Tunable properties of (Ho_xY_1-x)₂SiO₅ as damage self-monitoring environmental/thermal barrier coating candidates

Abstract

RE₂SiO₅ with low thermal conductivity, compatible thermal expansion coefficients and excellent high-temperature reliability in harsh environments are excellent candidates as advanced environmental/thermal barrier coating materials for high-efficiency gas turbine engines. A series of rare earth silicates (Ho_xY_1-x)₂SiO₅ are designed and their properties are comprehensively investigated in this paper. Through doping Ho into Y₂SiO₅, the thermal conductivity of Y₂SiO₅ is significantly decreased and the thermal expansion coefficient is also optimized closer to Si-based ceramics. High-temperature elastic stiffness and bending strength are increased with the enhancing of Ho content. Most important, doping Ho element provides (Ho_xY_1-x)₂SiO₅ with tunable luminescence characteristic. (Ho_xY_1-x)₂SiO₅ exhibit green, to yellow-green, then to orange-red luminescence color with increased Ho concentration. The results show that they can be used as damage self-monitoring environmental/thermal barrier coating materials for Si-based ceramics.

Introduction

Gas turbine engines have benefited from decades of development of nickel-based superalloys, however, operating temperatures are now reaching limits posed by the melting temperatures of these materials¹. Nowadays, higher operating temperatures are pursuing to achieve better efficiency and a reduction in environmentally harmful byproducts. Si-based ceramics such as SiC and SiC_f/SiC ceramic matrix composites are increasingly used as the hot-section components in gas turbine engines due to their excellent high-temperature properties. When exposed to high-speed water vapor they are subjected to severe recession and environmental barrier coating (EBC) materials are critically needed to protect them². Further increase in operating temperature, thermal barrier coating (TBC) is essential to allow a steep temperature gradient across it to lower the temperature of the matrix ceramics that for increasing the lifetime and efficiency of gas turbine engines². Therefore, an integrated environmental/thermal barrier coating (ETBC) material with the functions of corrosion resistance and thermal insulation is highly desirable. Rare earth (RE) silicates have been demonstrated to be the third generation of EBC for Si-based ceramics due to their excellent corrosion resistance to water vapor and molten silicates^3,4. Among them, Y₂SiO₅ is a promising EBC candidate with its merits of small density, low cost, abundant in the natural source, and relatively good thermal and mechanical properties^3,5. Previous work showed that thermal conductivity of Y₂SiO₅ is relatively higher than its RE₂SiO₅ counterparts and it needs modification on its performances for thermal insulation applications⁵. In addition, thermal expansion coefficients of Y₂SiO₅ is relatively larger than those of Si-based ceramics that may cause the exfoliation of coating material under longtime heating and cooling cycles. Consequently, it is urgent to lower the thermal conductivity and thermal expansion coefficients and improve the high-temperature mechanical properties of Y₂SiO₅ to realize its application as ETBC for Si-based ceramics.

In this work, the doping method was applied to improve the thermal and mechanical properties of Y₂SiO₅. Our prior work revealed that Ho can endow RE₂SiO₅ with excellent thermal and mechanical properties⁵. Therefore, Ho doping was adopted to modify the properties of Y₂SiO₅. (Ho_xY_1-x)₂SiO₅ solid solutions were prepared using the hot pressing method. Thermal conductivities, thermal expansion coefficients, bending strengths, elastic moduli and internal friction of (Ho_xY_1-x)₂SiO₅ were measured from room to high temperatures. The thermal conductivity of Y₂SiO₅ can be significantly reduced by introducing Ho into the crystal lattice and its high-temperature stiffness and bending strength can also be improved. In addition, (Ho_xY_1-x)₂SiO₅ were found to exhibit green, yellow-green and orange-red colors with various Ho doping concentration. This may contribute to damage self-detecting when the solid solution is adopted in the multi-layered or gradient ETBC coating. The present results show that (Ho_xY_1-x)₂SiO₅ solid solutions are promising damage self-monitoring ETBC candidates.

