Investigation of the tribological characteristics of Ta–Zr–Si–B–C–N coatings

Introduction

Tantalum disilicide is one of the promising materials in the family of high-temperature ceramics due to its high melting point (2300±100 °C) [1], electrical resistivity (50–70 μΩ·cm) [2], hardness (16 GPa) [3], strength at temperatures above 1000 °С, and good oxidation resistance [4]. TaSi₂ coatings are characterised by high thermal stability up to 500 °C and oxidation resistance at 800 °C due to the formation of a Ta₂O₅–SiO₂ oxide layer [5]. Due to their low resistivity (70 μΩ·cm) at t = 800÷900 °C, TaSi₂ coatings are often used in the semiconductor industry [6; 7].

To improve its mechanical and tribological properties and its oxidation resistance, tantalum silicide is doped with various elements such as C, N, B, Hf, and Zr [8–16]. The carbon-doped TaSi₂ coating demonstrates good resistance to high-temperature erosion at a thermal flux of 2.4 MW/m² [9] and oxidation resistance at t = 900 °C for more than 233 h [10]. The authors explain high oxidation resistance as due to the formation of a dense Ta₂O₅–SiO₂ oxide layer. The introduction of nitrogen also improves the mechanical properties and oxidation resistance of TaSi₂ coatings.

The study [11] found an extreme dependence of hardness and fracture toughness on the nitrogen content: their maximum values Н = 36 GPa and K_1с = 3.95 МPa·m\(^{0,5}\), respectively, were reached at a concentration of 35 at. % N. The Ta–Si–N coating is characterised by good oxidation resistance and thermal stability at t = 700 °С [12]. In previous studies, the authors studied the structure and properties of Ta–Si–N coatings [13]. The results showed that coatings with an optimal nitrogen concentration had the maximum values of hardness (24 GPa) and elastic recovery (77 %), and also demonstrated high oxidation resistance at t = 1200 °C. It is known that introducing nitrogen in Ta–Si–C coatings increases their tribological characteristics at temperatures up to 800 °C due to the formation of the ternary oxide TaSiO_x in the contact area [14].

Research on the effect of the introduction of additives of transition metal borides to the composition of TaSi₂-based coatings has been limited to a few studies. Doping Ta–Si–C coatings with zirconium boride [15] increases their adhesive and cohesive strength. The specimens demonstrate good oxidation resistance at t = 1500 °C, which may be related to the formation of a protective ZrO₂–SiO₂ oxide layer that prevents oxygen penetration. We have previously investigated the structure and oxidation resistance of Ta–Zr–Si–B–C–N coatings [16] obtained by magnetron sputtering in various gaseous media.

The main objective of this work is to study the tribotechnical characteristics of Ta–Zr–Si–B–C–N coatings exposed to abrasion and erosion impacts and in the sliding friction mode.

Materials and methods

The coatings were deposited by magnetron sputtering in a direct current mode. A sputtered composite target TaSi₂–Ta₃B₄–(Ta, Zr)B₂ (composition, at. %: 70.8 Ta, 18.6 Si, 7.4 Zr, 2.9 B) of diameter 120 mm and thickness 6 mm was obtained by hot pressing of crushed products of self-propagating high-temperature synthesis (SHS). Equipment based on the UVN-2M (УВН-2М) pumping system (JSC Kvarts, Russia) with a schematic diagram given in the work [17] was used to deposit the coatings. VOK-100-1 (ВОК-100-1) (JSC Polikor, Russia) aluminium oxide plates and disks were used as model substrates for the coatings. Before coating, the substrates were cleaned for 5 min in isopropyl alcohol using a UZDN-2T (УЗДН-2Т) unit (NPP UkrRosPribor, Ukraine) with an operating frequency of 22 kHz and in a vacuum using a gap-type ion source (Ar⁺ ions, 2 keV) for 20 min. Ar (99.9995 %) and its mixtures with N₂ (99.999 %) and C₂H₄ (99.95 %) were used as the working gas. The flow rate was controlled using a gas injection system (OOO Eltochpribor, Russia). The values are given in the table.

The coatings were deposited under the following conditions: the distance between the substrate and the target was 80 mm, the residual pressure was 10\(^{-}\)³ Pa, and the working pressure in the vacuum chamber was 0.1÷0.2 Pa. The magnetron power was kept constant at 1 kW using a Pinnacle+ power supply (Advanced Energy, USA), with a deposition time of 40 min.

