Phase Composition, Thermal Conductivity, and Toughness of TiO₂-Doped, Er₂O₃-Stabilized ZrO₂ for Thermal Barrier Coating Applications

Abstract

TiO₂ was doped into Er₂O₃-stabilized ZrO₂ (ErSZ) to obtain desirable properties for thermal barrier coating (TBC) applications. The phase composition, thermal conductivity, and mechanical properties of TiO₂-doped ErSZ were investigated. ErSZ had a non-transformable metastable tetragonal (t′) phase, the compound with 5 mol % TiO₂ consisted of t′ and cubic (c) phases, while 10 mol % TiO₂ doped ErSZ had t′, c, and about 3.5 mol % monoclinic (m) phases. Higher TiO₂ doping contents caused more m phase, and the compounds were composed of t′ and m phases. When the dopant content was below 10 mol %, TiO₂ doping could decrease the thermal conductivity and enhance the toughness of the compounds. At higher doping levels, the compounds exhibited an increased thermal conductivity and a reduction in the toughness, mainly attribable to the formation of the undesirable m phase. Hence, 10 mol % TiO₂-doped ErSZ could be a promising candidate for TBC applications.

1. Introduction

Thermal barrier coatings (TBCs) exert an increasingly significant role as a surface protection technology in the field of modern gas turbines. They can enhance the resistance to high-temperature oxidation, corrosion, and erosion of the hot sections and provide thermal insulation, enhancing the inlet gas temperature, and improving the efficiency of the engines [1,2,3]. Up to now, the most widely used TBCs are made of 7–8 wt.% Y₂O₃-stabilized ZrO₂ (YSZ), due to its low thermal conductivity, good phase stability, large thermal expansion, and desirable fracture toughness [1,2,3,4]. The primary methods to produce TBCs are electron beam physical vapor deposition (EB-PVD), air plasma spraying (APS), and plasma spray physical vapor deposition (PS-PVD) [1,2,5,6,7].

YSZ TBCs are composed of a non-transformable metastable tetragonal (t′) phase that is stable below 1200 °C. At temperatures higher than 1200 °C, the t′ phase becomes unstable, giving rise to a severe phase decomposition to cubic (c) and t′ phases [2,3,4,8,9,10]. During cooling, the latter transforms to an m phase accompanied by a large volume expansion, severely damaging the TBCs. Additionally, YSZ undergoes accelerated sintering at high temperatures, leading to the TBCs with an increased thermal conductivity and a reduced strain compliance [2,11,12]. Demands for enhancing gas turbine efficiency have necessitated a significant increase in the combustion temperature. Hence, alternative TBC materials with even better phase stability, lower thermal conductivity, and higher toughness are strongly required.

Two main approaches have been developed to seek for novel TBC materials. One is exploring high-temperature ceramics with excellent phase stability and low thermal conductivity. These ceramics include Gd₂Zr₂O₇, GdPO₄, La₂Ce₂O₇, doped CeO₂, and Y₃Al₅O₁₂ [13,14,15,16,17,18]. However, by comprehensive investigations, these materials are found to have some undesirable properties, like poor fracture toughness and low thermal expansion coefficients, which affect their practical applications. The other is on the ZrO₂-based systems with alternative stabilizers to Y₂O₃ [4,19,20,21,22,23,24]. Investigations have obtained remarkable success. It has been reported that, for rare earth oxides (RE₂O₃)-doped ZrO₂, with the decrease of the RE³⁺ size, the phase stability of the t′ phase increases, while the toughness decreases [4,13,19,20,24]. The rationale behind the phenomena might be related to the driving force for t′ phase partitioning and the lattice tetragonality of the t′ phase. Compared with YSZ, Yb₂O₃-doped ZrO₂ is found to have high phase stability but low toughness [13,24]. Since the size of Er³⁺ is between those of Yb³⁺ and Y³⁺, Er₂O₃-doped ZrO₂ (ErSZ) may reveal better comprehensive properties, i.e., good phase stability and high toughness.