Experiment

Sample preparation

Bulk (Ho_xY_1-x)₂SiO₅ (x = 0, 1/3, 2/3, and 1) samples were prepared by hot pressing method. At first, Y₂O₃, Ho₂O_3, and SiO₂ powders were mixed according to stoichiometric ratios by ball milling for 24 h. The obtained slurry was dried at 60 °C for 24 h. The mixture was annealed at 1550 °C for 1 h to synthesize (Ho_xY_1-x)₂SiO₅ powders. Dense (Ho_xY_1-x)₂SiO₅ ceramics were hot pressed at 1600 °C. The densities of the as-synthesized samples were determined by Archimedes method and the densities of as-synthesized (Ho_xY_1-x)₂SiO₅ samples were more than 94% of the theoretical values. The phase compositions of samples were identified using X-ray diffractometer with CuKα radiation (D/max-2400, Rigaku, Tokyo, Japan). Microstructures were observed with a SUPRA 55 scanning electron microscope (LEO, Oberkochen, Germany).

Mechanical properties measurements

The dynamic Young’s modulus and internal friction of the samples were evaluated through an impulse excitation technique using samples with dimensions of 3 mm × 15 mm × 40 mm. The samples were measured in a graphite furnace RFDA-HTVP 1750C (IMCE, Diepenbeek, Belgium) at a heating rate of 4 °C/min in an Ar atmosphere. Vibration signals captured by a laser vibrometer were analyzed by a resonance frequency and damping analyzer, and Young’s modulus was calculated from the flexural resonant frequency, according to ASTME 1876–97.

Bending strength was measured in a universal testing machine (CMT4204, SANS, Shenzhen, China) using samples with dimensions of 3 mm × 4 mm × 36 mm. The three-point bending method with a crosshead speed of 0.5 mm/min was applied, and three samples for each (Ho_xY_1-x)₂SiO₅ specimen were measured.

Thermal properties measurements

Experimental thermal conductivity was determined from the measurements of thermal diffusivity a, heat capacity C_p, and density ρ:

$$\kappa =a{C}_{p}\rho $$

(1)

Thermal diffusivities were measured using a laser flash analyzer (Netzsch LFA 457, Bavaria, Germany) in an argon atmosphere from room temperature to 1273 K. Both sides were sprayed with a thin layer of colloidal graphite to ensure complete and uniform absorption of the laser pulse. Isobaric heat capacity C_p of (Ho_xY_1-x)₂SiO₅ were obtained from literature data in terms of their constituent binary oxides (Y₂O₃, Ho₂O_3, and SiO₂) by Neumann-Kopp rule⁶.

Thermal expansion coefficients were determined by temperature-dependent changes in the length of specimens from room temperature to 1473 K by using a vertical high-temperature optical dilatometer (ODHT, Modena, Italy). The dimensions of the samples are 3 × 4 × 14 mm³.

Luminescence properties measurements

UV-Vis absorption spectral studies were carried out by UV-VIS-NIR spectrophotometer (SolidSpec-3700DUV, Shimazu, Japan). Photoluminescence (PL) spectra were measured on a steady-state fluorescence spectrophotometer (Fluorolog-3-TAU, Horiba Jobin Yvon, France).

