Effect of Spraying Distance on the Scratch Wear Behavior of 8YSZ and Gd-Yb-Y Co-Doped ZrO₂ TBCs

Abstract

This study investigates how torch standoff distance influences the microstructure, surface topography, and progressive-load scratch response of air plasma-sprayed 8YSZ and rare-earth co-doped GdYbYSZ thermal barrier coatings on an St-52 grade carbon steel substrate. Three nozzle-to-substrate spraying distances were examined: 80, 100, and 120 mm. X-ray diffraction revealed that the 8YSZ coatings possessed a predominantly tetragonal (t′) structure, with minor monoclinic fractions detected in the coatings obtained with the 80 mm and 100 mm distance parameters. The GdYbYSZ coatings, in contrast, exhibited a single-phase cubic defect-fluorite structure; their diffraction peaks appeared at lower 2θ angles relative to undoped cubic ZrO₂, consistent with lattice expansion caused by the substitution of Zr⁴⁺ by the larger Gd³⁺ and Yb³⁺ cations. Surface topography was quantified by non-contact laser profilometry, providing areal (Sa) and profile (Ra) roughness parameters for the as-sprayed condition as well as three-dimensional scratch-damage morphology after testing. Progressive-load scratch tests were performed using a Rockwell diamond indenter over a 2 mm track with the normal load ramped from 0.03 N to 30 N. Penetration depth, residual depth, tangential force, and acoustic emission were recorded continuously to identify critical damage transitions. Across all spraying distances, 8YSZ exhibited systematically shallower scratch grooves than GdYbYSZ; end-of-track maximum groove depths remained below 37 µm for 8YSZ, whereas GdYbYSZ reached up to 72 µm under identical loading conditions. The novelty of this study lies in combining torch standoff distance as a processing variable with multi-channel progressive-load scratch diagnostics, including in situ acoustic emission, depth profiling, and friction monitoring, to comparatively assess the scratch wear performance of 8YSZ and rare-earth co-doped zirconia TBCs for the first time.

1. Introduction

Wear has been identified as a dominant degradation phenomenon that impacts the useful life of engineering components exposed to mechanical interactions, thermal exposure, and chemically corrosive environments [1,2]. Within the aerospace industry, for instance, components of aero-turbines must withstand thermal, mechanical, and oxidative stresses. Therefore, the protection of metallic substrates from surface degradation has been identified as a critical design requirement. Among the various surface engineering techniques developed to combat this phenomenon, thermal barrier coatings have been identified as an essential solution for the thermal insulation of metallic substrates from hot gas flow, while also providing oxidation and wear resistance to the substrates [3,4,5]. Ceramics have been identified to be the dominant topcoat materials used for this purpose, based on their hardness, chemical inertness, and high-temperature oxidation resistance, making them suitable for protecting structural components from degradation in hostile service conditions [6,7]. Generally, applying ceramic coatings on metallic substrates has become a widespread solution [8]. The atmospheric plasma spray (APS) process is a predominant technique for obtaining high-quality coatings, owing to its elevated temperatures and deposition rates, which are used in the aerospace industry, as they provide excellent wear resistance to substrates possessing favorable adhesive bonding [9,10]. Beyond thermal insulation, however, thermal barrier coatings must withstand localized mechanical interactions arising from foreign object damage, erosion, and fretting, leading to chipping and delamination of the brittle topcoat [11,12]. Therefore, it has become essential to understand the scratch damage behavior of thermal barrier coatings for establishing reliable process-structure-performance relationships for these materials.

Among the plasma-sprayed ceramics family, yttria-stabilized zirconia (YSZ) has been the most widely used material as a thermal barrier coating as well as a wear-resistant coating [13,14]. Sliding wear investigations on YSZ coatings have revealed that the tribological properties of the material strongly depend on the counterpart material used during the sliding process. Brittle fracture and abrasive wear occur when the counterpart material is hard ceramic, whereas adhesive wear occurs when the counterpart material is metal [15]. In high-temperature applications, the wear resistance of YSZ can be improved to a considerable extent by the formation of metastable t’ phase during the thermal spraying process, which increases the hardness of the material and leads to a ductile–brittle transition during the wear process [16]. The microstructure of YSZ coatings has been reported to play a very important role in improving the wear resistance of the material. An increase in plasma power together with spraying distance results in a more compact coating exhibiting superior hardness and improved resistance to wear [17]. The morphology of the feedstock material used during the thermal spraying process has been reported to influence the coating’s ability to deform either in a ductile or brittle manner during the tribological process [18]. Rare-earth co-doped zirconia materials and gadolinium zirconate (Gd₂Zr₂O₇) have been reported to emerge as potential thermal barrier coating materials of the future. Scratch adhesion measurements performed on Gd₂Zr₂O₇ films deposited using the electron beam physical vapor deposition method revealed that the adhesion strength, hardness, and elastic recovery of the material depend on the deposition temperature [19]. Comparative studies on YSZ and lanthanum zirconate films revealed that YSZ has the highest wear resistance among the thermal barrier coating materials despite poor thermal conductivity [20].

Rare-earth co-doped zirconia compositions, including GdYbYSZ, are gaining considerable interest as next-generation TBC topcoat materials due to their excellent high-temperature phase stability and lower thermal conductivity compared with conventional 8YSZ [21,22]. Adding a high amount of rare-earth cations such as Gd³⁺ and Yb³⁺ into the zirconia lattice progressively stabilizes the cubic fluorite phase, which, unlike the metastable tetragonal t′ phase, is non-ferroelastic and therefore incapable of dissipating mechanical energy through domain switching under concentrated contact loading [23,24]. As a consequence, cubic-stabilized co-doped zirconia coatings exhibit intrinsically lower fracture toughness compared to t′-8YSZ, rendering them more susceptible to crack initiation, lateral crack coalescence, and brittle material removal under tribological loading, a mechanical liability that has not yet been systematically characterized through scratch-based diagnostics [25,26].

As reviewed above, the tribological performance of plasma-sprayed zirconia-based coatings has been investigated predominantly through sliding wear and erosion tests using pin-on-disk, ball-on-disk, or ball-on-flat configurations [15,16,17,18,19,20]. These geometries are effective for quantifying steady-state friction and volumetric material loss; however, they cannot resolve the progressive onset of discrete damage events such as crack initiation, chipping, and interfacial delamination that control the structural integrity of brittle ceramic coatings under concentrated contact. Progressive-load scratch testing addresses this limitation by ramping the normal load along a linear track while simultaneously recording penetration depth, residual depth, tangential force, and acoustic emission (AE) as real-time indicators of failure transitions. Despite this capability, scratch-based wear assessment remains scarce in the TBC literature. No comparative scratch study exists between conventional 8YSZ and emerging rare-earth co-doped systems such as GdYbYSZ, and the influence of torch-to-substrate standoff distance on scratch-induced damage evolution has not been explored. The present investigation addresses these gaps by providing the first systematic comparison of the progressive-load scratch wear behavior of APS 8YSZ and GdYbYSZ coatings deposited at three nozzle-to-substrate distances (80, 100, and 120 mm). The distinctive aspect of this research is the integration of four complementary diagnostic channels within a single scratch framework: AE monitoring, penetration and residual depth profiling, friction force recording, and post-test three-dimensional laser profilometry. This multi-signal approach enables a direct mechanistic correlation between in situ damage signatures and the phase constitution (tetragonal t′ versus cubic) and microstructural features imposed by spraying distance, providing process–structure–performance guidelines for advanced TBC topcoats.

