High-Temperature Effects on TGO Growth and Al Depletion in TBCs of Ni-Based Superalloy GTD111

Abstract

Thermal barrier coatings (TBCs) extend gas-turbine blade lifetime by improving high-temperature oxidation resistance and mechanical performance. We investigated the microstructural evolution, TGO growth, and Al depletion in air-plasma-sprayed (APS) single-layer YSZ top coat over a NiCrCoAlY bond coat on Ni-based superalloy circular plates, heat treated isothermally at 850 °C and 1000 °C for 50–5000 h. Cross-sectional SEM/EDS analysis showed TGO quadratic thickening kinetics at both temperatures, reaching ~10 µm at 1000 °C/5000 h, the growth rate of which was ~5.8 times higher than at 850 °C. On top of the single-layer TGO of Al₂O₃ observed from the onset, a NiCrCo oxide layer appeared and grew from ≥500 h at 850 °C, with increasing growth rate and cracking. The layer configuration of the YSZ top coat, the TGO of Al₂O₃, and the bond coat (comprising β-NiAl and γ-NiCr) on top of GTD111, showed an Al concentration gradient in the bond coat starting at 850 °C for 250 h, which intensified with increased duration and temperature. The decrease in Al concentration in the bond coat and the growth of TGO are due to the dissolution of β-NiAl and subsequent Al diffusion to the Al₂O₃ TGO.

1. Introduction

Thermal barrier coatings (TBCs) are used for the lifetime extension of superalloys with an additional layer of protection against high-temperature environments, reducing the thermal load on the underlying superalloy substrate and delaying the onset of oxidation and phase degradation [1,2]. These coatings are essential for gas turbines, where components are subjected to extreme temperatures due to the direct exposure to combustion gases. TBCs mitigate rapid oxidation and degradation, ensuring the durability of turbine components in harsh environments [3].

TBC systems typically consist of a ceramic top coat and a metallic bond coat. The bond coat, often composed of MCrAlY alloys (where M represents nickel and/or cobalt), enhances oxidation resistance and acts as a diffusion barrier [2,3,4]. This metallic layer consists of a solid solution γ matrix and other phases, such as β-NiAl and σ-Cr/Co phases.

The ceramic top coat, commonly YSZ, provides low thermal conductivity and high thermal expansion compatibility, reducing thermal stresses between the coating and the substrate [3,4,5]. This combination enhances engine efficiency by enabling higher operating temperatures, contributing to the overall performance of gas turbines [1,2,3]. In addition, recent works on novel coating systems such as high-entropy alloys and doped zirconia materials provide alternative strategies for TBC design [6,7,8]. Y was added to the top coat to improve adhesion [4].

Despite their advantages, TBCs are susceptible to degradation over time, primarily due to the formation of a TGO layer at the interface between the bond coat and the ceramic top coat. The TGO layer, composed predominantly of Al₂O₃, forms as the bond coat and oxidizes during high-temperature exposure [5,6]. After prolonged exposure, the TGO layer evolves from single-layer Al₂O₃ to a double-layer structure containing mixed oxides, accelerating crack formation and propagation [2,3,4].

Under isothermal high-temperature exposure, interfacial TGO thickness follows a quadratic law. In this regime, stresses arise mainly from growth strain in the TGO and are partly relaxed by high-temperature creep in the TGO and bond coat during the hold, and consequently, spallation propensity is lower than that under cyclic duty. Degradation is therefore dominated by Al reservoir depletion in the β-rich bond coat and gradual interfacial roughening rather than cycle-by-cycle rumpling [7,8,9].

