Al₂O₃-TiO₂ coatings deposition by intermixed and double injection SPS concepts

The study involved the use of two commercially available distilled water-based suspensions: (i) AuerCoat Al₂O₃ 25W, Treibacher Industrie, Althofen, Austria, and (ii) AuerCoat TiO₂ 25W, Treibacher Industrie, Althofen, Austria. The feed-stocks were used in as-produced state for double injection spray experiments. Both suspensions were also used for the formulation of three in-house intermixed feedstocks. They were intended to yield the same target ratio as for the double injection, by means of dilution, namely:

AT3 (Al₂O₃ + 3 wt.% TiO₂),

The general scope of study is summarized in the graphical abstract, while the details are provided in Table 1.

The raw powders used in commercial alumina and titania suspensions were of similar irregular morphology, built of polyhedral particles (Figure 4).

Fig. 4

The powder particle size analysis showed that the particles dried from both suspensions were of a similar size. The Al₂O₃ powders were characterized by d_v10 = 0.75 μm, d_v50 = 1.20 μm, and d_v90 = 2.46 μm, while the TiO₂ feedstock was characterized by d_v10 = 0.18 μm, d_v50 = 1.34 μm and d_v90 = 3.02 μm (Figure 5).

Fig. 5

The relationship between the viscosity and the shear rate of the intermixed AT3, AT13, and AT40 suspensions is shown in Figure 6a. Non-Newtonian shear thinning was observed for all feedstocks. According to the study of Tesař et al. [33], who compared stabilized and non-stabilized water-based alumina suspensions, the presence of stabilizing agents increases the suspension's viscosity. However, the information about the chemistry of such agents, their amount, and colloidal details were not specified for the used commercial feedstocks.

Fig. 6

Since the sizes of the injector orifice diameter were 0.2 for double injection and 0.35 mm for intermixed, the most important values of the viscosity were at the highest RPMs. This is where the highest shear rate was observed. In this case, all three suspensions had a similar viscosity. For example, for the shear rate of 100 s⁻¹, the average viscosity values were: η_AT3 = 7.3 ± 3.2 mPa·s, η_AT13 = 2.5±0.05 mPa·s, η_AT40 = 2.0±0.03 mPa·s. This is close to the viscosity of the water solvent: η_water = 1.0 mPa·s [33], thus all suspensions were able to be easily fed through the injection nozzles.

Figure 6c and Figure 6d refer to the as-delivered viscosity of the ready-to-spray alumina and titania suspensions. The viscosity of the AuerCoat Al₂O₃ suspension was already studied in the literature but for a solid content of 40 wt.% [33]. Its values ranged from 60 mPa·s at 10 RPM to 20 mPa·s at 100 RPM. In this paper, the alumina suspension contained 25 wt.% of the solid phase and, therefore, the viscosity was lower than presented in work [33]. The obtained values were in the range from 24.3±4.2 mPa·s at 10 RPM to 16.9±2.1 mPa·s at 100 RPM. The behavior of the titania suspension was significantly different from the previous one – the initial viscosity increased, and then reached a lower level. It may be caused by a higher sedimentation speed of the suspensions in comparison to other mixtures, which is discussed later.

The pH value is another important factor when discussing the suspension properties (see Figure 7). It should be between 4 and 10 ideally, to protect the hardware components against corrosion [1]. The measured pH values were within the extreme values of that recommended range. It can be also seen that the values of pH, obtained for the AT3, AT13, and AT40 suspensions, were only slightly lower than the one of Al₂O₃, but the pH value decreased when the rather acidic TiO₂ suspension was added.

Fig. 7

Figure 8 presents the levels of the sediment-solvent interfaces observed during the sedimentation tests. Investigations showed that each feedstock was stable without any mixing for at least 30 minutes. The TiO₂ and AT13 suspensions displayed a different sedimentation nature than the Al₂O₃, AT3, and AT40 ones. In the first one, the visible sediment-solvent interface was possible to be identified but after 5 hours (TiO₂) and 2 hours (AT13) of testing, the rapid sedimentation was observed in the form of a slurry at the bottom of the probe – the samples reached the saturated state [33]. The most distinguishable interfaces were observed for the AT40 and AT3 suspensions, where a gradual sedimentation was observed. Nevertheless, the spraying of all suspensions was successfully completed by using re-dispersing directly prior to spraying and additional continuous stirring during the deposition process.

Fig. 8

Shadowgraphy was used for the visualization of the fragmentation of the liquid droplets in the plasma jet (Figure 9). When comparing the AT3_i, AT13_i, and AT40_i suspensions, it was observed that the behavior of the droplets was similar. Furthermore, in the case of AT3_di, AT13_di, and AT40_di, the liquid droplets penetrated the core of the plasma stream and were quite uniformly fragmented. The injection angle of 25° provided a successful injection of the liquid only in the hot plasma core.

Fig. 9

The stability of the spraying was observed during the entire deposition. The double injection was controlled every 15 cycles of the deposition by means of shadowgraphy. As shown in Figure 10 for spray run AT40_di, no disturbances were observed and the fragmentation of the liquid feedstock was similar during the whole spray process.

