Physicochemical Properties of Yttria-Stabilized-Zirconia In-Flight Particles during Supersonic Atmospheric Plasma Spray

Abstract

In order to achieve better knowledge of the thermal barrier coatings (TBCs) by supersonic atmospheric plasma spraying (SAPS) process, an experimental study was carried out to elaborate the physicochemical properties of particles in-flight during the SAPS process. One type of commercially available agglomerated and sintered yttria-stabilized-zirconia (YSZ) powder was injected into the SAPS plasma jet and collected by the shock chilling method. The YSZ particles’ in-flight physicochemical properties during the SAPS process, including melting state, morphology, microstructure, particle size distribution, element composition changes, and phase transformation, have been systematically analyzed. The melting state, morphology, and microstructure of the collected particles were determined by scanning electron microscopy (SEM). The particle size distribution was measured by a laser diffraction particle size analyzer (LDPSA). Element compositions were quantitatively analyzed by an electron probe X-ray microanalyzer (EPMA). Additionally, the X-ray diffraction (XRD) method was used to analyze the phase transformation. The results showed that the original YSZ powder injected into the SAPS plasma jet was quickly heated and melted from the outer layer companied with breakup and collision-coalescence. The outer layer of the collected particles containing roughly hexagonal shaped grains exhibited a surface texture with high sphericity and the inside was dense with a hollow structure. The median particle size had decreased from 45.65 to 42.04 μm. In addition to this, phase transformation took place, and the content of the zirconium (Zr) and yttrium (Y) elements had decreased with the evaporation of ZrO₂ and Y₂O₃.

1. Introduction

Thermal barrier coatings (TBCs), which can provide thermal insulation to hot components of engines to protect them from corrosion and oxidation at high temperatures, play an important role in advanced gas-turbine and diesel engines [1,2,3]. A typical TBCs system consists of two layers over a super alloy substrate: a metallic bond-coat and a ceramic top-coat, where yttria-stabilized-zirconia (Y₂O₃ stabilized ZrO₂, YSZ) is a widely used material for the top coat owing to its low thermal conductivity, high thermal expansion coefficient, and good mechanical properties [4,5]. To date, electron beam physical vapor deposition (EB-PVD) and atmospheric plasma spraying (APS) are the two main processes to prepare YSZ coatings. In particular, APS is a commercial, high-efficiency, process and easy to operate, in which raw powder is injected into the plasma jet with a carrier gas and quickly heated and accelerated. Successive particles in-flight impinge onto the substrate followed by spreading and rapid solidification to form single splats. The as-sprayed coating develops with the stacking of these flattened splats [6,7].

At present, empirically tuning 3–6 main controllable parameters is the main method to find the relationship between the spray parameters and the coating [8,9]. However, APS is a complicated process in which more than 35 main parameters, both controllable parameters (such as the flow rate of gas, power, etc.) and some uncontrollable parameters (like the erosion of the electrode, and fluctuations in the particles’ injection geometry), are identified to influence the resulting microstructure and corresponding coating properties [10]. Therefore, it is a great challenge to produce reproducible and reliable plasma sprayed coatings. In recent years, particles’ in-flight properties have been studied theoretically, numerically, and experimentally and this has suggested that no matter how complicated the influence parameter is, the physicochemical properties of particles in-flight before impinging onto the substrate are fundamental parameters influencing the quality of the as-sprayed coatings [11,12,13,14,15,16]. Liu et al. [11] used the Box–Behnken Design experimental method to analyze the effect of spray parameters on the average velocity and surface temperature of La₂Ce₂O₇ particles in-flight and their influence on microstructure and mechanical properties. Bai et al. [12] adopted the one-factor-at-a-time method to investigate the effect of four spray parameters on the average velocity and temperature of nanostructured YSZ particles in-flight and to theoretically quantify the influence of in-flight particles’ melting state on the microstructure of as-sprayed coatings. Choudhury et al. [13] utilized an Artificial Neural Network (ANN) to quantify the relationship between processing parameters and Al₂O₃-TiO₂ particles’ in-flight characteristics, and to establish process controls by predicting the particles’ in-flight characteristics in the APS process. Suffner et al. [14] investigated the rapid solidification behavior of the Al₂O₃, ZrO₂, and Y₂O₃ molten droplets after impacting on the liquid nitrogen cooled substrate. Wei et al. [15] analyzed the melting state and refining behavior of particles’ flight during spraying using a numerical method. Tekmen et al. [16] investigated the oxidation behavior of Al particles in-flight and the effects of velocity and temperature on particles in-flight with an oxidation mechanism.

