Self-Lubricating Effect of WC/Y–TZP–Al 2 O 3 Hybrid Ceramic–Matrix Composites with Dispersed Hadﬁeld Steel Particles during High-Speed Sliding against an HSS Disk

: WC/Y–TZP–Al 2 O 3 hybrid ceramic–matrix composites (CMCs) with dispersed Hadﬁeld steel particles were sintered and then tested at sliding speeds in the range of 7–37 m/s and contact pressure 5 MPa. Fast and low-temperature sinter-forging allowed obtaining micron-sized WC grains, submicron-sized alumina-reinforced yttria partially stabilized polycrystalline tetragonal zirconia (Y–TZP–Al 2 O 3 ), and evenly distributed Hadﬁeld steel grains. Such a microstructure provided new hybrid characteristics combining high hardness with high fracture toughness and tribological adaptation. The CMCs demonstrated low friction and high wear resistance that were better than those demonstrated by other composite materials such as, for example, MAX-phase composites, zirconia-base ceramics, ZrB 2 -SiC ceramics, and metal matrix WC–(Fe–Mn–C) composites. These good tribological characteristics were obtained due to the in situ mechanochemical formation of iron tungstates FeWO 4 and Fe 2 WO 6 on the worn surfaces of composite samples. These mixed oxides were included in multilayer subsurface structures that provided the self-lubricating and self-healing effects in high-speed sliding because of their easy shear and quasi-viscous behavior.

