Aerosol-Deposited 8YSZ Coating for Thermal Shielding of 3YSZ/CNT Composites

Abstract

High-temperature conditions are harmful for carbon nanotube-based (CNT-based) composites, as CNTs are susceptible to oxidation. On the other hand, adding CNTs to ceramics with low electrical conductivity, such as 3YSZ, is beneficial because it allows the production of complex-shaped samples with spark plasma sintering (SPS). A shielding coating system may be applied to prevent CNT oxidation. In this work, the 8YSZ (yttria-stabilized zirconia) thermal shielding coating system was deposited by aerosol deposition (AD) to improve the composite’s resistance to CNT degradation without the use of bond-coat sublayers. Additionally, the influence of the annealing process on the mechanical properties and microstructure of the composite was evaluated by nanoindentation, scratch tests, scanning electron microscopy (SEM), X-ray diffraction (XRD), flame tests, and light microscopy (LM). Annealing at 1200 °C was the optimal temperature for heat treatment, improving the coating’s mechanical strength (the first critical load increased from 0.84 N to 3.69 N) and promoting diffusion bonding between the compacted powder particles and the substrate. The deposited coating of 8YSZ increased the composite’s thermal resistance by reducing the substrate’s heating rate and preventing the oxidation of CNTs.

1. Introduction

Thermal barrier coatings (TBCs) are widely used to improve the high-temperature performance and durability of critical components in a variety of applications, such as aerospace (fuel vaporizers, combustion chambers, afterburners), automotive (pistons), and nuclear (molten metal and salt-based reactors) [1,2,3].

The classic TBC system consists of (1) a metallic bond coat that ensures good adhesion and mitigates the stresses created by thermal expansion mismatch between the substrate and the coating [4]; (2) thermally grown oxides (TGO) that are created by the diffusion of oxygen during the manufacturing of the top layer and provide shielding from the oxidation of the substrate [1]; (3) a topcoat providing thermal insulating properties thanks to its low thermal conductivity and reduced radiative heat transfer [5].

One of the most common topcoat materials is zirconia (ZrO₂), which can be found in three allotropic forms: monoclinic (m), tetragonal (t), and cubic (c). The transformations take place at ~1173 °C (m − t) and ~2370 °C (t − c) [6]. To form tetragonal and cubic forms at room temperature (and prevent stresses caused by phase transformation-induced volume change during the heating and cooling cycle) [7], various oxides (CaO, MgO, CeO₂, etc.) may be added. Particularly common is the addition of Y₂O₃ at 3% mol. (3YSZ) for the tetragonal phase and 8% mol. (8YSZ) for the cubic phase. It is worth noting that 8YSZ is also commonly referred to as 8% wt. YSZ, which has a tetragonal structure.

YSZ coatings have been deposited using physical vapor deposition, electrophoretic deposition, and plasma spraying techniques, with the third of these being the most common due to its high deposition rates, coating adhesion, and low processing cost compared to the other methods [1,5]. Several works concerning plasma-sprayed YSZ TBCs have described the positive effect of the introduction of nanostructure and bimodal micro-nanostructure [7,8,9] in YSZ-based coatings, resulting in an improvement of mechanical properties and limited oxygen diffusion towards the substrate.

However, the plasma spraying of nanostructures is difficult due to the requirement to melt the sprayed particles to obtain sufficient interparticle bonding and consolidation of the coating. If all powder particles are fully molten in the thermal spray jet, the nanostructural characteristics of the powder particles (and those of the resulting coatings) will disappear [10]. Therefore, nanostructured feedstock powders tend to be quite sensitive to the spray parameters [11], and low-temperature deposition methods might be advantageous in fabricating coatings and preserving their feedstock microstructure.