Growth of α-Al₂O₃ in the TGO generates substantial growth-related and thermal-mismatch stresses that, as the scale thickens and roughens, sharpens the stress gradients and raises the likelihood of interfacial cracking, delamination, and spallation under engine-like thermal cycling [10,11,12,13]. In Ni-based bond coats, as this mechanical driving force is tightly coupled to chemistry, the phase constitution governs the Al activity required to sustain selective, adherent alumina. Al-rich β-NiAl functions as the primary Al reservoir, consumed by surface oxidation and interdiffusion. This depletion is tracked by the β→γ/γ′ transformation and the emergence of γ′-Ni₃Al precipitates, signaling a drop in Al chemical potential and a diminishing capacity to replenish the TGO, pushing the coating toward a critical Al threshold [14,15]. Consequently, coatings with a higher initial β fraction maintain the α-Al₂O₃ longer, whereas γ/γ′-strengthened microstructures trade some Al-supply capacity for improved ductility and thermal-fatigue resistance [13]. These trends are evident when β-NiAl-based bond coats are compared to γ–γ′ NiPtAl based ones. The selective alumina persists only while sufficient Al remains in γ/γ′-dominated systems, after which adherence and oxidation behavior degrade [16]. At elevated temperature, creep in the bond coat partially relaxes growth stress but amplifies interfacial undulation, concentrating tensile stresses at the TC/TGO interface and exacerbating spallation during cycling [11]. Alloying can modify this chemo-mechanical evolution. For example, Pt addition slows β→γ′ transformation and improves scale adhesion, effectively preserving the Al reservoir and delaying depletion [13]. To address these challenges, researchers have explored advanced TBC designs, such as multilayer systems and alternative top coat materials. Single-ceramic-layer coatings, incorporating gadolinium zirconate (GZ) alongside YSZ, have demonstrated improved thermal insulation and resistance to degradation due to their lower thermal conductivity and superior phase stability [17,18,19,20]. Functionally graded coatings and nanostructured top coats have also shown promise in reducing TGO growth and enhancing TBC performance under extreme conditions [13,14,15,16,17,20,21,22].

Unlike prior studies on GTD111 that examined bare-alloy oxidation at 900 °C for ≤452 h or focused on TBC failure modes without comprehensive, long-duration interfacial kinetics on GTD111, this work quantifies isothermal TGO growth on an APS YSZ/NiCrCoAlY system over 50–5000 h at 850 °C and 1000 °C, extracting K_p and an Arrhenius activation energy while simultaneously mapping β-NiAl depletion/γ′ precipitation after K2 etching. We further correlate TGO thickening with insulation-performance implications. This, combined with a GTD111-specific, time–temperature matrix and a microstructure–kinetics–performance linkage, addresses a gap in the literature [10,23].

2. Materials and Methods

The specimens used here were one-sided TBC-coated plates composed of a NiCrCoAlY bond coat applied by vacuum-plasma spraying (VPS) and a YSZ top coat applied by APS on a circular GTD111 alloy substrate of 13 mm in diameter and 3 mm thick (

Table 1. Chemical composition of the GTD111 specimens in this study (wt.%).

Elements	Cr	Co	Mo	W	Ta	Ti	Al	C	Zr	Ni
GTD111	14.0	10.0	1.5	4.3	4.7	2.7	4.0	0.08	0.03	Bal.

). The TBC-coated specimens were then isothermally heat treated at 850 °C and 1000 °C for 50–5000 h (

Table 2. Heat treatment conditions of the TBC-coated GTD111 specimens under atmospheric environment.

Temperature (°C)	Duration (h)
850	50	100	250	500	1000	3000	5000
1000	50	100	250	500	1000	3000	5000

).

For the cross-sectional analyses, the specimens were cut, mounted, and polished for microstructural and compositional analyses using SEM with a JEOL JSM-6510, 20 kV (JEOL Ltd., Japan) and JEOL JSM7900F, 15 kV (JEOL Ltd., Japan) equipped with EDS, Oxford X-Max 80 (Oxford Instruments, UK) equipped with an image analyzer (Image J 1.8.0).

3. Results and Discussion

Figure 1 shows the progressive growth and morphological evolution of the TGO layers (formed above the bond coat) at 850 °C after different exposure times. The respective average thicknesses of the top coat and bond coat were 529.77 ± 23.72 μm and 117.51 ± 3.22 μm at 850 °C; the corresponding standard errors were 8.96 and 1.22 μm, respectively.

After 50 h of exposure, the TGO layer, appearing as a crack-like and/or a thick black borderline (as indicated by the yellow arrow), was observed to be thin (~0.7 µm) and continuous (Figure 1d), with few microcracks or little delamination at the interface between the TGO and the top coat. The bond coat remained intact, and the top coat showed microstructural stability, indicating minimal thermal degradation at this early stage. By 1000 h, the TGO layer had thickened significantly (~1.96 µm), showing the onset of oxidation-driven growth (Figure 1e).

At 5000 h, the TGO layer exhibited a substantial increase in thickness to ~4 µm, accompanied by porosity and potential cracks (grain-boundary-like dark lines/microcracks indicated by black arrows in Figure 2 and Figure 5) near the TGO-top coat interface. However, no interfacial cracks were observed between the TGO-and the bond coat. Average TGO thickness was 1.87 ± 1.29 μm, with standard error of 0.49.