High-speed sliding usually means unlubricated sliding at speeds in the range of 5-90 m/s and only a limited number of materials can be capable of sustaining that harsh condition [7][8][9][10][11][12][13][14][15][16][17][18][19]. Those may be hybrid sliding pairs with ceramics, ceramic matrix (CMCs), or metallic matrix (MMCs) composites rubbed against hardened alloyed steel counterbodies [7][8][9][10][11][12][13][14][15][16][17][18][19]. It is worthwhile noting that despite temperatures as high as 1500-3000 • C that An XRD phase analysis of powder compositions (Table 1) clearly showed the presence of peaks belonging to their components such as HCP WC and FCC γ-iron, while XRD patterns of the Y-TZP-Al 2 O 3 powder blend allowed identifying peaks inherent with tetragonal ZrO 2 (t-ZrO 2 ) but having some peak tailings on their right shoulders (Figure 1g). No XRD peaks were found to detect the presence of an allotropic form of Al 2 O 3 . An inset in Figure 1e reproduces the low 2θ XRD pattern fragment obtained from the Y-TZP-Al 2 O 3 composition that allows deconvolution of the asymmetrical peak into "narrow" and "wide" components. Such a deconvolution may be the evidence in favor of the formation of a solid solution of the type ZrO 2 (3Y,Al) with variable concentration of Al 3+ ions in the tetragonal ZrO 2 lattice [41].
The source powder compositions were blended with ethyl alcohol and then subjected to planetary ball milling using ∅6 mm balls that allowed reducing the WC agglomerate size to 2.5 ± 0.5 µm. On milling, the ethyl alcohol was evaporated from the samples before cold pressing the dried source powders at 8.5 MPa. The resulting green samples were characterized by 55% of theoretical density. The temperature during sintering was controlled using an optical pyrometer and focusing on a graphite die whose inner surface was coated with the BN film to prevent diffusion of carbon.
Final consolidation was carried out using a sinter-forging machine and argon as a shielding gas by placing 13 × 7.2 × 6.5 mm 3 green samples between the top and bottom graphite ∅20 mm dies and applying induction heating to 1350 • C for 20 min. Then, these heated samples were held at that temperature for another 5 min for temperature equalization. Consolidation was performed by successively increasing the pressure at each stage in the order 5, 10, 15, 20, 25, and 30 MPa at 15 s intervals. The pressure was applied along the normal to the XOY plane as shown in Figure 2. Dwelling time at the final stage was limited to 10 min in order to avoid grain coarsening and retain the submicron structures. The next step was unloading, extracting the consolidated samples from the machine, and cooling them for 7-10 min in air. The bulk density of the sintered specimens was determined using the hydrostatic weighing method. The 25 W, 45 W, and 65 W sinter-forged samples had their densities equal to~98 ± 2% of the theoretical density. The source powder compositions were blended with ethyl alcohol and then subjected to planetary ball milling using 6 mm balls that allowed reducing the WC agglomerate size to 2.5 ± 0.5 μm. On milling, the ethyl alcohol was evaporated from the samples before cold pressing the dried source powders at 8.5 MPa. The resulting green samples were characterized by 55% of theoretical density. The temperature during sintering was controlled using an optical pyrometer and focusing on a graphite die whose inner surface was  Mean Vickers hardness and indentation fracture toughness values of as-sintered composites were determined from at least ten indentations using a Vickers pyramid hardness tester «Duramin-500» (Struers A/S, Ballerup, Denmark) obtained on a polished surface of the specimens at 10 Kg load and 10 s dwell time. Fracture toughness of samples was estimated by measuring the lengths of radial indentation cracks and calculating with the Niihara formula [42]. The elastic modulus (E) was calculated according to the Reuss rule of mixture [43], using stiffness values of 660 GPa, 200 GPa, 200 GPa, and 300 GPa for pure WC, Hadfield steel, ZrO 2, and Al 2 O 3 , respectively. heated samples were held at that temperature for another 5 zation. Consolidation was performed by successively increasi in the order 5, 10, 15, 20, 25, and 30 MPa at 15 s intervals. The the normal to the XOY plane as shown in Figure 2. Dwelling limited to 10 min in order to avoid grain coarsening and reta The next step was unloading, extracting the consolidated sam cooling them for 7-10 min in air. The bulk density of the sin mined using the hydrostatic weighing method. The 25 W, 45 samples had their densities equal to ~98 ± 2% of the theoretica Mean Vickers hardness and indentation fracture toug composites were determined from at least ten indentations us ness tester «Duramin-500» (Struers А/S, Ballerup, Denmark) face of the specimens at 10 Kg load and 10 s dwell time. Fra was estimated by measuring the lengths of radial indentation the Niihara formula [42]. The elastic modulus (E) was calcul rule of mixture [43], using stiffness values of 660 GPa, 200 GP pure WC, Hadfield steel, ZrO2, and Al2O3, respectively.
Tribological pin-on-disk testing of the composites was ca machine UMT-1 (Tochpribor, Ivanovo, Russia). A disk made (HSS) 63-65 HRC disk was used as a counterbody with slidin Chemical composition of an as-cast HSS steel was as follows 3.6 Mo, 1.8 Ni, 1.4 C, 1.0 Mn, Fe-balance).
Tests were conducted at nominal contact pressure of 5 M vided by total of three pins' contact area 63 mm 2 in the speed m/s, 20 m/s, 37 m/s) at ambient conditions of approximately humidity. The disk wear track diameter was 240 mm. Total o each with a wear path length of 1000 m was conducted at each To keep the wear path length constant, the test duration dim Tribological pin-on-disk testing of the composites was carried out using a sliding test machine UMT-1 (Tochpribor, Ivanovo, Russia). A disk made of as-cast high-speed steel (HSS) 63-65 HRC disk was used as a counterbody with sliding track diameter of 240 mm. Chemical composition of an as-cast HSS steel was as follows (wt.%): 15 V, 4.5 W, 3.6 Cr, 3.6 Mo, 1.8 Ni, 1.4 C, 1.0 Mn, Fe-balance).
Tests were conducted at nominal contact pressure of 5 MPa (normal load 315 N divided by total of three pins' contact area 63 mm 2 in the speed range of 7-37 m/s (7 m/s, 10 m/s, 20 m/s, 37 m/s) at ambient conditions of approximately 25 • C and 40-50% relative humidity. The disk wear track diameter was 240 mm. Total of three experiments (9 pins) each with a wear path length of 1000 m was conducted at each of the sliding speed values. To keep the wear path length constant, the test duration diminished with increasing the sliding speed so that, for speeds 7,10, 20, and 37 m/s, the test durations were~143 s,~100 s, 50 s, and~27 s, respectively.
Wear rate was evaluated by measuring the sample's height before and after the test and then calculating the wear rate as volume loss (mm 3 ) divided by sliding path length (m)-mm 3 /m.
Flash temperatures on the composite worn surfaces were estimated using the wellknown equation [44]: where µ is the CoF, P and V are the normal load and sliding speed, respectively, K sample and K disk are thermal conductivities of the pin and disk, respectively. J is the Joule's constant (in this case, J = 1) and α is the contact radius of the real contact area to be determined from an equation as follows (2): H sample is the sample's hardness.
The results of this estimation are presented in Table 2, together with the real temperatures of the sliding zone, as measured using a pyrometer (Promin-KH1, Prompribor, Russia).
Microstructural characterization of the subsurface layers was performed on crosssection polished and etched views prepared by breaking the previously notched samples.
A symmetrical Bragg-Brentano XRD configuration (θ/2θ) with the Cu-Kα radiation was carried out to identify phases formed in the as-sintered composites. An X-ray diffractometer DRON-3 (Burevestnik, Saint Petersburg, Russia), with a scan range of 20-80 • with (2θ) step size of 0.05 • was used. Grazing-incidence X-ray diffraction with the Co-Kα radiation was used to detect phases on the worn surfaces of the composites at beam incidence angle of 10 • . These experiments were carried out using a DRON-7 X-ray diffractometer (Burevestnik, Russia) with a scan range of 15-80 • with (2θ) step size of 0.05 • . Identification of the XRD reflections was performed using Crystal Impact's software "Match!" (Version 3.9, Crystal Impact, Bonn, Germany).