Aerosol deposition (AD) is a kinetic spraying technique that allows the fabrication of ceramic coatings based on room temperature impact consolidation (RTIC), where the ceramic particles bond with the substrate and consolidate as a result of plastic deformation and controllable cracking of fine particles, the latter of which contributes positively by refining the coating structure observed by other authors [12]. As a result, nanostructured coatings are obtainable, which may be desirable from the viewpoint of TBC applications. Several works have been presented for the fabrication of both thin and thick 8YSZ coatings by AD [13,14,15,16,17].

However, while most of these studies have focused on the application of 8YSZ coatings in electrochemical applications, such as solid oxide fuel cells (SOFCs) or gas sensors, this work introduces a novel application by extending the use of 8YSZ coatings into high-temperature industrial environments, specifically for TBCs. This represents a significant advancement in the field, as the effects of high-temperature post-processing on 8YSZ coatings have been less explored, mainly due to the temperature stability limitations of SOFC components (up to 900 °C) [13,14]. This study provides new insights by proving that an increased heat treatment temperature has a greater effect on the mechanical performance of coatings, making them more viable for TBC applications. In addition, this research introduces a novel approach by applying AD technology without the use of a bond layer, a method that reduces costs and processing time and improves sustainability, making it a significant contribution to the field.

The aim of this work is to study the influence of high-temperature post-processing for the fabricated 8YSZ AD coating and to test its capability as a means of 3YSZ/CNT composite oxidation protection without the use of bond-coat and TGO sublayers, enabled by relatively small differences in thermophysical properties between 3YSZ and 8YSZ [18]. To achieve this, a series of 3YSZ/CNT composites samples were coated with an 8YSZ layer and heat-treated at 1000–1200 °C, and then their structural properties were studied by scanning electron microscopy and X-ray diffraction. To assess the performance of the coatings, an assessment of the mechanical properties by indentation and scratch tests was made, as well as a flame-torch test to determine the high-temperature response of the fabricated coating.

2. Materials and Methods

For the spraying of coatings, 8% mol. yttria-stabilized zirconia powder (TZ-8Y, TOSOH Corporation, Tokyo, Japan) was used as feedstock. The powders were dried at 200 °C under vacuum for at least 8 h before spraying to ensure flowability during aerosol generation. The powder particle size distribution was analyzed using a Mastersizer 3000 (Malvern PANalytical, Malvern Worcestershire, UK). After spraying trials, the particle size distribution was remeasured to assess changes during aerosol generation.

The spraying trials were conducted using an Aerosol Cold Spray apparatus (Łukasiewicz Research Network—Poznań Institute of Technology, Poznań, Poland) [19], with the spraying parameters listed in

Table 1. Aerosol deposition parameters of 8YSZ coatings.

Spraying Parameter	Value
Vacuum chamber pressure during spraying	0.2 mbar
Aerosol gas type	Nitrogen
Aerosol gas flow	2.5–4 L/min
Sample traverse speed	2.5 mm/s
Nozzle standoff distance	10 mm
Process temperature	Room temperature

. The 3YSZ-10MWCNT composite used as the spraying substrate was spark plasma sintered using an HP D 25/3 (FCT Systeme GmbH, Frankenblick, Germany) device at 1350 °C with a holding time of 20 min. The powder mixtures of ZrO₂ and MWCNTs were prepared in isopropanol using a UP400S (Hielscher Ultrasonics GmbH, Teltow, Germany) device. The dies and punches used in the sintering process were made of high-strength graphite (grade 2334, Mersen, Courbevoie, France) with CFRC composite spacers (grade A015, Mersen, Courbevoie, France), which allowed for the almost complete densification of the samples. More detailed information on the material preparation and sintering process can be found in our previous paper [20]. The surface was grit-blasted using F40 alumina blasting media (MATT, Jaktorów-Kolonia, Poland) before deposition. The process was not assisted with a Low-Pressure Cold Spray heating source to prevent possible oxidation of MWCNTs [21] present inside the substrate’s microstructure, as well as possible erosion due to brittle particles’ impact with high kinetic energy [22]. The aerosol stream was injected axially through a diverging–converging nozzle with a 1.5 mm diameter throttle to the vacuum chamber.

