Application of High-Velocity Oxygen-Fuel (HVOF) Spraying to the Fabrication of Yb-Silicate Environmental Barrier Coatings

Article Application of High-Velocity Oxygen-Fuel (HVOF) Spraying to the Fabrication of Yb-Silicate Environmental Barrier Coatings

Emine Bakan 1,*, Georg Mauer 1, Yoo Jung Sohn 1, Dietmar Koch 2 and Robert Vaßen 1

1 Forschungszentrum Jülich GmbH, Institute of Energy and Climate Research (IEK-1), 52425 Jülich, Germany; [email protected] (G.M.); [email protected] (Y.J.S.); [email protected] (R.V.) 2 German Aerospace Center, Institute of Structures and Design, 70569 Stuttgart, Germany; [email protected] * Correspondence: [email protected]; Tel.: +49-2461-61-2785

Academic Editor: Yasutaka Ando Received: 22 February 2017; Accepted: 12 April 2017; Published: 18 April 2017

Abstract: From the literature, it is known that due to their glass formation tendency, it is not possible to deposit fully-crystalline silicate coatings when the conventional atmospheric plasma spraying (APS) process is employed. In APS, rapid quenching of the sprayed material on the substrate facilitates the amorphous deposit formation, which shrinks when exposed to heat and forms pores and/or cracks. This paper explores the feasibility of using a high-velocity oxygen-fuel (HVOF) process for the cost-effective fabrication of dense, stoichiometric, and crystalline Yb2Si2O7 environmental barrier coatings. We report our ﬁndings on the HVOF process optimization and its resultant inﬂuence on the microstructure development and crystallinity of the Yb2Si2O7 coatings. The results reveal that partially crystalline, dense, and vertical crack-free EBCs can be produced by the HVOF technique. However, the furnace thermal cycling results revealed that the bonding of the Yb2Si2O7 layer to the Silicon bond coat needs to be improved.

Keywords: environmental barrier coatings (EBC); ytterbium silicate; high-velocity oxygen-fuel (HVOF)

1. Introduction Silicon carbide (SiC) fiber reinforced SiC ceramic matrix composites (CMCs), having a relatively low density and higher temperature capability compared to their metallic superalloy counterparts, are notable materials for the future of gas turbine engine technology. A significant enhancement in gas turbine efficiency leading to lower fuel consumption and a lower emission of NOx are anticipated with the implementation of CMCs in the gas turbine engines [1–5]. However, the lack of durable CMCs in combustion environments interferes with this implementation, as water vapor reacts with the silica scale that forms on the CMCs, leading to the formation of gaseous reaction products such as Si(OH)4 [4,6]. An enhanced recession rate of the CMCs in high pressure and high gas velocity combustion atmospheres reveals the requirement of protective environmental barrier coatings (EBCs) for their long-term use under these conditions [7–9]. The establishment of environmentally stable (i) and well-adhering coatings on the CMC substrate (ii) that have a chemical compatibility within the layers (iii), and that develop minor stresses as a result of thermal expansion mismatch and phase transformations etc; (iv) are desired for this purpose [10]. EBC systems consisting of a silicon bond coat and a rare-earth (RE) silicate (e.g., Yb2Si2O7, Lu2Si2O7) top coat layer were successfully tested in gas turbine engines [10–14]. In these EBC systems, RE-silicates with a high water vapor stability impede the diffusion of oxygen and water vapor to the substrate. In the meantime, the silicon bond coat plays an important role in the oxidation protection

Coatings 2017, 7, 55; doi:10.3390/coatings7040055 www.mdpi.com/journal/coatings Coatings 2017, 7, 55 2 of 12 of the substrate, as well as for the adhesion of the coating on the substrate. While the RE-silicates were shown to be promising, the fabrication of crystalline and single-phase silicate coatings with a dense microstructure has been troublesome when using the conventional plasma spraying process. Modification in the plasma spraying process was thus proposed, suggesting the utilization of a furnace in which the substrate to be deposited is placed and heated up to 1200 ◦C[15]. By this adjustment, the deposition of dense and crystalline RE-silicate coatings was found to be ensured, yet, secondary phases in the coatings as a result of the evaporation of Si-containing species could not be avoided [16]. Furthermore, the feasibility and adaptation of such a process involving a furnace for the big and complex-shaped components are questionable. In a recent study [17], it was reported, based on initial trials, that the HVOF process can be an alternative method to achieve dense Yb2Si2O7 coatings with desired properties. This is because, in contrast with the APS process, HVOF offers a low heat transfer to the particles (flame temperatures up to 3000 ◦C), reducing the molten fraction of the sprayed powders which is expected to be favorable for obtaining crystalline and stoichiometric coatings. Moreover, the high particle velocities achieved in HVOF typically provide well-bonded particles resulting in high-density coatings. To this end, we further investigated the HVOF processing of Yb2Si2O7 coatings and address our findings in this study in terms of process optimization, microstructure, composition, and crystallinity. The crystallinity, as well as the adhesion of the Yb2Si2O7 coatings on the Si bond coat, are discussed in relation to the feedstock particle size and the furnace thermal cycling behavior of a Si/Yb2Si2O7 EBC system is reported.

