Bond coats were formed by using commercial coating powder of AMDRY 9624 and AMDRY962 (Sulzer Metco Holding AG, average weight ratio composition: Ni-22Cr-10Al-1.0Y, average particle sizes: 45 ~ 74μm (AMDRY 9624) and 53 ~ 106μm (AMDRY 962)), AMDRY 9951 (Sulzer Metco Holding AG, average weight ratio composition: Co-32Ni-21Cr-8Al-0.5Y, average particle size: 5 ~ 37 μm), and AMDRY 997 (Sulzer Metco Holding AG, average weight ratio composition: Ni-23Co-21Cr-9Al-0.6Y, average particle size: 5 ~ 38 μm). In formation of the bond coat, AMDRY 9624 and AMDRY 997 were applied to HVOF (Diamond Jet-2600 DJM, Sulzer Metco Holding AG, Switzerland) method, while AMDRY 962 was applied to APS (9MB, Sulzer Metco Holding AG, Switzerland) method. Top coats were formed by using APS method for the commercial powder of METCO 204 C-NS of general purity, and METCO 204C-XCL of high purity (METCO 204 C-NS: 8.0Y₂O₃-0.7SiO₂-0.2TiO₂-0.2Al₂ O₃-0.2Fe₂O₃ doped in ZrO₂ and METCO 204 C-XCL: 8.0Y₂O₃-0.05SiO₂-0.05TiO₂-0.05Al₂O₃-0.05Fe₂O₃ doped in ZrO₂, Sulzer Metco Holding AG, Switzerland, average particle size: 45~140μm, D50: 68 ~ 77 μm), METCO 206A with addition of rare earth elements such as Y, Yb, and Gd (Sulzer Metco Holding AG, Switzerland, ZrO₂ with addition of 9.5Y₂O₃-5.6Yb₂O₃-5.2Gd₂O₃, average particle size: 45 ~ 125 μm), and commercial La₂Zr₂O₇ powder (LZO, LAO-109-1, Praxair Surface Technologies, Indianapolis, IN). In addition, LZO powder and 204 C-NS powder were mixed and employed as a starting powder.

To evaluate the thermal barrier performance as a function of thickness, the total thicknesses of TBC systems were designed to be 250 and 1000 μm, and the thickness ratio between the top and bond coats was varied when the total thickness of TBC system was 250 μm. This was aimed at considering the effects of top coat thickness on thermal barrier performance for the same coating thickness and securing an optimum ratio in coating thickness. Accordingly, the thickness ratios between the top and bond coats for the structure were designed to be 150 : 100 and 200 : 50 for the thin TBC system, while the ratio was designed to be 850 : 150 for the thick TBC system. Detailed structural design is shown in Fig. 1. Namely, in the thin TBC system, the bond coats had compositions varied and were formed by HVOF method to the thicknesses of 100 and 50 μm, respectively, while the top coats were formed on each bond coat using 8YSZ by APS method to the thicknesses of 150 and 200 μm, respectively. In the thick TBC system, Ni-based bond coat was formed by APS method to the thickness of 150 μm, while the top coat was formed using 8YSZ by APS method to the thickness of 850 μm. In addition, the top coat with a bi-layer structure was designed to evaluate thermal barrier performance as functions of composition of bond coat and purity of buffer layer, which is shown in Fig. 2. In the bi-layer structure, the buffer layer was formed using a general purity (METCO 204 C-NS) and a high purity (METCO 204 C-XCL) by APS method to the thickness around 60 μm, and the top coat was formed using a low thermal conductivity powder of Yb-Gd-YSZ by APS method to the thickness around 540 μm. The bond coats in TBC systems of the bi-layer structure were formed using AMDRY 9951 and AMDRY 997, respectively, by HVOF to the thickness of 300 μm. Also, for evaluation of thermal barrier performance as functions of structure and composition in the top coat, the TBCs were designed as functions of the composition and number of the buffer layer as well as the composition and number of the top coat, which is shown in Fig. 3. Here, the bond coat was formed using AMDRY 962 by APS method to the thickness of 150 μm, and 8YSZ alone or LZO and 8YSZ mixed in 25 : 75 vol% was applied to the buffer layer. Meanwhile, top coats in the multilayer structure were formed using LZO and 8YSZ mixed in 50:50 vol% and 25:75 vol% as well as 8YSZ alone. Total thickness of the top coat was varied as shown in Fig. 3, and correlation between the thickness and number of coating layer was discussed.