Microstructures and Phases in the As-Sintered Composites
An XRD phase analysis showed that all the as-sintered samples contained WC, γ-iron, and t-ZrO 2 ( Figure 3). Only tiny Al 2 O 3 peaks were noticed in the 25 W sample while none of them were observed in samples with higher contents of WC, i.e., samples 45 W and 65 W. Such a finding can be explained by shielding the characteristic X-ray radiation from light elements Al and O by that of from high atomic number elements as Zr, and especially W. XRD patterns in Figure 3 have been obtained from the sample's surfaces parallel to ZOY and XOY planes ( Figure 2). The WC peak heights from sections parallel to XOY are close to that of the source WC powders so the mean peak ratios I (001) /I (100) and I (002) /I (110) were  SEM BSE images from the polished views of as-sintered samples (//XOY) demonstrated the presence of bright WC grain almost equiaxed agglomerates whose mean sizes were ~25 μm, ~55 μm, and ~100 μm in samples 25 W, 45 W, and 65 W, respectively (Figure 4a-c). SEM BSE images of the surfaces parallel to ZOY demonstrate bright WC grain agglomerates of higher aspect ratio ~1:5 resulting from sinter-forging with compression applied along the OZ axis ( Figure 2). SEM BSE images from the polished views of as-sintered samples (//XOY) demonstrated the presence of bright WC grain almost equiaxed agglomerates whose mean sizes were~25 µm,~55 µm, and~100 µm in samples 25 W, 45 W, and 65 W, respectively (Figure 4a-c). SEM BSE images of the surfaces parallel to ZOY demonstrate bright WC grain agglomerates of higher aspect ratio~1:5 resulting from sinter-forging with compression applied along the OZ axis ( Figure 2).

Hardness, Fracture Toughness and Crack Propagation
Both hardness and Niihara [42] fracture toughness grows with the WC content in the CMCs (Figure 9).
It is interesting that no traces of a monoclinic ZrO 2 phase (m-ZrO 2 ) were detected on the fracture surfaces of the composites after three-point flexural tests (Figure 10), which was evidence that no t-ZrO 2 → m-ZrO 2 strain-induced martensitic transformation occurred there.   Radial crack propagation from the Vickers indentations was examined to establish the fracture mechanism ( Figure 11). It turned out that the transcrystalline fracture was a dominating mechanism for coarse WC grains (Figure 11b

Hardness, Fracture Toughness and Crack Propagation
Both hardness and Niihara [42] fracture toughness grows with the WC content in the CMCs (Figure 9).