2. Experimental Procedure

2.1. Materials and Process

Commercially available Si and Yb2Si2O7 powders provided by Oerlikon Metco (US) Inc. (New York, NY, USA) were used in this study. The details about the two kinds of powders are given in Table1. The elemental analysis of the Si powder, determined by inductively coupled plasma optical emission spectroscopy (ICP-OES, TJA-Iris-Intrepid, Thermo Scientific GmbH, Kleve, Germany), revealed 98.7 wt % ± 0.7 wt % purity. The phase composition of the Yb2Si2O7 feedstock was determined with an X-ray diffractometer (XRD, D4 Endeavor, Bruker AXS GmbH, Karlsruhe, Germany) (Cu Kα radiation, operating voltage 40 kV, current 40 mA, step size 0.02◦, step time 0.75 s) over a 2θ-range of 10◦–80◦. The deconvolution of the XRD peaks for quantitative phase analysis was performed using Rietveld analysis (TOPAS Software, Bruker AXS GmbH, Karlsruhe, Germany). The XRD analysis of the Yb2Si2O7 powder, as shown in Figure1, yielded a crystalline structure with the presence of monoclinic Yb2Si2O7 (C2/m, JCPDS No 01-082-0734) and secondary monoclinic Yb2SiO5 (I2/a, JCPDS No 00-040-0386) phases. According to the Rietveld analysis, the powder was found to contain 88 wt % and 12 wt % Yb2Si2O7 and Yb2SiO5, respectively. Diamond Jet 2700 high-velocity oxy-fuel spraying equipment (Oerlikon-Metco, Wohlen, Switzerland) and a three-cathode TriplexPro-210 APS gun (Oerlikon-Metco, Wohlen, Switzerland) manipulated with a six-axis robot (IRB 2400, ABB, Zürich, Switzerland) in a Multicoat facility (Oerlikon-Metco, Wohlen, Switzerland) were used to fabricate the Yb2Si2O7 and Si coatings, respectively. The HVOF burner was fitted with a convergent cylindrical design nozzle (Type 2705) which yields lower particle velocities and thus longer dwell times in comparison to that of a convergent-divergent nozzle [18]. Methane has been used as the fuel gas in the HVOF process. For the optimization of the HVOF spraying conditions, the methane/oxygen ratios and stand-off distances (SOD) were varied. The flow rate of the methane was selected to be 190, 200, 215 slpm, whereas the oxygen flow rate was kept constant, at its maximum possible value of 395 slpm for the equipment. The more detailed spraying parameters used in the present study are listed in Tables2 and3. Ferritic 2 stainless steel (20 × 20 mm ) substrates were used for the initial spraying optimization of the Yb2Si2O7 coatings. The substrates were alumina grit blasted with an average particle size of 45–125 µm using 3 bar pressure for 20 s to reach an average roughness of 3.9 ± 0.1 µm. For the calculation of the area 2 specific weight (g/m ) of the Yb2Si2O7 coatings, the area of each substrate was measured and the Coatings 2017, 7, 55 3 of 12 weight of each sample was recorded before and after spraying (using five spraying cycles). In order to evaluate the splat morphologies, single splats of the Yb2Si2O7 were produced via a wipe test on single-side polished silicon wafers. The surface morphology of the splats was afterward measured Coatings 2017, 7, 55 3 of 12 by a 3D laser confocal microscope (Keyence VK-9700, Keyence, Osaka, Japan). For furnace cycling, SiC/SiCNIn CMC order substrates to evaluate providedthe splat morphologies, by DLR Stuttgart single splats [19] of were the Yb coated2Si2O7 withwere produced the optimized via a wipe Si/Yb 2Si2O7 EBC system.test on single‐side polished silicon wafers. The surface morphology of the splats was afterward measured by a 3D laser confocal microscope (Keyence VK‐9700, Keyence, Osaka, Japan). For furnace Tablecycling, 1. Manufacturing SiC/SiCN CMC method substrates and provided particle by size DLR of Stuttgart the feedstock [19] were powders coated with measured the optimized with laser Si/Yb2Si2O7 EBC system. diffraction (LA-950-V2, Horiba Ltd., Tokyo, Japan). Table 1. Manufacturing method and particle size of the feedstock powders measured with laser diffraction (LA‐950‐V2, Horiba Ltd., Tokyo, Japan). Particle Size (µm) Powder Manufacturing Method d Particle Size (dμm) d Powder Manufacturing Method 10 50 90 d10 d50 d90 SiSi Fused-crushedFused‐crushed 2828 40 40 59 59 Yb2Si2O7 Yb2Si2O7 Agglomerated-sinteredAgglomerated‐sintered 2222 34 34 52 52

Figure 1. X‐ray diffractogram of the Yb2Si2O7 feedstock. Figure 1. X-ray diffractogram of the Yb2Si2O7 feedstock.

Table 2. HVOF spray parameters for processing Yb2Si2O7 coatings. Table 2. HVOF spray parameters for processing Yb2Si2O7 coatings. Number Oxygen Methane Air Powder Spray Robot Substrate Equivalence of Spray Flow Flow Flow Feed Distance Velocity Temperature Ratio (φ) * Number SetsOxygen (slpm)Methane (slpm)Equivalence Air (slpm) PowderRate (%) (mm)Spray (mm/s) Robot (°C) Substrate of Spray #1Flow 395 Flow 190 Ratio 0.96 Flow 250 Feed Rate20 Distance350 1200 Velocity135 Temperature ◦ Sets (slpm)#2 395(slpm) 200 (φ)* 1.01 (slpm) 250 (%)20 350(mm) 1200 (mm/s) 150 ( C) #1#3 395 395 190215 0.961.08 250250 2020 350 350 1200 1200155 135 #2#4 395 395 200200 1.011.01 250250 2020 375 350 1200 1200130 150 #3#5 395 395 215200 1.081.01 250250 2020 400 350 1200 1200125 155 #4 395 200 1.01* φ = (fuel/oxygen)/(fuel/oxygen)stoichiometric. 250 20 375 1200 130 #5 395 200 1.01 250 20 400 1200 125 * φ Table= (fuel/oxygen)/(fuel/oxygen)stoichiometric. 3. APS spray parameters for Si deposition. Spray Current Ar Flow Rate Spray Distance Robot Velocity Substrate Feed Rate (%) (A) (slpm) (mm) (mm/s) Temperature (°C) 450 50Table 3. APS spray30 parameters100 for Si deposition.500 200

Spray Current2.2. CalculationsAr Flow of Gas Rate TemperatureFeed and Rate Velocity inSpray the HVOF Distance Nozzle Robot Velocity Substrate (A) (slpm) (%) (mm) (mm/s) Temperature (◦C) Chemical Equilibrium and Applications (CEA) [20] software was used to calculate the gas 450temperature and 50 velocity in the nozzle 30 throat using an 100 infinite‐area combustion 500 chamber model. 200 In this one‐dimensional model, the combustion conditions are obtained with the assumption of the 2.2. Calculationschemical of equilibrium Gas Temperature of the andcombustion Velocity products. in the HVOF Further Nozzle details of the model’s assumptions, Chemical Equilibrium and Applications (CEA) [20] software was used to calculate the gas temperature and velocity in the nozzle throat using an inﬁnite-area combustion chamber model. In this one-dimensional model, the combustion conditions are obtained with the assumption of the chemical equilibrium of the combustion products. Further details of the model’s assumptions, parameters, and procedures for obtaining the combustion and throat conditions can be found in the report of Gordon and McBride [20]. Coatings 2017, 7, 55 4 of 12

The calculations were made for three different oxygen/methane ﬂow rate combinations (#1, #2 and #3 in Table2). The individual nozzle entry pressure input for each oxygen/methane combination was located by iteration, according to the condition of the constant mass ﬂow rate (m˙ , kg/s),

. m = ρAu (1) where ρ, A, and u are the gas density (kg/m3), nozzle cross-sectional area (m2), and gas velocity (m/s), . respectively. m of the gas mixture and A of the nozzle throat (∅ = 7 mm) were known in this equation, while ρ and u of the gas mixture at the throat of the nozzle were calculated by the software for a given chamber pressure. After the nozzle entry pressures were correctly located for each mixture, the gas velocity and temperatures were ﬁnally calculated at the throat of the nozzle, which corresponds to the nozzle exit conditions.