The distribution profiles of the elements and the thickness-averaged composition of the coatings were determined using the glow-discharge optical emission spectroscopy (GDOES) method on a Profiler 2 device (Horiba Jobin Yvon, France) [18]. The structure of the coatings was examined by scanning electron microscopy (SEM) using an S-3400 microscope (Hitachi, Japan). X-ray diffraction (XRD) was performed on a D2 Phaser diffractometer (Bruker, Germany) using CuK_α-radiation. The X-ray photoelectron spectroscopy (XPS) studies were carried out on a PHI 5000 VersaProbe-II (ULVAC-PHI, USA) instrument. The excitation source was monochromatised AlK_α-radiation (hν = 1486.6 eV) with a power of 50 W and a diameter of 200 μm.

The thickness and abrasion resistance of the coatings were measured by a Calowear tester (JSC NII Tavtoprom, Russia) according to the ball-specimen set up as described in the methodology [19]. The material was exposed to an abrasive DiaPro suspension with a 1 μm polycrystalline diamond dispersion fed into the gap between a rotating ShKh-15 (ШХ-15) steel ball of diameter 27 mm and the surface of a stationary sample. The ball rotation speed was 13 rpm, and the load was 1.5 N. The volume of the coating material removed was determined using 2D microscopic images.

Abrasion tests were also used to determine the coating thickness using the formula

\[S = \frac{{{b^2} - {a^2}}}{{8R}},\]

where b is the wear scar diameter, μm; a is the substrate diameter, µm; R is the ball radius, µm.

The quantity of the coating material removed was calculated using the formula

\[V = \frac{\pi }{{64R}}\left( {{b^4} - {a^4}} \right),\]

where b and a are the outer and inner diameters of the crater, respectively, mm.

Erosion tests were carried out using a UZDN-2T (УЗДН-2Т) (NPP UkrRosPribor, Ukraine) ultrasonic disperser. A sample was placed in a container positioned in the working area, after 20 ml of water and 5 g of Si₃N₄ abrasive material were added. The distance from the waveguide to the substrate surface was 1 mm, and the frequency was set at 22 kHz. The experiment lasted 15–60 min. The change in the mass of the coating samples due to erosive impact was estimated using a GR202 (AND, Japan) analytical balance with an accuracy of 0.01 mg.

The coatings were tested for sliding friction on an HT Tribometer (CSM Instruments, Switzerland) automated friction machine using a 6-mm-diameter Al₂O₃ ball as a counter body. The load was 1 N. The change in the coefficient of friction was recorded during heating from a temperature of 25 °C to 500 °C. The contact zones after tribological tests in the abrasive wear and sliding friction modes were examined using a Wyko 1100 (Veeco, USA) optical profiler.

Result and discussion
Composition and microstructure of coatings

The table above shows the elemental composition of the coatings. It can be seen that the concentrations of nitrogen and carbon in the coatings increased with an increase in the N₂ and C₂H₄ gas flow rates, respectively.

Fig. 1, a shows the X-ray diffraction patterns of the coatings taken in the 2θ = 20÷50° range.

Besides the Al₂O₃ substrate peaks (card JCPDS 88–0107), the X-ray diffraction pattern of coating 1 showed peaks corresponding to the hexagonal h-TaSi₂ phase (JCPDS 89-2941). Note that the differences in the intensity of the peaks from the Al₂O₃ substrate may be associated with a change in the composition and amorphisation of the coatings as a result of the introduction of nitrogen or carbon. The size of the h-TaSi₂ crystallites, determined by the Scherrer equation, was 11 nm. The introduction of N₂ and C₂H₄ into the gaseous medium resulted in the formation of coatings with a highly dispersed or amorphous structure. For nitrogen- and carbon-containing coatings, the peak maxima positions in the 2θ = 25÷45° range were close to the positions of the most intense peaks of the TaN (JCPDS 89–5198) and TaC (JCPDS 89–3831) FCC phases.

The crystallite size of the h-TaSi₂ phase for reactive coatings 2–5 was estimated from minimally overlapping lines. For coatings 2 and 3 deposited at an N₂ flow rate of 5 and 10 cm³/min, the sizes were 6.0 and 4.5 nm, while for carbon-containing coatings 4 and 5, they were similar at 3.5 and 3.0 nm, respectively. The decrease in the size of h-TaSi₂ crystallites and the amorphisation of coatings upon transition to reaction media are associated with the formation of new TaN and TaC phases, which, apparently, interrupt the growth of h-TaSi₂ crystallites.