To increase the coating’s mechanical strength, pressureless annealing was applied. The coated samples were heated to 1000/1100/1200 °C (HT1000/1100/1200) with a 100 °C/h heating rate inside a R170/1000/13 tube furnace (Nabertherm GmbH, Lilienthal, Germany) under an argon gas flow of 100 L/h. The annealing time was set to 2 h.

The as-sprayed and heat-treated coatings were observed with a scanning electron microscope (SEM, Mira 3, Tescan, Brno, Czech Republic) equipped with secondary electrons and backscattered electron detectors to evaluate the specimens’ microstructure.

X-ray diffraction was conducted using the Aeris Research Diffractometer (Malvern PANalytical, Malvern Worcestershire, UK). The diffraction data were collected from a 20 to 90° 2θ diffraction angle range using Cu Kα (λ = 1.54 Å) radiation. The diffraction data were obtained from the PDF 5+ 2024 (International Centre for Diffraction Data, Newtown Square, PA, USA) database: cubic Yttrium-Zirconium Oxide (04-015-2373), tetragonal Yttrium Zirconium Oxide (04-005-4208), and monoclinic Yttrium-Zirconium Oxide (04-005-4252). The Rietveld refinement method [23] was used for microstrain and crystal size determination.

Nanoindentation results were obtained by using a Picodentor HM500 (Fischer Techniogy Inc., Windsor, CT, USA) with the following test parameters: a maximum load of 30 or 300 mN loaded during 20 s and dwell time of 5 s. Measurements were taken with a Vickers tip. To determine the hardness and modulus of the coatings, the Oliver–Pharr method was employed [24]. The Nix and Gao [25] equation was used as the indentation size effect (ISE) model for the calculation of dependence between hardness and penetration depth.

The dynamic scratch test was conducted on a Universal Mechanical Tester (Bruker, Bruker Nano Surface, San Jose, CA, USA) utilizing a high-accuracy zirconia bearing ball (class G10), with a Ra value of 0.025 µm and a diameter of 2 mm. The dynamic scratch test was conducted over a distance of 10 mm on the surface of the sample, with a linearly increasing load (0.3 N to 5 N) applied along the path length at a feed rate of 0.05 mm/s. During the measurement, the frictional load and acoustic emission were recorded with an appropriate sensor mounted near the bearing ball. The critical scratch load was determined by analyzing the friction, acoustic signal, and optical micrographs of the scratch, which were obtained with the Bruker Alicona RL apparatus (Bruker, Alicona Division, Graz, Austria) [26].

Flame tests were conducted by subjecting the samples to a butane-air flame torch (Dragon 220, Rocker, New Taipei City, Taiwan). The flame torch standoff distance was set at 20 mm and the samples were subjected to the flame for 15 min. The temperature was measured by a K-type thermocouple placed on the backside of the samples. Flame-tested sample cross-sections were observed using a VHX 7000 (Keyence, Mechelen, Belgium) digital microscope to assess the degree of degradation of the coating and substrates (and of the CNTs).

3. Results and Discussion

3.1. Feedstock Powder Characterization

The 8YSZ feedstock powder average particle size was measured at 56.2 µm. Sonification was applied to the powder for 4 min to destroy the possible agglomerates present in the powder material (Figure 1a). As a result of the powder deagglomeration, it was confirmed that the globular (Figure 1b) primary particles consisted mainly of smaller, micron, and submicron-sized particles similar to those reported by Mishra et al. [15]. The powders left from the fluidized bed aerosol generator after the spraying process (approximately 10 min of aerosol gas flow) contained trace amounts of smaller particle fractions, signifying that during fluidization, some fraction of the powder was deagglomerated.