Low thermal conductivity is a primary criterion for selecting materials for TBC applications. Research indicates that doping TiO₂ into RE₂O₃ stabilized ZrO₂ can reduce the thermal conductivity, enhance the toughness, and improve the phase stability [25,26,27,28]. Therefore, this study is conducted on a ZrO₂-Er₂O₃-TiO₂ system, investigating the effects of TiO₂ doping on the phase composition, thermal conductivity, and toughness of the compounds. The compositions are ErSZ doped with x mol % TiO₂ (x = 0, 5, 10, 15, 20). The ErSZ powders are synthesized by a chemical co-precipitation method, and the pellets for property measurements are produced by a solid state reaction and sintering method.

2. Experimental Procedures

Er₂O₃ (purity 99.99%) and ZrOCl₂·8H₂O (purity 99.95%) were employed as raw materials to produce 3.5 mol % Er₂O₃-doped Zirconia (ErSZ) powders by a chemical co-precipitation. Er₂O₃ powders and ZrOCl₂·8H₂O were dissolved in nitric acid and deionized water, respectively. The obtained solution was mixed according to the compositional requirement and stirred to generate a homogeneous solution, followed by adding it cautiously to excess ammonia water (pH > 12) until a gel-like precipitate was obtained. The precipitate was filtered with distilled water repeatedly and washed with alcohol until a neutral pH was acquired. After being dried at 120 °C for 10 h and calcinated at 800 °C for 5 h for crystallization, the resultant powders were ground and sieved to remove the coarse agglomerates.

TiO₂ doped ErSZ powders were synthesized by a solid state reaction method. First, the fabricated ErSZ powders were mixed with x mol % (x = 0, 5, 10, 15, 20) TiO₂ by ball milling (QM-3SP4, Nanjing University Instrument Factory, Nanjing, China) for 8 h. Afterward, the suspension liquid was dried at 150 °C for 12 h, followed by calcination at 1400 °C for 10 h. The obtained powders were cold pressed (NYL-2000D, Wuxi Jianyi Instrument & Machinery Co., Ltd., Wuxi, China) at ~250 MPa and sintered (SX-G05165, Tianjin Zhonghuan Experiment Electric Furnace Co., Ltd., Tianjin, China) at 1500 °C for 12 h to achieve pellets for thermal conductivity and toughness measurements.

The analysis of phase compositions was carried out by X-ray diffraction (XRD; D/MAX 2500, Rigaku Diffractometer, Tokyo, Japan) within a scanning range of 2θ = 20°–80°. Additionally, for the sake of distinguishing the t′ (t₁′, t₂′) and c phases, slow scans were performed over the range of 2θ = 72°–76° with a step size of 0.01° and a dwell time of 20 s.

Thermal diffusivity (α) was measured using a laser-flash apparatus (Netzsch LFA 427, Netzsch, Bavaria, Germany) from 20 to 1200 °C at an interval of 200 °C. Before the measurements, a thin graphite film was coated on the front and rear surfaces of the sample for thermal absorption of laser pulses. Each sample was measured three times at the concerned temperature. To obtain the thermal conductivity (λ), the following formula was employed:

$$\mathsf{\lambda }=\mathsf{\rho }\cdot \mathsf{\alpha }\cdot C_{\mathrm{p}}$$

(1)

where ρ is the density, which is measured by Archimedes method, and C_p is the specific heat capacity derived from the heat capacity values of the constituent oxides according to the Neumann-Kopp rule [29].

It was estimated that the uncertainty of the thermal conductivity was ±5%, considering the uncertainties for the density, the thermal diffusivity, and the specific heat capacity. Since the bulk is porous, the actual data was obtained by correcting the measured thermal conductivity using the following equation [30]:

$$\frac{\mathsf{\lambda }}{{\mathsf{\lambda }}_0}=1-\frac{4}{3}\mathsf{\varphi }$$

(2)

where φ is the fractional porosity.

The mechanical properties of the pellet was measured by an indentation method. Although this method to determine the mechanical property of a material is uncertain, it could be used for qualitatively comparing a series of samples produced by similar procedures [13,24]. Hence, it was employed in this study to reveal the variation trend of the mechanical properties with the TiO₂ content. Vickers hardness (H_V) was measured on the polished surface of the samples with 3 kg load dwelling 15 s by an automatic turret digital Vivtorinox hardness tester (HV-1000A, Laizhou Huayin Testing Instrument Co., Ltd., Yantai, China). At least 15 valid indentations were taken for each sample. Fracture toughness (K_IC) was calculated from the radial crack pattern of Vickers indentations by the following equation [31].

$$K_{\mathrm{IC}}=0.16H_{\mathrm{V}}{a}^2{c}^{-3/2}$$

(3)

where a represents the half length of the indent diagonal, c is the half-crack length measured from the middle of the indent to the tip of the crack.