Results and Discussion

Figure 1a shows the X-ray diffraction (XRD) patterns of as-sintered (Ho_xY_1-x)₂SiO₅ samples. They are consistent with the standard XRD spectrum of X2-Y₂SiO₅ (ICCD PDF No. 74-2011). All the samples are pure enough without detectable impurities. With the increase of Ho concentration, XRD patterns show a blue shift (inset in Fig. 1a) and it proves that a number of Y atoms were substituted by Ho atoms in the lattice of X2-Y₂SiO₅. RE₂SiO₅ has two polymorphs, both of which are monoclinic with space groups P2₁/c for larger RE elements (X1 phase) and C2/c for smaller RE elements (X2 phase). The coordination of RE for the X1 and X2 phase are 9, 7 and 7, 6 respectively⁷. Y₂SiO₅ possesses both X1 and X2 phases depending on the fabrication temperature. In this work, the as-prepared samples belong to X2-RE₂SiO₅. Figure 1b is the crystal structure of X2-RE₂SiO₅ and it contains two RE atomic sites, one Si atomic sites, and five oxygen atomic sites. The ionic radius of Ho³⁺ and Y³⁺ are 0.894 and 0.88 Å (six coordination)⁸, respectively and their difference is within 15% which allowed the materials being substitutional solid solutions⁹. The doped Ho occupies the two RE atomic sites and the mismatch of ionic radii of Ho³⁺ and Y³⁺ results in enhanced lattice parameters causing the blue shift of XRD patterns.

Figure 1

(a) XRD patterns of as-sintered (Ho_xY_1-x)₂SiO₅ samples and (b) crystal structure of X2-RE₂SiO₅.

Figure 2a presents Young’s moduli and shear moduli of (Ho_xY_1-x)₂SiO₅. Both of them decrease with the increase of Ho concentration. High-temperature Young’s moduli and internal friction were measured and shown in Fig. 2b. The Young’s moduli decrease linearly from room temperature to 1600 K. Then, they drop relatively faster accompanied by an exponential increase of internal friction. The quick decrease of Young’s moduli and increase of internal friction suggest the stiffness softening of the samples. The Young’s moduli of Y₂SiO₅ and (Ho_1/3Y_2/3)₂SiO₅ are close in magnitudes and those of (Ho_2/3Y_1/3)₂SiO₅ and Ho₂SiO₅ are nearly the same. Ho₂SiO₅ has excellent high-temperature stiffness retention compared with that of Y₂SiO₅ because its Young’s modulus can be measured up to 1857 K. For Y₂SiO₅, Young’s modulus can only be measured up to 1784 K and no signal can be detected at higher temperature due to weak response and strong noise. When doping with Ho, the high-temperature stiffness can be obviously improved. The Young’s moduli of doped samples can be retained at a higher temperature than that of Y₂SiO₅. Temperature dependence of Young’s modulus for crystalline materials can be attributed to the anharmonic effects of lattice vibrations and decreases in bond strengths at elevated temperature.

Figure 2

(a) Young’s moduli and shear moduli; (b) temperature dependent Young’s moduli and internal friction of (Ho_xY_1-x)₂SiO₅.

As the Ho doping effectively improves the high-temperature stiffness, the high temperature bending strength of (Ho_xY_1-x)₂SiO₅ were measured and exhibited in Fig. 3a. Both room temperature and high temperature bending strengths of Ho₂SiO₅ are higher than that of Y₂SiO₅. When doping with Ho, bending strengths of (Ho_1/3Y_2/3)₂SiO₅ and (Ho_2/3Y_1/3)₂SiO₅ at 1573 K are 234 ± 25 and 219 ± 28 MPa, respectively, they are higher than that of Y₂SiO₅ and lower than that of Ho₂SiO₅. The force constant is proportional to the strength of a chemical bond. Luo et al. and Li et al.^10,11 calculated the average interatomic force constants in RE₂SiO₅. Their theoretical results showed that average force constants of Y-O and Ho-O are 2.6 and 6.5 eV/Ų in Y₂SiO₅ and Ho₂SiO₅, respectively. Therefore, incorporating Y-O bond with smaller force constant corresponds to the relatively lower strengths of (Ho_1/3Y_2/3)₂SiO₅ and (Ho_2/3Y_1/3)₂SiO₅ than that of Ho₂SiO₅.

Figure 3

(a) Bending strengths of (Ho_xY_1-x)₂SiO₅ at room and high temperatures; (b–e) morphologies of the fracture surface of (Ho_xY_1-x)₂SiO₅ after bending strength at 1573 K.