3.2. Aerosol Deposition Process

The spraying process resulted in a successful coating of the 3YSZ/CNT substrate. The deposition efficiency (DE) of the spraying process was estimated to be in the 6%–9% range. This estimation was based on the weight difference of the powder in the fluidized bed and sample mass before and after the process. Such DE is higher than commonly reported in AD films [17], signifying that loose compaction of agglomerated particles might have taken place. This is in line with the previous primary particle assessment, where the presence of agglomerates was demonstrated (Figure 1a,b). Conducted wipe-tests showed that the as-sprayed coating is easily peeled off and its microstructure requires improvement. To do so, an annealing process of thin (Figure 2a) and thick (Figure 2b) coated coupons was employed to promote diffusion in the coating microstructure, as well as bonding between the substrate and the coating through the elimination of residual stresses after deposition [13].

3.3. Structural Analysis of Deposited and Annealed Coatings

3.3.1. X-Ray Diffraction

Typically for the AD process [13], the formation of the coating resulted in no phase changes in the 8YSZ material. However, a slight shift of peak positions was observed (Figure 3a), which may be attributed to changes in the lattice parameter of the cubic 8YSZ crystal structure. While AD films often exhibit microstrain in the 0.5% to 1% range [14], fabricated 8YSZ coatings do not reach such values (

Table 2. X-ray diffraction Rietveld refinement data.

	Powder	As-Sprayed	HT1000	HT1100	HT1200
Lattice parameter [Å]	5.139 ± 0.001	5.132 ± 0.001	5.136 ± 0.001	5.137 ± 0.001	5.138 ± 0.001
Microstrain [%]	0.355 ± 0.001	0.412 ± 0.001	0.258 ± 0.001	0.097 ± 0.001	0.066 ± 0.001
Crystallite size [Å]	329.2 ± 1.7	346.1 ± 2.1	392.6 ± 1.8	1924.0 ± 61.3	4480.7 ± 536.3

). This may be attributed to the cracking of the agglomerates during collision with the substrate mentioned in the previous paragraph, instead of the fracture of micron-sized particles. No drastic changes in crystallite size were detected, as nano-sized particles maintain their original size [27] during deposition by AD. Generally, the heat treatment between 1000 °C and 1200 °C resulted in the reduction in the microstrain, an increase in the crystallite size, and a change in the lattice parameter back to its pre-coating state. The heat treatment at 1200 °C resulted in the formation of a secondary, tetragonal YSZ phase (Figure 3b). This may be attributed to the thermally induced phase change in the 8YSZ, but also to possible diffusion between the coating and the 3YSZ-based substrate.

3.3.2. Scanning Electron Microscopy

The presence of individual particles within the deposited 8YSZ coating was observed following the heat treatment at 1000 °C (Figure 4a). Furthermore, the particles observed in the coating were of a smaller size than the particles in the powder material prior to spraying. This is consistent with the particles undergoing a process of refinement through fragmentation and powder deagglomeration during the deposition process. The occurrence of rebounding, bonding, and fragmentation with particle velocity has been previously described as a phenomenon of aerosol deposition at room temperature [28].

The application of elevated temperatures (1100 °C and 1200 °C) facilitated the formation of bonds between particles, resulting in necking (Figure 4b,c) and crystal growth, which was also proved by X-ray diffraction result analysis (Table 2). This process contributed to an improvement in their properties. This type of behavior is typical of pressureless sintering of bulk samples [29]. Moreover, the mass transport induced by the heat treatment resulted in a more uniform distribution of smaller pores within the resulting coatings. The heat treatment at 1200 °C yielded the most optimal results, as defined by the study. The cross-sectional view of the coating in Figure 4d illustrates the superior bonding quality of the coating produced through this method, in comparison to other coatings, particularly those in the as-deposited state, which exhibited visible cracks at the substrate and coating interface.

The thermal insulation properties of thermal barrier coatings are influenced by heat conduction as well as radiation [5]. The porosity present inside the 8YSZ coating may decrease infrared emissivity [30], improving the overall thermal insulation properties of the 8YSZ coating. Therefore, the nanoscale porosity of the fabricated coating may be beneficial for the overall performance of the coated elements.

