Thermal Spraying of Ultra-High Temperature Ceramics: A Review on Processing Routes and Performance

Ultra-high temperature ceramics (UHTCs) are materials deﬁned as having melting points over 3000 (cid:3) C and withstand temperatures beyond 2000 (cid:3) C without losing functionality. As service environments become even more extreme, such materials will be needed for the next generation of aeronautic vehicles. Whether it is atmospheric re-entry or sustained hypersonic ﬂight, materials with resistance to extreme temperature will be in demand. Due to the size and shape limitations encountered by current processing methods of bulk UHTCs research of UHTC coatings, speciﬁcally thermal spray UHTC coatings, is accelerating. This paper ﬁrst presents a general summary of UHTC properties, followed by a comprehensive summary of the processing routes and microstructures of current UHTC thermal spray coatings. Then, a detailed review of the oxidation and ablation resistance of UHTC thermal spray coatings is outlined. Finally, potential avenues for the development of new UHTC coating compositions are explored.


Introduction
Ultra-high temperature ceramics (UHTCs) are materials typified by melting points higher than 3000°C and stability above 2000°C. This group of ceramics is made up of carbides, borides and some nitrides of group four and five transition metals (Ti, Zr, Hf, V, Nb and Ta); they present strong covalent bonds, which are responsible for the elevated stability at high temperatures.
UHTCs combine stability at extreme temperatures with high hardness, thermal conductivity, elastic modulus, good wear resistance and low coefficient of thermal expansion. Due to the combinations of properties UHTCs possess, they have been under investigation for some time for use in extreme aerospace applications, where inevitably, materials are required to operate at extreme temperatures in oxidizing environments. These applications include rocket propulsion components, leading edges, control surfaces and nose cones for hypersonic flight and atmospheric re-entry craft (Ref [1][2][3][4][5]. During sustained hypersonic flight and atmospheric re-entry, operating temperatures can be as high as 2200°C ( Ref 6,7). With the modern proliferation of private spaceflight companies utilizing reusable craft and the desire to develop hypersonic flight technology for military and commercial purposes, UHTCs have remained materials of significant scientific interest (Ref 8).
While much research in UHTCs has focused on sintered bulk materials, UHTC coatings have also been investigated. UHTC coatings have the advantage of being near net shape while the size and shape of bulk UHTCs are limited by the processing routes needed to densify them (Ref 9,10). Using current processing methods, such as spark plasma sintering or hot pressing, such high temperatures and pressures are needed to densify UHTCs that only small, simple shaped components can be fabricated. UHTC coatings have been used to reduce wear in machine parts and bearings, provide oxidation resistance for C or SiCbased composites, provide corrosion resistance and act as diffusion barriers ( . Coatings can be deposited in numerous ways; for example, UHTC coatings have been produced using vapor deposition methods such as physical vapor deposition (PVD) and chemical vapor deposition (CVD) ( Ref 14,[16][17][18][19][20]. While vapor deposition techniques have been used to form UHTC coatings and have the advantage of creating dense coatings at temperatures below the melting points of UHTCs, these processes can be limited by coating thicknesses (*20 lm), deposition efficiency and size of the area that can be coated ( Ref 21). In order to deposit thick UHTC coatings, thermal spray methods have to be used; however, the extreme melting points and potential for oxidation pose some problems.
This review will focus on thermal spraying of UHTC borides and carbides, specifically TiB 2 , ZrB 2 , HfB 2 , TiC, ZrC, HfC and TaC. The use of UHTCs in cermet (ceramic with a metallic binder) coatings is beyond the scope of this work; however, ceramic-ceramic composites will be discussed. The first section will give a general overview of the physical, mechanical and thermodynamic properties of these bulk UHTCs. The following section will give a brief introduction to various thermal spray processes used to deposit UHTC coatings and how the parameters used within these processes affect the microstructure and properties of UHTC coatings. Of the properties discussed, particular attention will be paid to the high temperature performance of UHTC coatings; the effect of a range of particle reinforcements on the oxidation and ablation resistance of UHTC composite coatings will also be examined. Finally, pathways for the next generation of UHTC coatings will be discussed.

Physical, Mechanical and Thermodynamic Properties of UHTCs UHTC Borides
As early as the 1960s, at the height of the space race, UHTCs (specifically ZrB 2 and HfB 2 ) were investigated as solutions for the extreme temperatures encountered in the first generation spacecraft by Kaufman and Clougherty (Ref 22) at the United States Air Force Materials Laboratory. At the same time, in the Soviet Union, similar work was conducted by Samsonov at what is now the Frantsevich Institute for Problems in Materials Science in Kiev ( Ref 23,24). Owing to their excellent thermal and mechanical properties (especially high hardness, high modulus, high thermal conductivity, and low thermal expansion coefficient), UHTC materials were found to be of interest for heat shields, rocket and structural components in these early spacecraft. More recently, these compounds have become subject to increased research for wear resistant applications such as ball bearings, machine tools and engine valves ( Ref 25).
