Wear behavior at high temperature of ZrO₂–Y₂O₃ (YSZ) plasma-sprayed coatings

Atmospheric plasma-sprayed (APS) zirconia–yttria top coatings deposited on NiCrAlY bond coats have been widely used as thermal barrier coatings (TBCs) for gas turbine blades and combustion system components owing to their high thermal stability, low thermal conductivity, and relatively large thermal expansion, which is close to that of a metallic substrate [1‐3]. Under operating conditions, TBCs are exposed to high-temperature oxidation, hot corrosion, and severe wear, such as erosion, adhesion, abrasion, and fretting [3].

The mechanisms and chemical reactions that produce high-temperature oxidation and hot corrosion of yttria-stabilized zirconia (YSZ) TBCs manufactured by APS have been studied to identify ways to improve their performance and useful life [4‐7]. Recently, studies have been conducted to understand the tribological behavior of atmospheric-pressure plasma-sprayed YSZ thermal barrier coatings exposed to conditions similar to those in operation. Specifically, D. Shin et al. [8] have evaluated the erosion resistance of these coatings at temperatures between 537 and 980 °C, using an erosion tunnel to simulate the modern gas turbine operating conditions. The erodent particles were transported by gas at velocities between 122 and 305 m·s⁻¹, impacting the coating surface at angles between 20° and 90°. The effect of the coating structure porosity (porosities of 12.9 ± 0.5% and 19.5 ± 1.2%) on the erosion resistance was evaluated. These results demonstrate that higher wear rates due to erosion are associated with increased porosity in the YSZ coatings [8].

Similarly, Pakseresht et al. [3] studied the wear behavior of atmospheric plasma-sprayed YSZ coatings manufactured from Metco 204NS powder, with and without the addition of alumina whiskers, using a ball-on-disk microtribometer at room temperature to promote abrasive conditions on the contact surface between the coating and counterbody. An alumina ball with a diameter of 5 mm was used as the counterbody, on which normal loads of 7, 10, and 13 N were applied for each test. During the tests, the samples were rotated at a linear speed of 0.5 m·s⁻¹ up to a sliding distance of 500 m. The wear track analysis reported by the authors indicated that a smooth surface and abrasive detachment of particles were produced with wear rates between 4.1 \(\times\) 10^–2 and 7.3 \(\times\) 10^–2 mm³/N·m for the coating without adding alumina whiskers and between 3.5 \(\times\) 10^–2 and 5.3 \(\times\) 10^–2 mm³/N·m for specimens reinforced with alumina whiskers [3].

Liang et al. [9], Pawlowski [10], Shi et al. [11], Xiao et al. [12], and Lima et al. [13] reported that plasma thermally sprayed coatings using nanometric and submicrometric feedstock powders improved the mechanical properties by improving the coating’s structure. Additionally, H. Chen et al. [14] reported better wear performance of zirconia coatings manufactured from nanometric feedstock powders than that of the coatings sprayed from micrometric powders, which was attributed in the same way to the optimization of their structure, and therefore, the improvement of their mechanical properties.

L. Bai et al. [15‐17] studied the tribological performance of YSZ coatings exposed to sliding contact with an alumina ball from 25 to 800 °C. The results obtained by these researchers showed that the alumina ball used as a counterpart produces severe wear on the surface of these coatings, not only at room temperature [17], but also up to 800 °C [16]. Likewise, the results of these studies show that as the test temperature increases, the rate of coating wear tends to increase. However, the formation of a tribolayer produced on the surface of the coatings when they were tested at temperatures above 200 °C protects them from the damage produced by the alumina counterpart, reducing their wear rate [15, 16].

The topcoats of thermal barrier coatings are frequently manufactured by atmospheric plasma spraying from yttria-stabilized zirconia powders, which were previously fused and crushed, or agglomerated and sintered, using powder processing methods. Although fused and crushed powders are usually less expensive, agglomerated and sintered powders are commonly used to manufacture coatings that are exposed to high temperatures because their bimodal structure gives them higher thermal shock resistance than the monomodal structure obtained in coatings sprayed from fused and crushed powders [18]. At room temperature, the mechanical properties (hardness, elastic modulus, and fracture toughness) of monomodal and bimodal structure YSZ coatings could be statistically similar [19]. The phase transformations at high temperatures could change the mechanical performance of these coatings. Additionally, the yttria content used in the feedstock powders plays an essential role in the stability of the t′-ZrO₂ phase when zirconia-based materials are exposed to high temperatures for a long time. It has been reported that the complete and maximum stability of the t′-ZrO₂ phase is achieved when the content of Y₂O₃ is above 6 wt% [20].