3.4. Mechanical Properties Assessment

3.4.1. Indentation

Nano hardness tests (Figure 5a) were conducted using coating cross-sections, in order to minimize the influence of the coating roughness and thickness variations on the test results. Based on that data, the hardness and elastic modulus were calculated according to Equations (1) and (2):

$$H={\displaystyle \frac{F}{A}}$$

(1)

$$E_T={\displaystyle \frac{\sqrt{\pi }}{2\beta }}{\displaystyle \frac{S}{\sqrt{A}}}$$

(2)

where F is the indentation load, A is the contact area of the indent tip, β is a constant equal to 1.055 (for Vicker’s tip), and S is a tangent of the beginning of the unloading curves [24]. Based on the maximum indentation depth, these calculations were made only for a 30 mN indentation load. The results (Figure 5b) indicate an increase in both hardness and elastic modulus with an increasing heat treatment temperature, from 0.16 ± 0.01 GPa and 6.1 ± 0.3 GPa in the as-sprayed state to up to 1.33 ± 0.29 GPa and 35.8 ± 4.4 GPa at 1200 °C, respectively. No significant change in mechanical properties was observed in the as-sprayed and HT_1000 coatings, suggesting that 1000 °C temperature and 2 h dwelling time was not enough to cause significant structural change in the coatings’ structure, which was also confirmed by X-ray diffraction results (Table 2).

The nanoindentation mechanical behavior of the YSZ composites is believed to be evaluated based on the detailed characterization of the indentation size effect (ISE) parameter [31]. Their estimation on the basis of the nano hardness measurement of the YSZ ceramics allows for the description of their real inelastic behavior, which may be controlled by (i) bulk particle deformation and (ii) interparticle sliding. The graph (Figure 6) log H² = f(1/h − h_el) fit the linear approximation of the Nix and Gao [25] equation with simulation veracity ranging from R² = 0.97 to R² = 0.99. As such, the strain gradient plasticity (SGP) can describe the YSZ ceramic coating inelastic behavior.

As seen in Figure 6, during 30 mN indentation a decrease in log H² was visible, which suggests that the mechanical properties did not change throughout the mechanical tests (due to substrate or resin, which may have interfered with the mechanical response of the coating).

3.4.2. Scratch Tests

The scratch test shows a large scatter of coefficient of friction (COF) values (Figure 7a–d) during the test due to a large surface roughness and low coating thickness (<20 µm). A reduction in the COF scatter is observed for the YSZ coating heat-treated at 1200 °C. The wear tracks on the as-sprayed (Figure 7a) and 1000 °C heat-treated samples (Figure 7b) show visible amounts of debris left after the probe movement. In the case of the HT1100 sample (Figure 7c), no acoustic emission (AE) peaks were observed, even though penetration of the coating was seen. The critical load (Lc) values in that case were determined by the applied force (Fz) fluctuation with a corresponding drop in the coefficient of friction as well as presence. Scratch hardness was calculated (

Table 3. Scratch test results and calculation data.

	As-Sprayed	HT1000	HT1100	HT1200
Critical Load (P/Lc)	0.84 N	1.42 N	3.23 N	3.69 N
		2.38 N	3.56 N	4.44 N
			4.10 N
			4.25 N
Scratch width (w)	0.23 mm	0.21 mm	0.09 mm	0.09 mm
		0.23 mm	0.15 mm	0.14 mm
			0.18 mm
			0.13 mm
Scratch hardness (HSp)	5.05 MPa	10.25 MPa	126.93 MPa	145.01 MPa
		14.32 MPa	50.36 MPa	72.11 MPa
			40.28 MPa
			80.05 MPa

) using Equation (3) [26]:

HSp = P/πw²

(3)

where HSp is the scratch hardness, P is the normal force corresponding to the respective critical load, and w is the scratch width at the corresponding normal load.