2. Materials and Methods

2.1. Types and Properties of Substrate and Coatings

A carbon steel substrate (St 52 grade), supplied by Osmanlı Alaşımlı Çelikler San. ve Tic. Ltd. Şti. (İstanbul, Türkiye) was used in this study. The dimensions of the disk-shaped substrate are 25 mm and 4 mm in diameter and thickness, respectively. The raw powders of coatings in this experiment were all produced by Oerlikon Metco. The raw powders of the bond coat are Amdry 962-NiCrAlY (nominal particle size distribution of −106 + 53 μm), as shown in Figure 1a. The 8YSZ and GdYbYSZ powders (5.2 wt% Gd₂O₃ + 5.6 wt% Yb₂O₃ + 9.5 wt% Y₂O₃ + ZrO₂) were Metco 204NS (nominal particle size distribution: −125 + 11 μm) and Metco 206A (nominal particle size distribution: −106 + 22 μm), respectively, as shown in Figure 1b,c.

2.2. Fabrication of 8YSZ and GdYbYSZ TBCs

Before spraying, the substrate was sand-blasted using Al₂O₃ powder. First, the bond coat was deposited on the substrate by APS (Figure 2). Subsequently, the 8YSZ and GdYbYSZ ceramic topcoats were deposited by a robotically controlled APS system using a Metco F4-MB plasma torch. The spraying parameters of the bond coat and both topcoats (8YSZ and GdYbYSZ) are listed in

Table 1. Spray parameters for deposition of the bond coat and 8YSZ/GdYbYSZ top coats.

Parameters	Bond Coat	8YSZ Top Coat	GdYbYSZ Top Coat
Current (A)	575	630	650
Voltage (V)	73.4	62.2	60.9
Primary gas Ar (nlpm) *	45	45	35
Second gas H2 (nlpm)	12	6	10
Carrier gas (nlpm)	1.8	2.6	2
Gun traverse speed (mm·s−1)	300	250	250
Spray distance (mm)	130	80, 100, 120	80, 100, 120

* nlpm: normal liters per minute.

.

For standoff distances of 80 mm, 100 mm, and 120 mm, the selection was based on the practical operating window for the deposition of the topcoat of the ceramics as recommended by the manufacturer of the torch. The standoff distance of 100 mm represents the nominal condition that is most used in the practice of APS, while the 80 mm and 120 mm were selected to cover the effect of reduced or extended flight times on the melting state of the particles, the splat formation, and the microstructure of the coatings.

To ensure that observed differences in scratch response reflect microstructural and phase-related effects rather than thickness variations, the number of torch passes was kept constant for all standoff conditions within each coating type. Cross-sectional SEM micrographs confirmed that the ceramic topcoat thickness remained comparable across all spraying distances: approximately 54–59 µm for the 8YSZ coatings and 62–68 µm for the GdYbYSZ coatings, with an overall range of 54–68 µm. This level of thickness uniformity ensures that the standoff distance effects reported in this study are governed by differences in microstructure, porosity, and intersplat cohesion rather than by coating thickness.

Throughout this research, coated specimens are referred to by a naming convention that includes the type of topcoat material and the corresponding standoff distance. Thus, the 8YSZ-80, 8YSZ-100, and 8YSZ-120 specimens refer to the 8YSZ topcoats, whereas the GdYbYSZ-80, GdYbYSZ-100, and GdYbYSZ-120 specimens refer to the Gd–Yb–Y co-doped zirconia topcoats. However, it is important to note that all the specimens, regardless of the type of topcoat, are coated with the same NiCrAlY bond coat applied at a standoff distance of 130 mm.

2.3. Characterization of Samples with SEM and XRD

For the cross-section analysis, the specimens containing the coating were cut perpendicular to the coating surface using a low-speed diamond saw with continuous coolant. This was done to prevent both mechanical and thermal damage to the coating. The sections were then cold-mounted in epoxy resin to prevent any change in the microstructure due to thermal effects. The mounted sections were then progressively ground using SiC papers of grit sizes from 320 to 2400. This was followed by polishing the sections with a 3 µm and a 1 µm suspension of diamond to produce a polished finish. The specimens were ultrasonically cleaned in ethanol for 5 min between each stage.

Surface and cross-sectional morphologies of the coatings were examined using scanning electron microscopy (SEM, Jeol JSM-6060) (JEOL Ltd., Tokyo, Japan)). Phase composition analysis was carried out by X-ray diffraction (Rigaku Ultima+, Cu Kα radiation, λ = 1.5406 Å) using a θ–θ goniometer in conventional coupled 2θ/θ configuration with a standard sample holder. Data were collected in continuous scanning mode over the 2θ range of 10–80°, with a step size of 0.01° and a scan speed of 0.6 s/step. The mole fractions of m-ZrO₂, t′-ZrO₂ and c-ZrO₂ were determined from XRD using Miller’s equation [27]:

$$\frac{M_m}{M_{c,t^{\prime }}}=0.82\frac{I_m(11\stackrel{-}{1})+I_m(111)}{I_{c,t^{\prime }}\left(111\right)}$$

(1)

where M_m, M_t′, and M_c are the molar fractions of monoclinic, tetragonal prime, and cubic phases, respectively. I_m, I_t′, and I_c are the intensities of XRD peaks corresponding to monoclinic, tetragonal prime, and cubic phases.

2.4. Surface Roughness Characterization

A non-contact laser profilometer (Nanovea PS-50) was employed to characterize the surface topography of the 8YSZ and GdYbYSZ coatings after deposition and to quantify the three-dimensional damage morphology after the progressive-load scratch tests. For the characterization of surface roughness resulting from the as-sprayed process, a 5 × 5 mm scan area with a lateral step size of 5 µm in both the x and y directions and a laser spot size of 3.6 µm was utilized to adequately sample the surface features associated with APS ceramic coatings at the splat level. For post-scratch damage characterization, a 3 × 1 mm scan area was utilized with the center aligned with a 2 mm scratch track using the same lateral resolution as above to enable full-width coverage of the groove, edge chipping regions, and the undamaged surface from which scratch depth was initially determined. This optical technique avoids probe-to-surface interaction and enables direct reconstruction of the true surface height field, allowing extraction of roughness parameters from the coated surfaces as well as accurate geometric evaluation of the scratch scars. Raw height data were processed using DigitalSurf Mountains software (Digital Surf, Besancon, France) (application only), and the areal arithmetic mean height (Sa) was determined from three-dimensional maps in accordance with ISO 25178 [28], while the line-based arithmetic mean roughness (Ra) was obtained from representative profile analyses. Accordingly, post deposition surface roughness data (Sa and Ra) were acquired for each coating condition to document the as-sprayed surface state prior to scratching. After scratch testing, three-dimensional surface maps of the scratch tracks were recorded to visualize groove formation, edge chipping, and local spallation, and cross-sectional profiles extracted normal to the scratch direction were used to determine the scratch depth from the height difference between the intact reference surface and the groove bottom. For each coating condition, Sa and Ra values were obtained from three independent measurement areas on the as-sprayed surface, and the reported values represent the arithmetic mean of these three measurements. Post-scratch groove depth profiles were likewise extracted at three positions along the scratch track (near the beginning, middle, and end), and the reported depths correspond to the arithmetic mean of these readings.