A few studies have reported the performance of these coatings exposed to sliding contact with an alumina ball at temperatures up to 800 °C [16, 17]. However, their behavior at higher temperatures have not yet been published. For this reason, this work studied the wear performance of YSZ coatings exposed to abrasive conditions at temperatures between 25 and 1000 °C. The coatings studied were manufactured by atmospheric plasma spraying from agglomerated and sintered, as well as fused and crushed ZrO₂–Y₂O₃ feedstock powders, to produce bimodal and monomodal microstructures, respectively, which are specified for thermal barrier coatings in aircraft, stationary gas turbines, and engines with high thermal shock resistance, thermal insulating properties, and hot corrosion resistance [21].

To prepare the substrates, an INCONEL 718 bar was cut into discs with a diameter of 25 mm and a height of 7 mm. The surface to be coated was blasted using a corundum jet of particles, reaching an arithmetic average roughness (Ra) between 4 and 10 μm. Subsequently, the substrates were sonicated in an acetone bath to remove residues from the treatment previously carried out with abrasive particles and other dirt. NiCrAlY Sulzer–Metco Amdry 962^™ powder was atmospheric plasma sprayed as a bond coat on an INCONEL 718 substrate. Afterward, ZrO₂–Y₂O₃ top coatings were also manufactured by APS from the agglomerated and sintered H.C. Starck Amperit 827.423^™ and fused and crushed H.C. Starck Amperit 825^™ powders to produce bimodal and monomodal microstructures, respectively. The bond and top coatings are manufactured using a Sulzer–Metco PTF4^™ plasma torch according to the parameters listed in Table 1. These parameters were selected from preliminary tests carried out looking for coatings with the crystalline and amorphous phases, as well as the mechanical properties usually required for their use as topcoat in thermal barriers.

The chemical compositions of the feedstock powders were measured using a wavelength-dispersive X-ray fluorescence (WD-XRF) spectrometer with commercial reference: Thermo Fisher SCIENTIFIC ARL™ OPTIM'X. In the same way, Horiba PARTITA LA-950V2 laser diffraction (LD) equipment was used to characterize the particle size distribution of these powders. The crystallographic composition of the feedstock powders and the coatings was determined using an X-ray Cu Kα 1 radiation (DRX) Diffractometer with commercial reference: Bruker D8 ADVANCE and the X’Pert Highscore Plus Software following the COD cards: (1) t′-ZrO₂ (1525706), (2) c-ZrO₂ (1521753), and (3) m-ZrO₂ (1010912). Afterward, the Rietveld method was used to quantify the phases, following the same COD cards and the Material Analysis Using Diffraction (MAUD) software. In addition, a JEOL JSM IT-300 LV scanning electronic microscopy (SEM) equipment was used to characterize the morphological features of the ZrO₂–Y₂O₃ feedstock powders particles, the surfaces, and the cross sections of the coatings, as well as the surface of wear tracks. The surfaces and the cross sections of the coatings were ground and polished according to the ASTM E1920 standard [22] to obtain an arithmetic average roughness (Ra) lower than 0.2 µm. The porosity was determined on the cross sections of the coatings from images taken by SEM according to the indications of the ASTM E2109 standard [23] and using the Image J software. On the other hand, the hardness, the elastic modulus, and the fracture toughness of the YSZ coatings were determined from indentations performed on the polished surfaces of all samples using a Shimadzu HMV-G20 equipment following the specifications of ASTM C-1327 [24] and ASTM E-384 [25] standards. The hardness, the elastic modulus, and the fracture toughness were calculated according to Eqs. (1)–(3), respectively:where \({H}_{{\text{V}}}\) is the Vickers microhardness [GPa], \({P}_{{\text{N}}}\) is the normal load applied to the indenter [N], and \(d\) is the average length of the two diagonals produced during indentation [mm].where \(E\) is Young’s modulus [GPa], \(\alpha\) is a constant \((\alpha =0.45)\), \({H}_{{\text{k}}}\) is the Knoop microhardness [Pa], \({a}{\prime}\) and \({b}{\prime}\) are the longer and shorter diagonals, respectively, produced by the indentation [µm], and \(a\) and \(b\) are the geometric constants of the indenter \((b/a=1/7.11)\), as in Fig. 1a.where \({K}_{{\text{IC}}}\) is the fracture toughness [MPa·m^1/2],\(E\) is the Young’s modulus [GPa], \(H\) is the Vickers microhardness [GPa], \({P}_{{\text{N}}}\) is the applied normal load on the indenter [N], and \(C\) is the longest radial crack produced during the indentation [mm], as shown in Fig. 1b.