2.3. Characterization of the Coatings The microstructures of the as-sprayed and thermally cycled coatings were analyzed using a scanning electron microscope (SEM, TM3030, Hitachi High-Technologies, Chiyoda, Tokyo, Japan) coupled with an energy dispersive X-ray spectrophotometer (EDX, Quantax70, Bruker Nano GmbH, Berlin, Germany); secondary electron (SE) and backscattered electron (BSE) signals were collected. Acquired BSE-SEM images were also employed to assess the volume fractions of pores in the coatings by image analysis, using an image thresholding procedure with the analySIS pro software (Olympus Soft Imaging Solutions GmbH, Münster, Germany). The analysis was performed on 10 SEM micrographs (×2000 magnification) per sample, each having a resolution of 1280 × 1100 pixels and covering a horizontal field width of 126 µm. A metallographic cross-section of an as-sprayed coating was also investigated with FEG-SEM (LEO Gemini 1530 from Carl Zeiss NTS GmbH, Oberkochen, Germany) equipped with an electron backscatter diffraction (EBSD) detector (NordlysNano, Oxford Instruments, Oxfordshire, UK) to demonstrate the amorphous and crystalline areas, as well as to generate a phase map of the crystalline regions. The phase composition of the coatings was determined with the same measurement parameters for the feedstock given above (D4 Endeavor, Bruker AXS GmbH, Karlsruhe, Germany). The PONKCS (partial or no known crystal structure) method was used to determine the amorphous content of the as-sprayed coatings using the XRD data, as described in the previous work [17]. Further details about the procedure can also be found in [21]. The roughness (arithmetic mean deviation of the surface, Ra) of the substrates and the as-sprayed coatings was measured with an optical profilometer using a chromatic white light sensor (CHR 1000) (CyberScan CT 300, CyberTechnologies, Ingolstadt, Germany). An evaluation of the roughness measurements was made in accordance with the recommendations of the standards DIN EN ISO 4288 [22] and DIN EN ISO 3274 [23]. Accordingly, the cut-off wavelength (λc) was selected, depending on the expected roughness values. At the same time, the evaluation length (In) and the corresponding traverse length (It) were defined, according to the standards.

2.4. Furnace Thermal Cycling Test

The Si/Yb2Si2O7 EBCs deposited on the SiC/SiCN substrates were placed in a mufﬂe furnace and heated in ambient atmosphere to 1200 ◦C, with a rate of 10 K/min. The sample was held for 20 h at this high temperature and then cooled to room temperature with a rate of 10 K/min, and this thermal cycle was repeated four times.

3. Results and Discussion

3.1. Effect of Methane/Oxygen Flow Rates and Stand-Off Distance For the assessment of the effect of the chosen oxygen/methane ﬂow rates (#1, #2, and #3 in Table2), the area speciﬁc weight (ASW) and average roughness ( Ra) of the coatings were evaluated. Coatings 2017, 7, 55 5 of 12

Coatings 2017, 7, 55 5 of 12 Coatings 2017, 7, 55 5 of 12 A correlation was found between the ASW and the roughness values of the as-sprayed coatings; correlation was found between the ASW and the roughness values of the as‐sprayed coatings; the the maximum ASW was calculated for the coating with the minimum Ra and vice versa, and was maximumcorrelation ASWwas found was calculatedbetween the for ASW the coatingand the withroughness the minimum values of Rthea and as‐ sprayedvice versa, coatings; and was the employedemployedmaximum as as anASW an indirect indirectwas calculated method method for for optimization optimizationthe coating with(Figure the 22a,b). a,b).minimum It It is is reasonable reasonableRa and vice to to ascribeversa, ascribe andthe the highwas high ASW—lowASW—lowemployed roughness as roughness an indirect combination combination method for of of theoptimization the coating coating to (Figure the well well-molten, 2a,b).‐molten, It is flattened, reasonable ﬂattened, and andto bonded ascribe bonded particles,the particles, high whichwhichASW—low can can be be translated roughness translated to combination to the the high high particle ofparticle the coating temperature temperature to the well and and‐ velocitiesmolten, velocities flattened, delivered delivered and by bonded the combustion particles, of theofwhich fuel/oxygen the fuel/oxygencan be translated mixture. mixture. The to the particlesThe high particles particle will will otherwise temperature otherwise remain andremain invelocities the in non-molten/partially-moltenthe nondelivered‐molten/partially by the combustion‐molten state andstateof they the and willfuel/oxygen they either will bounce-off eithermixture. bounce The the ‐particlesoff substrate the substrate will resulting otherwise resulting in a remain reduction in a reduction in the in non the in‐ ASW molten/partiallythe ASW or will or will stick‐ stickmolten to it, to but theit,state latter but and the will theylatter lead will will to either a lead rougher bounceto a rougher surface‐off the ﬁnish.surface substrate For finish. easeresulting For of ease understanding, in ofa reductionunderstanding, in a comparison the ASWa comparison or will of the stick of smooth the to andsmoothit, rough but the and surface latter rough will morphologies surface lead to morphologies a rougher of HVOF surface of sprayed HVOF finish. sprayed molten For ease molten and of understanding, non-molten/partially and non‐molten/partially a comparison molten molten of singlethe smooth and rough surface morphologies of HVOF sprayed molten and non‐molten/partially molten Yb2singleSi2O7 Ybsplats,2Si2O7 respectively, splats, respectively, is given is in given Figure in Figure3a,b. 3a,b. single Yb2Si2O7 splats, respectively, is given in Figure 3a,b.

(a) (b) (a) (b) Figure 2. ASW and roughness of the HVOF sprayed Yb2Si2O7 coatings as a function of the total inlet Figure 2. ASW and roughness of the HVOF sprayed Yb2Si2O7 coatings as a function of the total inlet gasFigure flow 2. ( aASW) and and (b) theroughness stand‐off of distance.the HVOF Error sprayed bars ofYb the2Si2 ASWO7 coatings values as calculated a function from of the the total standard inlet gas ﬂow (a) and (b) the stand-off distance. Error bars of the ASW values calculated from the standard deviationgas flow ( aof) and four ( bsubstrate) the stand area‐off measurementsdistance. Error andbars errorof the bars ASW of values the roughness calculated show from thethe standard deviation of four substrate area measurements and error bars of the roughness show the standard deviation ofof 10four profile substrate measurements area measurements (λc = 2.5 mm). and error bars of the roughness show the standard deviation of 10 proﬁle measurements (λ = 2.5 mm). deviation of 10 profile measurements (λc c = 2.5 mm).