According to the SEM images, base coating 1 had a columnar structure (see Fig. 1, b). It is important to note that this phenomenon adversely affects the mechanical properties and oxidation resistance of coatings [20; 21]. All reactive coatings showed an identical structure. The introduction of N₂ and C₂H₄ into the gas medium led to the suppression of columnar growth and the formation of highly dispersed crystallites.

The Calowear tester measurements showed coatings 1 and 2 as having a similar thickness of 7.2 and 7.0 µm, respectively (Fig. 2). An increase in nitrogen concentration led to its growth by 25 %. In [11], a similar result was obtained related to an increase in the thickness of coatings with an increase in the N₂ gas flow rate. Increasing the C₂H₄ flow rate to 5 and 10 cm³/ min led to a decrease in thickness by 18 and 10 %, respectively. The Calowear tester measurements were compared with the thickness values determined from the fracture cross-section SEM images of the coatings (Fig. 2). The results obtained were similar. This method can therefore be used for the rapid assessment of coating thicknesses.

Erosion resistance

The trial tests allowed determining the optimal mode in which the wear of the coatings (abrasive material – Si₃N₄, its mass – 5 g, liquid volume – 20 ml) was observed. The graph in Fig. 3 shows the dependence of the change in mass on the time of exposure to abrasive particles.

Coating 1 obtained in the Ar medium was observed to have the minimum mass loss Δm = –0.2 mg throughout the experiment. For coating 2, the value of Δm increased to 0.2 mg over a 0–30 min interval, which is probably due to the adhesion of wear products and abrasive particles on the sample surface. The subsequent decrease in mass by 3.8 mg over a 30–60 min interval is due to coating wear (Fig. 3, b). Coating 3 had the value Δm = –0.3 mg over a 60 min interval, which corresponds to the data obtained for the non-reactive coating. Unstable behaviour was observed for the carbon-containing sample 4: the Δm value increased by 1.0 mg over a 0–15 min interval, after which at 15–30 min of exposure it dropped to the initial values. Over a 30–60 min interval, Δm = 0.8 mg. Coating 5 with the maximum carbon concentration had Δm = 1.5 mg over a 0–15 min interval, after which the sample mass gradually decreased and by the 60\(^{{\rm{th}}}\) minute of the test approached the initial value (Δm ≈ –0.1 mg).

Visual inspection of the samples revealed no signs of wear on the coating surface 1 (Fig. 3, b). The coatings obtained in nitrogen had a clear circular wear boundary with noticeable areas of the substrate, whereas the samples obtained in ethylene had less pronounced wear marks and no areas corresponding to the substrate.

The samples obtained in Ar and Ar + C₂H₄ media therefore showed the best erosion resistance. The high erosion resistance of carbon-containing coatings can be explained by the increased hardness of the TaC carbide phase compared to the TaN and TaSi₂ phases [22; 23].

Abrasion resistance

The results of the abrasive tests showed that scratches from the impact of abrasive particles were observed on the surfaces of all samples. Fig. 4 shows the depths (H) and thickness (h) of wear craters under abrasive action for coatings 1–5.

Fig. 5 shows the dependence of the volume of material removed (V) on the abrasive exposure time (1- and 3-min) for the studied samples. Coatings 1, 3–5 showed similar values of H = 4÷5 µm and V = 5÷6·10\(^{-}\)⁴ mm³, respectively, differing within the error margin. The nitrogen-containing coating 2 has maximum values of H = 5 µm and V = 11·10\(^{-}\)⁴ mm³. With an increase in the exposure time to 3 min, coatings 1–3 had crater depths in the range of 6–7 µm, and the volume of the removed material was 24·10\(^{-}\)⁴ mm³. Note that the crater depths did not exceed the thickness of the samples 1–3. After a 3-minute exposure, coating 4 had an H value of 7 μm at a thickness of 6 μm, which is indicative of wear. In this case, the sample was characterised by a lower value of V = 18·10\(^{-}\)⁴ mm³ compared to coatings 1–3, which may be due to the influence of the solid Al₂O₃ substrate. Coating 5 with the lowest carbon concentration showed the minimum results (Н = 5.5 μm and V = 15·10\(^{-}\)⁴ mm³).

Summing up the data obtained, it can be concluded that the coating deposited at the maximum concentration of ethylene has better abrasive resistance, which may be due to the positive role of carbon in the friction process [24].