2.5. Scratch Test Procedure

Progressive-load scratch tests were carried out on the investigated coatings using an instrumented scratch tester, Micro Combi Tester MCT3 (Anton Paar GmbH, Austria). A Rockwell-type diamond indenter with a 100 µm tip radius was employed, and linear scratches with a 2 mm track length were performed at a constant scratching speed of 1 mm/min while the normal load was ramped from 0.03 N to 30 N, corresponding to a loading rate of 14.98 N/min. The indenter approach speed was set to 2%/s of the full Z-range, ensuring a controlled contact detection prior to load application; the normal force application rate was set to 0.1 N/s, and the force removal rate was set to 10 N/s. Pre-scan and post-scan passes were conducted at a constant scanning load of 0.03 N, with a return speed of 1.2 mm/min, to quantify the residual groove profile. During scratching, the normal force, tangential force, penetration depth, and acoustic emission were recorded continuously, and acoustic emission monitoring was performed with a sensitivity setting of 5 to enable identification of damage onset and critical load levels associated with coating cracking, chipping, and interfacial failure. For each coating condition, three replicate scratch tests were performed on the same specimen surface at parallel, non-overlapping locations, and the results presented correspond to representative traces selected from these replicates. The repeatability of the critical load and depth responses was confirmed by a coefficient of variation below 5% across the three replicates.

3. Results

3.1. Phase Analysis of the As-Sprayed Coatings

Figure 3 illustrates the XRD patterns of the 8YSZ and GdYbYSZ coatings. The 8YSZ coating is predominantly composed of tetragonal zirconia, as confirmed by the characteristic splitting of diffraction peaks associated with the tetragonal crystal structure, namely the (004) and (400) reflections in the 2θ range of 72–76°, which was observed in all 8YSZ coatings. In the (111) region (2θ = 27–33°), the monoclinic (m) phase was found in the 8YSZ-80 and 8YSZ-100 coatings. The monoclinic phase contents in these coatings were calculated to be 1.88 wt% and 0.98 wt%, respectively. In contrast, no monoclinic phase was detected in the 8YSZ-120 coating. On the other hand, the GdYbYSZ coatings exhibit a cubic-phase structure, and no monoclinic phase was observed. Furthermore, the absence of distinct peaks for Y₂O₃, Yb₂O₃, and Gd₂O₃ confirms the complete dissolution of these rare-earth dopants into the zirconium oxide lattice.

Quantitative analysis of the intensity of the primary (111) peak from the raw data also confirmed this trend. The GdYbYSZ (111) peak position shifted to 2θ = 29.68° and 29.74° for the 80 mm and 100 mm conditions, respectively, compared to the undoped cubic ZrO₂ reference peak position at approximately 30.10°. This equates to peak shifts of 0.43° and 0.36° to lower 2θ positions, respectively. Applying Bragg’s law with Cu Kα radiation and a wavelength of 1.5406 Å, the calculated cubic lattice parameter values were a = 5.210 Å for the GdYbYSZ-80 and a = 5.198 Å for the GdYbYSZ-100, reflecting a 1.2%–1.4% expansion relative to the undoped cubic reference with a = 5.138 Å. This expansion arises due to the substitution of the Zr⁴⁺ ion (0.84 Å ionic radius in an eight-coordinate arrangement) with the larger Gd³⁺ (1.053 Å) and Yb³⁺ (0.985 Å) cations, which systematically increase the interplanar spacing throughout the cubic fluorite-type lattice [21,29]. However, the GdYbYSZ-120 condition showed a (111) peak position at 2θ = 30.13°, very close to the undoped cubic reference peak position, suggesting the longer dwell time at 120 mm might reduce the incorporation of the rare earth dopants into the rapidly solidified material.

3.2. Microstructure Analysis of the As-Sprayed Coatings

As shown in Figure 4, 8YSZ (Figure 4a) and GdYbYSZ (Figure 4b) coatings exhibit a typical plasma spray surface morphology characterized by fully melted powders, pores, and cracks. The coating surface consists of two distinct regions: a rough region associated with semi-melted powder particles/agglomerates and a relatively smooth region formed by fully molten particles. In the smooth region, a discontinuous and irregular network of microcracks is observed, which is attributed to the large tensile stresses generated by the fast cooling experienced during plasma spraying.

The Ra values of the 8YSZ-80 coating and the GdYbYSZ-80 coating were found to be 5.73 μm and 7.52 μm, respectively (

Table 2. Roughness analysis of 8YSZ/GdYbYSZ top coat surfaces (mean ± SD, n = 3).

Parameters	8YSZ-80	8YSZ-100	8YSZ-120	GdYbYSZ-80	GdYbYSZ-100	GdYbYSZ-120
Ra (μm)	5.73 ± 0.19	5.85 ± 0.19	5.54 ± 0.14	7.52 ± 0.26	7.90 ± 0.45	7.97 ± 0.26
Rq (μm)	7.20 ± 0.28	7.38 ± 0.19	6.90 ± 0.38	9.28 ± 0.43	9.87 ± 0.37	9.96 ± 0.23
Sa (μm)	11.8 ± 0.8	9.84 ± 0.61	11.3 ± 1.1	11.9 ± 1.5	11.9 ± 0.6	16.9 ± 1.3
Sq (μm)	15 ± 1	12.5 ± 0.5	14.2 ± 1.2	14.9 ± 0.9	14.9 ± 0.7	21.3 ± 1.3

). The increase in the roughness of the coatings is due to the different powder shapes and the varying thermal conductivity of the powders. The 8YSZ powder used in the current study is prepared using the hollow spherical (HOSP) method. However, the GdYbYSZ powder is prepared using the agglomerated-and-sintered (A&S) method. GdYbYSZ powder is known to have lower thermal conductivity. Gd₂O₃ and Yb₂O₃ are incorporated into the YSZ powder. GdYbYSZ has lower thermal conductivity compared to the 8YSZ powder [21]. The 8YSZ HOSP feedstock was composed of hollow spherical particles with thin, dense outer shells. The melting efficiency of the 8YSZ powder is better compared to the GdYbYSZ powder. The melting efficiency of the 8YSZ powder is responsible for the reduction in the roughness of the coating. A reduction in the roughness of the coating is noticed compared to the GdYbYSZ coating.

The Ra values measured for the 8YSZ coatings (5.54–5.85 µm) fall within the range of 4–8 µm commonly reported for APS YSZ top coats in the literature [17,18], confirming that the present spraying conditions produce surface finishes representative of industrial practice. The consistently higher Ra values of GdYbYSZ (7.52–7.97 µm) agree with Guo et al. [21], who attributed the rougher surface of Gd₂O₃–Yb₂O₃ co-doped zirconia to the lower melting efficiency of agglomerated-and-sintered feedstock powders compared with the hollow spherical (HOSP) morphology typically used for conventional YSZ.

