresholded cross-sectional SEM images, the reported values should be considered two-dimensional comparative area-fraction estimates rather than absolute volumetric porosity.

The best combination of mechanical and tribological properties was obtained for sample b. This coating showed the highest microhardness of 1826 ± 38.4 HV 0.5, the highest apparent pull-off adhesion strength of 18.4 MPa and the lowest friction coefficient of μ ≈ 0.54 ± 0.01. Quantitative wear assessment showed that sample b also had the lowest wear volume and specific wear rate. These improvements are attributed to the combined influence of oxygen pressure, more effective lamellar consolidation, increased coating thickness, reduced porosity and limited oxidation.