(a) (b) (a) (b) Figure 3. 3D‐surface morphology of HVOF‐sprayed (a) molten (φ = 1.01) and (b) non‐molten (φ = Figure 3. 3D‐surface morphology of HVOF‐sprayed (a) molten (φ = 1.01) and (b) non‐molten (φ = Figure0.96) 3. Yb3D-surface2Si2O7 single morphology splats on Si ofwafer. HVOF-sprayed The images were (a) molten produced (φ = using 1.01) a and confocal (b) non-molten laser microscope. (φ = 0.96) 0.96) Yb2Si2O7 single splats on Si wafer. The images were produced using a confocal laser microscope. Yb2Si2O7 single splats on Si wafer. The images were produced using a confocal laser microscope. Based on these assumptions, it can be interpreted from Figure 2a that the maximum particle temperatureBased on and these velocities assumptions, were reached it can beat theinterpreted total gas fromflow Figureof 595 slpm,2a that while the maximumthe increased particle total flowtemperatureBased rate on(610 theseand slpm) velocities assumptions, induced were a reduction. reached it can be atThe interpretedthe equivalence total gas fromflow ratio of Figure of595 the slpm,2 fuel/oxygena that while the the maximum mixtures, increased which particle total temperatureisflow described rate (610 and asslpm) velocitiesthe inducedactual were fuel/oxygena reduction. reached Theratio at the equivalence divided total gas by ratio ﬂowthe of stoichiometric of the 595 fuel/oxygen slpm, whilefuel/oxygen mixtures, the increased which ratio total(isφ ﬂowdescribed= (fuel/oxygen)/(fuel/oxygen) rate (610 as the slpm) actual induced fuel/oxygenstoichiometric a reduction., fuelratio rich Thedivided if equivalence φ > by1, fuelthe lean ratiostoichiometric if φ of the< 1 and fuel/oxygen fuel/oxygen stoichiometric mixtures, ratio if whichφ (φ == is 1)(fuel/oxygen)/(fuel/oxygen) described is also noted as the on actual the graph, fuel/oxygenstoichiometric indicating, fuel ratio rich that dividedif the φ >selected 1, byfuel the leanfuel/oxygen stoichiometric if φ < 1 andratios stoichiometric fuel/oxygen are close to ratioifa (φ =φ (fuel/oxygen)/(fuel/oxygen)= 1) is also noted on the graph, indicating, fuel that rich the if selectedφ > 1, fuel fuel/oxygen lean if φ

Coatings 2017, 7, 55 6 of 12 to a stoichiometric mixture. The stoichiometric mixture (φ = 1) is suggested by the combustion thermodynamicsstoichiometric mixture. to obtain The a maximumstoichiometric adiabatic mixture flame (φ = temperature 1) is suggested from by the the combustion combustion of a hydrocarbonthermodynamics fuel at to a givenobtain chamber a maximum pressure adiabatic [24]. Furthermore,flame temperature the relevance from the of combustion the stoichiometric of a fuel/oxygenhydrocarbon ratios fuel (or at a slightly given chamber fuel-rich pressure mixtures) [24]. to Furthermore, the maximum the particle relevance temperature of the stoichiometric and velocities infuel/oxygen the HVOF processratios (or has slightly been fuel revealed‐rich mixtures) in the to literature the maximum [25–27 particle]. On temperature the other hand, and velocities no general agreementin the HVOF was foundprocess regarding has been therevealed effect in of the totalliterature inlet [25–27]. gas flow On on the the other particle hand, temperature no general and velocities.agreement Zhao was et found al. and regarding Turunen the et effect al. [28 of,29 the] reported total inlet an gas increase flow on in the both particle the particle temperature temperature and andvelocities. velocities Zhao with et increasingal. and Turunen total et gas al. flow [28,29] rates reported associated an increase with higherin both burningthe particle enthalpy temperature and gas expansion.and velocities However, with increasing Picas et al.total [26 gas] showed flow rates a decrease associated in with particle higher temperatures burning enthalpy when and fuel gas rates exceededexpansion. a certain However, level. Picas Moreover, et al. [26] Guo showed et al. a [ 30decrease], who in worked particle with temperatures methane fuelwhen similar fuel rates to the currentexceeded study, a recordedcertain level. increasing Moreover, particle Guo velocitieset al. [30], with who rising worked fuel with flow methane rates and fuel yet similar they found to the that thecurrent particle study, temperatures recorded graduallyincreasing reducedparticle velocities at the same with conditions. rising fuel Theflow decreasingrates and yet temperatures they found of thethat particles the particle at higher temperatures fuel flow rates gradually were thus reduced attributed at the to their same reducing conditions. dwell The time decreasing in the flame, temperatures of the particles at higher fuel flow rates were thus attributed to their reducing dwell which can also explain the abovementioned changes in the ASW and roughness values at the highest time in the flame, which can also explain the abovementioned changes in the ASW and roughness total flow rate. To further check this, the gas temperature and velocities were calculated for the chosen values at the highest total flow rate. To further check this, the gas temperature and velocities were constant oxygen and increasing methane flow rates (#1, #2, and #3 in Table2) in this study and the calculated for the chosen constant oxygen and increasing methane flow rates (#1, #2, and #3 in Table results2) in are this shown study inand Figure the results4. Accordingly, are shown the in Figure highest 4. gas Accordingly, velocity and the the highest highest gas gas velocity temperature and the are foundhighest at the gas highest temperature total gas are flow found (φ at= the 1.08). highest It can total be expected gas flow that(φ = the1.08 temperature). It can be expected would dropthat the again if φtemperaturewas further would increased. drop again Considering if φ was thatfurther the increased. optimum Considering roughness andthat the ASW optimum values roughness are obtained applyingand ASW a gas values mixture are obtained of an almost applying stoichiometric a gas mixture composition of an almost (φ = stoichiometric 1.01) in this work, composition it is clear that(φ the= 1.01 maxima) in this of work, the roughness it is clear and that ASW the maxima are seen of at the specific roughness particle and velocities ASW are andseen temperatures, at specific andparticle beyond velocities these, begin and totemperatures, decrease. Thus, and thebeyond highest these, temperature begin to anddecrease. velocity Thus, values the arehighest not the optimumtemperature parameters and velocity in this values respect. are not the optimum parameters in this respect.

FigureFigure 4. 4.Calculated Calculated gas gas temperature temperature andand velocities in in the the throat throat of of the the nozzle nozzle at atdifferent different total total gas gas ﬂowflow rates. rates. The The oxygen oxygen ﬂow flow is is constant constant atat 395395 slpm.

Using this optimum fuel/oxygen mixture, the Yb2Si2O7 coatings were also sprayed at different standUsing‐off thisdistances optimum (SODs) fuel/oxygen (Figure 2b). mixture, The SOD the < 350 Yb mm2Si2 yieldedO7 coatings lower were ASW also values sprayed of the atcoatings different stand-offand thus distances these results (SODs) are (Figure not shown2b). in The the SOD graph. < 350 Furthermore, mm yielded almost lower no ASW difference values was of theobserved coatings andby thus increasing these results the SOD are from not shown 350 mm in to the 375 graph. mm. Furthermore,However, a slight almost decrease no difference in the ASW was observedand a bysimultaneous increasing the increase SOD in from the roughness 350 mm to were 375 obtained mm. However, at 400 mm a SOD. slight The decrease effect of in the the SOD ASW will andbe a simultaneousfurther discussed increase below. in the roughness were obtained at 400 mm SOD. The effect of the SOD will be further discussed below. 3.2. Microstructure, Crystallinity and Phase Composition 3.2. Microstructure, Crystallinity and Phase Composition Figure 5a–e shows the low magnification cross‐section microstructures of five Yb2Si2O7 coatings

sprayedFigure with5a–e different shows the total low flow magnification rates and SODs, cross-section and the microstructuresmeasured porosity of fiveof these Yb 2coatingsSi2O7 coatings via sprayedimage withanalysis different are given total in. flow The ratesmicrostructures and SODs, of and all coatings the measured were found porosity to be of fairly these dense coatings and via imagesimilar. analysis Nevertheless, are given the in. lowest The microstructuresporosity content was of all measured coatings at were φ = 1.01 found with to 350 be mm fairly SOD dense (#2). and