Therefore, high-temperature mechanical properties can be improved through doping with Ho. To investigate the mechanism of high-temperature enhancement, the morphologies of the fracture surface after bending strength test at 1573 K are illustrated in Fig. 3b–e. Grains can be obviously found in the fracture surface of Y₂SiO₅ indicating an intergranular fracture. On the contrary, (Ho_1/3Y_2/3)₂SiO₅, (Ho_2/3Y_1/3)₂SiO₅ and Ho₂SiO₅ present mostly transgranular fracture. As we know, grain boundaries are weakly bounded interfaces where fractures usually generate. In transgranular fracture, cracks go through the crystals instead of only along grain boundaries. Therefore, Ho doping can modify the high-temperature mechanical properties through the change of fracture modes at elevated temperature.

Thermal conductivity is one of the key parameters for ETBC candidates. Figure 4a is the temperature dependent thermal diffusivities of (Ho_xY_1-x)₂SiO₅. They decrease with the increase of temperature and then gradually approach steady at high temperature. Thermal diffusivity of Y₂SiO₅ is much higher than that of Ho₂SiO₅ and the values could be decreased with the rise of Ho content. With the increase of temperature, the gap of thermal diffusivity between Y₂SiO₅ and Ho₂SiO₅ is reduced. Thermal conductivity can be obtained based on Eq. 1 and the results are shown in Fig. 4b. Thermal conductivities of (Ho_1/3Y_2/3)₂SiO₅ and (Ho_2/3Y_1/3)₂SiO₅ are lower than that of Y₂SiO₅. The room temperature thermal conductivity of (Ho_1/3Y_2/3)₂SiO₅ is slightly higher than that of (Ho_2/3Y_1/3)₂SiO₅. They tend to be close as temperature increases. Callaway suggested that the thermal conductivities of a solid containing defects can be calculated by¹²:

$$\frac{\kappa }{{\kappa }_{P}}=\frac{{\tan }^{-1}(\beta )}{\beta }$$

(2)

where κ_P and κ are the lattice thermal conductivities of the parent and defected solids, respectively, and β is defined by:

$$\beta =(\frac{{\pi }^{2}{{\rm{\Theta }}}_{D}{\rm{\Omega }}}{h{\upsilon }_{m}^{2}}{\kappa }_{P}{\rm{\Gamma }})$$

(3)

where Ω is average volume per atom, h is Planck’s constant, Θ_D is Debye temperature, υ_m is average sound velocity and Γ is scattering coefficient. Scattering coefficient Γ is consist of two parts: one is mass fluctuation and the other is strain field fluctuation which is defined as:

$${\rm{\Gamma }}={x}_{i}[(\frac{{M}_{i}-\bar{M}}{\bar{M}})+\varepsilon {(\frac{{\delta }_{i}-\bar{\delta }}{\bar{\delta }})}^{2}]$$

(4)

where x_i is the concentration of defects, M_i is the atomic mass of the dopant, δ_i is the ionic radius of the dopant, $\bar{M}$ and $\bar{\delta }$ are the average atomic mass and mean ionic radius in the specific site in the solid solutions and ε is strain field factor. Therefore, we can find that the decrease of thermal conductivity of (Ho_xY_1-x)₂SiO₅ mainly origin from the fluctuations of mass $(\frac{{M}_{i}-\bar{M}}{\bar{M}})$ and strain filed $(\frac{{\delta }_{i}-\bar{\delta }}{\bar{\delta }})$. The molar mass and ionic radius of Ho and Y are 164.9 and 88.9 g/mol, and 0.894 and 0.88 Å, respectively⁸. The large difference between molar mass and ionic radius would cause obvious lattice distortion and enhance phonon scattering, and finally reduce the thermal conductivity. Figure 4c shows the temperature dependent thermal conductivities of several promising TBC candidates¹³. The thermal conductivity of (Ho_1/3Y_2/3)₂SiO₅ and (Ho_2/3Y_1/3)₂SiO₅ are much lower than most of the candidates and they are closed to that of Gd₂Zr₂O₇. The thermal conductivity of the widely used thermal barrier coating material 7YSZ is nearly 1 W/(m·K) higher than that of (Ho_1/3Y_2/3)₂SiO₅ and (Ho_2/3Y_1/3)₂SiO₅. Therefore, it is an efficient method to decrease the thermal conductivity Y₂SiO₅ by doping Ho.