Wear tests were performed at 25, 500, 750, and 1000 °C using a ball-on-disk tribometer under dry sliding contact without eliminating the formed debris. An alumina ball 6 mm in diameter, with a hardness Vickers of 18.0 ± 0.5 GPa, was used as a counter-body, on which a normal load of 5 N was applied. The samples were rotated during 20000 cycles reaching a relative linear speed of 0.1 m·s⁻¹ with respect to the alumina ball, according to some recommendations of the ASTM G-99 standard [26]. Morphological characterization of the wear tracks produced during the tribological tests was performed using SEM with the aforementioned equipment. The wear rate was calculated from the profile curves of the wear tracks measured on the samples (Fig. 2) using a Surtronic S125 profilometer and Eq. (4).where \({WR}_{{\text{Sample}}}\) denotes the wear rate [mm³/N·m],\({A}_{s}\) is the wear track cross-sectional area [µm²], \({r}_{{\text{wt}}}\) is the radius of the wear track [mm], \({P}_{{\text{N}}}\) is the applied normal load [N]; and \({N}_{{\text{c}}}\) is the total number of cycles.

In the same way, to calculate the wear rate produced for each counter-body, an electronic micrometer with commercial reference: Mitutoyo and Eq. (5) were used.where \({WR}_{{\text{Counterbody}}}\)is the wear rate [mm³/N·m], \(h\) is the spherical cap height [mm], \(R\) is the radius of the counter-body [mm], \({P}_{{\text{N}}}\) is the normal load applied [N], and \({D}_{{\text{T}}}\) is the total distance of the test [m].

After the wear tests, the porosity, crystallographic phases, hardness, elastic modulus, and fracture toughness were reevaluated using the same equipment, standards, and equations mentioned above to compare the values with those obtained before the wear tests. The porosity, mechanical properties, and worn area were measured for three samples, ten times each, guaranteeing statistical reproducibility and repeatability for all measurements.

In ceramic materials, as YSZ coatings manufactured by APS, the tribological performance is influenced mainly by their hardness and fracture toughness [41, 42], which depend, among other factors, on the crystalline and amorphous phases of which they are composed. It is important to note in Table 4, the increase of the t′-ZrO₂ phase at the expense of the amorphous phase after the wear tests developed at 1000 °C (p-value < 0.05), which other researchers have previously reported, but for Al₂O₃–ZrO₂ plasma sprayed coatings [43]. However, the quantity of the m-ZrO₂ and c-ZrO₂ phases did not change with the heating of the samples during the tribological tests at high temperatures. This could be linked to both the temperature and time at which these tests were carried out are not enough for the transformation of these phases. During the tribological evaluation, the samples were exposed to each test temperature for 4 h, which could be insufficient for the diffusion of an additional amount of Y₂O₃ in the t′-ZrO₂ phase, as well as for Y₂O₃ diffusion from t′-ZrO₂ to produce a mixture of stable tetragonal phases with monoclinic or cubic phases. Other authors have reported the transformation of t′-ZrO₂ to m-ZrO₂ in ZrO₂-8 wt% Y₂O₃ coatings manufactured by APS and heated at 1100 °C for more than 800 h [20], as well as the transformation from t′-ZrO₂ to c-ZrO₂ in a single crystal of similar chemical composition heated at 1600 °C during 50 h [44]. Likewise, it has been reported that the transformation from t′-ZrO₂ to m-ZrO₂ in coatings manufactured by APS from powders with less than 6 wt% Y₂O₃ occurred slowly at 1300 °C [20]. In this order of ideas, since the m-ZrO₂ and c-ZrO₂ phases did not change and any decrease in porosity of the coatings was evidenced by sintering during tribological tests (Table 3), it was possible to establish that the key for the good wear behavior of coatings tested at 1000 °C, was the transformation from amorphous to t′-ZrO₂ phase, thanks the higher hardness of this crystalline phase regarding the amorphous one [45].