Coatings 2017, 7, 55 7 of 12

φ similar.Coatings Nevertheless, 2017, 7, 55 the lowest porosity content was measured at = 1.01 with 350 mm7 of SOD 12 (#2). This finding confirms the prior discussions regarding the highest ASW and the lowest roughness recordedThis due finding to well-molten confirms the particles prior discussions with the regarding same spray the parameters.highest ASW On and the the other lowest hand, roughness the highest porosityrecorded was measured due to well‐ atmoltenφ = 1.01 particles with with 400 the mm same SOD spray (#5). parameters. Obviously, On the particles other hand, start the to highest significantly cool down/slowporosity was downmeasured beyond at φ = 3751.01 mmwith and400 mm more SOD particles (#5). Obviously, bounce particles off the substrate.start to significantly Supportively, cool down/slow down beyond 375 mm and more particles bounce off the substrate. Supportively, Figure5e reveals a thinner coating with a rougher surface finish due to the presence of a high amount Figure 5e reveals a thinner coating with a rougher surface finish due to the presence of a high amount of non-moltenof non‐molten or partially or partially molten molten particles particles in in the the microstructure.microstructure.

(a)

(b)

(c)

(d)

(e)

Figure 5. Microstructure of HVOF sprayed Yb2Si2O7 coatings with the spray settings of (a) #1; (b) #2; Figure 5. Microstructure of HVOF sprayed Yb2Si2O7 coatings with the spray settings of (a) #1; (b) #2; (c) #3; (d) #4; (e) #5. (c) #3; (d) #4; (e) #5.

Table 3. Measured porosity and amorphous content of the Yb2Si2O7 coatings.

Figure6 showsNumber the of XRD Spray patterns Sets of thePorosity HVOF (%) sprayed ± SD YbAmorphous2Si2O7 deposits. Content The(wt %) indexed peaks in the patterns suggest the#1 presence of monoclinic8.0 Yb ± 0.82Si 2O7 (C2/m, JCPDS72 No ± 5 01-082-0734) and Yb2SiO5 (I2/a, JCPDS No 00-040-0386)#2 phases in the feedstock.7.7 ± 0.8 The broad humps73 at ± 2 5θ angle ranges of 24◦–38◦ and 40◦–70◦ imply that the#3 as-sprayed deposit8.3 contains ± 0.6 amorphous phases,69 ± as5 expected and discussed #4 8.1 ± 0.5 70 ± 5 in the previous study [17#5]. A quantitative comparison9.1 ± 1.2 of the amorphous65 ± contents 5 of the as-sprayed coatings determined by the PONKCS method is shown in Table4. The results were found to be quite similar to eachFigure other 6 shows within the XRD the estimatedpatterns of the margin HVOF of sprayed error of Yb 5%,2Si2O as7 deposits. well as higherThe indexed than peaks in the in earlier work (48%the patterns amorphous suggest content the presence at 325 of mmmonoclinic SOD [ Yb17]).2Si2 ThisO7 (C2/m, is probably JCPDS No due 01‐ to082 the‐0734) smaller and Yb particle2SiO5 size distribution(I2/a, JCPDS of the No feedstock 00‐040‐0386) in the phases current in the work, feedstock. as well The broad as the humps longer at SOD, 2θ angle which ranges is selectedof 24°–38° to be a and 40°–70° imply that the as‐sprayed deposit contains amorphous phases, as expected and discussed minimum of 350 mm to obtain higher ASWs. in the previous study [17]. A quantitative comparison of the amorphous contents of the as‐sprayed Thecoatings crystalline/non-crystalline determined by the PONKCS areas method in the is coatingsshown in Table were 4. also The investigated results were found by an to EBSD be quite analysis. The analysis was performed on the cross-section of the HVOF sprayed coating #2 as an example (Figure 7a,b). The non-molten/partially molten particles embedded in a molten and well-bonded coating matrix are observed in Figure7a, and Figure7b shows a phase map of the same coating area. Coatings 2017, 7, 55 8 of 12 CoatingsCoatings2017, 20177, 55, 7, 55 8 of 12 8 of 12 similar to each other within the estimated margin of error of 5%, as well as higher than in the earlier similar to each other within the estimated margin of error of 5%, as well as higher than in the earlier work (48% amorphous content at 325 mm SOD [17]). This is probably due to the smaller particle size The latterwork clearly(48% amorphous demonstrates content that at the325 moltenmm SOD matrix [17]). This is non-crystalline is probably due and to the hence smaller yields particle no diffraction size distribution of the feedstock in the current work, as well as the longer SOD, which is selected to be a (blackdistribution regions), of whereas the feedstock the non-molten in the current particles work, as preservedwell as the longer their crystallinity.SOD, which is selected The Yb 2toSi be2O a7 and minimum of 350 mm to obtain higher ASWs. Yb2SiOminimum5 phases of were350 mm indexed to obtain for higher these ASWs. non-molten particles, supporting the XRD analysis results.

Figure 6. XRD patterns of HVOF deposited coatings with the given process parameters (#1, #2, #3, #4, #5). FigureFigure 6. 6.XRD XRD patterns patterns of of HVOF HVOF deposited deposited coatingscoatings with the the given given process process parameters parameters (#1, (#1, #2, #2, #3, #3, #4, #4, #5). #5). The crystalline/non‐crystalline areas in the coatings were also investigated by an EBSD analysis. Table 4. Measured porosity and amorphous content of the Yb Si O coatings. TheThe analysis crystalline/non was performed‐crystalline on the areas cross in‐section the coatings of the wereHVOF also sprayed investigated2 coating2 7 by #2 an as EBSD an example analysis. The(Figure analysis 7a,b). was The performed non‐molten/partially on the cross molten‐section particles of the embeddedHVOF sprayed in a moltencoating and #2 wellas an‐bonded example ± (Figurecoating 7a,b).Number matrix The are of non Sprayobserved‐molten/partially Sets in Figure 7a, Porosity molten and Figure (%) particles 7bSD shows embedded a phase Amorphous mapin a ofmolten the Content same and coating (wt well %)‐ bondedarea. coatingThe latter matrix clearly are #1observed demonstrates in Figure that 7a, the and molten 8.0 Figure± 0.8 matrix 7b shows is non a phase‐crystalline map of 72and the± 5hencesame coatingyields noarea. Thediffraction latter clearly (black #2demonstrates regions), whereas that the nonmolten 7.7‐molten± 0.8 matrix particles is non preserved‐crystalline their 73 and± crystallinity.5 hence yields The no ± ± diffractionYb2Si2O7 and(black Yb #32regions),SiO5 phases whereas were indexedthe non 8.3 for‐molten these0.6 nonparticles‐molten preserved particles, 69 theirsupporting5 crystallinity. the XRD The #4 8.1 ± 0.5 70 ± 5 Ybanalysis2Si2O7 and results. Yb2 SiO5 phases were indexed for these non‐molten particles, supporting the XRD #5 9.1 ± 1.2 65 ± 5 analysis results.