Figure 4

(a) Thermal diffusivities of (Ho_xY_1-x)₂SiO₅ and thermal conductivities of (b) (Ho_xY_1-x)₂SiO₅ and (c) several TBC candidates.

Compatible thermal expansion coefficients with Si-based substrates is of great significance for ETBC. Figure 5 illustrates the temperature dependent thermal expansion coefficients of (Ho_xY_1-x)₂SiO₅. The thermal expansion coefficients are shown from 500 K and the lower temperature values are omitted due to strain release between the samples and the test fixture¹⁴. Thermal expansion coefficients of Y₂SiO₅ and Ho₂SiO₅ are close and they are obviously larger than that of Si-based ceramics. Through doping with 2/3 Ho, thermal expansion coefficients can be slightly decreased, such as 6.86 × 10⁻⁶/K at 1473 K. When doped with 1/3 Ho, thermal expansions coefficient can be significantly reduced and are much closer to that of SiC ceramic, especially at high temperature, for instance, 5.89 × 10⁻⁶/K at 1473 K. Reduction of thermal expansion coefficients may originate from the doping induced distortion of crystal lattice. Thermal expansion of RE₂SiO₅ is a consequence of the anharmonicity of lattice vibrations and is related to the magnitude and sign of Grüneisen constants of phonons. Luo et al.¹⁰ found two groups of phonons in X2-Y₂SiO₅, one group with positive Grüneisen constants, and the other with negative Grüneisen constants. These two species of phonons contribute to positive and negative thermal expansion, respectively; and thermal expansion coefficient is a compromise of these two contributing contents. Li et al.¹¹ also found that low-frequency phonons in X2-RE₂SiO₅ extensively have negative mode Grüneisen constants and these phonons contribute to negative thermal expansion. Doping in RE₂SiO₅ may cause more low-frequency phonons having negative mode Grüneisen constants and then contributing more to negative content of thermal expansion. As a result, macroscopic thermal expansion coefficients of solid solutions have smaller magnitude compared with two end members, Y₂SiO₅ and Ho₂SiO₅. The reduced thermal expansion coefficients of (Ho_1/3Y_2/3)₂SiO₅ makes it a promising candidate of ETBC.

Figure 5

Thermal expansion coefficients of (Ho_xY_1-x)₂SiO₅ and some Si-based ceramics.

ETBC are susceptible to the extremely high-temperature combustion environment. Degradation of the coating can occur by gradual thinning and erosion from the surface ultimately leading to complete failure. Visual examination through a borescope cannot identify such gradual degradation; only total delamination can be detected. Therefore, a quantitative nondestructive diagnostic approach should be developed to provide an adequate warning before the decreased environmental or thermal protection due to reductions in ETBC thickness or safety threatening thresholds¹⁵. Non-contact luminescence method was first put forward by Amano et al. to monitor the health of TBCs¹⁶. Trivalent lanthanide ions were introduced into the crystal structure, local information can be collected through conveyed by the luminescence emissions from optically excited luminescent layers. It is suggested that embedding luminescent sublayers into ETBCs to serve as erosion markers, so that when erosion exposes the luminescent “marker” layer, the luminescence characteristic of that layer will be produced when illuminated by the appropriate excitation wavelength¹⁷. To extend our knowledge of RE₂SiO₅ as self-monitoring ETBCs, the luminescence properties of (Ho_xY_1-x)₂SiO₅ were comprehensively investigated.

