Surface and Coatings Technology 168 (2003) 275–280

Evaluation of functionally graded thermal barrier coatings fabricated by detonation gun spray technique J.H. Kima, M.C. Kimb, C.G. Parka,* a

Center for Advanced Aerospace Materials, Pohang University of Science and Technology, Pohang 790784, South Korea b Research Institute for Science and Technology (RIST), Pohang 790-784, South Korea Received 30 October 2002; accepted in revised form 19 December 2002

Abstract In a new approach, an excellent functionally graded thermal barrier coating (FGM TBC) has been fabricated by the detonation gun spray process in conjunction with a newly proposed ‘shot-control method’. FGM TBCs were sprayed in the form of multilayered coatings with a compositional gradient along the thickness direction. The gradient ranged from 100% NiCrAlY metal on the substrate to a 100% ZrO2 –8 wt.% Y2O3 ceramic for the topcoat, and consisted of a finely mixed microstructure of metals and ceramics with no obvious interfaces between the layers. In the FGM layer of the FGM TBCs, the ceramics and metals maintained their individual properties without any phase transformation during the spraying process. Thermal shock properties of FGM TBCs were also investigated and the data obtained were compared with those for traditional duplex TBCs. 䊚 2003 Elsevier Science B.V. All rights reserved. Keywords: Thermal barrier coating; Functionally gradient material; Detonation gun spray technique

1. Introduction Thermal barrier coatings (TBCs) are being used in the fabrication of high temperature components of gas turbine engines for achieving improved performance, efficiency andyor extended component life. The major limitation to expanding the application range of TBCs lies in their premature failure and, as a result, their poor reliability. Most TBC systems used in aircraft engines consist of a 120–250 mm thermally insulating ceramic layer and a 50–120 mm metallic bond coat layer between the ceramic layer and the metal component surface, which are commonly referred to as ‘duplex TBCs’, and it has been reported that the failure of these duplex TBC systems is mainly caused by the thermal expansion mismatch between the ceramic and metal coating layers of the systems w1–3x. One way to overcome this problem is to introduce the concept of functionally gradient material (FGM) into TBCs, which are referred to as ‘FGM TBCs’. FGM TBCs are sprayed in the form of multi-layered coatings, *Corresponding author. Tel.: q82-54-279-2139; fax: q82-54-2792399. E-mail address: [email protected] (C.G. Park).

the composition of which varies in the thickness direction from 100% metal, applied directly to the substrate, to 100% ceramic for the topcoat. While the concept of FGM TBC itself is rather intuitive and simple, the fabrication of a fine mixture of ceramics and metals with a compositional gradient is quite difficult. Several processing techniques have been explored, e.g. plasma spraying, powder metallurgy, in situ synthesis, etc. but the optimum process for the fabrication of FGM TBCs still remains a challenging task w2–5x. In the present study, a detonation gun (d-gun) spray technique was applied to the fabrication of FGM TBCs. Since the d-gun spraying method provides inherently higher adhesiveycohesive strength than other thermalspray coatings because of the higher kinetic energy of the powder particles w6x, d-gun spraying represents a promising thermal spray technique for high quality coatings. A new approach, a ‘shot control method’ for the preparation of an excellent FGM TBC is also proposed in this study. For characterization of the d-gun sprayed FGM TBCs, X-ray diffraction (XRD), optical and scanning electron microscopy (SEM) were utilized. The thermal shock resistance of the FGM TBCs were

0257-8972/03/$ - see front matter 䊚 2003 Elsevier Science B.V. All rights reserved. doi:10.1016/S0257-8972(03)00011-2

J.H. Kim et al. / Surface and Coatings Technology 168 (2003) 275–280

276

Fig. 1. Schematic illustration showing the basic concept of the ‘shot control method’.

also evaluated by means of a burner rig tester, and compared with those for duplex TBCs. 2. Shot control method In d-gun spraying, the coatings are deposited in a discrete manner; i.e. by each shot (or each explosion), molten particles spread out and splatter as they strike the surface to form an approximately 10 mm thick coating deposit with a diameter of 20–25 mm, as shown in Fig. 1. The nominal working rate of the machine is four shots per second, and for spraying wide areas, the substrate is moved using a specimen manipulator as shown in this figure. The ‘shot control method’ proposed in the present study utilizes this unique feature of d-gun spraying, and cannot be applied to the plasma spray techniques, in which case the spray process is continuous. The basic concept of the method involves the alternate spraying of the ceramic and metal powders using two powder feeders to deposit a 1:1 mixture of ceramic and metal, as shown in Fig. 1. In the area overlapped by the successive shots, an excellent mixture of ceramics and metals can be produced on the level of each particle splat with this method. The area overlapped by each shot can be controlled by adjusting the moving speed of the specimen manipulator. In the present study, the optimum speed of the manipulator for a fine mixture of ceramics and metals was set at 2 cmys, as determined by several pre-experiments. Various spraying parameters of the d-gun machine used in the present study can be precisely controlled for each shot. Therefore, the ‘shot control method’ has an excellent advantage of spraying both ceramic and metal powders with their optimum spraying parameters since, with this method, basically only one type of powder is deposited per individual shot to produce a coating mixture of ceramics and metals. In other fabrication