(a) (b)

Figure 7. EBSD analysis(a results) of the Yb2Si2O7 coating (#2): (a) Forescatter(b diode) (FSD) image and (b) phase map. Figure 7. EBSD analysis results of the Yb2Si2O7 coating (#2): (a) Forescatter diode (FSD) image and Figure 7. EBSD analysis results of the Yb2Si2O7 coating (#2): (a) Forescatter diode (FSD) image and 3.3.(b Adhesion) phase map. of HVOF Sprayed Top Coat on Si Bond Coat (b) phase map. The Si coatings with a thickness of 60 μm were sprayed by the APS with the given parameters 3.3. Adhesion of HVOF Sprayed Top Coat on Si Bond Coat 3.3. Adhesionin Table of3 on HVOF SiC/SiCN Sprayed CMC Top substrates. Coat on Si Afterward, Bond Coat the Yb2Si2O7 was sprayed using the HVOF parametersThe Si coatings (#2) on withthe Si a bondthickness coat andof 60 Figure μm were 8a shows sprayed the crossby the‐section APS with micrograph the given of thisparameters trial. The Si coatings with a thickness of 60 µm were sprayed by the APS with the given parameters in ATable decreased 3 on thicknessSiC/SiCN of CMC the Si bondsubstrates. coat and Afterward, the Yb2Si2O the7 coated/uncoated Yb2Si2O7 was patchessprayed on using top of itthe clearly HVOF inparameters Tablesuggest3 on that SiC/SiCN(#2) the on sprayed the CMCSi Ybbond2Si substrates.2 Ocoat7 particles and Figure Afterward, eroded 8a the shows brittle the the Yb Si layer.cross2Si2O ‐Asection7 depositionwas sprayedmicrograph could using be ofachieved this the trial. HVOF parametersA decreasedlater by (#2)reducing thickness on the the Siof particle the bond Si bondsize coat of coat and the and Figurefeedstock, the8 Yba shows 2dueSi2O to7 coated/uncoated the ease cross-section of melting patches micrographfine powders, on top of of whichit this clearly trial. A decreasedsuggest that thickness the sprayed of Yb the2Si Si2O bond7 particles coat eroded and the the Yb brittle2Si2O Si7 coated/uncoatedlayer. A deposition patchescould be onachieved top of it clearlylater suggestby reducing that the the particle sprayed size Yb 2ofSi 2theO7 feedstock,particles erodeddue to the ease brittle of Simelting layer. fine A deposition powders, couldwhich be achieved later by reducing the particle size of the feedstock, due to the ease of melting ﬁne powders, which results in a superior wetting behavior (Figure8b). A higher density of the coating is also evident Coatings 2017,, 7,, 55 9 of 12 results in a superior wetting behavior (Figure 8b). A higher density of the coating is also evident from fromthe microstructure. the microstructure. However, However, as shown as shown with withthe XRD the XRDpatterns patterns in Figure in Figure 8c, reducing8c, reducing the particle the particle size sizeis also is alsoaccompanied accompanied by a by decrease a decrease in the in the coating coating crystallinity. crystallinity. According According to to the the PONCKS PONCKS analysis, analysis, this coatingcoating has has a morea more than than 90% 90% amorphous amorphous content content and yet and no yet vertical no vertical crack formation crack formation was observed was inobserved the as-sprayed in the as state,‐sprayed as well state, as the as afterwell thermalas the after cycling, thermal as shown cycling, in Figure as shown9a. This in Figure behaviour 9a. wasThis found to be different to that of the APS deposited highly amorphous Yb Si O coatings, as they are behaviour was found to be different to that of the APS deposited highly amorphous2 2 7 Yb2Si2O7coatings, verticallyas they are cracked, vertically even cracked, after the even deposition after the (such deposition a microstructure (such a microstructure can be seen in can [17 be]). seen The cracking in [17]). inThe the cracking APS coating in the is attributedAPS coating to itsis higherattributed thermal to its expansion higher thermal (particularly expansion if the (particularly Yb2SiO5 is present) if the than the substrate, which results in the development of the tensile stresses in the coating during cooling Yb2SiO5 is present) than the substrate, which results in the development of the tensile stresses in the ◦ fromcoating high during deposition cooling temperatures from high (abovedeposition 500 temperaturesC) to room temperature. (above 500 Possible °C) to room reasons temperature. preventing thePossible similar reasons crack formationpreventing in the HVOFsimilar sprayed crack formation coatings in in the the as HVOF sprayed sprayed state and coatings after the in thermal the as cyclingsprayed are: state (i) and a higher after the porosity thermal in thecycling HVOF are: coatings (i) a higher than porosity the APS in coatings the HVOF (in the coatings range than of 5–8% the higherAPS coatings in the former, (in the based range on of the 5%–8% image analysis),higher in whichthe former, provides based the HVOFon the coatings image analysis), with some which strain ◦ tolerance;provides the (ii) HVOF lower HVOFcoatings deposition with some temperatures strain tolerance; (150–170 (ii) lowerC) which HVOF diminishes deposition the temperatures magnitude of(150–170 stresses °C) due which to the diminishes reduced cooling the magnitude rates; (iii) of the stresses possibly due lower to the content reduced of Yb cooling2SiO5 rates;with a (iii) higher the thermal expansion in the HVOF coatings than the APS coatings due to lower deposition temperatures, possibly lower content of Yb2SiO5 with a higher thermal expansion in the HVOF coatings than the whichAPS coatings hinders due the evaporationto lower deposition of Si-bearing temperatures, species; and which ﬁnally hinders (iv) the the possible evaporation compressive of Si‐bearing stress growthspecies; in and the finally HVOF (iv) coatings the possible as a result compressive of the peening stress effect growth of the in particles the HVOF on coatings the substrate as a orresult on the of previouslythe peening deposited effect of the layer. particles on the substrate or on the previously deposited layer. Nevertheless, although the HVOF coating remained attached to the bond coat after cycling and built no vertical crackscracks (Figure(Figure9 a),9a), in in a a few few locations, locations, where where a particularlya particularly loose loose non-molten non‐molten particle particle is inis in contact contact with with the the bond bond coat, coat, partial partial delaminations delaminations were were observed observed (Figure (Figure9c). 9c). It is It probable is probable that that the reducedthe reduced interfacial interfacial bonding bonding strength strength by the by presence the presence of these of these loosely loosely bonded bonded particles particles eases eases the crack the propagationcrack propagation and hence and hence the delamination. the delamination.