To affirm the status of cracking inside coating layer, cracking and delamination at interfaces which can be included upon formation of TBC system, microstructure was observed as functions of thickness, composition, and incorporation status of the buffer layer. The specimens were mounted with a fluid epoxy resin, polished using silicon carbide paper, then final polishing using 3 μm and 1 μm diamond pastes. The cross-sectional microstructure of TBC specimens was observed using a scanning electron microscope (SEM, JEOL Model JSM-5610, Japan). Also, for differentiation of interfaces as resulting from incorporation of the buffer layer and conversion to multilayers, a composition analysis was conducted for each top coat by using an energy dispersive spectrometer (EDS, Oxford Instruments, Oxford, UK).

To evaluate thermal barrier performance as functions of thickness, composition, and incorporation status of the buffer layer in TBC system, Jet Engine Thermal Shock (JETS) test was employed, and the equipment used for evaluation of thermal barrier performance is shown in Fig. 4. In JETS test, the surface of top coat was heated for 20 seconds using flame at 1400°C with a mixture of LPG and oxygen, and then cooling for 20 seconds using nitrogen gas was conducted, which constituted one cycle, until 2000 cycles. JETS test is a universally applied method by OEM (Original Equipment Manufacturing) industry to evaluate thermal stability of TBCs, which can evaluate interface stability under drastic temperature changes. Also, in JETS test, the flame was prevented from being applied to other parts of the specimen except for the top coat surface by wrapping surroundings of the specimen with a ceramic holder. Pyrometer (Infrared thermometer, Wavelength : 3.9 μm, CTlaser MT, Optris, Germany) was installed on the surface of top coat and the back face of substrate, and temperatures were measured by using the pyrometer for 20 seconds during which the flame was applied, with calibration of the pyrometer being conducted by using a thermocouple whenever composition, thickness, and structure was changed. The surface temperature of TBC was increased to about 1000°C by heating for 20 seconds, at which time the temperature on the back face of substrate reached a range of 560 ~ 610°C. Meanwhile, the surface temperature of the top coat and the back face temperature of substrate cooled down to about 550°C and about 350°C, respedively, during cooling by nitrogen gas for 20 seconds. Thermal barrier performance of TBC was evaluated by measuring temperatures for the total of 4 sections after setting up according to the number of cycles (Section 1 : 0 ~ 50 cycles, Section 2 : 500 ~ 1000 cycles, Section 3 : 1000 ~ 1500 cycles, and Section 4: 1500 ~ 200 0cycles). To evaluate thermal barrier performance as a function of thickness, temperature difference between the surface of top coat and the back face of substrate was affirmed first for the 4 types of TBC systems having a thin thickness, and a comparison was made with the thick TBC system after selecting 1 type with the excellent thermal barrier performance. Thermal barrier performance was also compared and considered for composition, incorporation status of the buffer layer, and conversion to multilayers in TBC system by the same method of measuring temperatures of the surface of each top coat and the back face of substrate.

The cross-sectional microstructures of TBC systems formed in accordance with Fig. 1 are shown in Fig. 5, which was designed by varying thickness ratio in the bond and top coats and thickness in the top coat. In the thin TBC system of Figs. 5(a) - (d), the top coat thicknesses were formed to be 110 ~ 154 (Figs. 5(a) and 5(b)) and 176 ~ 185 μm (Figs. 5(c) and 5(d)), respectively, while those of the bond coat were formed to be 76 ~ 56 and 31 ~ 57 μm, respectively, with the thicknesses being formed to be somewhat thinner than the designed values. Meanwhile, in the thick TBC system of Fig. 5(e), the thicknesses of the bond and top coats were formed to be about 131 ~ 145 and 820 ~ 878 μm, respectively. In TBC systems, no cracking and delamination at interfaces between coating layers were affirmed irrespective of the thickness ratio between coating layers (thickness ratio between the bond and top coats) and the thickness, with clear differences of thickness being affirmed in the microstructures. In the top coat, inherent defects such as pore, splat boundary, and unmelted particles, etc. appearing in the microstructures formed by APS method were observed, and relatively porous microstructures were observed in the top coat of all specimens.