methods for FGM TBCs, such as the mixed powder method, it is difficult to optimize the spraying parameters for a mixed coating of ceramics and metals due to the large difference in their material properties, e.g. melting temperature w3x. By changing the ratio of the ceramic powder to the metal powder that is shot, it is possible to control the volume ratio of the coating mixtures of ceramics and metals. For example, spraying with a shot ratio of 1:3, which means the spraying sequence consisting of three shots of metal powder followed by one shot of ceramic powder, resulted in a mixture of ceramics and metals with a ceramic to metal volume ratio of approximately 25:75. Therefore, an FGM coating having a compositional gradient in the thickness direction can be produced by spraying several coating layers with increasing ratios of ceramic to metal shots in a sequence as shown in Fig. 2. 3. Experimental details TBC specimens were produced with a Russian OB dgun machine w7x using an explosive gas mixture of acetylene and oxygen. Yttria-stabilized zirconia (YSZ) powder (ZrO2 –8 wt.% Y2O3) was used for the ceramic coating and a NiCrAlY powder (Ni–22Cr–10Al–1Y) for the metal coating. The porosity of the as-sprayed YSZ ceramic coating was 10.9% as measured by an image analysis method w8x. The duplex TBC specimens were prepared on Nibased superalloy (INCO-HX) substrates by spraying of a YSZ layer (550–600 mm thick) on top of a NiCrAlY layer (200–250 mm thick). In the case of the FGM TBC specimens, a functionally graded coating of a YSZyNiCrAlY mixture (the FGM layer) with a thickness of approximately 300 mm was added between the YSZ and NiCrAlY layers, which were the same as of the duplex TBCs. The FGM layer was produced by

J.H. Kim et al. / Surface and Coatings Technology 168 (2003) 275–280

277

Fig. 2. Schematic illustration showing the deposition scheme to produce a 7-layer FGM TBC by applying the ‘shot control method’.

spraying five coating layers in which the metal to ceramic shot ratios were 3:1, 2:1, 1:1, 1:2 and 1:3, respectively, as shown in Fig. 2. Coating samples for characterization were prepared by the classical metallographic technique and the samples were analyzed using optical microscopy, SEM and XRD. The micro-hardness profile in the thickness direction of the FGM coating was obtained using a Vickers hardness tester with a load of 300 g and at a distance of 100 mm from the NiCrAlY bond layer through to the YSZ top layer. In order to evaluate and compare the thermal shock resistance of both FGM and duplex TBCs, thermal shock tests were conducted using a burner rig test system. The surfaces of the coated specimens were heated in an acetylene–oxygen combustion flame, and a cycle of 10 min of heating followed by 10 min of a forced air cooling to room temperature was repeated

with the thermal load stepwise raised in a step of 100 8C for each cycle from 1000 to 1600 8C. 4. Results and discussion 4.1. FGM TBCs fabricated by ‘shot control method’ Fig. 3 shows the cross-sectional microstructure of the FGM layer of d-gun sprayed FGM TBCs using the ‘shot control method’. The NiCrAlY deposits have a light contrast while the YSZ deposits have a dark contrast under an optical microscope. As is clearly shown in this figure, the FGM layer had a finely mixed microstructure and the volume ratio of YSZ to NiCrAlY increased gradually inside the FGM layer in the thickness direction of the coating. In addition, despite the fact that the FGM layer was produced by spraying five coating layers and each layer was different in the volume ratios of metal

Fig. 3. (a) Cross-sectional micrograph showing the FGM coating layer of the d-gun sprayed FGM TBC fabricated by using the ‘shot control method’. (b) and (c) are the enlarged micrographs showing ceramic-rich and metal-rich areas of the FGM layer, respectively. (d) and (e) show the surface microstructure of 75% ceramic q25% metal (ceramic-rich) and 25% ceramic q75% metal (metal-rich) coating layers, respectively. The ceramic (YSZ) deposits have a dark contrast while the metal (NiCrAlY) deposits have a light contrast.

278

J.H. Kim et al. / Surface and Coatings Technology 168 (2003) 275–280

are fed simultaneously into a plasma heat source in the plasma spray methods w5x. The ‘shot control method’ proposed in the present study, therefore, has a definite advantage over the other thermal spray methods for FGM TBC fabrication; that is to spray both ceramic and metal powders with their optimum spraying parameters, since only one type of powder is deposited per individual shot to produce a coating mixture. As clearly shown in the results of the present study, the resultant benefit from the advantage of the ‘shot control method’ is to produce an excellent coating mixture of finely distributed ceramics and metals, maintaining their individual properties without any degradation such as oxidation and phase transformation. Fig. 4. Hardness values vs. distance from the NiCrAlY bond coating layer through to the SZ top coating layer of the FGM TBC specimens.