The fracture toughness of the C_A–S coating was slightly higher than that of the C_F–C coating at all the temperatures evaluated. This indicates that the fine particles remaining inside the partially molten granules detected in the structure of the C_A–S coating (Fig. 4a) can relax the stress and then arrest the cracks produced by the microindentations carried out to measure this mechanical property, as reported by other authors [18, 46]. Additionally, the fracture toughness of the C_A–S coating remained constant after the wear tests performed at different temperatures. At the same time, for the C_F–C coating, this property decreased slightly after the wear test was carried out at 1000 °C. Although in Eq. (3), the increase in microhardness (\(H\)) could promote the decrease in the \(E/H\) ratio, the decrease in the crack length (\({C}^{3/2}\)) owing to the presence of partially molten particles in the C_A–S coating, in turn prompted an increase in the \({P}_{N}/{C}^{3/2}\) ratio, maintaining the fracture toughness at the end. On the other hand, in the C_F–C coating, the absence of partially molten particles in its structure does not allow the reduction of the \(E/H\) ratio to be compensated by the increase of the \({P}_{N}/{C}^{3/2}\) ratio, due to the reduction of crack length after the tribological test carried out at 1000 °C.

The typical tribological mechanisms of ceramic materials under sliding contact conditions are functions of speed and load [46]. It is essential to highlight that when a ceramic material can withstand the mechanical stress applied by the counterbody, it produces a wear mechanism called “ductile deformation,” which usually shows features such as friction marks, and plastic flow, and, therefore, low wear rates, as obtained for coatings tested at 25 and 1000 °C [47]. Moreover, suppose the ceramic material cannot withstand the mechanical stress applied by the counterbody. In that case, it produces another wear mechanism called “brittle deformation,” which typically shows features such as fracture, cracks, and excessive detachment of particles, and therefore, high wear rates, as obtained for coatings tested at 500 and 750 °C [47]. Ductile and brittle deformation are wear mechanisms applicable only to ceramic materials [47].

It is important to note that despite using two powders with different chemical compositions (especially in terms of Y₂O₃ content) and morphologies, the results in Table 4–5 show that the manufactured coatings have similarities in the type and percentage of crystalline and amorphous phases, as well as in their mechanical properties, which gives them similar tribological behaviors at each temperature evaluated. However, this does not rule out a possible difference between the two types of coatings if they are evaluated at high temperatures for longer durations. In the coating manufactured from the powder containing 3 wt% Y₂O₃, the t′-ZrO₂ phase could reach a transformation to the m-ZrO₂ phase [37], while in that containing 8 wt% Y₂O₃, the t′-ZrO₂ phase will be completely stable [20]. Thus, it is possible to establish that the decrease in the wear resistance of the two coatings is due to the increase in thermal stress with the temperature of the tests, promoting cracks and the detachment of particles by brittle deformation. However, when the two coatings were tested at 1000 °C, the amorphous to t′-ZrO₂ phase transformation occurred. Their hardness increased, and a protective debris layer was produced, promoting the wear by ductile deformation.