Coefficient of Friction and Wear Rate
Sliding was carried out on composite surfaces parallel to the ZOY plane ( Figure 2). Samples 25 W, 45 W, and 65 W demonstrated rather steady sliding with moderate friction force oscillation amplitudes (Figure 13a,b). Tribological tests allowed obtaining the dependencies of wear rate and CoF on sliding speed for all the composites (Figure 13c It is interesting that no traces of a monoclinic ZrO2 phase (m-ZrO2) were detected on the fracture surfaces of the composites after three-point flexural tests (Figure 10), which was evidence that no t-ZrO2→ m-ZrO2 strain-induced martensitic transformation occurred there. Figure 10. The XRD patterns obtained from the fracture surfaces parallel to the ZOY plane. It is interesting that no traces of a monoclinic ZrO2 phase (m-ZrO2) were detected on the fracture surfaces of the composites after three-point flexural tests (Figure 10), which was evidence that no t-ZrO2→ m-ZrO2 strain-induced martensitic transformation occurred there. Figure 10. The XRD patterns obtained from the fracture surfaces parallel to the ZOY plane. Figure 10. The XRD patterns obtained from the fracture surfaces parallel to the ZOY plane.
Al2O3 component of the composite allows exposing either a transcrystallite fracture (F ure 11a,c area 1) or the propagating cracks were deflected by the coarse Y-TZP-Al spherical agglomerates (Figure 11b area 2). Both of these mechanisms can provide fractu toughness improvements from the viewpoint of increasing the energy spent on cra propagation.  The fracture surfaces resulting from the bending test allowed observing th scrystallite fracture mode on the WC grains (Figure 12a-e, area 1) and Y-TZP spherical agglomerates with eutectic structures (Figure 12a,b, area 2). The Hadfie particles demonstrated ductile cleavage (Figure 12a-e, area 3).

Coefficient of Friction and Wear Rate
Sliding was carried out on composite surfaces parallel to the ZOY plane ( Fig  Samples 25   demonstrated CoF maximums at sliding speeds in the range of 7-10 m/s, while minimum CoF values were achieved at 37 m/s speed (Figure 13c). The wear rate for all composites only slightly increased in sliding at speeds from 7 to 37 m/s (Figure 13d). The minimal wear rate and absolute minimum CoF = 0.07 were achieved on the 65 W composite at 37 m/s.

Microstructures and EDS of Worn Surfaces
Along with detecting WC and t-ZrO2, the XRD of worn surfaces allowed for the detection of the presence of mixed oxides FeWO4 and Fe2WO6 (Figure 14a) whose contents increased with the content of WC in the composites (Figure 14b).