Lately, an advanced high efficiency and low energy consumption (lower than 80 kW) supersonic atmospheric plasma spraying (SAPS) process has been successfully developed by the Academy of Armored Forces Engineering (Beijing, China) [17,18]. The main components of this novel system are a plasma gun, a powder feeder, a gas-supply system, a cooling water system, and a power supply and control system. Figure 1 shows a schematic diagram of the supersonic plasma spraying system. The key to this system is a novel SAPS gun with a Laval nozzle and internal feedstock injection mode, which means the feedstock particles can be heated and accelerated adequately with no limitation to the melting temperature of the sprayed materials, and it has greater advantage in fabricating high performance ceramic, cermet, or metallic coatings than APS. However, the physicochemical properties of YSZ particles in-flight during the SAPS process have not yet been systematically studied.

Motivated by the aforementioned considerations, this study is dedicated to understanding the physicochemical properties of YSZ particles in-flight by using experimental methods to achieve better knowledge of TBCs using the SAPS process. We note that it is difficult to directly observe the physicochemical properties of YSZ particles in-flight during the SAPS process due to their high temperature and velocity. In this paper, a shock chilling method was employed to trap the YSZ in-flight particles at the usual spraying distance during the SAPS process. The YSZ particles in-flight physicochemical properties including melting state, morphology, microstructure, particle size distribution, element composition changes, and phase transformation were investigated in detail during the SAPS process.

2. Experimental Material and Methods

Experiments were carried out on a SAPS system (HEPJ-II, Beijing, China) and the plasma gun was mounted on a 6-axis robot (MOTOMAN-2010, Ysakawa, Japan). Details of the spraying parameters are listed in

Table 1. Spray parameters of yttria-stabilized-zirconia (YSZ) by supersonic atmospheric plasma spraying (SAPS).

Current (A)	Voltage (V)	Ar Gas (slpm #)	H2 Gas (slpm)	Carrier Gas (slpm)	Powder Feeding Rate (g/min)	Spray Distance (mm)	Injector Diameter (mm)	Spraying Angle (°)
430	150	120	24.2	7	35	100	2.2	90

# slpm: standard liter per minute.

, in which argon and hydrogen were used as the primary and assistant gases, respectively. One type of commercially available agglomerated and sintered (AS) 6–8 wt. % Y₂O₃-ZrO₂ (YSZ) particles was used as feedstock. To collect particles in-flight before impinging onto the substrate, an experimental apparatus was developed (as illustrated in Figure 2). During the SAPS process, YSZ particles in-flight were rapidly solidified and collected in a collection apparatus containing liquid nitrogen in which the cooling rate of the spraying particles exceeded 10⁶ K/s that enabled a rapid cooling rate [19]. Furthermore, liquid nitrogen, an inert cooling medium, was used to prevent chemical reaction from occurring between the YSZ particles in-flight and the cooling medium. Thus, this shock chilling method could be used to characterize the physicochemical properties of YSZ particles in-flight. The collection apparatus was positioned vertically to the SAPS gun to enable the YSZ particles in-flight to fly into the liquid nitrogen. The trapped particles were extracted from the resulting suspension by a membrane filter with a pore diameter of 0.1 μm and then dried in a vacuum drying box. The external and internal morphology of the original AS YSZ powders (referred as O-AS) and collected YSZ particles (referred as C-AS) were determined by a scanning electron microscope (SEM, GeminiSEM 500, ZEISS, Oberkochen, Germany). The particle size distribution of the O-AS powders and C-AS particles was measured by a laser diffraction particle size analyzer (LDPSA, Horiba LA-950, Kyoto, Japan). Element compositions of the O-AS powders and C-AS particles were quantitatively analyzed by using wavelength dispersive X-ray spectroscopy (WDS) in an electron probe X-ray microanalyzer (EPMA, EMAX-1770, Kyoto, Japan) with a resolution of 110 ppm and operating voltage at 20 kV. In order to determine the element compositions of the original O-AS powders and C-AS particles accurately, five measurement points were randomly obtained in one image during analysis. The final value was calculated using 15 images selected randomly. The phase transformation of the O-AS powders was characterized by X-ray diffraction (XRD, Philips X’Pert Pro, Almelo, Holland) under room temperature with Cu Kα radiation of wavelength 1.5418 Å. The scanning 2θ angle ranged from 10° to 90° with a step scanning rate of 0.02 °/min, while the voltage and electric current were held at 40 kV and 30 mA, respectively. The samples were determined three times, respectively.