3. Results and Discussions

Figure 1 shows the XRD patterns of the ErSZ pellet and the compounds doped with different amounts of TiO₂. In all the patterns, no diffraction peak ascribed to an Er₂O₃ or a TiO₂ phase could be detected, indicating that a ZrO₂-based solid solution has been produced. All the compounds are mainly composed of t′ phase. However, when the TiO₂ content reaches to 10 mol %, an m phase is formed. With the increase of the TiO₂ content, the m phase to t′ phase peaks intensity ratio increases, suggesting that more m phase forms. Since many diffraction peaks of t′ and c phases are often overlapped in the XRD patterns, it is difficult to preclude the c phase just according to Figure 1.

To further identify the phase constitution, {004} diffraction peaks at 2θ = 72°–76° were examined in detail by slow scans. As oberved in Figure 2, ErSZ only consists of a t′ phase, which decomposes to t₁’ and t₂’ phases in other specimens. Here, t′, t₁’ and t₂’ phases all have a tetragonal structure, but with different dopant contents. In the XRD patterns of the compositions with 5 mol % and 10 mol % TiO₂, a (400)_c peak appears indicative of the formation of a c phase. Higher TiO₂ doping causes the disappearance of the c phase. Therefore, according to the XRD analysis, it could be concluded that ErSZ is composed of a t′ phase, the compounds with 5 mol % consist of t′ (t₁’ + t₂’) and c phases, 10 mol % TiO₂ doping leads to the formation of t′ (t₁’ + t₂’), c and m phases. While at higher doping levels, the compounds have m and t′ phases. The phase contents of the compounds are listed in

Table 1. Phase compositions of 10 mol %, 15 mol %, and 20 mol % TiO₂-doped ErSZ.

TiO2 Content (mol %)	c	t1′ + t2′	m
10	7.8	88.7	3.5
15	0	91.8	8.2
20	0	87.6	12.4

. It could be found that 10 mol % TiO₂ doped ErSZ has the lowest m phase content, about 3.5 mol %.

Previous research has demonstrated that, in the TiO₂-Yb₂O₃-ZrO₂ solid solution, Ti⁴⁺ first occupies the interstitial sites when the content is ≤2.5 mol %, followed by substituting for Zr⁴⁺ at higher doping contents [28]. In TiO₂-doped YbSZ series, YbSZ consists of a t′ phase, small amounts of TiO₂ doping causes a c phase, and higher TiO₂ doping decreases the c phase content and produces an m phase. The solid-solution mechanism provides a reasonable explanation for the phase structure evolution of the TiO₂ doped compounds. In this study, the smallest doping content is 5 mol %, larger than the critical value for creating interstitial ions. Therefore, the evidence for the presence of interstitial ions cannot be observed through the XRD analysis. However, the phase evolution of TiO₂-doped ErSZ can be explained by the solid-solution mechanism proposed in [28]. The formation of the c phase is due to the decreased average coordination numbers of cations caused by the presence of Ti⁴⁺ interstitial ions. Since Ti⁴⁺ ions are not evenly distributed in the solid solution, c and t′ phases are both present in the compounds. At higher doping levels, the substitution of Ti⁴⁺ ion for Zr⁴⁺ generates overcrowding problems of O²⁻ ions, giving rising to a strain field in the lattice. Thus, the c and t′ phases become unstable, producing an m phase.

Figure 3 shows thermal diffusivities of different amounts of TiO₂-doped ErSZ in a temperature range of 20–1200 °C. Error bars are omitted due to their small size. The thermal diffusivity decreases with the temperature. At a selected temperature, the thermal diffusivity does not decrease monotonically with the TiO₂ content. Instead, it declines, followed by a growing tendency with increasing the TiO₂ content. The calculated thermal conductivity is shown in Figure 4. For each sample, the thermal conductivity first decreases with the temperature, and then rises when the temperature is above 1000 °C. There is a strong correlation between the thermal conductivity and the TiO₂ content. TiO₂ doping initially decreases the thermal conductivity, while at higher doping levels, the samples reveal an increased thermal conductivity. In TiO₂-doped ErSZ series, the composition with 10 mol % TiO₂ has the lowest thermal conductivity, with a value of 1.71 W/mK at 1200 °C.

