**5.1. EB-PVD process**

system. Although extensive research has been initiated to study the effect of residual stress on TBC life, there is still ample scope to carry out this study on novel TBC systems involving novel compositions [9]. High residual stresses are induced in the TBC due to thermal expansion mismatch and bond coat (BC) oxidation leading to failure by spalling and delamination. An analysis of the stress distributions in TBC systems, which is a prerequisite for the understand‐ ing of failure mechanisms, was performed by Sfar et al. [10] using the finite element method (FEM). Cracks in the interface region were considered in the FE models in order to determine the loading conditions for their propagation and thus, the failure criteria of the TBCs as cracking usually occurs at or near the interfaces between BC/TGO and TBC/TGO depending on the processing mode of the TBC. The modified crack closure integral (MCCI) method combined with an FE analysis led to highly accurate energy release rate values. Moreover, this method enables the determination of mode-dependent energy release rates. TBC failure models could be developed and verified using this tool and appropriate crack propagation criteria [10]. Yang et al. [11] investigated the residual stress evolution in air plasma-sprayed yttria-stabilized zirconia (YSZ) TBCs after thermal treatments at 1,150°C. The residual stress in the YSZ layer was measured using Raman spectroscopy and the curvature method. Generally, as-deposited YSZ layer was under compressive stress and subsequently after thermal treatment for 30 h it was under tensile stress partly due to the monoclinic to tetragonal phase transformation in the YSZ layer. Sintering of the YSZ layer occurred with prolonged thermal treatment resulting in the gradual transformation of the residual stress, from tensile to compressive stress. Further, β-NiAl to γ/γ′-Ni3Al phase transformation in the bond coat also

plays an important role on the stress development in the top coat [11].

Top coat degradation is another parameter that governs TBC failure. The ceramic top coat has a tendency to crack due to stress generated from thermal expansion mismatch between the three layers of the TBC system. When the top coat cracks, oxygen easily diffuses to the bond coat leading to the catastrophic failure of the TBC system. Significant research is being carried out to improve the microstructure, mechanical properties, and stability of the ceramic top coat [12]. TBCs are subject to many kinds of degradation, e.g., erosion, foreign object damage (FOD), oxidation, etc., which deteriorate the integrity and mechanical properties of the whole system. Moreover, a new type of damage has been highlighted, i.e., corrosion by molten Calcium-Magnesium-Alumino Silicates, known as CMAS with the aim to increase the turbine inlet temperature. Basu et al. studied interactions between YSZ materials synthesized via the solgel process and synthetic CMAS powder via a step-by-step methodology. However, CMAS can cause faster sintering of the ceramic and thereby, leading to loss of strain tolerance in the protective coating. Further, a dissolution/re-precipitation mechanism between YSZ and CMAS resulted in the transformation of the initial tetragonal YSZ into globular particles of monoclinic zirconia. In addition, CMAS infiltrated both EB-PVD and sol-gel YSZ coatings at 1,250°C for 1 h [12]. Thompson and Clyne [13] deposited a vacuum plasma spray (VPS) MCrAlY bond coat and atmospheric plasma spray (APS) zirconia top coat onto a nickel superalloy substrate. They measured the stiffness of detached top coats by cantilever bending and also by nanoin‐ dentation technique. Measurements were made on as-sprayed specimens and after various

**4.3. Top coat degradation**

114 Advanced Ceramic Processing

In the EB-PVD process, the source material is heated with an electron beam, vapors are produced, and the evaporated atoms condense on the substrate. Crystal nuclei form on favored sites and grow laterally and in thickness to form individual columns that provide in-plane compliance [15]. A TGO layer often forms on the bond coat in these TBC systems and increases the residual stress. Further, brittleness of the top coat increases with the sintering of the coating. Consequently, the adhesion of the bond coat to the top coat becomes weak at high tempera‐ tures. Therefore, the TGO layer is very detrimental for TBC performance [16]. Movchan and Yakovchuk [17] described the design of a new generation of electron beam units for the deposition of the TBCs and cost-effectiveness of the one-step deposition process. They produced variants of graded TBC, which consist of bond coats of NiAl or MCrAlY+NiAl and YSZ-based outer ceramic layer in a one-step cycle by evaporation of a composite ingot. The composition and structure of the bond coats, outer ceramic layer, and the transition barrier zones of the substrate/bond coat and bond coat/outer ceramic layers was controlled in a broad range. They have shown distributions of chemical elements in the coating/substrate system and microstructure after deposition and after heat treatment. Various types of graded TBCs were subjected to thermal cycling tests at 1,150°C and their thermal cyclic resistance was monitored [17]. Current numerical approaches in modeling the intrinsic failure of TBC relies largely on the fact that spallation occurs when the accumulating strain energy stored in the coating exceeds a fixed critical value resembling interfacial adhesion. If this is to be entirely correct, one would expect that this critical value of interfacial adhesion varies with different materials, but stays independent of their thermal exposure history. Wu et al. [18] characterized the adhesion of oxide-bond coat interface among five systematically prepared material systems using a unique cross-sectional indentation technique. The results re-confirmed that interfacial adhesion is a material-specific property and the adhesion is dynamic, particularly with time and temperature. Certain parameters such as the oxide growth rate, rumpling of the oxidebond coat interface, and phase transformation of bond coat were studied as a function of thermal exposure to understand the dynamics. They clearly indicated that the oxide-bond coat interfacial adhesion depends strongly on the phase distribution of the bond coats and TGO growth rate while having little effect from TGO rumpling and residual stress [18].