Microstructures and EDS of Worn Surfaces
Along with detecting WC and t-ZrO 2 , the XRD of worn surfaces allowed for the detection of the presence of mixed oxides FeWO 4 and Fe 2 WO 6 ( Figure 14a) whose contents increased with the content of WC in the composites (Figure 14b). The mean WC XRD peak ratios such as I(001)/I(100) and I(002)/I(110) obtained from composites 25 W, 45 W, and 65 W were 0.22, 0.26, and 0.3, respectively, i.e., practically coincided with those from the corresponding polished surfaces parallel to ZOY plane ( Figure 3). Therefore, sliding did not result in any redistribution of the texture components. Such a finding may be explained by reducing the mechanical stresses because of the self-lubrication effect.
Sliding at 37 m/s resulted in the generation of a tribological layer (TL) on the worn surfaces of samples (Figures 15 and 16) whose total area depends on the content of WC in the composite. Only small TL patches can be observed on the worn surfaces of the 25 W sample together with bright WC agglomerate islets, while almost full TL coverage of the worn surfaces occurred on the 65 W sample. The mean WC XRD peak ratios such as I (001) /I (100) and I (002) /I (110) obtained from composites 25 W, 45 W, and 65 W were 0.22, 0.26, and 0.3, respectively, i.e., practically coincided with those from the corresponding polished surfaces parallel to ZOY plane ( Figure 3). Therefore, sliding did not result in any redistribution of the texture components. Such a finding may be explained by reducing the mechanical stresses because of the self-lubrication effect.
Sliding at 37 m/s resulted in the generation of a tribological layer (TL) on the worn surfaces of samples (Figures 15 and 16) whose total area depends on the content of WC in the composite. Only small TL patches can be observed on the worn surfaces of the 25 W sample together with bright WC agglomerate islets, while almost full TL coverage of the worn surfaces occurred on the 65 W sample. A semi-quantitative EDS mapping of the worn surfaces shows them enriched with oxygen (Figures 15 and 16). Additionally, it can be seen in Figure 15 that iron is concentrated within the WC islets.  A semi-quantitative EDS mapping of the worn surfaces shows them enriched with oxygen (Figures 15 and 16). Additionally, it can be seen in Figure 15 that iron is concentrated within the WC islets.

Tribological Layer Microstructures and Phases
Subsurface section views were prepared from the worn samples that allowed obtaining SEM images of tribological layers (TLs) generated on them in the course of high-speed sliding (Figure 17a-c). The subsurface section of the 25 W sample did not allow the demonstration of any TL except for some fragmentation in the top part of the WC agglomerate (Figure 17a).
Clearly observed is the TL on both 45 W and 65 W samples (Figure 17b,c). Almost the entire worn surface area of sample 65 W was covered by the TL (Figure 17c) and below the TL there is a transition zone (TrZ) where less intense plastic deformation, as well as WC and ceramic grain fracture and oxidation, occurred.

Tribological Layer Microstructures and Phases
Subsurface section views were prepared from the worn samples that allowed obtaining SEM images of tribological layers (TLs) generated on them in the course of high-speed sliding (Figure 17a-c). The subsurface section of the 25 W sample did not allow the demonstration of any TL except for some fragmentation in the top part of the WC agglomerate (Figure 17a).
Clearly observed is the TL on both 45 W and 65 W samples (Figure 17b,c). Almost the entire worn surface area of sample 65 W was covered by the TL (Figure 17c) and below the TL there is a transition zone (TrZ) where less intense plastic deformation, as well as WC and ceramic grain fracture and oxidation, occurred. Table 3, Figures 17 and 18 demonstrate the results obtained from the EDS examination of these zones. The TL is microstructurally represented by two different components   Figures 17 and 18 demonstrate the results obtained from the EDS examination of these zones. The TL is microstructurally represented by two different components such as homogeneous dark grey regions or TLA1 (Figure 17c, areas 3,4) and light grey regions composed of bright nanosized fragments (TLA2) (Figure 17c, areas 1,2). The EDS data in Figures 18 and 19, and Table 3 shows these regions as those enriched with tungsten, iron, chromium, manganum, and oxygen. The difference between TLA1 and TLA2 regions is that TLA1 contains more vanadium, chromium, and tungsten and less manganum as compared to the composition of TLA2 (Table 3, areas 1-4, Figures 18 and 19), i.e., the TLA1 is enriched with oxidized elements transferred from the HSS disk. On the other hand, higher concentrations of manganum (Table 3, areas 1-4) as compared to that of TLA1 and the presence of bright fine WC particles may be evidence in favor of TLA2 generation by tribooxidizing of the composite's structural components. eration by tribooxidizing of the composite's structural components.
From the viewpoint of its chemical composition, the TrZ looks like a transition from undeformed as-sintered base material to the TL. Apart from the relatively coarse but already fragmented WC (Table 3, area 9) and Y-TZP-Al2O3 grains ( Table 3, area 10), there are TrZ regions that are compositionally similar to those of TLA1, but have fewer concentrations of oxygen (Table 3, areas 6-8). In addition, the dark grey particles chemically similar to the TLA1 can be observed at the TL/TrZ boundary (Table 3,