Fig. 10

The SEM studies on the coating morphology showed that all samples, excluding AT3_di, had a finely-grained, cauliflower-like topography (Figure 11). The AT3_di sample contained a lot of loosely bonded deposits of micrometer size, which was easily observed even during macroscopic observation. However, higher magnification images (see the inserts in Figure 11) revealed that for the other samples (AT3_i, AT13_di, AT40_i), loose particles of submicrometer size remained between the splats. As already reported in the literature [2], this phenomenon is observed for the deposition based on the radial manner of the feedstock injection. During the injection and break-up of the liquid feedstock, some part of it cannot be directly introduced into the hottest part of the plasma jet. This also means that not all of the powder particles will undergo a complete thermal treatment by the plasma, especially those travelling in the jet peripheries [34]. The shadowgraphy observations, presented in Figure 9, confirm this issue. Although the suspension injection pressure was set so that the mean particle trajectory was along the plasma jet axis, it is visible that some of the feedstock encountered the plasma periphery areas. It should also be noted that similar topography was observed in the alumina SPS coatings presented in the previous work [18]. Moreover, inhomogeneous melting of agglomerated powder particles has been noted also by other researchers [35]. This clearly shows that, on the one hand, a significant advantage of the proposed spraying is the possibility of tailoring any chemical composition but, on the other, the complexity of the process itself is increased, and obtaining a coating with the intended dense structure is also more difficult [36] and will demand further optimization.

Fig. 11

Fig. 12

Among all coatings, the AT40_di sample contained the most homogeneously distributed and melted splats. This may result from the fact that the AT40 suspension contained a higher content of TiO₂ powder, which was characterized by (i) smaller melting temperature and (ii) slightly greater particles size in comparison with Al₂O₃. Lighter and smaller particles build up the coatings with a columnar structure more easily mainly because they follow the gas swirls parallel to the bond coat/substrate and stick to the surface asperities [37].

The microstructural observations of the cross-sections confirmed the top view observations. It was revealed that all coatings, regardless of their chemical composition and type of feedstock injection, had a similar columnar-like microstructure (see Figure 12). Additionally, a significant content of intercolumnar gaps were observed in each coating. The width of the intercolumnar spacing was the smallest for the AT40 coatings (Figure 12e and 12f). Additionally, it was observed that the increased content of TiO₂ provided a higher density of columns and their uniformity. This could be combined with a low viscosity of the AT40 suspension (see section 3.1), which probably had a beneficial impact on the fragmentation of the suspension during transfer via plasma.

Spraying with the intermixed suspension as well as with the double injection showed that the range of the proposed parameters had a limited influence on the coating microstructure. An important issue seems to be the type of the used solvent. Water was preferred due to the safety of storage and environmental purposes. Nevertheless, it requires almost three times more energy for vaporization than the ethanol solvent [38]. In most cases, it is claimed that water-based suspensions provide coatings characterized by a high density and homogeneous microstructure [39], while ethanol-based ones promote the formation of columnar and porous structures [40]. However, in work [33], the authors showed that the resulting microstructure is related not only to the solvent but that the atomization of the suspension also matters. As already discussed in Figure 10, the fragmentation of the feedstock was undisturbed during the whole spray process. Therefore, it may be suspected that the obtained columnar microstructures were rather a consequence of the fine powder size but neither of the water-based solvent nor the disturbances in the suspension fragmentation.

It should be emphasized here that an extensive study on the phase composition of the aluminatitania coatings has been carried out already. Particular attention has been directed toward the stabilization of α–Al₂O₃ [18], which is a more desirable phase when compared to the metastable γ–Al₂O₃ and δ–Al₂O₃ phases [41]. However, the SPS alumina coatings generally tend to possess only a limited content of the alpha phase, because of the preferred formation of the gamma and delta phases during rapid solidification, as discussed by [42] and [43]. Indeed, in the investigated coatings, according to the XRD patterns shown in Figure 13, both the stable α–Al₂O₃ and the metastable γ–Al₂O₃ and δ–Al₂O₃ phases were observed. Unfortunately, the presence of α–Al₂O₃ in the studied coatings is mainly attributed to the presence of nonmelted powder particles, which were observed in the topography of the surfaces of the coatings. However, the content of this phase can be also increased by the relatively high temperature of the substrate in the spray process (the substrate was preheated prior to spraying) [44], as well as the use of a high enthalpy plasma torch [4].

Fig. 13

Tialite Al₂TiO₅ and rutile-TiO₂ were identified for the AT13 and AT40 coatings. As suspected, tialite was present especially in the AT40 (both AT40_i and AT40_di) coatings – for 40 wt.% TiO₂ content, this is the major phase in the Al₂O₃-TiO₂ phase diagram. Additionally, based on work [45], it is suggested that some of the initial rutile-TiO₂ in the feedstock titania powder reacted with alumina during spraying and formed the tialite phase. This could be promoted especially for the intermixed suspensions (AT13_i and AT40_i), due to the already discussed increased rate of the contact area, leading to congruent melting and mutual phase transformation [4].

The investigations showed that the differences observed in the studied alumina-titania coatings were influenced not only by different initial material properties but also by the way of feedstock introduction into the plasma jet. The summary on using the double injection or intermixed concept is provided in Table 4.

Another important advantage of the intermixed suspension spraying is the reduction of the process time and increasing the coating growth rate (Figure 14).

Fig. 14

It should be emphasized that both the double injection and the intermixed spraying provide the possibility to formulate almost any content ratio of the used feedstocks. The tailored chemical compositions of the coatings (not limited to Al₂O₃-TiO₂ system) are of strategic importance to many industries. Consequently, it seems that the effect of the suspension characteristics on the coating properties is worth being investigated further. In the near future, it could bring measurable economic benefits, such as the extended lifetime of the components, which are key factors in the development of many new applications, like ion cells [46], solid oxide fuel cells [47], wear-resistant coatings [34], and many others. Nevertheless, due to the complexity of the process, obtaining an intended coating is a demanding task – as the obtained columnar microstructures were originally expected to be dense coatings.