The microstructural changes become apparent with irregularities at the interface in the vicinity of the TGO and the bond coat, suggesting the presence of localized stresses due to the differential thermal expansion and localized growth of oxides. The top of the bond coat showed continuous γ belt growth into the bond beneath the TGO layer, which is an early stage of Al depletion adjacent to the interface (Figure 1e), due to the diffusion of Al to form the TGO [17,24]. Keeping the thickness of the TGO layer below 6 μm is critical for maintaining the strength of the protective Al₂O₃ layer, which is chemically stable but mechanically weak, especially when it is thicker than 6 μm [24].

These features highlight the advanced stage of oxidation, where the diffusion of oxygen through the TGO and the depletion of Al in the bond coat accelerate the degradation process.

Figure 2 shows the microstructural degradation of the specimens at 1000 °C, demonstrating the challenges faced by the TBC system at elevated temperatures. The roles of Ru-containing superalloys and advanced NiAl coatings have also been highlighted in the literature, and support the present findings regarding bond coat evolution [25,26,27]. At 1000 °C, the average top-coating layer thickness was 525.54 ± 11.10 μm (SE = 4.20), and the bond-coat thickness was 124.36 ± 2.88 μm (SE = 1.09).

Initially, the thickness of the TGO layer after 50 h, ~2.14 µm (Figure 2d), was notably higher than that of the 850 °C specimen after the same exposure duration. This rapid growth reflects enhanced diffusion of oxygen and Al, promoting faster oxidation in the bond coat. The TGO layer at this stage appeared relatively uniform and continuous, effectively maintaining the protective properties of the Al₂O₃ scale without delamination at the interface.

Under the present isothermal exposures, TGO growth and rumpling increased interfacial stresses, and as a result, microcracks were occasionally noted along the TGO/YSZ interface. Nevertheless, these did not coalesce into macroscopic spallation. This behavior is consistent with the higher spallation propensity under thermal cycling, where repeated expansion–contraction raises the strain-energy release rate at the top-coat and bond-coat interface beyond the adhesion toughness. When the TGO thickness approached ~9.9 µm at 1000 °C for 5000 h (Figure 3), microcrack density increased, indicating elevated susceptibility; thus, while gross delamination was absent here, a lower spallation threshold would be expected under non-steady engine operation [8,9].

Figure 4 presents the linear growth of the TGO layer as a function of exposure time at 850 °C and 1000 °C. Over 50–5000 h, the squared thickness ${\mathrm{h}}^2$ increases linearly with time, consistent with diffusion-controlled (quadratic) oxidation. Accordingly, the thickness $\mathrm{h}\left(\mathrm{t}\right)$ follows Equation (1), where ${\mathrm{K}}_{\mathrm{p}}$ is the quadratic rate constant, and C accounts for any initial oxide thickness/formation.

$${\mathrm{h}}^2={\mathrm{K}}_{\mathrm{p}}\mathrm{t}+\mathrm{C}$$

(1)

$${\mathrm{K}}_{\mathrm{p}}={\displaystyle \frac{{\mathrm{d}(\mathrm{h}}^2)}{\mathrm{d}\mathrm{t}}}$$

(2)

Equations (1) and (2) yield ${\mathrm{K}}_{\mathrm{p}}$ rate constant values for the (Zr,Y)O top coating and NiCrCoAlY bond coating of $8.87\times {10}^{-7}{\mathsf{\mu }\mathrm{m}}^2/\mathrm{s}$ at 850 °C and $5.10\times {10}^{-6}{\mathsf{\mu }\mathrm{m}}^2/\mathrm{s}$ at 1000 °C. As expected, the rate constant increases with temperature, indicating faster TGO growth at 1000 °C. This behavior is mathematically expressed by the Arrhenius relation in Equation (3), which can be cast into the linear form of Equation (4).

$$\mathrm{l}\mathrm{n}{\mathrm{K}}_{\mathrm{p}}={\mathrm{l}\mathrm{n}\mathrm{K}}_0-{\displaystyle \frac{\mathrm{Q}}{\mathrm{R}\mathrm{T}}}$$

(3)

$$\mathrm{y}=\mathrm{a}\mathrm{x}+\mathrm{C}$$

(4)