Structure of As-Sintered Composites. Fracture Toughening Mechanism оf WC/Y-TZP-Al2O3 Hybrid Ceramic-Matrix Composite with Dispersed Hadfield Steel Particles
Relatively fast and low-temperature sinter-forging allowed obtaining a dense CMC composed of coarse WC grains, Y-TZP-Al2O3 microcomposite spheres, and grains as with dispersed Hadfield steel particles (Figures 4 and 8), i.e., exactly phases the same as those of source powders before any consolidation (Figure 3). Images in Figures 3 and 4, and Figures 4 and 8 allow observing that the composites had high aspect ratio WC and Y-TZP-Al2O3 components as expected from sinter-forging. It is plausible that efficient pore displacement to the periphery of the sample occurred there, allowing obtaining a dense WC/Y-TZP-Al2O3 hybrid ceramic-matrix composite with dispersed Hadfield steel particles. Let us note that these as-sintered composites revealed some specialness related to preferential concentration of HS in the intergrain spaces of WC grain agglomerates while no HS particles were found in areas occupied by Y-TZP-Al2O3 (Figures 5-7).
It was reported earlier [45,46] that the WC grains may form agglomerates when sintering the WC-ZrO2 and WC-Al2O3 composites even if the source powder blend was thoroughly homogenized. In our case, there is another factor facilitating the agglomeration because sinter-forging was carried out at 1350 °C , i.e., above the 1143 °C that is the WC-Fe eutectic point [37]. Therefore, some liquid phase sintering might have occurred when Hadfield steel preferentially interacted with the WC grains, and possibly partially infiltrated into the already formed WC grain agglomerates due to its better wettability of the WC [7,10,47] as compared to that of ZrO2 [48] or Al2O3 [49], i.e., of Y-TZP-Al2O3 components, which formed their own agglomerates outside of the regions occupied by WC-Hadfield steel cermet. From the viewpoint of its chemical composition, the TrZ looks like a transition from undeformed as-sintered base material to the TL. Apart from the relatively coarse but already fragmented WC (Table 3, area 9) and Y-TZP-Al 2 O 3 grains (Table 3, area 10), there are TrZ regions that are compositionally similar to those of TLA1, but have fewer concentrations of oxygen (Table 3, areas 6-8). In addition, the dark grey particles chemically similar to the TLA1 can be observed at the TL/TrZ boundary (Table 3,   Relatively fast and low-temperature sinter-forging allowed obtaining a dense CMC composed of coarse WC grains, Y-TZP-Al 2 O 3 microcomposite spheres, and grains as with dispersed Hadfield steel particles (Figures 4 and 8), i.e., exactly phases the same as those of source powders before any consolidation (Figure 3). Images in Figures 3 and 4, and Figures 4 and 8 allow observing that the composites had high aspect ratio WC and Y-TZP-Al 2 O 3 components as expected from sinter-forging. It is plausible that efficient pore displacement to the periphery of the sample occurred there, allowing obtaining a dense WC/Y-TZP-Al 2 O 3 hybrid ceramic-matrix composite with dispersed Hadfield steel particles. Let us note that these as-sintered composites revealed some specialness related to preferential concentration of HS in the intergrain spaces of WC grain agglomerates while no HS particles were found in areas occupied by Y-TZP-Al 2 O 3 (Figures 5-7) .
It was reported earlier [45,46] that the WC grains may form agglomerates when sintering the WC-ZrO 2 and WC-Al 2 O 3 composites even if the source powder blend was thoroughly homogenized. In our case, there is another factor facilitating the agglomeration because sinter-forging was carried out at 1350 • C, i.e., above the 1143 • C that is the WC-Fe eutectic point [37]. Therefore, some liquid phase sintering might have occurred when Hadfield steel preferentially interacted with the WC grains, and possibly partially infiltrated into the already formed WC grain agglomerates due to its better wettability of the WC [7,10,47] as compared to that of ZrO 2 [48] or Al 2 O 3 [49], i.e., of Y-TZP-Al 2 O 3 components, which formed their own agglomerates outside of the regions occupied by WC-Hadfield steel cermet.
Another interesting fact is that there was not any grain growth in the CMC areas occupied by the Y-TZP-Al 2 O 3 matrix, i.e., neither ceramic grains nor microcomposite eutectic spheres increased their sizes, and therefore all as-sintered composites 25 W, 45 W, and 65 W did not experience any t-ZrO 2 → m-ZrO 2 martensitic transformation during either flexural test (Figure 10) or high-speed sliding ( Figure 14).
All of this allowed preparing the WC/Y-TZP-Al 2 O 3 hybrid ceramic-matrix composites with a unique structure combining regions structurally similar to the WC-Hadfield steel MMCs [10] with microcomposite eutectic Y-TZP-Al 2 O 3 core structures inside the micron-thickness spherical shells similar to those found in the source ceramic powders (Figure 1c-f).
These materials possess simultaneously high hardness and fracture toughness ( Figure 10) achieved by combining the crack deflection on the coarse WC and Y-TZP-Al 2 O 3 grains ( Figure 11) with transcrystallite fracture of these grains (Figure 12a-e, area 1, Figure 12a,b, area 2) as well as improving the bulk fracture toughness due to the presence of a ductile steel component (Figure 12a-e, area 3). The complex eutectic core structure of Y-TZP-Al 2 O 3 spherical particles inside the agglomerates may be another factor in favor of enhanced fracture toughness (Figure 12a,b, area 2).
The preferential distribution of steel particles inside the WC agglomerates had an effect on the fracture toughness of the composites. Smaller content of the ductile steel component in the 25 W CMC resulted in the fact that the crack was propagating mainly via Y-TZP-Al 2 O 3 areas (Figure 11a,b). Along with that, even if having minimal fracture toughness at the level of~7.5 MPa m 1/2 this 25 W composite was still tough as compared to common ceramics with fracture toughness at the level of~2-3 MPa m 1/2 .