Given the thermal conductivity theory in an electrical insulating substance, thermal conduction mainly depends on the phonons scattering. The stronger the phonon scattering, the shorter the phonon mean free path. A decline in the phonon mean free path leads to decreased thermal conductivity. The phonon mean free path is comprised of two parts, which can be expressed as follows [32,33]:

$$\frac{1}{l\left(\mathsf{\omega },T\right)}=\frac{1}{l_i\left(\mathsf{\omega },T\right)}+\frac{1}{l_{\mathrm{p}}\left(\mathsf{\omega },T\right)}$$

(4)

where l(ω, T) is the phonon mean free path, l_i(ω, T) and l_p(ω, T) are the specific phonon mean free path due to the phonon-phonon scattering and the defect scattering, respectively, and ω and T refer to frequency and the corresponding temperature.

The ability of phonon scattering has a relationship with the phonon scattering coefficient (Γ), which is determined by the difference of atomic mass (Γ_M) and that of the ionic radius (Γ_δ). Considering the effect of the anharmonicity in a solid solution expressed by an influence factor ε, the relation of Γ, Γ_M, and Γ_δ can be expressed in the following equations [27,33,34]:

$$\Gamma ={\Gamma }_{\mathrm{M}}+\mathsf{\epsilon }{\Gamma }_{\mathsf{\delta }}$$

(5)

$${\Gamma }_{\mathrm{M}}=\frac{{\displaystyle \sum f_i{(M_i-\overline{M})}^2}}{{(\overline{M})}^2}$$

(6)

$${\Gamma }_{\mathsf{\delta }}=\frac{{\displaystyle \sum f_i{({\mathsf{\delta }}_i-\overline{\mathsf{\delta }})}^2}}{{(\overline{\mathsf{\delta }})}^2}$$

(7)

where M_i, δ_i, and f_i represent the atomic mass, the ionic radius, and the fractional content of the imperfection, respectively, $\overline{M}$ and $\overline{\mathsf{\delta }}$ are the average atomic mass and mean ionic radius of a specific site in the lattice, respectively.

Zr⁴⁺, Er³⁺, and Ti⁴⁺ have Shannon effective ionic radius of 0.84 Å, 1.004 Å, and 0.74 Å, respectively [35]. The substitution of Ti⁴⁺ for Zr⁴⁺ could cause elastic field in the ZrO₂ lattice. Additionally, due to the mass difference between Ti⁴⁺ and Zr⁴⁺, mass fluctuation can also be introduced in the lattice. As a result, a large phonon scattering coefficient is obtained in the compounds, giving rise to stronger phonon scattering ability and a lower phonon mean free path [27,28,36]. Hence, the thermal conductivity of TiO₂-doped ErSZ decreases with the increase of the TiO₂ content. At high doping levels (>10 mol % TiO₂), the compounds have an increased thermal conductivity, resulting from the m phase formation. Since the m phase has higher thermal conductivity than those of c and t′ phases [2,20,21,37,38], it is reasonable to observe increased thermal conductivities for ErSZ doped with high TiO₂ contents.

The relative densities of the samples for mechanical measurements are above 95%. Vicker hardness (H_v) and toughness of the samples are both presented in Figure 5. It is possible to observe that TiO₂-doped ErSZ reveals a nearly composition independent hardness. However, the toughness is sensitive to the composition. Increasing the TiO₂ doping content, the toughness of TiO₂ doped ErSZ first exhibits an upward variation trend, and then decreases when the doping content is above 10 mol %. Specially, 10 mol % TiO₂ doped ErSZ has the highest toughness, with a value of ~6.09 MPa·m^1/2, nearly 35% and 30% higher than that of ErSZ and YSZ, respectively [13].