(a) (b)

(c)

Figure 8. (a,b) Backscattered SEM micrographs of EBC cross‐sections. (a) HVOF sprayed Yb2Si2O7 (#2) Figure 8. (a,b) Backscattered SEM micrographs of EBC cross-sections. (a) HVOF sprayed Yb2Si2O7 (#2) on Si bond coat; (b) HVOF sprayed Yb2Si2O7 (#2) using reduced particle size (d10 = 17 μm, d50 = 25 μm, on Si bond coat; (b) HVOF sprayed Yb2Si2O7 (#2) using reduced particle size (d10 = 17 µm, d50 = 25 µm, d90 = 34 μm) on Si bond coat. (c) Comparison of the XRD patterns of the Yb2Si2O7 coatings deposited d90 = 34 µm) on Si bond coat. (c) Comparison of the XRD patterns of the Yb2Si2O7 coatings deposited using same spray parameters (#2) but different particleparticle sizes.sizes. (#2) and (#2) * denote original and reduced particle sizes, respectively.

Coatings 2017, 7, 55 10 of 12 Coatings 2017, 7, 55 10 of 12

(a) (c)

(b) (d)

Figure 9. Backscattered SEM micrographs of EBC cross‐section after five cycles between 1200 °C and Figure 9. Backscattered SEM micrographs of EBC cross-section after ﬁve cycles between 1200 ◦C and room temperature with 20 h high temperature dwell time. Images ( a,,c)) are taken from the different regions of the same sample. White rectangles in ( a,c) indicate indicate the the region region of high high-magniﬁcation‐magnification images shown below (b,d), in which oxidation of the silicon coating at the Si/Yb2Si2O7 interface as well as on shown below (b,d), in which oxidation of the silicon coating at the Si/Yb2Si2O7 interface as well as on the pore surfaces is clearly visible via darker imageimage contrast.contrast.

4. Conclusions 4. Conclusions The HVOF process was used for the deposition of the Yb2Si2O7 coatings, and the effects of the The HVOF process was used for the deposition of the Yb2Si2O7 coatings, and the effects of oxygen/methane ratio and stand‐off distance on the coating microstructure and crystallinity were the oxygen/methane ratio and stand-off distance on the coating microstructure and crystallinity discussed. The concurrent changes in the area specific weight and arithmetic mean roughness were discussed. The concurrent changes in the area speciﬁc weight and arithmetic mean roughness associated with the presence of molten/non‐molten particles in the coatings at different process associated with the presence of molten/non-molten particles in the coatings at different process conditions (e.g., fuel lean or rich conditions) were interpreted. Accordingly, an equivalence ratio of conditions (e.g., fuel lean or rich conditions) were interpreted. Accordingly, an equivalence ratio of 1.01 1.01 was found to yield the highest area specific weight when it was combined with a stand‐off was found to yield the highest area speciﬁc weight when it was combined with a stand-off distance distance of 350–375 mm. The amorphous content of the HVOF sprayed coatings was found to be in of 350–375 mm. The amorphous content of the HVOF sprayed coatings was found to be in the range the range of 65%–73% by PONKCS analysis, and the EBSD analysis clearly demonstrates that the of 65%–73% by PONKCS analysis, and the EBSD analysis clearly demonstrates that the non-molten non‐molten particles explain this partial crystallinity. However, despite the fact that the depositions particles explain this partial crystallinity. However, despite the fact that the depositions successfully successfully worked on the metallic substrates, under the same deposition conditions, the silicon worked on the metallic substrates, under the same deposition conditions, the silicon bond coat layer bond coat layer was found to be eroded. It was shown that this could be avoided by decreasing the was found to be eroded. It was shown that this could be avoided by decreasing the particle size of the particle size of the Yb2Si2O7 feedstock; however, by doing that, the amorphous content of the coating Yb2Si2O7 feedstock; however, by doing that, the amorphous content of the coating is further increased. is further increased. Nevertheless, no vertical cracks were observed in this highly amorphous and Nevertheless, no vertical cracks were observed in this highly amorphous and dense as-deposited dense as‐deposited coating, revealing the advantage of HVOF deposition over the APS, which was coating, revealing the advantage of HVOF deposition over the APS, which was elaborated in the elaborated in the results and discussion section. The thermal cycling test of this highly amorphous results and discussion section. The thermal cycling test of this highly amorphous coating, however, coating, however, suggested that non‐molten particles at the Si/Yb2Si2O7 interface may be unfavorable suggested that non-molten particles at the Si/Yb2Si2O7 interface may be unfavorable for the adhesion, for the adhesion, as their presence at the interface coincided with the delamination cracks. as their presence at the interface coincided with the delamination cracks. Consequently, if the bonding Consequently, if the bonding of the HVOF top coat to the Si bond coat can be improved, the of the HVOF top coat to the Si bond coat can be improved, the conventional HVOF method can be conventional HVOF method can be used for the manufacturing of silicate EBCs without requiring used for the manufacturing of silicate EBCs without requiring very high deposition temperatures as is very high deposition temperatures as is the case in the plasma spraying process. Current the case in the plasma spraying process. Current investigations are looking at different methods to investigations are looking at different methods to develop a well‐adhered interface and steam cycling develop a well-adhered interface and steam cycling tests are planned for the validation of the EBCs. tests are planned for the validation of the EBCs.

Acknowledgments: TheThe authors thank K.H.R., R.L. R.L. and and F.K. F.K. (all in IEK-1,IEK‐1, JÜLICH) for their support on manufacturing the coatings. Special thanks to E. Wessel (IEK-2,(IEK‐2, JÜLICH)JÜLICH) forfor performingperforming EBSDEBSD analysis.analysis. Author Contributions: E.B., G.M., and R.V. conceived and designed the experiments, D.K. contributed the Author Contributions: E.B., G.M., and R.V. conceived and designed the experiments, D.K. contributed the materials, E.B. performed the experiments, E.B., G.M., and Y.S. analyzed the data, and E.B. wrote the paper. materials, E.B. performed the experiments, E.B., G.M., and Y.S. analyzed the data, and E.B. wrote the paper. Conﬂicts of Interest: The authors declare no conﬂict of interest. Conflicts of Interest: The authors declare no conflict of interest.