Tribological tests in the sliding friction mode

Fig. 6 shows the results of tribological testing of coatings in the sliding friction mode during heating from a temperature of 25 °C to 500 °C.

Coating 1 showed a stable coefficient of friction of µ ~0.3 up to t = 225 °C. Above this temperature, the value of µ increased to >0.8, which is indicative of coating wear. Sample 2 had an unstable coefficient of friction over the entire temperature range. Over the 25–110 °C range, there was a rapid increase of the µ value from 0.2 to 0.82, which may be associated with the formation of friction wear products. A further decrease in μ to 0.3 is due to the removal of wear products from the tribocontact zone. After a stable interval from 150 to 210 °C, the value of µ gradually increased until it exceeded the 0.8 value at t = 400 °C.

Coating 3 with the maximum nitrogen content showed a sharp increase in µ to ~1 at t = 25÷50 °C. The effect of an increase in the coefficient of friction to values close to 1 may be associated with the exit to the substrate and the friction of the counter body material (Al₂O₃) over the Al₂O₃ substrate, accompanied by adhesive interaction. A similar process was described in detail in [25] in the example of the emergence of a steel-to-steel tribocontact. Sample 4 with the minimum carbon content showed a stable value of μ ~0.3 up to a temperature of 250 °C. In the range t = 250÷350 °C, an increase in µ to 0.9–1.0 was observed. Coating 5 with the maximum carbon concentration demonstrated the best result, with its coefficient of friction having a stable value of 0.25 up to a temperature of 350 °C. According to the literature data, the Ta–Si–C–N coating is characterised by a high coefficient of friction of 0.6 at t = 300÷400 °C [26]. Note that the value of µ = 0.25 for sample 5 is half the value obtained earlier for the Ta–Si–C–N coating.

Thus, the coating obtained at a 10 cm³/min flow rate of C₂H₄ has the minimum coefficient of friction μ = 0.25 and the maximum operating temperature of 350 °C. To determine the reason for the decrease in the coefficient of friction with increasing carbon concentration, coating 5 was studied by X-ray photoelectron spectroscopy (see Fig. 6). In the C1s spectrum, peaks were observed at a binding energy of 282.9 and 284.4 eV, indicating the presence of Ta–C and C–C bonds, respectively [27; 28]. The reduced coefficient of friction may be associated with the positive role of free carbon, which in some cases can be released during supersaturation of the crystalline carbide phase and acts as a solid lubricant during friction [29]. The influence of the MeC carbide phase with a lower coefficient of friction compared to the MeN nitride phase also cannot be ruled out [30].

Conclusion

In this work, Ta–Zr–Si–B–C–N system coatings were obtained by magnetron sputtering method using a TaSi₂–Ta₃B₄–(Ta, Zr)B₂ target. Ar, as well as Ar + N₂ and Ar + C₂H₄ mixtures, were used as the working gas. The nonreactive Ta–Zr–Si–B coating was characterised by a columnar structure with a crystallite size of the h-TaSi₂ hexagonal phase of about 11 nm. When N₂ and C₂H₄ were introduced into the working medium, a change in the columnar structure to an equiaxed one with an h-TaSi₂ grain size of about 3–6 nm was observed. The thickness of the coatings was between 6.0 and 8.1 μm.

Abrasive tests showed that, when exposed for 1–3 min, the sample obtained at the maximum concentration of ethylene has the best abrasive resistance. This effect is associated with the positive role of carbon, which functions as a solid lubricant during friction.

Erosion tests showed that the base sample has the minimum mass change of –0.2 mg. The introduction of nitrogen did not affect the erosion resistance, and the weight loss values for samples 2 and 3 were –0.2 and –0.3 mg, respectively. The introduction of С₂H₄ into the working medium promoted the growth of Δm to 1.1–1.5 mg. No wear was observed on the surface of carbon-containing samples, which indicates their better erosion resistance.

The sliding friction tests showed that coating 1 has a stable coefficient of friction μ = 0.3 up to the maximum working temperature of 225 °C. The introduction of nitrogen led to an increase in the μ values of the coatings up to 0.8–1.0 and a decrease of the maximum operating temperature to 50–110 °С. The coating with the minimum carbon concentration was characterised by a coefficient of friction of ~0.3 up to 250 °C, which is close to the values for the non-reactive coating. Sample 5 containing the maximum content of carbon showed the best result, with its coefficient of friction remaining at the 0.25 level up to a temperature of 350 °C.