Given the success of Kaufman and Clougherty in characterizing the high temperature properties of UHTC borides, much work into UHTCs over the subsequent years was focused on these compounds. Fahrenholtz et al. (Ref 26) provided a detailed summary of the properties of ZrB 2, and HfB 2 while work by Munro (Ref 27) provides similar information for TiB 2 . Key physical, mechanical and thermal properties for these materials are outlined in Table 1, where the high melting temperature and hardness can be appreciated (Fig. 1).
The After the value of UHTCs unique combination of properties had been determined, in the 1970s researchers began studies in an effort to understand the oxidation behavior of these materials, with much of the early work in this area again emanating from the Frantsevich Institute in Kiev and the USA ( Ref 31,32). UHTC borides undergo stoichiometric oxidation according to the reaction shown in Eq 1, where M is a group four or five transition metal ( Ref 33,34). UHTC borides form, at temperatures below 1200°C, a protective liquid B 2 O 3 layer. Oxygen diffusion through this protective liquid limits further oxidation. At higher temperatures, the B 2 O 3 evaporates, leaving a nonprotective porous, metal oxide skeleton leading to rapid oxidation. Due to the higher melting point and low vapor pressure of Zr and Hf oxides (2715 and 2758°C, respectively), ZrB 2 and HfB 2 have more high temperature resistance than other UHTC borides ( Ref 35). To further increase the oxidation resistance of these materials, the addition of silicon carbide (SiC), or other silicon containing compounds (such as MoSi 2 or TaSi 2 ) creates a borosilicate glass outer layer which is stable up to temperatures of *1600°C (Ref 33).

UHTC Carbides
Like UHTC borides, the UHTC carbides were investigated in the 1960s by NASA and various defence agencies and continued through to the 1990s and 2000s for use at high temperatures (Ref [36][37][38][39]. ZrC has been investigated for various nuclear fuel applications ( Ref 40). Carbides, in general, are renowned for their excellent hardness at high temperatures; in fact Miyoshi and Hara (Ref 41) showed that even at 800°C TiC maintained a microhardness of *1700 Hv (*17 GPa). Due to their high hot hardness, UHTC carbides have also been used in cutting tool applications ( Ref 42,43). Key physical, mechanical and thermal properties for the UHTC carbides covered in this review are listed in Table 2. As with the UHTC borides, the hardness and melting points stand out as being extreme. Compared to UHTC borides, the carbides have lower thermal conductivities meaning despite having higher melting temperatures, they are less attractive for use in heat shield applications at ultra-high temperatures. Although UHTC carbides have lower elastic moduli than borides at room temperature, they do maintain their strength at elevated temperatures ([ 1000°C) better than the borides. This means carbides are preferred in applications where higher thermal and mechanical loads are encountered ( Ref 44).
Unlike borides, the UHTC carbides are stable across a range of stoichiometries as can be seen in the phase diagrams in Fig. 3 (Ref 53, 54). TiC, ZrC and HfC are all stable between *37.5 and up to 50 at. % C, while TaC is stable between *47.5 and 50 at. % C. This range of stable stoichiometries means UHTC carbides have potentially tailorable physical and mechanical properties. As can be appreciated from the data and Table 2 and the phase diagram in Fig. 3, HfC and TaC have some of the highest melting points of all materials.
UHTCs will generally oxidize following the reaction in Eq 2, where M is Ti, Zr or Hf, and Eq 3, where M is Ta ( . In environments with low oxygen pressure, carbon may remain un-oxidized. Oxidation of these compounds can be affected by a number of variables such as chemical composition (it can be seen from the phase diagrams in the previous section that these carbides are not line compounds can present a variety of stoichiometries), grain size and porosity.

Thermal Spraying of UHTCs
As described previously, the current processing routes for bulk UHTCs, such as spark plasma sintering and hot pressing, limit the size and shapes of components that can be produced. Thermal spraying techniques are already widely used in many industries to coat large areas relatively quickly. This section of the review will focus on the thermal spray processes used in research and their effect on the microstructure, mechanical properties, wear resistance, oxidation and ablation resistance of UHTC coatings.

UHTC Boride Coatings
Deposition and Microstructure of UHTC Boride Coatings Atmospheric Plasma Spraying Arguably the most versatile thermal spray process is atmospheric plasma spraying (APS). APS uses a radio frequency or, more commonly, direct current arcs to ionize process gases creating a plasma jet. As these unstable plasma ions reform into their gaseous states, a large amount of thermal energy is released, creating extremely high temperatures, up to 14,000 K, within the plasma jet. The primary process gas typically used in APS is argon, with hydrogen, nitrogen, helium or a combination thereof being used as secondary gases to modify the properties of the thermal plasma. Feedstock particles are injected into the gas stream, where particle velocities can be between 20 and 500 mm/s depending on the size of the particle (Ref 63  Coefficient of thermal expansion (K -1 ) 6.7 x 10 -6 7.7 x 10 -6 6.6 x 10 -6 6.3 x 10 -6 Thermal conductivity (W m -1 K -1 ) 2 1 2 4 3 0 2 2 showed the boride phases to be dominant but with some Zr/ HfO 2 . The influence of spraying power on ZrB 2 coatings has been studied by Hu Fig. 4. At 95 kW, the residual stresses in the coating caused a certain degree of peeling; hence, 75 kW was found to be the optimum spray power. No difference in phase composition was reported for different coatings with ZrB 2 being the main phase detected, but ZrO 2 and ZrO were also present. Conversely, Feng et al. found that when depositing a ZrB 2 -SiC coating at 30, 75 and 97 kW all the coatings were highly porous (58, 43 and 53 % porosity, respectively) regardless of spray power. The coatings deposited at the two higher powers showed a higher degree of fully melted feedstock. ZrO 2 was also detected in the coating deposited at 97 kW while it was not present in the other two coatings.