3. Results and Discussion

3.1. Characterization of the O-AS Powders and C-AS Particles

Characteristics of the O-AS powders are shown in Figure 3. The overall surface morphology of the O-AS powders and the high magnification SEM micrograph of a single powder are shown in Figure 3a,b, respectively. It can be seen that the O-AS powders are composed of agglomerated primary granules with 1–2 μm diameter. The O-AS powders with a rough surface are near spherical morphology which can ensure good flow ability. Figure 3c,d reveals the cross-sectional morphology of the O-AS powders, which indicates that the inside of the particles has a relatively loose and porous microstructure. Figure 4 and Figure 5 show the surface and cross-sectional SEM micrographs of the C-AS particles passed through the SAPS plasma jet. Compared with the O-AS powders (Figure 3), the C-AS particles after the SAPS plasma jet exhibited a smoother surface with high sphericity (Figure 4a) due to the surface of the O-AS powders being immediately melted by the high temperature of the SAPS plasma jet when injected into the SAPS plasma jet and the surface of the melted particles becoming spherical under the effect of surface tension [20]. The surface texture of the C-AS particles could be clearly seen when the C-AS particles were more closely examined (see Figure 4b), which may be due to the molten primary granules with 1–2 μm diameter being rapidly cooled and some roughly hexagonal shaped grains being formed. Figure 4c,d displays the cross-sectional morphology of the C-AS particles, which indicates that the C-AS particles with a hollow structure are denser than the O-AS powders with a relatively loose and porous microstructure. This is because the formation of a liquid outer layer prevented further escape of the gas from the O-AS powders and the inside of the particles in-flight continued to melt with the increase of flight distance [21]. As can be seen in Figure 4, there are many smaller and exploded C-AS particles due to the trapped gas expanding when heated, resulting in an explosion and breakup, which can be attributed to the effect of surface tension, shear force, and pressure inside the particles in-flight. There are also some bigger particles, which may be attributed to the collision and coalescence of the particles in-flight during the SAPS process (see Figure 5a) [22]. Figure 5b shows the molten surface microstructure of the particles in-flight at the initial SAPS process stage.

3.2. Particle Size Distribution of the O-AS Powders and C-AS Particles

The particle size distribution of the O-AS powders and C-AS particles is given in Figure 6. The median particle sizes (D₅₀) of the original AS YSZ particles and collected YSZ particles were 45.65 μm and 42.04 μm, respectively. The values of D₁₀, D₅₀, and D₉₀ are summarized in

Table 2. Particle size distribution of the O-AS powders and C-AS particles.

State of the YSZ Particles	Particle Size (μm)
	D10	D50	D90
O-AS powders	32.22	45.65	63.86
C-AS particles	30.16	42.04	60.34

. As can be seen from the result, the particle size of the O-AS powders which were injected into the plasma jet obviously decreased during the SAPS process. The possible causes of the particles’ in-flight refining were analyzed as follows: 1) as the particles in-flight melt during the SAPS process, the particles in-flight become smaller and denser; 2) The molten particles in-flight may breakup during the SAPS process that further results in the reduced particle size. The smaller particles in-flight are able to form coatings with more complex maze structures which is beneficial for thermal insulation and anti-oxidation [23].

3.3. Element Composition Changes of the O-AS Powders

Element composition of the O-AS powders and C-AS particles measured by EPMA are shown in Figure 7. It can be seen in Figure 7 that the element composition of the O-AS powders and C-AS particles were the same. However, the content of zirconium (Zr, ~65.17 wt. %) and yttrium (Y, ~7.07 wt. %) in the O-AS powders’ particles was higher than the C-AS particles (Zr, ~63.58 wt. %; Y, ~4.29 wt. %), while the content of the oxygen element in the O-AS powders was lower. This was probably related to the partial evaporation of ZrO₂ and Y₂O₃ from the original AS YSZ particles during the SAPS process. When the O-AS powders were injected into the plasma jet, in which the temperature was about 10⁴ K and much higher than the boiling point of ZrO₂ (~4300 K) and Y₂O₃ (~4300 K), the evaporation of ZrO₂ and Y₂O₃ occurred [24]. As a result, the content of Zr and Y elements decreased. Furthermore, the decrease of Y (~2.78 wt. %) was higher than Zr (~1.59 wt. %), which was ascribed to a higher vapor pressure of Y₂O₃ than ZrO₂ at the elevated temperature (see Figure 8) leading to a faster loss of Y₂O₃ than ZrO₂ during the SAPS process [25].