In Figure 5a, the 8YSZ top coat produced at an 80 mm torch standoff exhibits a heterogeneous three-dimensional height field with scattered asperity clusters and pore-related depressions, reflecting splat-to-splat height discontinuities typical of APS ceramic surfaces, and this morphology is consistent with its measured roughness level (Ra 5.73 μm and Sa 11.8 μm). In Figure 5b, increasing the standoff to 100 mm yields a comparatively more leveled 8YSZ surface with a reduced density of extreme peaks and a more uniform relief, matching the lowest Sa among the 8YSZ conditions (Sa 9.84 μm and Sq 12.5 μm) and indicating more effective particle melting and splat flattening at this distance. In Figure 5c, the 8YSZ 120 mm torch standoff condition shows renewed development of localized mound-like features and sharper asperities attributable to partially melted agglomerates, which is consistent with the increase in areal roughness relative to 100 mm (Sa 11.3 μm) and provides additional stress concentrators that can promote brittle microcracking during concentrated contact loading. In Figure 5d, the GdYbYSZ 80 mm torch standoff surface appears more irregular than the corresponding 8YSZ condition, with more pronounced asperities consistent with less complete melting of the A&S type feedstock and its higher roughness (Ra 7.52 μm, Sa 11.9 μm) compared with 8YSZ at the same standoff. As depicted in Figure 5e, the GdYbYSZ 100 mm torch standoff condition maintains a similar level of areal roughness but with a slightly more even distribution of micropeaks, suggesting that increasing the standoff distance improves spreading but does not completely remove remnants of unmelted particles. As depicted in Figure 5f, the GdYbYSZ 120 mm torch standoff condition possesses the highest level of peak-to-valley relief, where valleys are deep, and asperities are large. This condition also possesses the highest values of roughness metrics measured across the surface (Ra 7.97 μm, Sa 16.9 μm, Sq 21.3 μm) and provides physical validation of the increased likelihood of localized brittle damage initiation under scratch and sliding conditions.

3.3. Comparative Evaluation of Scratch Test and Wear Performance

In Figure 6a, where the sample was an 8YSZ-80 with a spray distance of 80 mm, it can be seen that the scratch track shows microcracking in the initial low-load region, progressing to edge chipping along the groove shoulders in the mid-section, and culminating in delamination near the track end, which are signs of an unstable material removal process during progressive loading. This observation is consistent with the acoustic emission, where it was shown that there were frequent high-amplitude bursts throughout the scratch, with the most significant events occurring during the mid- to late-stage region of the scratch (0.9–1.6 mm). These bursts can be mechanistically associated with the brittle damage features identified on the scratch micrograph (Figure 6a), namely microcracking at the groove onset, lateral cracking and edge chipping in the mid-section, and delamination with localized spallation near the track terminus. The penetration depth curve shows a progressive increase in indentation depth with distance, as expected from the ramped normal load, but a distinct catastrophic excursion occurs around 1.5 mm, where the penetration depth abruptly deepens to about 55 µm, and the residual depth simultaneously increases in magnitude to about 40 µm. The concurrent acoustic emission (AE) maxima and the abrupt increases in both penetration and residual depth indicate a transition from predominantly microcracking and compaction to a major fracture event, most plausibly local delamination and coating breakthrough, which reduces the load-bearing capacity of the coating and permits deeper indenter penetration. Relative to the longer standoff conditions, the 80 mm coating therefore shows earlier and more severe brittle instabilities during scratching, which is consistent with a microstructure where high thermal input and rapid splat solidification can elevate residual stresses and promote crack coalescence under concentrated contact loading.

In Figure 6b (8YSZ-100, 100 mm standoff), the scratch morphology is comparatively uniform, with isolated microcracking near the groove onset, stable plowing through the mid-section, and only minor chipping at the groove shoulders toward the end, suggesting a more stable damage evolution. The acoustic emission signal remains close to the baseline for most of the scratch length, interrupted by a limited number of discrete bursts, including an early cluster near 0.2 mm to 0.3 mm and isolated peaks at later distances. These AE bursts are consistent with localized microfracture events, such as the opening of pre-existing pores or weak interfaces and short lateral crack segments that do not develop into sustained spallation. The penetration depth increases gradually toward the end of the track, reaching about 50 µm, while the residual depth remains consistently less negative, typically in the range of about 5 µm to 20 µm, indicating appreciable elastic recovery and the absence of a single dominant delamination event. Importantly, the depth curves do not show the abrupt step change observed for 8YSZ-80, supporting the interpretation that the 100 mm standoff produced a coating with better tolerance to progressive contact loading, expressed as reduced AE activity, smoother groove development, and a more gradual transition from compaction and microcracking to limited fracture.

In Figure 6c (8YSZ-120, 120 mm standoff), the scratch track shows microcracking in the early portion, followed by progressive chipping in the mid-to-late section, and severe spallation accompanied by debris accumulation near the track end, which is mirrored by the acoustic emission trace. In this condition, AE activity is relatively modest through the mid-section of the scratch, but a strong cluster of high-amplitude bursts occurs in the late stage (approximately 1.7 mm to 2.0 mm), indicating rapid accumulation of fracture events once the normal load reaches the upper range. This late-stage AE cluster is characteristic of lateral cracking and chipping evolving into partial spallation, as clearly identified on the scratch micrograph (Figure 6c), where crack networks formed within a more compliant or porous lamellar structure begin to coalesce and eject fragments from the groove shoulders. The penetration depth increases progressively and reaches about 58 µm near the end of the track, while the residual depth remains relatively shallow through most of the scratch but drops abruptly in the final region, from roughly 30 µm to 35 µm. The divergence between a steadily deepening penetration depth and a suddenly increasing residual depth at the end indicates that the coating transitions from elastic plus recoverable deformation to irreversible material removal associated with fracture and debonding. When compared across standoff distance, the 120 mm coating delays major damage until higher loads, but once the critical damage state is reached, it exhibits a pronounced late-stage instability, consistent with a microstructure expected to contain higher porosity and weaker intersplat cohesion due to increased in-flight cooling at longer spray distance. Overall, the approximate critical load values given in

Table 3. Approximate critical loads (Lc) corresponding to major damage transitions identified from acoustic emission bursts and depth excursions during progressive-load scratch testing.