The particle size of powder feedstocks typically utilized in HVOF thermal spraying and APS is limited between 10 and 100 lm. Using powders of this size ensures the powder particles have enough momentum upon injection to penetrate the middle of the jet, where the highest temperatures are to be found, yet are small enough to melt completely in a very short period of time (Ref 74). Using nano-and submicron scale feedstocks can lead to reducing splat size, reduced porosity and improved properties. To get around this, a technique called suspension thermal spraying has been developed. This is where small particles (\10 lm) are suspended in a liquid, which can flow through the feed system and has sufficient momentum to penetrate the high temperature region of the flame.
Using suspension plasma spraying (SPS), Yvenou et al. (Ref 75) deposited a TiB 2 feedstock with a median particle size of 1.4 lm. XRD results showed no oxide phases present in the coating; however, porosity was high as particles were not melting within the plasma plume.
Plasma Spraying in Inert Atmospheres As discussed in the previous section, when using APS to spray boridebased feedstocks, many researchers have reported the presence of oxide phases in the deposited coatings ( used APS and CAPS systems to spray ZrB 2 powder. After it had been sprayed into the water (to retain the feedstock as powder after spraying) via both systems, XRD diffractograms of the powder showed the APS technique to have large peak intensity for ZrO 2 phases, indicating a high degree of oxidation during the spraying process. Comparatively, XRD analysis of the powder sprayed by CAPS was shown to have large peak intensity for ZrB 2 phases while only some ZrO phase was detected. Depending on the spraying parameters used, the microhardness of the coatings deposited using CAPS was in the range of 9.8 to 15.7 GPa with microhardness generally increasing with the power of the torch and pressure inside the spraying vessel. The use of the CAPS system also ensured that the coating microstructures were all dense with minimal porosity. Similarly, Rietveld refinement was used by Kahl et al. to identify and quantify the phases present in APS and CAPS ZrB 2 coatings. Using the CAPS system, with an argon atmosphere, reduces the amount of total oxide phases by 45.7 wt. % compared to the APS coating. The average hardness of the coating was increased from 14.0 to 18. ) compared an APS coating to one produced by VPS using a ZrB 2 ? 20 vol. % MoSi 2 composite feedstock. XRD diffractograms of the two coatings showed the presence of ZrO 2 phase in the coating deposited by APS; the VPS coating showed no oxide phase. The microstructure of the APS coating showed interconnected porosity; meanwhile, the VPS coating had smaller, closed porosities. The porosity was measured as being 9.3 and 6.8 %, respectively. Like CAPS, ZrB 2 -based coatings deposited using VPS show no oxidation of the feedstock during spraying; however, these studies measured porosity in the coatings to be as high as *10 % ( Ref 85,86).
A comparison between LPPS and HPPS ZrB 2 -based coatings was made by Bartuli et al. (Ref 87). Characterization of single splats showed distinct morphologies for each process, as shown in Fig. 5. The splats deposited using HPPS show disc-like morphology while the LPPS splats have a branched structure, indicating particles were fully molten when they impacted the substrate. The difference in morphology was due to the higher particle velocities achieved in LPPS, which created splashing as the particles impinged the substrate. The authors suggest that the splats created by HPPS would offer improved cohesive and adhesive strength.
Shrouded Plasma Spraying: In an effort to maintain the inert atmosphere of VPS and CAPS while reducing the cost, some researchers have utilized a technique called shrouded plasma spraying to spray ZrB 2 -based coatings (Ref [88][89][90]. Instead of the expensive vacuum and furnace systems required in CAPS and VPS, shroud plasma spraying creates a contained or un-contained Ar or N curtain via an attachment on the end of the plasma torch, limiting the interaction between air and particles within the plasma jet. A detailed study on the effect of various shroud gas flow rates was conducted by Torabi et al. (Ref 90). This work found that increasing the Ar flow rate from 0 l/min (unshrouded) to 30 l/min and finally 150 l/min reduced the ZrO 2 phase content from 41.6 wt. % to 14.5 wt. % to 4.8 wt. %, respectively. Increasing the shroud gas flow also altered the microstructure and splat morphology of the coatings. The unshrouded coating featured many un-melted particles and had a porous microstructure, while increasing shroud gas flow led to a combination of fully melted splats and partially melted particles, as shown in Fig. 6, as well as less porous microstructures ( Fig. 7 and 8).