### **5.2. APS process**

In the APS process, ceramic powders are introduced into a high temperature plasma plume, melted inside the plume, and accelerated towards the substrate wherein molten droplets spread and form splats that are rapidly quenched. In one pass, several successive splats are deposited on the substrate and the coating thickness is increased by means of several passes [19]. A typical fractured cross-section of the plasma sprayed ceramic coating show layers of splats along with interlamellar pores, cracks, and globular pores [15]. Coating compliance is increased by the presence of the cracks and thereby, extending their lifetimes [19]. The resulting coating microstructure is strongly dependent on processing conditions such as spray param‐ eters (e.g., torch current, plasma gas flow rate, carrier gas flow rate, torch traverse velocity, and stand off distance) and feedstock materials (e.g., size, temperature, and velocity). Splat morphologies are changed with the angle of impact of impinging particle [15]. Higher substrate temperatures lead to lower porosity and improved inter-splat contact resulting in enhanced coating properties [20]. During service operations at high temperatures, a TGO layer, mainly an Al2O3 layer, is developed between the bond coat and the top coat due to the oxidation of the bond coat. This is the most important factor that determines the lifetime of the TBC system. The thickness of this layer increases with increasing operation time. High stresses are present at the bond coat and TGO interface because of oxide layer growth, thermal expansion misfit, and applied loads. As a result, crack initiates and propagates resulting in spallation of the ceramic layer, and finally, system degradation [3]. During thermal exposure at ≥1,000°C, Ni(Cr, Al)2O4 (spinels) and NiO clusters are also formed at the interface of the Al2O3 layer and the ceramic coating in the TBC system with MCrAlY (M=Ni, Co) bond coat. Cracks were nucleated on these oxide clusters and grew into the ceramic coating leading to premature TBC separation. A heat treatment in a low pressure oxygen environment was found to promote the formation of a uniform, thin protective layer of Al2O3 at the ceramic-bond coat interface and can reduce these detrimental oxides [21].

Thermo-mechanical properties of TBCs have been studied to improve TBC performance. The Young's modulus of the ceramic top coat is an important factor that affects the thermal stress distribution in TBCs and thus, thermal fatigue behavior. Apparent Young's modulus (Eap) indicates the macro-elastic properties of the coatings. Eap of the top coat is usually much lower than the value for dense YSZ due to the porous microstructure. The extremely low Eap values are also attributed to the weak bonding between the particles because of the extremely high cooling rate. Tang and Schoenung [22] conducted bending tests of the TBC specimens exposed to thermal cycling to determine their Eap. The Eap decreased with increasing thermal cycles, up to certain thermal cycles, and then remained unchanged for increased thermal cycles. The breaking of the bonds at the splat boundaries or the formation of new cracks caused by thermal strain is the reason for the decrease in Eap with increasing thermal cycles. Effect of heat treatment on the elastic properties of the separated porous plasma sprayed zirconia TBCs was investigated by D. Basu et al. [23]. The depth-sensitive indentation technique was employed to determine the elastic moduli of the coatings. The characteristic moduli were dependent on the indentation load. The increase of moduli with decreased indentation load was attributed to the presence of small pores and micro-cracks at the subsurface. Heat treatment of the coatings at 1,100°C increased the elastic moduli appreciably due to the formation of sintering necks and the elimination of the micro-pores within the lamellae.