Thermal barrier performance, i.e. for temperature difference, evaluated for the TBC systems having cross-sectional microstructure of Fig. 5 is shown in Table 1 and Figs. 6 and 7. For both the thin and thick TBC systems, the temperature difference was measured in Section 3 (1000 ~ 1500 cycles), and the temperature differences as a function of thickness ratio between the bond and top coats were measured and shown in Fig. 6 for the relatively thin TBC systems. In the case of TBC systems where the bond and top coats were designed and formed to be 150 and 100 μm, respectively, the TBC system having Ni-Co-based bond coat showed a thermal barrier performance (temperature difference) of 110 ~ 132°C, and the TBC system having Ni-based bond coat a thermal barrier performance of 115 ~ 155°C as shown in Figs. 6(a) and 6(b), respectively. Meanwhile, in the case of TBC systems where the bond and top coats were formed to be 200 and 50 μm, respectively, the TBC system having Ni-Co-based bond coat showed a temperature difference of 132 ~ 179°C, and the TBC system having Ni-based bond coat a temperature difference of 119 ~ 193°C as shown in Figs. 6(c) and 6(d), respectively. Based on such results, thermal barrier performance could also be affirmed to be improved as the thickness of the top cost increases in relatively thin TBC system. Also, composition of the bond coat could be seen to have no effects on the temperature difference, i.e. thermal barrier performance. The results of comparative consideration for thermal barrier performance are shown in Fig. 7 as a function of the thickness of TBC system, and compared with the results of Fig. 6(c) as the case showing the largest temperature difference in Fig. 6 for understanding improvement in thermal barrier performance in the thick TBC system. Temperature differences in the thick TBC were 248 ~ 699°C, showing excellent thermal barrier performance compared with the thin TBC system (Fig. 7(b)). This is a result in agreement with the report that the temperatures experienced by high-temperature components for gas turbines were lowered by 4 ~ 9°C whenever the coating thicknes was increased by 25 mm.²⁴⁾ In the present study, similar results were also observed in Section 1, Section 2, and Section 3 not shown here, and no delamination or fracture was observed in the specimens even after 2000 cycles.

The cross-sectional microstructures of TBC systems formed in accordance with Fig. 2 are shown in Fig. 8, which was designed with incorporation of the 8YSZ material as the buffer layer to supplement the low CTE and the low mechanical properties of the low thermal conductivity materials. The thickness of the top coat was formed to be 578 ~ 761 μm, and that of bond coat 185 ~ 318 μm, while the specimen c (Fig. 8(c)) was formed to be relatively thick. The buffer layer was 67 ~ 119 μm thick and formed to be thicker than the designed thickness as a whole. Meanwhile, since the buffer layer thicknesses were designed and formed to allow improvement to be produced in the same range as fracture or delamination in TBC system generally occurred around 100 μm from the interface between the bond and top coats.^25,26) As mentioned earlier, only inherent defects visible in the microstructures formed by APS method were observed in the top coat, and no cracking and delamination were observed at the interface in the bond and top coats, and between the buffer layers.