Coating	Lc1 (N)	Damage Mode at Lc1	Lc2 (N)	Damage Mode at Lc2
8YSZ-80	13.5	Onset of repetitive AE bursts (microcracking)	22.5	Catastrophic depth excursion (delamination)
8YSZ-100	3.0	Isolated AE cluster (pore collapse)	—	No catastrophic failure up to 30 N
8YSZ-120	25.5	Late-stage AE cluster (lateral cracking)	28.5	Chipping and partial spallation
GdYbYSZ-80	12.0	Intermittent AE bursts (microcracking)	22.5	Major depth excursion (local debonding)
GdYbYSZ-100	9.0	Distributed AE spikes (intersplat cracking)	21.0	Penetration collapse (delamination)
GdYbYSZ-120	4.5	Early high-amplitude AE (brittle fracture)	22.5	Severe groove deepening (spallation)

provide a quantitative description of the transitions in the damage processes as indicated by the AE and depth traces. Here, Lc₁ denotes the onset of initial damage, corresponding to the initiation of microcracking, intersplat cracking, and/or brittle fracture, whereas Lc₂ represents the catastrophic damage threshold, associated with debonding, delamination, and spallation. For 8YSZ, Lc₁ increased from 13.5 N at 80 mm to 25.5 N at 120 mm, while 8YSZ-100 sustained up to 30 N without the Lc₂ transition. These results confirm the superior resistance of 8YSZ-100 to progressive loading. For GdYbYSZ, the values of Lc₁ were consistently lower at all standoff distances. The values decreased to 4.5 N at a scratch distance of 120 mm. The Lc₂ values for GdYbYSZ remained in the range of 21–23 N for all standoff distances. These data show that once microcracking is initiated in the GdYbYSZ system, the transition to delamination or spallation is in the same range of loading. These critical load values demonstrate that the increase in standoff from 80 to 100 mm has the effect of stabilizing the scratch response by increasing the critical load for crack initiation as well as the critical load for catastrophic failure. The increase in standoff to 120 mm from 100 mm significantly reduces the Lc₁ values for all materials, consistent with the loss of intersplat cohesion for the longer standoff distance.

For Figure 7a, GdYbYSZ-80, 80 mm torch standoff, it can be noted that the scratch track exhibits cracking in the initial region and chipping along the groove edges in the mid-section, while a zone of local debonding is evident near the track end where the groove morphology becomes markedly disrupted. This indicates that damage is accumulated progressively due to microfracture, as opposed to material removal. It can also be noted that the acoustic emission signal is primarily composed of high-intensity bursts superimposed on a low-level background, which is a typical characteristic of intermittent brittle fracture in a lamellar ceramic coating. The most intense burst, occurring in the mid- to late-region of the scratch track near 1.5 mm, is where the greatest disturbance in the depth profiles is noted, where both penetration and residual depths show a sharp deepening. This coupled response supports a transition from stable compaction and intersplat microcracking to a localized fracture episode, most plausibly lateral cracking followed by chipping and partial local debonding within the coating, consistent with the damage features labeled on the micrograph (Figure 7a), which momentarily reduces the load-bearing integrity and allows deeper indenter sink-in. Away from this region, the residual depth remains consistently less negative than the penetration depth, indicating appreciable elastic recovery and suggesting that most of the track deformation is recoverable until the critical damage event.

In Figure 7b (GdYbYSZ-100, 100 mm torch standoff), the scratch track shows a plowing groove with chipping at locally disturbed shoulders in the mid-section, and a delamination zone in the late-stage region, yet the overall morphology remains comparatively stable over a large fraction of the track length. The acoustic emission response exhibits frequent low-to-moderate spikes rather than repeated large bursts, indicating distributed microdamage, such as short intersplat crack initiation, pore collapse, and limited chipping, without continuous spallation. The depth curves initially evolve gradually with increasing distance, consistent with the progressive loading regime, then a pronounced penetration-depth drop occurs in the later stage of the scratch (around 1.4 mm to 1.7 mm), where the penetration depth reaches its maximum magnitude while the residual depth simultaneously becomes markedly more negative. This signature points to a localized fracture or delamination-assisted collapse event, in which a sudden loss of cohesive support beneath the indenter produces a transient deep penetration and an increased permanent groove. Following this event, partial recovery of the residual depth together with continued oscillations indicates that damage remains localized and does not propagate as a long, unstable spall region along the full track.

In Figure 7c (GdYbYSZ-120, 120 mm torch standoff), the scratch track shows the most conspicuous surface disruption, with brittle fracture initiating early in the track, chipping intensifying through the mid-section, and large-scale spallation accompanied by debris formation toward the track end, consistent with the most severe damage evolution among all conditions. The acoustic emission trace contains a high initial burst near the scratch onset, followed by numerous spikes distributed along the track, which is indicative of early-stage microfracture and repeated crack events under the progressively increasing normal load. The depth profiles show stronger oscillations than the other standoff conditions, with several abrupt excursions and a pronounced deepening in the late stage where the penetration depth approaches its maximum magnitude, accompanied by a substantial increase in residual depth. This coupling suggests that damage evolves from early compaction plus microcracking into more severe mechanisms at higher loads, including lateral cracking, edge chipping, and spallation with interfacial debonding, as identified on the scratch micrograph (Figure 7c), which generates the clustered AE activity and drives the irreversible groove formation. Considering that the 80, 100, and 120 labels represent torch-to-substrate distance in millimeters, the 120 mm condition exhibits the most damage-prone scratch response overall, consistent with reduced cohesive strength and higher defect sensitivity expected when particle cooling in-flight increases at longer standoff, leading to earlier and more frequent brittle events and a larger permanent deformation component under scratch loading.

Figure 8 shows that the coefficient of friction (COF) evolves in a distinctly non-linear manner along the 2 mm progressive-load scratch path, reflecting sequential transitions from run-in to stable plowing, and further to damage-induced friction instabilities for both 8YSZ (Figure 8a) and GdYbYSZ coatings (Figure 8b) deposited at 80, 100, and 120 mm torch standoff. For 8YSZ, the 80 mm condition exhibits repeated abrupt COF collapses to near zero in the early and mid track, followed by intermittent peaks approaching about 0.5 in the late stage, which is indicative of unstable contact governed by episodic cracking and chipping events that intermittently reduce tangential resistance and then regenerate high friction through debris-assisted plowing. In contrast, 8YSZ at 100 mm displays a comparatively more coherent friction response with a prominent critical spike reaching about 0.6 near 1.0 mm, followed by recovery to a moderate COF level and a gradual increase toward the end of the scratch, suggesting a localized transition to a more severe damage mode, such as intensified microcracking with debris compaction, superimposed on the expected load driven friction rise. The 8YSZ 120 mm coating shows a pronounced initial peak of about 0.6 at the scratch onset and frequent early fluctuations, then a more stable mid-track regime around 0.25 to 0.35, implying that initial asperity fracture and early debris formation dominate the run-in, whereas the later response is governed by steadier plowing under increasing load. For GdYbYSZ, the 80 mm coating maintains a generally higher and more sustained COF level, commonly around 0.35 to 0.5, with a distinct near-zero drop around 0.45 mm and subsequent recovery, consistent with a single major instability followed by reestablishment of a load-bearing groove. The 100 mm GdYbYSZ curve contains the most clearly defined late-stage critical event, namely a sharp rise to about 0.6 around 1.55 mm, followed by a sudden collapse near 1.8 mm, which is characteristic of a fracture or local spallation episode that temporarily reduces tangential force after a debris and plowing intensified friction peak. The 120 mm GdYbYSZ condition exhibits step-like friction evolution with an early peak near 0.55, a mid-track drop toward about 0.1, and repeated recoveries toward 0.4 to 0.5, indicating frequent transitions between debris-controlled sliding and renewed plowing. Because the scratch tester records normal and tangential forces continuously during the progressive loading, these COF critical points can be directly correlated with the damage events already evidenced by acoustic emission bursts and depth excursions, enabling a consistent mechanistic interpretation across coating type and spraying distance.