Reactive Plasma Spraying Some researchers have combined self-propagating high temperature synthesis (where constituent elements of a compound are reacted together at high temperatures) or reduction reactions with thermal spraying techniques in what is known as reactive plasma spraying (RPS). During RPS, reactions between precursor particles inside the plasma jet create the desired coating material in situ.
An % B 4 C feedstock resulted in the highest relative peak intensity of the ZrB 2 phase, however, both 15 and 30 wt. % feedstocks showed the presence of residual B 4 C. The ZrB 2 coating had a microhardness of 1.6 GPa, much lower than a ZrO 2 coating sprayed using similar parameters. The low hardness is linked to the highly porous coating microstructures; the authors suggest two reasons for this: unmolten ZrB 2 particles, a consequence of the ZrB 2 particles being formed in situ and having a short residence time in the high temperature plasma jet, or the boron carbide reduction reaction continuing after the coating has been deposited, releasing gases. reaction between constituent elements, in this case Ti and B, becomes thermodynamically favorable in inert atmospheres. SHS relies on the ability of these highly exothermic reactions to be self-sustaining and, therefore, energetically efficient (Ref 94). The use of LPPS eliminated oxidation of the feedstock; the coating had a high degree of porosity; meanwhile, the APS coating had improved density due to the use of Cr as a binder. In terms of composition, the coating produced using APS was made up of TiB 2 and TiN phases with Ti 2 O 3 and TiO 2 as well. Comparatively, the LPPS coating was mainly comprised of TiC 0.3 N 0.7 and TiB 2 with no oxide phases (the authors suggested residual N remained in the atmosphere despite the low-pressure vacuum). Microhardness values for the LPPS coating were measured to be 4.9 GPa with the low hardness being attributed to the level of porosity in the coating; the corresponding value for the APS coating was 7.1 GPa.
High Velocity Oxy Fuel Thermal Spraying High velocity oxy fuel (HVOF) thermal spraying is a form of flame spraying whereby a gas or liquid fuel (for example, hydrogen, kerosene, acetylene, propylene or natural gas) is ignited in the presence of oxygen. This creates a high temperature, highly pressurized mixture of gases within the combustion chamber into which the feedstock is injected either radially or axially. The feedstock is heated to the molten or semi-molten state within the hot gas stream. A small diameter nozzle accelerates the particles and gas stream to supersonic velocities and directs them towards the substrate. In HVOF thermal spraying, particle velocities can reach 1000 m/s with jet temperatures of approximately 3000 K ( Ref 63). Coatings produced by HVOF thermal spraying typically present a lower amount of oxidized phases than coatings produced by plasma spray since the temperatures are lower and the particle velocities are higher. The high impact velocity means HVOF thermal spraying can create coatings with higher densities than other thermal spray processes.
Attempting to prevent oxidation of the feedstock, Cheng et al. used an HVOF thermal spray system to produce a ZrB 2 ? 20 vol. % SiC ? 10 vol. % MoSi 2 composite coating ( Ref 95). XRD of the coating showed the presence of no oxide phases. This could be due to the hydrogen/ oxygen ratio used in the combustion, where excess hydrogen (3:1 as opposed to stoichiometric 2:1) created a reducing flame ( Ref 96). The surface of the coating showed poorly bonded particles indicating the feedstock was not fully melted during spraying. Table 3 outlines the spraying systems and parameters used in the APS, CAPS, VPS and HVOF thermal spraying studies discussed in this section. Despite various feedstocks, spraying systems and spraying parameters employed, what is clear is that obtaining a dense, oxide free diboride coating is very difficult to achieve without using vacuums or controlled atmospheres.

High Temperature Properties of UHTC Boride Coatings
Many researchers have attempted to characterize the oxidation mechanisms of boride coatings over the years. In one of the earliest studies, TGA analysis of an LPPS coating by Bartuli Ref 78,85) reported that the same mechanism could be applied to VPS ZrB 2 -MoSi 2 and ZrB 2 -Si coatings, with a thick, protective SiO 2 layer being detected after 6 hours at 1773 K. In comparison, a ZrB 2 -MoSi 2 coating deposited by APS was found to totally fail after 6 hours; the authors suggested this failure was due to increased porosity within the as-sprayed APS coating, meaning a continuous SiO 2 protective layer could not form. The poor oxidation resistance of APS coatings was further characterized in work by Feng et al. (Ref 66). In this study, three ZrB 2 -SiC coatings were deposited using various plasma spray parameters and equipment. Oxidation products were detected after 9 hours at 873 K, with the authors suggesting complete evaporation of B 2 O 3 due to its vapor pressure. While after oxidation at 1273 K, the coatings had totally failed.