Coefficient of Friction and Normalized Wear Rate of WC/Y-TZP-Al 2 O 3 Hybrid Ceramic-matrix Composites with Dispersed Hadfield Steel Particles. Comparison to Other Composites
The CoF values of all WC/Y-TZP-Al 2 O 3 hybrid ceramic-matrix composites with dispersed Hadfield steel particles obtained at sliding speed 37 m/s and contact pressure 5 MPa were lower than those of Y-TZP-20 wt.% Al 2 O 3 [11] and ZrB 2 -20 vol.% SiC [8] ( Figure 20a). In particular, the CoF of the 65 W composite was at the level of that of WC-30 wt.% Hadfield steel MMC [10].
To provide a meaningful comparison of the wear characteristics of composites, the normalized wear rates were obtained as wear rates divided by the corresponding appearing contact areas [50]. It can be observed from Figure 20b that 25 W, 45 W, and 65 W composites demonstrated rather low wear rates as compared to that of pure ceramic composite Y-TZP-Al 2 O 3 while the wear rate of WC-30 wt.% Hadfield steel MMC was about an order of magnitude higher (Figure 20b).
These results obviously show that the WC/Y-TZP-Al 2 O 3 hybrid ceramic-matrix composites with dispersed Hadfield steel particles allowed for achieving a hybrid effect when the composites inherited the self-lubrication effect from the WC-30 wt.% Hadfield steel MMCs whose poor wear resistance was improved by the addition of wear-resistant TZP-Al 2 O 3 matrix.
normalized wear rates were obtained as wear rates divided by the corresponding appearing contact areas [50]. It can be observed from Figure 20b that 25 W, 45 W, and 65 W composites demonstrated rather low wear rates as compared to that of pure ceramic composite Y-TZP-Al2O3 while the wear rate of WC-30 wt.% Hadfield steel MMC was about an order of magnitude higher (Figure 20b).  When comparing the tribological characteristics of the 25 W, 45 W, and 65 W composites with those of ultra-high temperature ZrB 2 -20 vol.% SiC CMC one can see that the hybrid composites showed a comparable level of wear resistance and essentially lower CoFs (Figure 20a,b).