The results of temperature differences at Section 1 (0 ~ 500 cycles) for all 4 types of specimens for the TBC systems prepared by application of Yb-Gd-YSZ starting powder and incorporation of 8YSZ buffer layer, which were designed to secure thermal durability, are shown in Table 2 and Fig. 9. First, when the buffer layer was incorporated, the larger temperature difference, i.e. the more excellent thermal barrier performance was manifested, the higher the purity of buffer layer (purity of starting powder). When the starting powders of general purity were used, the TBC systems with the Ni-Co-based and Ni-based bond coat compositions showed temperature differences of 284 ~ 369°C (Fig. 9(a)) and 380 ~ 435°C (Fig, 9(b)), respectively. When the buffer layer was formed using a starting powder of high purity, the TBC systems with the Ni-Co-based and Ni-based bond coat compositions showed temperature differences of 193 ~ 504°C (Fig. 9(c)) and 276 ~ 632°C (Fig. 9(d)), respectively. Also, temperature differences were measured with composition of the bond coat, and excellent thermal barrier performance was manifested in the TBC system with application of Ni-based bond coat (Figs. 9(b) and 9(d)), unlike Fig. 6 for the thin TBC system. Such large temperature differences are considered attributable to the improvement in thermal barrier performance due to Yb-Gd-YSZ coating layer of low thermal conductivity and the effect of thermal dispersion at interfaces due to formation of additional interfaces through incorporation of the buffer layer. Consequently, manifestation of thermal barrier performance can be affirmed through application of low thermal conductivity materials and incorporation of the buffer layer which was similar to the case where the thickness of coating layer was increased to 850 μm. Also, the improvement in thermal barrier performance as a result of applying high-purity starting powder in the buffer layer may be considered to be a result of delaying extinction of pores with excellent sintering resistance under high-temperature environments. For the TBC system with the Co-Ni-based bond coat, delamination occurred at 720 cycles in the case where the buffer layer of general purity was incorporated, and at 218 cycles in the case where the buffer layer of high purity was incorporated. Meanwhile, for the TBC system with the Ni-based bond coat, delamination occurred at 1127 cycles in the case where the buffer layer of general purity was incorporated, whereas no delamination or fracture was observed up to 2000 cycles when the buffer layer of high purity was incorporated. Consequently, the improved thermal barrier performance and thermal durability could be secured together upon application of the Ni-based bond coat as well as incorporation of the high-purity buffer layer in design and formation of TBC system.

The cross-sectional microstructures of TBC systems formed in accordance with Fig. 3 are shown in Fig. 10, which was designed with variation in the composition and number of the buffer layer and in the composition and number of the top coat. The thicknesses of top coat, buffer layer, and bond coat were formed in accordance with the design concept. In the case of Fig. 10(a), the thicknesses of top coat, buffer layer, and bond coat were formed to be 244 ~ 294, 28 ~ 70, and 115 ~ 159 μm, respectively, while the thicknesses of top coat, buffer layer, and bond coat in Fig. 10(b) were formed to be 306 ~ 328, 54 ~ 92, and 135 ~ 173 μm, respectively. Meanwhile, the thickness of the buffer layer with mixed composition in Fig. 10(b) was calculated as a top coat layer. The thicknesses of the top coats in Figs. 10(c) and (d) were formed to be 437 ~ 517 and 495 ~ 555 μm, the buffer layer 37 ~ 87 and 38 ~ 64 μm, and the bond coat 161 ~ 175 and 145 ~ 163 μm, respectively. Also, the thickness of the buffer layer with mixed composition in Fig. 10(d) was calculated as a top coat layer. In the surfaces of top coat, the microstructure of mixed composition of LZO and YSZ (Figs. 10(a) and 10(b)) and the microstructure of YSZ alone (Figs. 10(c) and 10(d)) can be detected, respectively. The interfaces of the first buffer layer (porous 8YSZ layer) and the second layer (top coat layer or the second buffer layer) are shown by dotted lines. Continuous microstructures were observed without differentiation of interfaces within the buffer layer and the top coat showing gradient composition. Meanwhile, incorporation of the YSZ buffer layer and the mixed layer (mixed layer of YSZ and LZO) was aimed at relaxation of a difference CTE from that of the bond coat which could occur when the low thermal conductivity material, LZO, was applied along with supplementation of low mechanical characteristics. Since the brighter colors could be observed to be displayed in the microstructures of the top coat and buffer layer as the added amount of the LZO was increased, proper formation of microstructures according to the design concept could be verified.