The COF values recorded during progressive-load scratching in this study (generally 0.25–0.60) are notably higher than the steady-state COF of 0.15–0.40 typically reported for YSZ coatings under ball-on-disk sliding conditions [15,16]. This difference is expected because the scratch test geometry involves a sharp diamond indenter that penetrates into the coating under continuously increasing load, producing plowing-dominated friction rather than the surface-sliding regime of conventional tribometers. Importantly, the observation that GdYbYSZ maintained a generally higher and more sustained COF than 8YSZ across all standoff distances is consistent with the higher surface roughness and greater propensity for brittle debris formation in the cubic co-doped system. In a similar vein, Deng et al. [15] also found that the coefficient of friction (COF) for the YSZ sliding experiments showed a sensitivity to the counterface material and the dominant wear mechanism. Specifically, the contact with a brittle fracture mechanism showed a higher and less stable friction trace compared to the plastic deformation mechanism, a phenomenon similar to the AE-rich high COF seen for the GdYbYSZ coatings presented here.

Figure 9a plots the maximum groove depth profiles obtained using a profilometer at the initiation and end of scratching. All coatings exhibit increased end-position maximum groove depths. A distinction between the tougher group of 8YSZ and the damage-prone group of GdYbYSZ coatings can also be noted. The end-position maximum groove depths range from 19.7 to 22.9 µm (8YSZ-80), from 19.4 to 29.5 µm (8YSZ-100), and from 22.1 to 38.2 µm (8YSZ-120) for 8YSZ. In contrast, these depths range from 53.3 to 57.1 µm (GdYbYSZ-80), from 50.8 to 53.6 µm (GdYbYSZ-100), and from 35.6 to 74.5 µm (GdYbYSZ-120) for GdYbYSZ. In Figure 9b, the cross-section profile of 8YSZ-80 indicates a shallow and relatively uniform cross-section profile at the initial position. A slight increase in maximum groove depth at the end position can also be noted. This indicates that this coating exhibits stable plowing action with limited breakout under increased applied pressure. In Figure 9c, 8YSZ-100 indicates a smooth cross-section profile with a slight increase in maximum groove depth at the end position. As can be observed from Figure 9d, the groove resulting from the 8YSZ-120 specimen has the largest dimension compared to the other 8YSZ specimens, not only in terms of the damage caused by scratching, which is indicated by the general results, but also in terms of the groove width and depth at the terminal point. Figure 9e shows that the GdYbYSZ-80 specimen created the deepest and broadest groove at the initiation position, which persisted at the end position, consistent with the more brittle removal mechanism and lower strain accommodation, as indicated by the overall scratch damage results. Figure 9f shows that the GdYbYSZ-100 specimen retains high scratch depths while having a slightly more stable groove profile compared to the GdYbYSZ-80 specimen, consistent with the lower maximum scratch depths of the GdYbYSZ series. Figure 9g shows that the GdYbYSZ-120 specimen had a moderate groove profile at the initiation position, which dramatically increases at the end position, making it the largest of the series, consistent with the instability indicated by the overall scratch damage results, which suggest crack coalescence and spallation at the highest normal load of the scratch process.

The scratch-induced damage evolution observed in both 8YSZ and GdYbYSZ coatings is consistent with established deformation mechanisms of zirconia ceramics. The comparatively stable groove development and fewer catastrophic acoustic emission bursts in 8YSZ can be rationalized by ferroelastic domain switching in tetragonal (t′) zirconia, which dissipates mechanical energy and retards crack coalescence under concentrated contact, an energy-absorbing route not available to cubic-stabilized analogs such as GdYbYSZ [23]. Furthermore, rare-earth oxide additions that drive zirconia toward a more cubic, non-ferroelastic state have been shown to suppress transformation/ferroelastic toughening and diminish crack-growth resistance, which aligns with the larger residual groove depths and more frequent chipping recorded for GdYbYSZ in our progressive-load scratches [24]. Complementarily, composition- and processing-dependent studies indicate that maintaining tetragonal domains is essential for achieving a favorable hardness–toughness balance; in contrast, stronger cubic stabilization tends to trade toughness for hardness, thereby increasing susceptibility to brittle removal during scratching consistent with our AE-intensive responses in GdYbYSZ [30]. This interpretation is further supported by indentation-cracking analyses on cubic zirconia, which document pronounced crack initiation and lower apparent toughness relative to microstructures capable of ferroelastic accommodation, mirroring the abrupt depth excursions and localized spallation we observed for GdYbYSZ at higher normal loads [31]. Beyond phase constitution alone, the interplay between microstructural length scales and ferroelastic variant switching provides a direct bridge between the acoustic emission (AE) signatures and groove-growth trends recorded in this study: phase field modeling of t′-YSZ shows that grain size and interfacial energetics shift the critical load for variant switching, increasing elastic energy dissipation and stabilizing damage evolution, an effect that rationalizes the smoother depth/AE responses of our 8YSZ relative to GdYbYSZ under progressive scratching [32]. Complementarily, composition–property maps reveal that increasing yttria content (and the associated drift toward cubic fractions) suppresses transformability/ferroelastic accommodation and lowers fracture resistance compared with partially stabilized compositions, predisposing the material to brittle microcracking and larger permanent scratch grooves under concentrated loads consistent with the behavior we measured for GdYbYSZ [33]. Finally, studies on multivariate rare-earth co-doped zirconia akin to GdYbYSZ report single-phase c-ZrO₂ with excellent high-temperature phase stability but poor sintering resistance manifested by pronounced grain coarsening; such coarsening is consistent with reduced interlamellar cohesion in thermally sprayed lamellae and helps explain the late-stage, AE-intensive depth excursions and local spallation observed at higher normal loads in our cubic-like coatings [22]. The maximum end-of-track groove depths for GdYbYSZ (46–72 µm) are higher than those for 8YSZ (19–37 µm) by a factor of approximately 2–3. This factor corresponds to the fracture toughness difference. Jung et al. [25] reported that the fracture toughness (K_IC) values for YGYZ coatings ranged from 0.95 to 1.25 MPa·m^1/2 compared to those for 8YSZ (1.85–2.23 MPa·m^1/2), which showed a degradation of approximately 40%–50% in fracture. The roughly twofold increase in scratch groove depth observed in the present work for GdYbYSZ relative to 8YSZ is therefore proportionally consistent with this toughness deficit, supporting the interpretation that the ferroelastic toughening mechanism operative in t′-8YSZ is the dominant factor controlling scratch damage resistance. Furthermore, the finding that the 100 mm standoff condition yielded the most stable scratch response for both coating types is in qualitative agreement with Lamuta et al. [17], who reported that intermediate APS process parameters produced the densest microstructure with the highest hardness and lowest wear rate, reflecting an optimum balance between particle melting and intersplat cohesion.