Here, $\mathrm{Q}$ represents the activation energy for TGO growth, $\mathrm{R}$ is the gas constant and $\mathrm{T}$ temperature. ${\mathrm{K}}_0$ is the pre-exponential factor obtained from the Arrhenius plot. Using the two temperatures (${\mathrm{T}}_{850°\mathrm{C}}=1123.15 \mathrm{K}$ and ${\mathrm{T}}_{1000°\mathrm{C}}=1273.15 \mathrm{K}$), the slope of $\mathrm{l}\mathrm{n}{\mathrm{K}}_{\mathrm{p}}$ as plotted versus $1/\mathrm{T}$ yields the activation energy and the intercept yields the pre-exponential factor: $\mathrm{Q}=138.6 \mathrm{k}\mathrm{J}/\mathrm{m}\mathrm{o}\mathrm{l}$, ${\mathrm{K}}_0=2.48\times {10}^{-12}{ \mathrm{m}}^2/\mathrm{s}$ for the present NiCrCoAlY system. The activation energy depends on the deposition method used for the TBCs as well as on the exposure time and conditions. Other reported values for TGO growth kinetics are 188 kJ/mol for APS YSZ/APS NiCrAlY [28], and 347 kJ/mol for EB-PVD YSZ/NiCoCrAlY system [29].

Our obtained value for the activation energy shows that the alumina growth was diffusion-assisted rather than purely lattice-controlled [30]. These derived values, together with the obvious linear dependence of h² on t and the single value for the Arrhenius activation energy indicate that TGO layer thickening during the long-term exposure is governed predominantly by diffusion through the oxide scale. The K_p that is ~5.8 times higher at 1000 °C than at 850 °C indicates a reduced oxidation margin at the higher temperature.

The bond coat shows extensive Al depletion, impairing its ability to regenerate the protective Al₂O₃ layer and mixed oxides between the TGO and the top coat.

Selective formation of α-Al₂O₃ is favored because the Gibbs free energy of formation of alumina (${∆\mathrm{G}}_{{\mathrm{A}\mathrm{l}}_2{\mathrm{O}}_3}^{\mathrm{o}}\approx -1329 \mathrm{k}\mathrm{J}·{\mathrm{m}\mathrm{o}\mathrm{l}}^{-1} \mathrm{a}\mathrm{t} 1100 \mathrm{K})$ is far more negative than that of NiO (${∆\mathrm{G}}_{\mathrm{N}\mathrm{i}\mathrm{O}}^{\mathrm{o}}\approx -141 \mathrm{k}\mathrm{J}·{\mathrm{m}\mathrm{o}\mathrm{l}}^{-1} \mathrm{a}\mathrm{t} 1100 \mathrm{K})$ or Cr₂O₃ (${∆\mathrm{G}}_{{\mathrm{C}\mathrm{r}}_2{\mathrm{O}}_3}^{\mathrm{o}}\approx -852 \mathrm{k}\mathrm{J}{·\mathrm{m}\mathrm{o}\mathrm{l}}^{-1} \mathrm{a}\mathrm{t} 1100 \mathrm{K})$. As long as the local Al activity, ${\mathrm{a}}_{\mathrm{A}\mathrm{l}}$ remains above a critical threshold, alumina is preferentially sustained. As oxidation and interdiffusion proceed, β becomes unstable, and the microstructure evolves along β → γ tie-lines, marking a reduced ability to replenish alumina. This behavior can be framed by the thermodynamic flux (Equation (5)):

$${\mathrm{J}}_{\mathrm{A}\mathrm{l}}\propto {\mathsf{\mu }}_{\mathrm{A}\mathrm{l}}$$

(5)

Here, the surface oxide imposes a low ${\mathsf{\mu }}_{\mathrm{A}\mathrm{l}}$ condition, and interdiffusion into the γ/γ′ substrate further draws Al from the β-phase in the bond coat. For the proportion constant calculation of (Equation (5)), typical values in the literature for Al activity, about 0.30 in the β-rich interior and 10−10 to 10−12 at the alumina-equilibrated interface are used [31,32]. The thermodynamic driving-force drop across the coating is about ~247 kJ/mol at 850 °C and ~280 kJ/mol at 1000 °C. This corresponds to average driving-force gradients of roughly ~2.10 kJ/mol·µm at 850 °C and ~2.25 kJ/mol·μm at 1000 °C. Under isothermal conditions, this persistent driving force is maintained by ongoing scale growth and interdiffusion, so Al continues to flow from the β-phase in the bond coating layer toward the TGO surface. As the reservoir drains, β becomes unstable, which is the microstructural signature of falling Al availability. This interpretation is consistent with CALPHAD/DICTRA studies of β-depletion in MCrAlY and with quaternary Ni–Al–Cr–Co thermodynamic descriptions used to map phase fields at 850–1000 °C [32,33,34,35,36]. Long-term exposure can promote substrate creep and rumpling of the interface between substrate and bond coat. The measured β depletion and γ/γ′ enrichment near the interface plausibly modify local creep response, enabling morphology changes that concentrate tensile stresses in the TGO and top coat during thermal excursions, cooperating with oxidation to accumulate interfacial damage over long durations.