The particle detachment evidenced on the wear tracks of samples tested at 500 °C, which was more notorious for the coating performed at 750 °C, is characteristic of wear produced by brittle deformation because the stresses applied by the hard alumina ball (~ 19 GPa) used as a counterbody during sliding contact are substantially higher than the mechanical resistance of the coatings, whose hardness is ~ 9 GPa  and their fracture toughness is 2.9–3.2 MPa·m^1/2 [47]. On the other hand, the microcracks produced by the sliding of the counterbody on the surface of the samples are due to fatigue fracture, as has been previously reported by other authors [48, 49]. These cracks were more evident in the samples tested at 500 and 750 °C owing to the thermal stresses produced as the test temperature increased. In the coatings tested at 1000 °C, the fine debris could have decreased the stress contact between their surface and the alumina counterbody [50] and produced a continuous layer densified by this contact.

The transformation of the wear mechanism from ductile deformation produced at room temperature to brittle deformation at temperatures up to 800 °C was reported by J. H. Ouyang et al. [48] for ZrO₂–Y₂O₃ coatings manufactured by low-pressure plasma spraying; however, the new transformation of wear mechanism toward ductile deformation occurring at 1000 °C that is presented in this study for ZrO₂–Y₂O₃ coatings manufactured by atmospheric plasma spraying is unpublished. This transformation from wear by brittle deformation produced in the samples tested between 500 and 750 °C to wear by ductile deformation at 1000 °C (Fig. 6m–p) was mainly due to the increase in hardness produced by the crystallization of the amorphous phase, as well as the formation of a protective layer from debris. Figure 6n–p indicates the protective layer produced on the wear tracks of coatings tested at 1000 °C, evidenced by plastic flow described by other authors for wear with ductile deformation [47].

From Fig. 6a–p, it is possible to see similar wear behaviors for both coatings, starting with wear by ductile deformation, followed by wear by fragile deformation, and finally wear by ductile deformation. These wear behaviors have been reported in [51, 52] for ZrO₂–Al₂O₃ and Al₂O₃ coatings, respectively. It is also important to note that the wear rate values reported in Fig. 7a are comparable to those reported in [16] for YSZ coatings tested at similar tribological conditions at 25 and 500 °C, as well as those reported in [51, 52] at high temperatures for Al₂O₃ and Al₂O₃–ZrO₂ coatings, respectively.

As shown in Fig. 7a, the samples tested at 25 °C showed the lowest wear rate, which was likely due to the slight damage caused by ductile deformation during the tribological contact between the surface of the coating, the alumina ball, and the low quantity of debris (Fig. 6a–d). At 500 and 750 °C, the wear rates increased with the tribological test temperatures, probably because of the contribution of thermal stresses to the wear tests (Fig. 6e–l). For the samples tested at 1000 °C, the wear rate was lower than that at 500 °C and 750 °C, probably due to the reduction in the severity of the tribological contact promoted by the protective layer formed from debris. In the same way, Fig. 7b shows the increase of the counterbodies wear rates as the temperature increased in all samples (Fig. 8a–h), which was probably related to: (i) the hardness decrease of alumina ball when exposed to high temperatures [53], (ii) the increase in the severity of the contact conditions due to the increase in thermal stresses, and (iii) the increase of hardness in both coatings heated at 1000 °C owing to their crystallographic changes mentioned above.

The highest wear rate in the alumina ball used as the counter-body in tests performed at 1000 °C confirms that the increase in the hardness of the coatings was a driving factor for the change in the wear mechanism from brittle deformation at 750 °C to ductile deformation at 1000 °C. Wear-by-ductile deformation occurs when the tested material has sufficient mechanical resistance to withstand the contact conditions to which it is exposed, and a controlled quantity of rounded debris produces a layer that protects the sample [47, 50]. However, it can increase the severity of the damage produced to the counter-body by harder particles. The decrease in the friction coefficient measured in the tests performed at 1000 °C indicated that the debris that acted as a third body tended to be more rounded than those produced in the tests performed at 500 and 750 °C, where the wear was due to brittle deformation.

The friction coefficient values (Fig. 9) measured during the tribological tests for both coatings increased with the temperature increase until 750 °C. They then decreased for the samples evaluated at 1000 °C, which could be linked to the fine particles of debris in the protective layer, whose morphology tended to be mainly spherical, thus reducing this coefficient. The obtained coefficients of friction were similar to those previously reported for YSZ and PSZ materials against alumina [16, 54].