The present findings are also consistent with the broader trends reported in the recent literature on rare-earth co-doped zirconia TBCs. Chen et al. [34] prepared La₂O₃, Gd₂O₃, and Yb₂O₃ co-doped YSZ (LGYYSZ) coatings by APS and confirmed that the co-doped system exhibited a pure cubic phase structure. They explicitly noted the absence of both phase-transformation toughening and ferroelastic toughening mechanisms in this cubic co-doped ceramic, and showed that a double-ceramic-layer architecture with a YSZ transition layer was required to achieve a thermal cycling lifetime 2.7 times that of single-layer YSZ at 1400 °C, compensating for the intrinsically lower mechanical damage tolerance. In more recent studies, Pi et al. [35] used classical nucleation theory and first-principles computations to investigate the impact of six different RE dopants (Yb, Y, Dy, Gd, Eu, and Sm) on the ferroelastic switching barrier of tetragonal ZrO₂. The results of their analysis show that the critical stress required to cause material switching follows the order σ_Sm > σ_Eu > σ_Gd > σ_Dy > σ_Y > σ_Yb. This demonstrates that larger RE³⁺ cation size causes increased tetragonality and increases the energy barrier to switching, reducing the material’s ability to toughen under ferroelastic stress. This theoretical model verifies the experimental observations of more frequent AE bursts and deeper scratch grooves in our cubic GdYbYSZ coatings. In another study on thermal shock resistance of APS Gd and Yb co-doped YSZ coatings, Zhao et al. [36] reported that vertical cracks occur because of tensile stresses in the top coating layer, while the growth of thermally grown oxide (TGO) causes stress concentrations at the top coat (TC)/bond coat (BC) interface, which leads to transverse interfacial cracks that worsen with increasing thermal shock temperature. The brittle crack propagation patterns they described are mechanistically analogous to the chipping, spallation, and interfacial failure signatures captured by our progressive-load scratch diagnostics, reinforcing the conclusion that the absence of ferroelastic energy dissipation in cubic co-doped zirconia is the primary factor limiting its mechanical damage tolerance relative to t′-8YSZ.

3.4. Influence of Phase and Microstructure on Scratch Performance

Zirconia-based coatings doped with rare earth elements are observed to have significantly lower fracture toughness as opposed to 8YSZ. GdYbYSZ coatings are reported to have a range of 0.95 to 1.25 MPa·m^1/2, while 8YSZ is reported to range from 1.85 to 2.23 MPa·m^1/2. Due to the reduced fracture toughness of GdYbYSZ coatings, the stress caused by the indenter induces a more brittle failure mode. As a result, crack initiation occurs quickly, which increases the probability of cohesive coating failure and scratch tracks due to easier material removal [25]. As opposed to 8YSZ coatings, the tetragonal phase of zirconia is observed to provide a distinct toughening mechanism through the domains of the ferroelastic phase. Consequently, the application of an external stress, such as the compressive stresses exerted by the indenter of the scratch test, causes the domains of the crystal lattice to be able to change orientation, thus absorbing energy. Therefore, crack propagation is delayed. GdYbYSZ coatings are observed to contain a high amount of stabilizers such as Gd₂O₃ and Yb₂O₃. These stabilizers result in the zirconia-based coatings being stabilized to the cubic phase, which is non-ferroelastic. As a result of the cubic phase being non-ferroelastic, the GdYbYSZ coatings are more brittle than the 8YSZ coatings. This leads to the high fracture toughness of 8YSZ coatings, which can withstand the coalescence of lateral cracks as well as radial cracks that result from the scratch test. Hence, the 8YSZ coatings can withstand chipping as well as chunking of the material. As opposed to 8YSZ coatings, the low fracture toughness of GdYbYSZ coatings results in the rapid coalescence of microcracks at the boundaries of the splats as well as intra-splat, thus leading to material removal at a low energy input [26,37,38].

The fundamental reason 8YSZ exhibits better scratch resistance than GdYbYSZ is the presence of the metastable tetragonal (t′) zirconia, which provides a ferroelastic toughening capability. Under the combined compressive–shear stress field beneath a moving scratch indenter, ferroelastic domains in t′-ZrO₂ can reorient (domain switching) and accommodate strain in a partially recoverable way. This mechanism dissipates a portion of the input mechanical energy and reduces the local driving force for rapid crack growth, thereby delaying the transition from distributed microcracking/compaction to large, unstable fracture events. In contrast, rare-earth co-doped zirconias such as GdYbYSZ typically stabilize zirconia toward a cubic or “cubic-like” non-ferroelastic structure, which lacks domain-switching-mediated energy absorption. Accordingly, the deformation response is more dominated by brittle microfracture, and the coating is more susceptible to crack initiation at lamellar interfaces and defects, followed by faster crack coalescence under progressive loading, consistent with lower fracture toughness commonly reported for co-doped systems.

This intrinsic phase-dependent difference is reflected in the progressive-load scratch outputs reported. Across the same standoff distance conditions, the maximum scratch depths extracted from profilometer profiles are systematically lower for 8YSZ than for GdYbYSZ (e.g., beginning/end depths of 14.6–14.64–14.5/19.2–22.5–36.7 µm for 8YSZ versus 35.5–38.5–32.6/46.0–47.7–71.9 µm for GdYbYSZ). In the depth and acoustic emission (AE) traces, the 8YSZ condition deposited at the intermediate standoff distance displays a comparatively smoother evolution of penetration and residual depth with fewer pronounced AE bursts, indicating a more stable progression of damage under increasing normal load. Conversely, GdYbYSZ coatings show a response characterized by more frequent burst-type AE activity and stronger depth excursions—features that are consistent with intermittent brittle fracture events in a lamellar ceramic where the absence of ferroelasticity limits the capacity for stress accommodation. Importantly, while torch standoff distance modulates the severity and the location along the scratch at which major instabilities appear—by influencing in-flight particle thermal/kinetic state and thus intersplat cohesion—the phase constitution sets the baseline damage tolerance: 8YSZ benefits from ferroelastic toughening in t′, whereas GdYbYSZ, lacking this mechanism, is more prone to cracking and localized delamination/chipping under tribological loading.