This depletion leaves behind a zone enriched with Ni and Cr, which fosters the formation of secondary mixed oxides (Figure 5). EDS analysis revealed that these mixed oxides on the TGO layer contain Ni, Cr, and Co. These mixed oxides are referred to as (Ni, Co)(Cr, Al)₂O₄ spinel-type oxides [37,38,39].

Observed mixed oxides can generate substantial growth stresses in the ceramic top coat, leading to an increased risk of cracking and accelerating the failure of the TBC layer. As a result, the system becomes increasingly susceptible to the formation of brittle oxides and the propagation of defects, further exacerbating structural and mechanical degradation (Figure 6).

Ultimately, the combination of accelerated TGO growth, Al depletion, and the formation of less protective oxides compromises the integrity of the TBC system. The increased TGO thickness also raises the thermal conductivity of the coating system, reducing its thermal insulation effectiveness and exposing the underlying alloy to higher temperatures. These factors culminate in spallation of the ceramic top coat, marking the failure of the TBC system under prolonged high-temperature exposure. These findings emphasize the critical need for optimizing bond-coat compositions to improve Al retention and oxidation resistance, as well as the development of advanced coating-application techniques to mitigate stress concentrations and delay failure [40]. Enhanced materials and process control can significantly extend the service life of TBC systems in high-temperature environments.

4. Conclusions

This study investigates the effect of isothermal heat treatment on the microstructural evolution of TBCs on Ni-based superalloy GTD111. SEM/EDS analysis reveals the following:

• The top coat, bond coat, and thermally grown oxide (TGO) are comprised of (Zr, Y)O, Ni-Cr-Co-Al, and Al₂O₃, respectively. TGO thickness increases from 0.7 µm at 850 °C/50 h to 9.9 µm at 1000–5000 h, showing accelerated growth at higher temperature and longer exposure.
• From 850 °C to 250 h onward, the Al in the β-NiAl phase of the bond coat depletes, leading to the thickening of the Al₂O₃ layer and the formation of a less protective γ-phase, which weakens the overall bond coat.
• In all the specimens exposed from 850 °C for 500 h to 1000 °C for 5000 h, spinel mixed metal oxides, such as Ni-Cr-Co-O, formed in the TGO layer due to the diffusion of elements like Ni, Cr, and Co, deteriorating the bond coat’s condition, as evidenced by the increased cracks and defects in the TBCs due to the high stress at the interface between the TGO layer and the top coat. While no gross spallation occurs during isothermal exposure up to 5000 h, the observed TGO thickening and interfacial microcracking indicate heightened spallation risk under cyclic thermal loading.
• The TGO layer also thickened steadily with higher temperatures and longer exposure, reaching about ~9.9 μm. This thickening, along with the cracks, increases the risk of TBC system failure. Increased time and temperature lead to a higher growth rate in the TGO layer. Quantitatively, ${\mathrm{h}}^2$ increases linearly with time (quadratic law), giving ${\mathrm{K}}_{\mathrm{p},850\mathrm{℃}}=8.87\times {10}^{-19}{ \mathrm{m}}^2/\mathrm{s}$ and ${\mathrm{K}}_{\mathrm{p},1000\mathrm{℃}}=5.10\times {10}^{-18}{ \mathrm{m}}^2/\mathrm{s}$. The latter (or the ${\mathrm{K}}_{\mathrm{p}}$ at 1000 °C) is ~5.8 times higher at 1000 °C. The progressive TGO thickening therefore raises growth and thermal mismatch stresses at the TGO/YSZ interface, accelerating crack initiation.

These overall findings in this study underscore the importance of understanding oxidation retardation and microstructure stabilization to enhance the durability of TBC-coated components under high-temperature exposure.