Figure 6a displays the UV-Vis absorption spectra of (Ho_1/3Y_2/3)₂SiO₅, (Ho_2/3Y_1/3)₂SiO₅, Ho₂SiO₅ and Ho₂O₃. They exhibit abundant peaks from 300 to 700 nm. The strongest peaks are located around 450 nm. The intensities of peaks are in direct proportion to the concentration of Ho. With the increase of Ho content, the intensities of peaks enhance gradually. Figure 6b shows the excitation spectra of (Ho_1/3Y_2/3)₂SiO₅, (Ho_2/3Y_1/3)₂SiO₅, and Ho₂SiO₅. They present the same excitation peaks and strong peaks appear at 448 nm. Photoluminescence spectra of (Ho_1/3Y_2/3)₂SiO₅, (Ho_2/3Y_1/3)₂SiO₅, and Ho₂SiO₅ were measured using 448 nm excitation light and displayed in Fig. 7. The main emission band of Ho₂SiO₅ is centered at 664 nm corresponding to the ⁵F₅ to ⁵I₈ transition. Another weak emission band is located at 549 nm induced by ⁵S₂ + ⁵F₄ to ⁵I₈ transition. (Ho_2/3Y_1/3)₂SiO₅ contains two emission band, and emission band at 549 nm is stronger than that of Ho₂SiO₅. (Ho_1/3Y_2/3)₂SiO₅ also possesses two main emission bands. Emission band at 664 nm is weaker than those of (Ho_2/3Y_1/3)₂SiO₅ and Ho₂SiO₅ and it exhibits stronger emission band at 549 nm.

Figure 6

(a) Absorption and (b) excitation spectra of (Ho_xY_1-x)₂SiO₅.

Figure 7

(a) Emission spectra and (b) energy level transition diagram of (Ho_xY_1-x)₂SiO₅.

The CIE (Commission Internationale de L’Eclairage) chromaticity diagram of (Ho_1/3Y_2/3)₂SiO₅, (Ho_2/3Y_1/3)₂SiO₅, Ho₂SiO₅ are presented in Fig. 8. With the increase of Ho concentration, luminescence color changes from green ((Ho_1/3Y_2/3)₂SiO₅) to yellow-green ((Ho_2/3Y_1/3)₂SiO₅), then to orange-red (Ho₂SiO₅). When (Ho_xY_1-x)₂SiO₅ silicates are used as different players in ETBCs, quantitative assessment of coating thickness degradation and erosion can be made using these luminescent layers of known position and thickness. Therefore, through the design of Ho concentration in a multi-layered or gradient ETBC coating, the luminescence color can be modified in a wide range. Above all, (Ho_xY_1-x)₂SiO₅ solid solutions with tunable luminescence color, low thermal conductivity, and matchable thermal expansion coefficients are novel damage self-monitoring ETBC candidates.

Figure 8

CIE chromaticity diagram of (H_1/3Y_2/3)₂SiO₅, (Ho_2/3Y_1/3)₂SiO₅ and Ho₂SiO₅.

Conclusions

(Ho_xY_1-x)₂SiO₅ solid solutions are prepared by the hot pressing method. The impact of Ho substitution on the thermophysical properties of (Ho_xY_1-x)₂SiO₅ solid solutions is investigated. Young’s modulus can be slightly decreased and high-temperature stiffness can be enhanced. The thermal conductivity of Y₂SiO₅ is significantly reduced by doping Ho due to the multiple enhanced phonon scattering mechanisms. Thermal expansion coefficients can also be modified closer to Si-based ceramics. In addition, (Ho_xY_1-x)₂SiO₅ exhibits tunable luminescence colors. These silicates present green, to yellow-green, then to orange-red colors with increased Ho concentration. This work highlights (Ho_xY_1-x)₂SiO₅ a novel damage self-monitoring ETBC candidates for Si-based ceramics.