using a unique cross-sectional indentation technique. The results re-confirmed that interfacial adhesion is a material-specific property and the adhesion is dynamic, particularly with time and temperature. Certain parameters such as the oxide growth rate, rumpling of the oxidebond coat interface, and phase transformation of bond coat were studied as a function of thermal exposure to understand the dynamics. They clearly indicated that the oxide-bond coat interfacial adhesion depends strongly on the phase distribution of the bond coats and TGO

In the APS process, ceramic powders are introduced into a high temperature plasma plume, melted inside the plume, and accelerated towards the substrate wherein molten droplets spread and form splats that are rapidly quenched. In one pass, several successive splats are deposited on the substrate and the coating thickness is increased by means of several passes [19]. A typical fractured cross-section of the plasma sprayed ceramic coating show layers of splats along with interlamellar pores, cracks, and globular pores [15]. Coating compliance is increased by the presence of the cracks and thereby, extending their lifetimes [19]. The resulting coating microstructure is strongly dependent on processing conditions such as spray param‐ eters (e.g., torch current, plasma gas flow rate, carrier gas flow rate, torch traverse velocity, and stand off distance) and feedstock materials (e.g., size, temperature, and velocity). Splat morphologies are changed with the angle of impact of impinging particle [15]. Higher substrate temperatures lead to lower porosity and improved inter-splat contact resulting in enhanced coating properties [20]. During service operations at high temperatures, a TGO layer, mainly an Al2O3 layer, is developed between the bond coat and the top coat due to the oxidation of the bond coat. This is the most important factor that determines the lifetime of the TBC system. The thickness of this layer increases with increasing operation time. High stresses are present at the bond coat and TGO interface because of oxide layer growth, thermal expansion misfit, and applied loads. As a result, crack initiates and propagates resulting in spallation of the ceramic layer, and finally, system degradation [3]. During thermal exposure at ≥1,000°C, Ni(Cr, Al)2O4 (spinels) and NiO clusters are also formed at the interface of the Al2O3 layer and the ceramic coating in the TBC system with MCrAlY (M=Ni, Co) bond coat. Cracks were nucleated on these oxide clusters and grew into the ceramic coating leading to premature TBC separation. A heat treatment in a low pressure oxygen environment was found to promote the formation of a uniform, thin protective layer of Al2O3 at the ceramic-bond coat interface and can reduce

Thermo-mechanical properties of TBCs have been studied to improve TBC performance. The Young's modulus of the ceramic top coat is an important factor that affects the thermal stress distribution in TBCs and thus, thermal fatigue behavior. Apparent Young's modulus (Eap) indicates the macro-elastic properties of the coatings. Eap of the top coat is usually much lower than the value for dense YSZ due to the porous microstructure. The extremely low Eap values are also attributed to the weak bonding between the particles because of the extremely high cooling rate. Tang and Schoenung [22] conducted bending tests of the TBC specimens exposed to thermal cycling to determine their Eap. The Eap decreased with increasing thermal cycles, up

growth rate while having little effect from TGO rumpling and residual stress [18].

**5.2. APS process**

116 Advanced Ceramic Processing

these detrimental oxides [21].

Functionally graded Al2O3–ZrO2 TBC was prepared by plasma spraying technique and reported elsewhere [24]. Functionally-graded TBC was found to reduce the oxidation rate of the TBC system. Thus, large residual stress associated with the formation of TGO was minimized. The Al2O3 interlayer should be very thin to increase the adhesion of the layers. However, low fracture toughness of Al2O3 might lead to TBC failure. In addition, phase transformation of γ-Al2O3 to α-Al2O3 could induce additional residual stress, which should be minimized to get reliable TBC systems. Thick thermal barrier coatings (thickness >1 mm) have been developed for increased thermal protection by using the APS method [25]. However, low thermal shock resistance is the problem with the thick coating. Certain degrees of porosity and micro-cracks, preferably segmentation cracks, in TBCs favor to achieve high thermal shock resistance. Chen et al. [26] prepared a new functionally-graded thermal barrier coating based on LaMgAl11O19 (LaMA)/YSZ by using air plasma spraying technique. The coefficient of thermal expansion (CTE) of the functionally-graded coating varied gradually from the YSZ bottom layer to the LaMA top layer, resulting in the decrease in residual stress level than that of the LaMA/YSZ double ceramic layered TBC system. Excellent thermal cycling lifetime (~11,749 cycles at~ 1,372°C) of the functionally graded TBC proved the potential of these TBCs for advanced applications [26].