In Figure 10a (8YSZ, 80 mm standoff), the cross-section seems a APS lamellar shape with intersplat pores and locally connected microcrack paths, indicating that the short spray distance promotes strong splat flattening but also high quench tensile stresses that can generate an irregular microcrack network within the top coat, which is consistent with brittle, crack dominated scratch damage observed for this condition. The accompanying EDS spectrum and elemental maps confirm that the ceramic layer is chemically dominated by Zr, Y, and O, with a spatially uniform Zr rich skeleton and a finely dispersed Y signal across the coating thickness, indicating successful stabilizer incorporation rather than macroscopic Y rich segregation, the quantified chemistry in the selected region is close to the expected 8YSZ stoichiometry (about 71.1 wt% for Zr, about 8.6 wt% for Y, about 20.3 wt% for O). In Figure 10b (8YSZ, 100 mm standoff), the splat stacking appears comparatively more homogeneous at the same magnification, with fewer conspicuously large void clusters and a more continuous ceramic ligament between pores, implying improved intersplat cohesion at the intermediate standoff distance. This microstructural integrity provides a mechanistic basis for the comparatively more stable scratch response reported for 8YSZ at 100 mm, since a better-connected lamella network delays intersplat crack coalescence and reduces the probability of large-scale chipping events under progressive loading. The EDS results again show a consistent, well mixed Zr–Y–O distribution and similar composition (about 70.6 wt% for Zr, about 8.5 wt% for Y, and about 20.9 wt% for O), supporting that the performance differences among 80 and 100 mm are primarily microstructure controlled rather than chemistry controlled, while XRD indicates the coating is mainly tetragonal zirconia with only a minor monoclinic fraction at 80 and 100 mm. In Figure 10c (8YSZ, 120 mm standoff), the coating shows more pronounced interlamellar discontinuities and pore bands that are consistent with increased in-flight cooling and partial melting at longer spray distance, which weakens splat to splat bonding and facilitates preferential crack propagation along lamellar boundaries during scratching, thereby rationalizing the larger penetration and residual damage trends typically associated with more porous APS top coats. The elemental maps remain compositionally uniform and the local EDS chemistry is still close to nominal 8YSZ (about 69.7 wt% for Zr, about 8.4 wt% for Y, about 21.9 wt% for O), and, consistent with the phase analysis, no monoclinic phase is detected for 8YSZ at 120 mm, indicating that the standoff dependent scratch response is governed mainly by the standoff induced changes in lamellar cohesion, porosity, and microcrack topology rather than a change in the stabilizer distribution.

In Figure 11a (GdYbYSZ, 80 mm standoff), the cross-section shows a lamellar APS top coat with pronounced intersplat porosity and local microcrack paths, and the EDS spectrum plus maps confirm a Zr–O matrix uniformly decorated by Y together with clearly detectable Gd and Yb, with the quantified region indicating relatively higher dopant levels (about 11.75 wt% for Y, 4.91 wt% for Gd, 6.01 wt% for Yb) that are spatially well dispersed rather than concentrated in a continuous segregated band, which is consistent with the strong but intermittent brittle damage signatures observed during scratching. In Figure 11b (GdYbYSZ, 100 mm standoff), the splat stacking appears more coherent and the pore network is less interconnected at the same magnification, while the elemental maps still show a chemically homogeneous coating where Zr and O form the continuous ceramic skeleton and Y, Gd, and Yb remain distributed across the thickness, and the lower measured Gd and Yb in the selected spot (about 2.34 wt% for Gd and 2.75 wt% for Yb) is interpreted as local sampling variability within a heterogeneous lamellar microstructure rather than a process driven loss of dopants, supporting that the comparatively more stable scratch response at 100 mm is governed mainly by improved intersplat cohesion. In Figure 11c (GdYbYSZ, 120 mm standoff), the coating exhibits more pronounced interlamellar discontinuities and large void features, consistent with increased in-flight cooling at longer spray distance, and although the maps again indicate broadly uniform distribution of the constituent elements, the combination of a more defective microstructure with measurable dopant presence (about 10.82 wt% for Y, 4.79 wt% for Gd, 4.41 wt% for Yb) provides a mechanistic basis for the severe late stage scratch instability and the largest permanent groove formation reported for this condition.

4. Conclusions

This study systematically evaluated air plasma-sprayed 8YSZ and GdYbYSZ top coats produced at torch-to-substrate standoff distances of 80, 100, and 120 mm, with emphasis on how phase constitution and as-sprayed surface topography govern progressive-load scratch damage evolution as captured by instrumented scratch signals and three-dimensional profilometry. The principal conclusions can be summarized as follows.

(1) X-ray diffraction confirmed that the 8YSZ coatings were predominantly tetragonal, with a small monoclinic fraction at 80 mm and 100 mm (1.88 wt% and 0.98 wt%, respectively), while the 120 mm 8YSZ condition showed no detectable monoclinic phase. In contrast, GdYbYSZ exhibited a cubic phase structure without monoclinic peaks, and the systematic peak shift toward lower angles supports lattice expansion due to rare-earth dopant dissolution into the ZrO₂ lattice.

(2) Three-dimensional surface roughness measurements showed consistently smoother 8YSZ top coat surfaces relative to GdYbYSZ, with Ra of 5.54 to 5.85 µm and Sa of 9.84 to 11.8 µm for 8YSZ, versus Ra of 7.52 to 7.97 µm and Sa of 11.9 to 16.9 µm for GdYbYSZ. Among the investigated conditions, 8YSZ at 100 mm yielded the lowest Sa (9.84 µm), whereas GdYbYSZ at 120 mm exhibited the highest areal roughness (Sa 16.9 µm), indicating a stronger propensity for asperity-driven stress concentration under concentrated contact loading.

(3) From the results of the progressive-load scratch measurements, there is a noticeable coating-type influence on damage tolerance, with 8YSZ consistently exhibiting shallower grooves compared to GdYbYSZ at the initial and terminal points of the scratch path. The increase in the terminal depth was significantly magnified with a 120 mm standoff for both coatings. To be specific, the maximum end depth of the scratch was between 19.2 and 36.7 μm for 8YSZ, while the maximum end depth was between 46.0 and 71.9 μm for GdYbYSZ, thus exhibiting a much higher tendency towards permanent material removal with GdYbYSZ as the normal loading is increased. Scratch signal interpretation indicates that an increase in the standoff from 80 mm to 100 mm results in a reduction in acoustic emission signals while stabilizing penetration and residual depth evolution. On the other hand, a 120 mm standoff results in a dominant fracture mode shifting to the high-loading portion of the scratch track, consistent with the chipping and spallation observed in the mechanically weaker lamellar structure.

(4) The coefficient of friction responses exhibited coating-dependent instabilities that were consistent with the inferred fracture and debris formation events, where GdYbYSZ showed generally higher sustained friction and more pronounced critical excursions, including a peak near 0.6 around 1.55 mm for GdYbYSZ at 100 mm, followed by a sharp collapse, indicative of a localized fracture or spallation episode. These friction critical points support the mechanistic linkage between microdamage accumulation, debris-assisted plowing transitions, and abrupt changes in groove formation captured by depth and acoustic emission signals.

Overall, the combined phase analysis, roughness quantification, and instrumented scratch diagnostics demonstrate that torch standoff distance is a primary processing lever controlling as-sprayed surface relief and progressive contact damage tolerance, with 120 mm promoting the most severe late-stage groove deepening and brittle damage, particularly for GdYbYSZ, where the highest Sa and the deepest end track were obtained. Future work should extend this framework to repeated statistics based scratch testing and elevated temperature contact loading, and should quantitatively couple porosity and intersplat cohesion metrics with critical damage transitions to further strengthen process to structure to performance design rules for advanced TBC top coats.

From a broader perspective, this study demonstrates that progressive-load scratch testing combined with multi-channel in situ diagnostics offers a more discriminating assessment of TBC damage tolerance than conventional sliding wear methods, and that the ferroelastic toughening capacity inherent to the t′ phase remains the single most critical material-level factor governing resistance to concentrated mechanical contact in zirconia-based thermal barrier coatings.