Coatings 2017, 7, 55 11 of 12

References

1. Gupta, A.K.; Lilley, D.G. Combustion and environmental challenges for gas turbines in the 1990s. J. Propuls. Power 1994, 10, 137–147. [CrossRef] 2. Jacobson, N.S.; Smialek, J.L.; Fox, D.S.; Opila, E.J. Durability of Silica-Protected Ceramics in Combustion Atmospheres; American Ceramic Society: Westerville, OH, USA, 1995. 3. Ohnabe, H.; Masaki, S.; Onozuka, M.; Miyahara, K.; Sasa, T. Potential application of ceramic matrix composites to aero-engine components. Compos. Part A Appl. Sci. Manuf. 1999, 30, 489–496. [CrossRef] 4. Bansal, N.P. Handbook of Ceramic Composites; Bansal, N.P., Ed.; Kluwer: Boston, MA, USA, 2005. 5. Bansal, N.P.; Lamon, J. Ceramic Matrix Composites: Materials, Modeling and Technology; Wiley: Hoboken, NJ, USA, 2014. 6. Opila, E.J.; Hann, R.E. Paralinear oxidation of CVD SiC in water vapor. J. Am. Ceram. Soc. 1997, 80, 197–205. [CrossRef]

7. Robinson, R.C.; Smialek, J.L. SiC recession caused by SiO2 scale volatility under combustion conditions: I, experimental results and empirical model. J. Am. Ceram. Soc. 1999, 82, 1817–1825. [CrossRef]

8. Smialek, J.L.; Robinson, R.C.; Opila, E.J.; Fox, D.S.; Jacobson, N.S. SiC and Si3N4 recession due to SiO2 scale volatility under combustor conditions. Adv. Compos. Mater. 1999, 8, 33–45. [CrossRef]

9. Opila, E.J.; Smialek, J.L.; Robinson, R.C.; Fox, D.S.; Jacobson, N.S. SiC recession caused by SiO2 scale volatility under combustion conditions: II, thermodynamics and gaseous-diffusion model. J. Am. Ceram. Soc. 1999, 82, 1826–1834. [CrossRef] 10. Lee, K.N. Current status of environmental barrier coatings for Si-based ceramics. Surf. Coat. Technol. 2000, 133, 1–7. [CrossRef] 11. Eaton, H.E.; Linsey, G.D.; More, K.L.; Kimmel, J.B.; Price, J.R.; Miriyala, N. EBC Protection of SiC/SiC Composites in the Gas Turbine Combustion Environment. In Proceedings of the ASME Turbo Expo 2000: Power for Land, Sea, and Air, Munich, Germany, 8–11 May 2000. 12. Eaton, H.E.; Linsey, G.D.; Sun, E.Y.; More, K.L.; Kimmel, J.B.; Price, J.R.; Miriyala, N. EBC Protection of SiC/SiC Composites in the Gas Turbine Combustion Environment: Continuing Evaluation and Refurbishment Considerations. In Proceedings of the ASME Turbo Expo 2001: Power for Land, Sea, and Air, New Orleans, LA, USA, 4–7 June 2001. 13. Lee, K.N.; Fox, D.S.; Eldridge, J.I.; Zhu, D.; Robinson, R.C.; Bansal, N.P.; Miller, R.A. Upper temperature limit of environmental barrier coatings based on mullite and BSAS. J. Am. Ceram. Soc. 2003, 86, 1299–1306. [CrossRef] 14. Lee, K.N.; Fox, D.S.; Bansal, N.P. Rare earth silicate environmental barrier coatings for SiC/SiC composites

and Si3N4 ceramics. J. Eur. Ceram. Soc. 2005, 25, 1705–1715. [CrossRef] 15. Richards, B.T.; Zhao, H.; Wadley, H.N.G. Structure, composition, and defect control during plasma spray deposition of ytterbium silicate coatings. J. Mater. Sci. 2015, 50, 7939–7957. [CrossRef] 16. Richards, B.T.; Young, K.A.; de Francqueville, F.; Sehr, S.; Begley, M.R.; Wadley, H.N.G. Response of ytterbium disilicate–silicon environmental barrier coatings to thermal cycling in water vapor. Acta Mater. 2016, 106, 1–14. [CrossRef]

17. Bakan, E.; Marcano, D.; Zhou, D.; Sohn, Y.J.; Mauer, G.; Vaßen, R. Yb2Si2O7 coatings deposited by various thermal spray techniques: A preliminary comparative study. J. Therm. Spray Technol. 2017, submitted. 18. Korpiola, K.; Hirvonen, J.P.; Laas, L.; Rossi, F. The influence of nozzle design on HVOF exit gas velocity and coating microstructure. J. Therm. Spray Technol. 1997, 6, 469–474. [CrossRef] 19. Mainzer, B.; Friess, M.; Jemmali, R.; Koch, D. Development of polyvinylsilazane-derived ceramic matrix composites based on Tyranno SA3 fibers. J. Ceram. Soc. Jpn. 2016, 124, 1035–1041. [CrossRef] 20. Gordon, S.; MeBride, B.J. Computer Program for Calculation of Complex Chemical Equilibrium Compositions and Applications. I. Analysis; NASA Reference Publications; Lewis Research Center: Cleveland, OH, USA, 1994; pp. 25–32. 21. Scarlett, N.V.Y.; Madsen, I.C. Quantification of phases with partial or no known crystal structures. Powder Diffr. 2006, 21, 278–284. [CrossRef] 22. ISO 4288:1996–Geometrical Product Specifications (GPS)–Surface Texture: Profile Method–Rules and Procedures for the Assessment of Surface Texture; International Organization for Standardization: Geneva, Switzerland, 1996. Coatings 2017, 7, 55 12 of 12

23. ISO 3274:1996–Geometrical Product Specifications (GPS)–Surface Texture: Profile Method–Nominal Characteristics of Contact (Stylus) Instruments; International Organization for Standardization: Geneva, Switzerland, 1996. 24. Glassman, I. Combustion; Elsevier Science: Amsterdam, The Netherlands, 1997. 25. Li, M.; Christofides, P.D. Multi-scale modeling and analysis of an industrial HVOF thermal spray process. Chem. Eng. Sci. 2005, 60, 3649–3669. [CrossRef] 26. Picas, J.A.; Punset, M.; Baile, M.T.; Martín, E.; Forn, A. Effect of oxygen/fuel ratio on the in-flight particle parameters and properties of HVOF WC-CoCr coatings. Surf. Coat. Technol. 2011, 205, S364–S368. [CrossRef] 27. Cheng, D.; Xu, Q.; Tapaga, G.; Lavernia, E.J. A numerical study of high-velocity oxygen fuel thermal spraying process. Part I: Gas phase dynamics. Metall. Mater. Trans. A 2001, 32, 1609–1620. [CrossRef] 28. Zhao, L.; Maurer, M.; Fischer, F.; Lugscheider, E. Study of HVOF spraying of WC–CoCr using on-line particle monitoring. Surf. Coat. Technol. 2004, 185, 160–165. [CrossRef] 29. Turunen, E.; Varis, T.; Hannula, S.P.; Vaidya, A.; Kulkarni, A.; Gutleber, J.; Sampath, S.; Herman, H. On the role of particle state and deposition procedure on mechanical, tribological and dielectric response of high velocity oxy-fuel sprayed alumina coatings. Mater. Sci. Eng. A 2006, 415, 1–11. [CrossRef] 30. Guo, X.; Planche, M.-P.; Chen, J.; Liao, H. Relationships between in-flight particle characteristics and properties of HVOF sprayed WC-CoCr coatings. J. Mater. Process. Technol. 2014, 214, 456–461. [CrossRef]

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).