Subsurface Multi-Layer Tribooxidation Structure of WC/Y-TZP-Al 2 O 3 Hybrid
Ceramic-matrix Composite with Dispersed Hadfield Steel Particles That Provided the Self-Lubricating Effect and Self-Healing Effect in High-Speed Sliding The above-discussed behavior can be related to the generation of multilayer subsurface structures on the worn surfaces of the WC/Y-TZP-Al 2 O 3 hybrid ceramic-matrix composite with dispersed Hadfield steel particles. Such a subsurface structural evolution of the composites may be represented in the following manner.
The composites are intensely heated during sliding at 37 m/s so that subsurface temperatures may be close to that of Hadfield steel melting. Such a thermal plasticization facilitates the transfer of the HSS components from the counterbody onto the composite as well as the outcome of Hadfield steel components from the bulk of the composite. All these components may intermix and oxidize according to reactions as follows: Subsurface fragmentation of the WC grains give fine fragments that can be easily oxidized to WO 3 starting from 800 • C [51]: WC + 2O 2 = WO 3 + CO The positive effect of such an oxidization is that conditions are created for synthetizing self-lubricating iron tungstates from iron oxides (reactions (3) and (4)) and WO 3 (reaction (5)) starting from~600 • C and according to the reactions as follows [52,53]: Fe 2 O 3 + WO 3 → Fe 2 WO 6 (7) The higher the WC content in the composite, the more iron tungstate generated in the subsurface layers (Figure 14b), and the more worn surface area is covered by the tribological layer ( Figure 17). Another reason may stem from the fact that the heat conductivity increased with the WC content as well as the contact zone temperature (Table 2), which then resulted in easy flow-out of the quasi-viscous TL.
Let us note that the presence of nanosized particles in the TLA2 served as a marker to reveal the TL quasi-viscous flow pattern when more viscous TLA2 did not fully intermix with TLA1 while involved in sliding-induced motion. Nevertheless, no voids were formed during such a motion, and on the contrary, even some self-healing behavior was demonstrated when this quasi-viscous mass filled up the interfragmentary spaces and cracks (Figure 17c). Such an effect is similar to that demonstrated in high-speed sliding on ultrahigh temperature ZrB 2 -20 vol.% SiC ceramics by the in situ formed melted borosilicate glass [8,9].
The above-discussed results allow claiming that WC/Y-TZP-Al 2 O 3 hybrid ceramicmatrix composite with dispersed Hadfield steel demonstrated structural adaptation to high-speed sliding, and simultaneously improved wear resistance and the friction reduction mechanism due to the unique combination of its components.

Conclusions
New WC/Y-TZP-Al 2 O 3 hybrid ceramic-matrix composites with dispersed Hadfield steel have been prepared by sinter-forging for high-temperature (high-speed) sliding against steel and composite counterbodies. All composites demonstrated a high aspect ratio WC and microcomposite Y-TZP-Al 2 O 3 grains resulting from sinter-forging. The structural design of such a composite allows combining high hardness with good fracture toughness, and acceptable wear resistance with self-lubrication self-healing behavior due to the tribochemical generation of multilayer quasi-viscous and thick tribological layers. This layer served to cover and protect the underlying composite grains against direct mechanical fragmentation and removal of the fragments as well as create conditions for slow oxidizing of the WC fragments and incorporating them with the tribological layer. Such a synergetic adaptation effect was achieved due to the in-situ synthesis of iron tungstates FeWO 4 and Fe 2 WO 6 that are capable of encapsulating hard wear particles and fragments, filling voids and cavities, thus simultaneously reducing wear and friction in the hybrid composite/steel sliding pair.