The addition of AlN to a ZrB 2 -SiC coating was investigated by Grigoriev et al. (Ref 67). The coating was subjected to a thermocycling test where the sample was heated to *2273 K, held for 2 min and then allowed to air cool for 10 min; this was repeated for 15 cycles. The addition of AlN drastically altered the oxidation mechanism of the coating. The authors reported the formation of an Al 2 SiO 5based solid solution around spheroidal ZrO 2 particles, on top of this a protective SiO 2 -Al 2 O 3 solid solution layer formed, which acted as an effective barrier to the diffusion of O 2 . The authors suggested this coating showed excellent stability above 2173 K and offered more protection than typical UHTC coatings.    One area where ZrB 2 -based coatings have been researched heavily over recent years is to protect carbon-based composites from high temperature oxidation (Ref 99). These composites are ideal for use as high temperature structural components for atmospheric re-entry vehicles due to their excellent high temperature mechanical properties. In use, these components will undergo thermochemical ablation due to oxidation at very high temperatures ([1800°C) and high gas flow rates. However, carbon-based composites will oxidize readily at temperatures above 500°C; thus, protective, oxidation-resistant coatings are required. Due to the high melting points of their oxides (2700 and 2800°C, respectively), Zr-and Hfbased ultra-high temperature ceramics have been the main focus of research, as any liquid phases will be removed by the high gas flow rates, reducing the protection of the underlying component.
As explained in previously, the addition of SiC and other Si containing ceramics to ZrB 2 improves the oxidation resistance of the composite. ZrB 2 -SiC composite coatings have been produced using thermal spraying, and these coatings are explored for use in protecting graphite, carbon/carbon (C/C) and carbon/silicon carbide (C/SiC) composites.   greatly reduced the ablation rates of the coatings, the coating deposited using no shroud had a mass ablation rate of 1857 mg/s while increasing the shroud gas flow rate 150 l/min reduced the ablation rate to 39.3 mg/s. As the shroud gas flow rate was increased, the oxide phase content and porosity of the coating were reduced leading to the greater ablation resistance. The mechanism of ablation from this study is shown in Fig. 10; note how the SiC interlayer also oxidizes and liquid SiO 2 fills the pores created by the oxidation of ZrB 2 .
Using LPPS as the deposition method, Wang et al. (Ref 86) found that the addition of TaSi 2 to a ZrB 2 -SiC composited could effectively reduce the ablation rate. The reasons for the reduction in ablation rate were twofold, the addition of TaSi 2 produced a denser coating, and during ablation, a higher fraction of protective glassy SiO 2 phase was produced, which could fill any pores in the oxide scale and prevent subsequent oxidation.
A summary of the ablation tests conducted on UHTC boride coatings is shown in Table 4 where possible the heat flux, surface temperatures and ablation rates have been reported.

Tribology and Wear of UHTC Boride Coatings
The tribology of bulk UHTC borides has been researched widely (Ref 101-106

UHTC Carbide Coatings
Deposition and Microstructure of UHTC Carbide Coatings Atmospheric Plasma Spraying As with the boride coatings discussed earlier, due to the extreme melting points of UHTC carbides, plasma spraying is the most popular deposition technique. In the 1980s and 1990s, APS TiC coatings were investigated to protect nuclear fusion device components from thermal shock (Ref [107][108][109][110][111][112]. Some of these early coatings suffered from high porosity, oxidation and decarburization (Ref 110, 113, 114).
More recently, a detailed characterization of a TiC APS coating was carried out by Hong et al. (Ref 68, 115). The phases present in the coating were quantified as being 87 wt. % TiC, 9 wt. % TiO 2 (rutile) and 4 wt. % TiO. Porosity was measured at 8.0 %. As with previous studies, the assprayed surface showed melted and un-melted particles while the microstructure was largely dense and well bonded with some microcracks caused by stresses upon cooling. Hardness and elastic modulus were also measured for this coating, 7.7 GPa and 189.7 GPa, respectively. The authors suggested that these mechanical properties were lower than reported for bulk ceramics because of porosity levels, inter-splat strength and phase composition.
Mahade et al. (Ref 116) deposited a TiC feedstock with a median particle size of 2.21 lm using SPS. The XRD diffractogram of the coating showed the main phases were titanium oxycarbide (TiC 0.1 O 0.9 ), TiC and Ti 2 O 3 with smaller peak intensities of TiO 2 (both anatase and rutile). The as-sprayed surface of the coating showed very fine (*3 lm) melted splats and some un-melted particles. The microstructure revealed uniformly distributed porosity, a few un-melted particles and good adhesion between splats, see Fig. 11.
When depositing ZrC coatings with APS, researchers have typically found a small degree of oxidation with ZrC forming monoclinic and tetragonal ZrO 2 with small relative peak intensities relative to ZrC when characterized with XRD (Ref 117-120). Generally, decarburization has been minimal; however, other works have found more severe oxidation of ZrC with relatively large peak intensities of ZrO 2 and other oxidation products detected (Ref 121,122). Interestingly, in a study by Wu et al. (Ref 123), XRD detected small peak intensities of cubic ZrO 2 . Cubic ZrO 2 is formed above 2370°C, whereas between 1170 and 2370°C tetragonal is the stable phase (monoclinic being formed below 1170°C). The presence of this phase could indicate higher temperatures were achieved in the plasma plume using this set of parameters compared to the other studies. The coating microstructures produced in all these studies are similar, with the surface showing a combination of melted and un-melted splats and the cross-sectional microstructure appearing fairly dense with minimal pores; a typical example from Wu et al. is shown in Fig. 12. Fewer studies have investigated APS of HfC coatings, but the results were similar (Ref 124 -126). During spraying, some oxidation of HfC was reported, the microstructures of the coatings were dense, and the as-sprayed surfaces showed some melted and un-melted splats.