For a ZrO₂-based ceramic consisting of t′, c, and m phases, the toughness could be considered as the weighted average of each phase [39]. Therefore, a qualitative comparison among the toughness values of TiO₂-doped ErSZ can be made. ErSZ and the compositions with 5 mol % and 10 mol % TiO₂ contents are mainly composed of t′, while higher TiO₂ doping contents lead to the formation of an m phase. Since the t′ phase has higher toughness than that of the m phase [2,8,11,24], it is reasonable to obtain lower toughness when the doping content is at high levels.

As can be seen in Figure 1 and Figure 2, 5 mol % TiO₂-doped ErSZ is composed of t′ and c phases, while ErSZ only has a t′ phase. Since the c phase has a lower toughness than the t′ phase, the 5 mol % TiO₂-doped ErSZ should have lower toughness compared with ErSZ based on the aforementioned qualitative comparison, which is inconsistent with the result in Figure 5. Hence, another mechanism may contribute to the variation of the toughness. It is known that the high toughness of the t′ phase is due to the ferroelastic toughening. Based on the mechanism, the ferroelastic domain existing in the compounds would switch around the crack tip when a high stress is applied, which absorbs the energy and thus induces the toughening effect. The increased fracture energy due to domain switching can be expressed as follow [39,40,41]:

$$\Delta \Gamma =2fh{\mathsf{\tau }}_c\mathsf{\gamma }$$

(8)

where h represents the width of the domain switching process zone, f is the volume fraction of the domain that undergoes switching, τ_c is the coercive stress required for switching, and γ is the related ferroelastic strain, which is determined by the lattice tetragonality [39,40,41].

$$\mathsf{\gamma }=\frac{2}{3}(\frac{c}{a}-1)$$

(9)

where c and a are lattice constants of the t phase and c/a is the tetragonality. Assuming that the f and h values are identical for a series of compounds, the fracture energy increases with the increase of the lattice tetragonality.

Figure 6 shows the tetragonality of the t′ phase (t₁′ and t₂′) in TiO₂ doped ErSZ. The value increases with the increase of the TiO₂ content. As a result, the ferroelastic strain increases, causing higher fracture energy and thus better toughness if only lattice variation is considered. Taking into consideration the effects of the tetragonality and the amount of the t′ phase, the higher toughness of 5 mol % TiO₂-doped ErSZ compared with ErSZ suggests that the tetragonality of the t′ phase play a major role in affecting the toughness. However, the toughness increase rate from ErSZ to 5 mol % TiO₂-doped ErSZ is slower than that from 5 mol % TiO₂-doped ErSZ to the composition with 10 mol % TiO₂, indicating that the effect of the c phase formation cannot be neglected. When the TiO₂ doping content is larger than 10%, although the tetragonality of t′ phase still increases, the toughness shows a reduction trend, mainly attributable to the formation of the m phase and the obvious decrease in the t′ phase amount. In this case, the t′ phase content play a key role in affecting the toughness.

TiO₂ doping has visible effects on the phase composition, thermal conductivity, and toughness of ErSZ. Among the compositions in the present study, it could be found that 10 mol % TiO₂-doped ErSZ reveals the lowest thermal conductivity and the highest toughness, which are desirable attributes for TBC applications. Based on this study, our future work will focus on producing 10 mol % TiO₂-doped ErbSZ TBCs to exploit advantages of this TBC candidate material to the full.

4. Conclusions

Compounds in a TiO₂-Er₂O₃-ZrO₂ system were fabricated, and their phase composition, thermal conductivity, and toughness were investigated. ErSZ exhibited a t′ phase, 5 mol % TiO₂-doped ErSZ consisted of t′ and c phases, whereas the composition with 10 mol % TiO₂ had had t′, c, and about 3.5 mol % m phases. At higher TiO₂ doping levels, the compounds are composed of t′ and m phases. When the doping content is below 10 mol %, TiO₂ doping reduced the thermal conductivity and improved the toughness. The thermal conductivity and toughness of 10 mol % TiO₂-doped ErSZ were 1.71 W/mK (1200 °C) and ~6.09 MPa·m^1/2, respectively, obviously superior to those of ErSZ. However, higher TiO₂ doping contents had an adverse effect on the thermal conductivity and toughness of the compounds, which were attributed to the formation of the m phase. Given the low thermal conductivity and good toughness, 10 mol % TiO₂-doped ErSZ might be a promising TBC material.