3.4. Phase Transformation of the O-AS Powders

Figure 9 shows the XRD patterns of the O-AS powders and C-AS particles. According to the XRD patterns shown in Figure 9a, Y_O.15Zr_0.85O_1.93, monoclinic zirconia (m-ZrO₂), and Cubic Zirconia (c-ZrO₂) were the main phases of the O-AS powders. It can be seen from Figure 9b that the C-AS particles were mainly composed of Y_O.15Zr_0.85O_1.93 and c-ZrO₂ phases. The disappearance of the m-ZrO₂ phase in the O-AS powders during the SAPS process was attributed to phase transformation. As the O-AS powders were passed through the high temperature plasma jet, m-ZrO₂ in the O-AS powders changed into t-ZrO₂ over 1443 K, and then t-ZrO₂ transformed into c-ZrO₂ above 2643 K [26]. The phase transformation of pure ZrO₂ is reversible. However, with the addition of Y₂O₃ stabilizing oxides in the crystalline of ZrO₂ combined with relatively high cooling rates, the amount of phase transformation crystalline ZrO₂ can undergo is limited during the SAPS process. Pure ZrO₂ experiences the following phase transformation:

$$\mathrm{m}\text{-}\mathrm{Z}\mathrm{r}{\mathrm{O}}_2\underset{1223\mathrm{K}}{\overset{1443\mathrm{K}}{\rightleftarrows }}\mathrm{t}\text{-}\mathrm{Z}\mathrm{r}{\mathrm{O}}_2\stackrel{2643\mathrm{K}}{\rightleftarrows }\mathrm{c}\text{-}\mathrm{Z}\mathrm{r}{\mathrm{O}}_2\stackrel{2953\mathrm{K}}{\rightleftarrows }\mathrm{l}\mathrm{i}\mathrm{q}\text{-}\mathrm{Z}\mathrm{r}{\mathrm{O}}_2$$

Figure 10 shows the surface and cross-sectional morphologies of C-AS particles. It can be observed that larger pores appear inside and on the surface of spraying particles compared with O-AS particles given in Figure 3. A schematic diagram representing the development of physicochemical properties of YSZ particles in-flight during supersonic atmospheric plasma spray is shown in Figure 11. The physicochemical properties of YSZ particles in-flight before impinging onto the substrate can significantly influence the microstructure and phase composition of the as-sprayed coatings. The relationship between the physicochemical properties of particles in-flight and the properties of high efficiency supersonic atmospheric plasma-sprayed YSZ coatings still need to be investigated.

4. Conclusions

In this study, YSZ particles in-flight were successfully collected by a shock chilling method using a collection apparatus containing liquid nitrogen, and the physicochemical properties of the melting state, morphology, microstructure, particle size, chemical composition changes, and phase transformation during the SAPS process have been systematically investigated and the following conclusions can be drawn:

(1)

The O-AS powder injected into the SAPS plasma jet was quickly heated and melted from the outer layer companied with breakup and collision-coalescence. The outer layer of the C-AS particles containing roughly hexagonal shaped grains exhibited a surface texture with high sphericity and the inside was dense with a hollow structure.

(2)

The molten YSZ particles in-flight became smaller and denser and may have broken up during the SAPS process. As a result, there was a refining phenomenon of particles in-flight during the SAPS process. The median particle size decreased from 45.65 to 42.04 μm.

(3)

The element composition of the O-AS powder and C-AS particles were the same. However, the content of Zr and Y (wt. %) in the O-AS powder particles was higher than the C-AS particles. The content of Zr and Y (wt. %) decreased with the evaporation of ZrO₂ and Y₂O₃ during the SAPS process, respectively. Furthermore, the decrease of Y (~2.78 wt. %) was higher than Zr (~1.59 wt. %).

(4)

Phase transformation took place during the SAPS process. The O-AS powders were mainly composed of Y_0.15Zr_0.85O_1.93, m-ZrO₂, and c-ZrO₂ phases; the corresponding C-AS particles were mainly composed of Y_0.15Zr_0.85O_1.93 and c-ZrO₂ phases. The m-ZrO₂ phase vanished during the SAPS process.