Controlled Atmosphere and Vacuum Plasma Spraying As with thermal spraying of most non-oxide ceramics, researchers have turned to spraying in inert atmospheres or vacuums to protect the feedstock from oxidation. A comparison between APS and VPS ZrC coatings was made by Compared to agglomerated powder prepared by spray drying (SD), with the use of IPS a higher degree of melting was observed on the as-sprayed coating surface, porosity was reduced from 10.7 to 4.6 %, and deposition efficiency was increased.
In an early study, Varacelle et al. (Ref 133) investigated the effect of three VPS parameters on TiC coatings, specifically arc current, primary gas flow and secondary gas flow, using a Taguchi style design of experiment. The lowest porosity (0.49 %) and highest hardness (9.4 GPa) were found in the coating deposited using the highest power to gas flow volume ratio, meaning high spray powers and relatively low primary gas flows led to a greater degree of melting of the TiC feedstock, better deposition efficiency and less porosity. In another early study, the effect of Ar and N 2 atmospheres on CAPS TiC coatings was investigated (Ref 134). Minimal differences were noted between the two atmospheres; the microstructures appeared similar, the hardness of the coatings was similar (12.5 GPa for Ar and 12.75 for N 2 ), and the decarburization was minimal in both cases. This led the authors to believe, when spraying TiC in a controlled atmosphere, the cheaper N 2 gas could be used. to deposit a HfC feedstock with a median particle size of 7.08 lm. Due to the density of HfC; the powder had to be further crushed to *200 nm particle size in order to make a stable suspension. Using a suspension with 20 wt. % solid loading a coating of *50 lm was produced. Despite spraying in a vacuum, the XRD diffractogram of the coating presented with large relative peak intensities for HfO 2, which was attributed to oxygen present in the ethanol in which the HfC particles were suspended.
VPS has also been used to deposit TaC and TaC-based composite coatings. Researchers have noted the formation  ) also used HVOF thermal spray to deposit a TiC coating, this time using a suspension of TiC particles between 2 and 3 lm in size in water. This study used three water-based suspensions; one comprised of 20 wt. % of TiC powder, the second containing 20 wt. % milled TiC powder, and the final containing 20 wt. % of the powder with an added dispersant and the pH adjusted in an effort to make a more stable suspension. During spraying all the feedstocks experienced significant oxidation, XRD diffractograms identified the main phases present in all of the coatings as being TiO 2 (rutile and anatase) and TiC. SEM images of the as-sprayed surface also showed a combination of melted and un-melted particles. The microstructure was mainly dense with some carbide pullout and microcracking. The coating produced from the first suspension (TiC powder and water) had the lowest porosity, 1.9 %, and the highest hardness, 5.2 GPa. Table 5 outlines the spraying systems and parameters used in the APS, CAPS, VPS and HVOF thermal spraying studies discussed in this section.

High Temperature Properties of UHTC Carbide Coatings
As with UHTC boride coatings, one area where UHTC carbide coatings have potential applications is in the protection of carbon-containing composites. Thus, the high temperature properties, namely the ablation resistance, of these carbide coatings have been widely researched.
The behavior of ZrC coatings, when subjected to ablation by oxyacetylene torch, has been studied by Wu et al. . Despite different deposition methods, the mechanism of ablation described by the authors was largely similar. In all cases, the only phase detected after ablation testing was monoclinic ZrO 2 . At high temperatures, ZrO 2 will have a tetragonal or even cubic crystal structure, but upon cooling, it will transition to the monoclinic phase; the volume change associated with this phase change has resulted in the formation of cracks after testing, while escaping CO and CO 2 gases due to the oxidation process created pores.
While the mechanism reported in these studies was similar, interestingly, some of the results were different.  A comparison between the ablation resistance of VPS and APS deposited ZrC coatings was made by Hu et al.
( Ref 97). The VPS coating offered better protection to ablation due to its less porous microstructure and lower oxidation during spraying, allowing a dense ZrO 2 layer to be formed during ablation.
In order to improve the ablation resistance of ZrC coatings, many researchers have focused on the additions of other materials to form composites. Similar to UHTC boride coatings, Si-containing materials such as SiC and MoSi 2 are common additives to carbide composite coatings, as it forms a protective SiO 2 layer at high temperature. Jia  Due to extreme temperatures, the ZrC became molten and exposed the ZrC-SiC layer below, causing volatilization of SiO and the formation of many pores on the surface of the coating.