Temperature differences were measured for the case with incorporation of compositional continuity to control the thermal and residual stresses generated at the interfaces due to low CTE of low thermal conductivity materials, and for the multilayer structure aimed at performance improvement of surfaces, with the results shown in Table 3 and Fig. 11. Figs. 11(a) and 11(b) are the results of temperature difference measured in Section 3 (1000 ~ 1500 cycles), and Figs. 11(c) and 11(d) are the results of temperature difference measured in Section 1 (0 ~ 500 cycles).

When the specimens in Fig. 11 had 2 layers (Fig. 11(a)), 3 layers (Fig. 11(b)), and 4 layers (Fig. 11(c)) as the top coat layers, a sound state was maintained up to 2000 cycles, while delamination occurred at 102 cycles in the case of 5 layers (Fig. 11(d)). First, when only the buffer layer of 8YSZ was incorporated (Figs. 11(a) and 11(c)), temperature differences were measured to be 325 ~ 391°C for the top coat (2 layers, Fig. 11(a)) with LZO and YSZ being mixed in 50 : 50 vol%, and 501 ~ 539°C for the multilayer top coat (4 layers, Fig. 11(c)). Consequently, improvement of thermal barrier performance could be affirmed when the number of interfaces between coating layers (number of layers) was increased in the multilayered structure. Namely, it is considered to be the scattering effect upon thermal diffusion at interfaces resulting from an increase in the number of interfaces and the result caused by a relatively thick coating thickness. Meanwhile, in the case of TBC system with the multilayered buffer layers (Figs. 11(b) and 11(d)), the temperature differences were shown to be 427 ~ 447°C for the top coat (3 layers, Fig. 11(b)) with LZO and YSZ being mixed in 50 : 50 vol%, and 572 ~ 631°C for the multilayer top coat (5 layers, Fig. 11(d)).

Although the low thermal conductivities of top coat and the thermal barrier performance at the multilayered structure were to be comparatively considered by making the thicknesses of Figs. 3(a) and 3(b) to be the same, and the thicknesses of Figs. 3(c) and 3(d) similar in the design concept for TBC systems, the specimen b (Fig. 10(b)) was formed to be thicker than the specimen a (Fig. 10(a)) in the actually formed TBCs. Therefore, the thickness of coating layer together with the previously mentioned scattering effect at interfaces is considered to have effects as the factors determining thermal barrier performance. Meanwhile, when a thickness similar to that of design concept was formed, a larger temperature difference was measured for the specimen d (Fig. 10(d)) than the specimen c (Fig. 10(c)) of TBC systems having the multilayered structure at the interface. Based on this observation, an increase in the number of interfaces, i.e. the scattering effect at interfaces could be more efficient in improving thermal barrier performance. However, although an increase in the number of top coat layer, i.e. the number of interfaces in TBC system can produce improvement of thermal barrier performance, it is not considered to be desirable from the aspect of thermal durability, in view of the artificial defects which can be included in the forming process, resulting in the requirement for additional process control to secure interface stability. ²⁷⁾ Meanwhile, the thermal barrier performance evaluated in the present study showed larger temperature differences than the existing report where the TBCs with a thickness of about 100 ~ 500 μm generally exhibited the thermal barrier performance of about 100 ~ 300°C,⁹⁾ because the metal substrate was cooled by the air at 400 ~ 600°C through cooling flow path under the environments where actual high-temperature components were applied. Consequently, although temperature difference in the present study may not be considered as the absolute thermal barrier performance for TBC systems, it could be utilized as a useful method for comparison of relative thermal barrier performance in developing and applying new materials, including structures.

TBCs with the single-layer buffer layer (Fig. 10(a)) and the bi-layer buffer layer (Fig. 10(b)) were completed up to 2000 cycles. While tests were conducted through 2000 cycles in Fig. 10(c) designed as a multilayered structure, delamination occurred at 102 cycles in Fig. 10(d). Consequently, quick delamination could be affirmed to occur upon incorporation of a large number of buffer layers in design and formation of TBC systems due to the defects between interfaces.