Both the HT1100 (Figure 7c) and HT1200 (Figure 7d) samples exhibited an increase in the first Lc as well as a decrease in COF and wear track width, suggesting improvement of the mechanical resistance of the 8YSZ coatings after being subjected to high temperatures [5]. An increase in the mechanical properties was also observed during nanoindentation (Figure 5a) and was attributed to observed necking between deposited 8YSZ particles (Figure 4b,c) and crystallite growth (Table 2) [32].

3.5. Flame Resistance Test

3.5.1. Heating–Cooling Curves

The flame tests of uncoated samples, as well as the samples with thin coating, resulted in immediate cracking due to a high temperature gradient ending the tests prematurely. Only the thick samples (25.4 mm diameter, 5.5 mm substrate thickness) with a thick 8YSZ layer heat-treated at 1200 °C did not break during the initial stage of the flame test, signifying that the thermal shock cracking of the substrate was averted. After about 12 min of the test, the 550 µm thick coating was delaminated from the sample (Figure 8). The precise cause of the delamination remains unclear to the author. One potential explanation for this phenomenon is the difference in the coefficient of thermal expansion (CTE) between the 3YSZ-10MWCNT substrate and the 8YSZ coating. This thermal mismatch can give rise to interface stresses, particularly during rapid heating, as observed in the conducted experiment. The degradation of the thermal barrier coatings was previously described by other authors [33].

3.5.2. Flame-Tested Surface Cross-Sections

The LM micrograph presents the surface topography after flame tests (Figure 9a). The flame tests resulted in damage to the coating, showing both cohesive (Figure 9c) and adhesive failure (Figure 9b) of the coating. The area of cohesive fracture covered most of the coated surface, signifying that the cohesion strength between sublayers of the coating was lower than the coating–substrate adhesion. Even after the delamination of the coating, some mechanically interlocked particle areas were visible (Figure 9b). While the coating was damaged during the flame test, the surface underneath the coating exhibited no signs of carbon nanotube oxidation.

3.6. Coating of Shell-Shaped 3YSZ-CNT Composite

In order to demonstrate the scalability of the aerosol deposition process for the fabrication of end-components, the proposed spraying methodology was implemented to coat Ø75 mm diameter shell-shaped 3YSZ/CNT composites, which serve as an illustrative example of components used in this form in the aerospace industry, as described in our prior paper [20]. The coating was applied both to the face and sides of the substrate (Figure 10a) in a two-step coating process (face and sides coated separately), showing no delamination prior to or after the heat treatment process on most of the coated area due to good adhesion and the special holder that covered the previously coated area. The edges of the substrate (Figure 10b) proved more difficult to shield from erosion caused by rebounded particles from the substrate-holding elements.

4. Conclusions

The aim of this research was to achieve effective oxidation protection for a 3YSZ/CNT composite substrate by applying an 8YSZ coating without the use of bond-coat and TGO sublayers.

The conducted studies demonstrated a considerable enhancement in the mechanical properties following the application of a heat treatment. The heat treatment at 1200 °C was identified as the most beneficial, as evidenced by the X-ray diffraction and SEM results, which indicated a reduction in diffusion and crystallite growth, as well as a decrease in residual stress within the coating. This resulted in an increase in the first critical load, hardness, and elastic modulus, from 0.84 N, 0.16 GPa, and 6.1 GPa in the coating’s as-sprayed state to 3.69 N, 1.33 GPa, and 35.8 GPa in the 1200 °C heat-treated state, respectively. The objective of achieving oxidation protection was accomplished not only through the application of the 8YSZ coating, which exhibits inherent thermal insulating properties, but also by enhancing the coating’s adhesion to the substrate. Furthermore, the fact that the coating predominantly exhibited cohesive failure after testing indicates that it maintained its integrity as an oxidation barrier, thereby ensuring continued protection of the substrate.

The developed technology, though requiring further optimization in composition and process, represents a significant step towards the cost-efficient, simplified manufacturing of bond-free thermal barrier coatings for extreme environments such as aerospace components (e.g., turbine engines), automotive components (e.g., low-heat-rejection engines), and energy sectors (e.g., advanced nuclear components) [1,2,3].