A more thorough investigation into the mechanism by which SiC addition can improve the ablation resistance of ZrC-based coatings was conducted by Jia et al. (Ref 120). In this work, a ZrC composite coating containing 20 vol. % SiC was subjected to ablation testing at three temperatures under a heat flux of 2.4 MW/m 2 . At 2011 K, a glassy SiO 2 phase was formed, encapsulating the ZrO 2 and protecting the structure from further oxidation. When the temperature was increased to 2378 K, SiO 2 evaporated, leaving behind a porous, unprotective ZrO 2 coating, and the linear ablation rate increased to 2.5 lm/s, and the mass ablation rate was 0.49 mg/s. However, as the temperature was increased further to 2543 K, the authors suggest the temperature was high enough for the composite oxide ZrO 2 -SiO 2 to be semimolten, even as SiO 2 was evaporated. The semi-molten phase offers protection from further oxidation and is viscous enough not to be removed mechanically by the gas stream. In one final experiment, the authors increased the heat flux to 4.2 MW/m 2 ; the coating failed completely with the increased heat flux.
The pre-treatment of a ZrC-SiC feedstock was examined using induction plasma spheroidization (IPS) by Pan et al. (Ref 132). A coating made with this feedstock showed lower consumption during ablation testing compared to coating produced with a spray dried (SD) agglomerated feedstock. The authors suggested that this was due to the reduced porosity in the coating produced with the IPS treated feedstock, allowing a dense, protective oxide scale to form.
Liu et al. (Ref 127) compared the ablation resistance of ZrC-SiC, ZrC-MoS i2 and multilayer ZrC-SiC/ZrC-MoSi 2 coatings. Both the single-layer coatings were found to offer insufficient protection. While a protective, liquid SiO 2 layer was formed, which filled pores and bonded ZrO 2 on the surface of the ZrC-SiC coating, this caused a layer underneath to become porous as active oxidation of SiC caused SiO to diffuse towards the surface of the coating. The authors believed this would lead to weakened adhesion between the oxidized coating layers and any remaining material beneath, eventually causing failure of the coating. As for the MoSi 2 containing coating, the build-up of the oxidation product MoO 3 , which unlike other oxidation products CO and CO 2 was unable to pass through the ZrO 2 layer, created a bubble which, when the pressure was high enough, burst and ruptured the coating. In comparison with the single-layer coatings, the multilayer coating performed very well. The outer ZrC-SiC layer was able to form protective SiO 2 , which prevented the formation of destructive MoO 3 in the ZrC-MoSi 2 inner layer. Oxidation of the inner layer produced Si, which was able to diffuse upwards, oxidize and eliminate the porous lower layer seen in the ZrC-SiC coating. Diagrams for all three of these ablation mechanisms are shown in Fig. 17. In another work looking at ZrC-MoSi 2 coatings, by reducing the heat flux from 3.01 to 1.94 MW/m 2 , the authors suggested that MoSi 2 could be a suitable additive for ablation resistance coatings (Ref 147). The rate of SiO 2 evaporation from the surface was lower than the rate of formation of SiO 2 from the oxidation of MoSi 2 . A stable SiO 2 layer in turn would prevent the formation of the destructive MoO 3 species, preserving the coating. As MoSi 2 content was increased from 0 to 20 to 40 vol. %, the mass ablation rate reduced from -2.80 to -0.92 to -0.68 mg/s, respectively.
While some researchers have had success using SiC containing composites, they are limited by how rapid they can be depleted when active oxidation of SiC occurs and SiO 2 vaporizes and leaves behind a porous structure. Instead of SiC, recent research has focused on the addition Table 5 A summary of process parameters for thermal spraying of UHTC carbide ceramics       125,126) found that a single-phase HfC coating was not enough to protect from ablation. Similar to the behavior of ZrO 2 , the authors reported that when the HfC was oxidized, the HfO 2 became porous and loose, allowing oxygen to diffuse into the coating. In these studies, the authors added 10, 20 and 30 vol. % TaC to HfC coatings. Under ablation, the coatings oxidized to form liquid Ta 2 O 5 and solid HfO 2 and Hf 6 Ta 2 O 17 . At 10 vol. % TaC addition, Ta 2 O 5 was able to seal any cracks and pores on the oxide surface, as shown in Fig. 18. As TaC content was increased a composite Ta 2 O 5 -HfO 2 liquid oxide was formed and subsequently   144). Both works found that TaC coatings with SiC additions provided the best protection from ablation. Singlephase TaC coatings oxidized to liquid Ta 2 O 5 , which was removed by the shearing effect of the gas flow. When SiC was added, a Ta 2 O 5 -SiO 2 mixed oxide was formed which had a higher density and could withstand erosion.
A summary of the ablation tests conducted on UHTC boride coatings is shown in Table 6. Where possible, the heat flux, surface temperatures and ablation rates have been reported.

Tribology and Wear of UHTC Carbide Coatings
Due to having the highest hardness of the UHTC carbides, TiC is the most widely researched for wear resistant applications. In fact, it is the only thermal spray coating material of all the UHTC carbides to have its tribological properties investigated thoroughly. Hong et al. (Ref 68) prepared a TiC coating using APS, which was subjected to wear test under 20 and 50 N loads against a WC-Co ball. Giving COF of 0.53 and 0.49 and wear rates of 0.07 x 10 -5 and 2.42 x 10 -5 mm 3 N -1 m -1, , respectively, the wear mechanisms were described as fatigue and tribo-oxidation under both sets of conditions. The TiC coating showed much lower wear rates under both loads than a TiB 2 coating tested under the same conditions. In a further study, the authors tested the same TiC coating against a range of different ball materials under 50 N load; specifically, WC-Co, 304 stainless steel and Si 3 N 4 balls were used ( Ref 115). Against the steel ball, the coating showed a low wear rate of 2.55 x 10 -6 mm 3 N -1 m -1 due to the relative softness of the ball. A COF of 0.65 was attributed to the wear debris of the coating acting as an abrasive and ploughing the softer steel ball; some evidence of adhesive wear was also detected. When tested with the Si 3 N 4 ball, a low COF of 0.46 and wear rate of 9.76 x 10 -6 mm 3 N -1 m -1 were reported, due to the oxidation of the ball to form SiO 2 . The fluctuation of COF was high, however, due to the spallation of this oxide. Due to the high hardness of the WC-Co ball, the wear rate was much higher (2.42 x 10 -5 mm 3 N -1 m -1 ).
The tribological properties of VPS TiC coatings against WC-Co balls were also tested by Guo et al. (Ref 135).
After testing under loads of 20 and 50 N, the authors found that the addition of Mo to the coating reduced the wear rate and COF at both conditions. The added ductility of the Mo also helped change the wear mechanism from particle pullout and fatigue wear to abrasive wear.
An SPS TiC coating was deposited by Mahade et al. (Ref 116); SPS allows the deposition of feedstocks with extremely fine particle sizes, potentially improving wear resistance by reducing splat and pore size. The coating was subjected to a sliding wear test against a WC-Co pin under 5 kgf, which resulted in a 0.2129 mm 3   coated material and tested at three loads, 0.5, 1 and 2 kg loads. The wear rates, however, were high, which was attributed to poor bonding between the TiC and C r2 O 3 .

Reinforced UHTC Coatings
As explored in previous sections, a range of particle reinforcements (SiC, MoSi 2 , etc.) to UHTC coatings have already been investigated by researchers, with the primary aim of improving the high temperature performance. As with many ceramics, however, UHTCs suffer from intrinsic brittleness, which can limit their application. Research into sintered UHTCs over the years has covered various toughening mechanisms that can be incorporated into a UHTC composite, largely focussed on continuous fiber reinforcement with C or SiC fibers (Ref 150

High Entropy UHTC Coatings
Borrowing from previous work on high entropy alloys (HEAs) and highentropy ceramics (HECs), high entropy ultra-high temperature ceramics (HE-UHTCs) have garnered significant interest over the last five years (Ref 164,165  showed that up to 1200°C weight gain was much lower than some of the constituent borides, for example, TiB 2 and ZrB 2 . While HEC and HE-UHTC thermal spray coatings are yet to be developed, HEA coatings have been deposited, using a variety of thermal spray processes, to provide wear, corrosion and oxidation resistance (Ref 169).

Summary
As the next generation of spacecraft and hypersonic flight applications is developed, UHTCs will become materials of great importance because of their high melting points and good mechanical properties. Due to the limitations of current processing methods, only small, simple shaped bulk UHTC components can be formed. To alleviate this problem, UHTC coatings can be employed, and as C-and SiC-based composites become more widely as structural components in aeronautics, protective coatings will be required to protect them from the most extreme of environments. While much work on UHTC coatings has been done outside of the public domain, close collaboration with industrial partners must be sought for future research. Due to the applications UHTC coatings are suited to, this will help produce viable processing conditions that can be achieved on an industrial scale and testing procedures that will represent expected service environments. This paper has presented a detailed review of UHTC coatings produced by various thermal spray processes. Because of the ultra-high melting temperatures, plasmabased thermal spray techniques have been found to be the most popular for depositing UHTC coatings due to the temperatures which can be reached within the plasma plume itself. To prevent oxidation of UHTC feedstocks, spray systems have often been contained within inert atmospheres or vacuums. While successful at eliminating oxide phases within the coatings, such setups remain expensive. To this end, shrouded plasma spray systems have shown promise as a lower cost alternative; however, further development is needed to deposit completely oxidefree coatings.
The oxidation and ablation resistance of UHTC-based coatings has been widely reported, and the mechanisms are largely understood. Various UHTC composite coatings have been investigated as a means to improve oxidation and ablation resistance, and composite coatings with Si containing materials (such as SiC and MoSi 2 ) have proved to be particularly effective at this. Despite widespread research on the tribology of bulk UHTCs, investigations into the wear resistance of UHTC thermal spray coatings have been sporadic. For example, thin film TiB 2 , TiC and Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons. org/licenses/by/4.0/.