4.2. Thermal shock properties

to ceramic, no detectable interface was observed inside the FGM layer. The FGM layer of the d-gun sprayed FGM TBCs exhibits a lamellar structure of NiCrAlY and YSZ deposits. A lamellar structure is one of the typical features of thermally sprayed coatings and plasma sprayed FGM TBCs have also been reported to exhibit this type of structure w2–4x. The ceramic-rich area of the FGM layer appears to be a poor mixture of YSZ and NiCrAlY with stripes of NiCrAlY deposits due to the lamellar characteristic of d-gun sprayed coatings. However, the enlarged micrographs of both areas taken from both cross-sectional and planar viewpoints show that both the ceramic-rich and metal-rich areas of the FGM layer are also finely mixed on the level of each particle splat, as shown in Fig. 3 b–e. Measurements of the micro-hardness were made from the inner part of the NiCrAlY bond layer through to the YSZ top layer (Fig. 4). It was observed that hardness values increase linearly from the NiCrAlY layer through the FGM layer. Since the FGM layer of this specimen was produced to have a linear gradient of the ceramic volume fraction, this result indicates that both the ceramic and metal deposits maintain their own hardness values in the coating mixture fabricated by using the ‘shot control method’. Fig. 5 shows the results of XRD analyses from the dgun sprayed FGM TBC specimens. The intensity of both YSZ and NiCrAlY peaks changed proportionally with their volume fractions, but no other peaks were detected, suggesting that neither phase transformation nor severe oxidation had occurred during the d-gun spraying process. However, in the case of the plasma sprayed FGM TBC, it has been reported that severe oxidation and even phase transformation of metallic phases occurred as a result of the high processing temperature, because both metal and ceramic powders

In order to evaluate and compare the thermal shock resistance of both FGM and duplex TBCs, thermal shock tests were conducted using a burner rig test system. During the tests, both the input flame temperature at the front surface and the back surface temperature of the coating were measured in real time; the input flame temperature was monitored by a thermocouple placed at the surface of the coating, and the back surface temperature of the coating were calculated using three thermocouples which were inserted into the center of the substrate and spaced 5 mm apart behind the interface between the coating and the substrate. The difference between these two temperatures when the steady state had been achieved (after heating for f10 min) was defined as DT for the coated specimen at that input flame temperature, as shown in Fig. 6a. With a 600 mm thick YSZ top coating layer, d-gun sprayed FGM TBC specimens showed an excellent thermal barrier property with DT values reaching as high as approximately 800 8C. In addition, FGM TBCs exhibited a larger value for the DT than the duplex TBCs at all the input flame temperatures, due to the thermal barrier effect of the YSZ ceramic in the FGM layer, as shown in Fig. 6b. During the thermal shock tests, no visible damage of the coating was observed in either the FGM or the duplex TBC specimens until the input temperature reached 1600 8C. Microscopic observation of the FGM TBC specimens after the tests also revealed that FGM TBCs were apparently stable after the repeated and raised thermal loads reaching as high as 1600 8C, as shown in Fig. 7a. However, in the case of the duplex TBC specimens, a considerable number of defects and even a partial spalling were observed inside the YSZ coating layer after the thermal shock tests, as shown in Fig. 7b. It is considered that this loss of ceramic coat by partial spalling during the tests resulted in the rapid degradation of the thermal barrier property (DT) of the duplex TBC specimens, as shown in Fig. 6b. Since all the other factors are identical in both the FGM and

J.H. Kim et al. / Surface and Coatings Technology 168 (2003) 275–280

279

Fig. 5. Variation in the intensity of XRD peaks as a function of the YSZ and NiCrAlY volume fractions.

Fig. 6. (a) Temperature profiles of the d-gun sprayed FGM TBC specimens measured using the burner rig test system. (b) Comparison of the thermal barrier properties (DT) between FGM and duplex TBCs, as measured after 10 min of heating at each step of input flame temperatures during the thermal shock tests.

Fig. 7. SEM micrographs of (a) FGM and (b) duplex TBC specimens after the thermal shock tests, showing the damaged microstructures due to the repeated and raised thermal loads reaching 1600 8C. The arrows indicate the locations of horizontal crack inside the coatings.

280

J.H. Kim et al. / Surface and Coatings Technology 168 (2003) 275–280

duplex TBC specimens except introduction of FGM layer to the FGM TBC specimens, these results clearly demonstrate that FGM TBCs have higher thermal shock resistance than duplex TBCs with the effective aid of the functionally graded structure. In addition, it is noted that the degradation behavior of duplex and FGM TBCs were also different. In the case of the duplex TBC specimens, a considerable amount of horizontal cracks were formed inside the YSZ layer near the interface between the YSZ and NiCrAlY layers, as indicated by arrows in Fig. 7b, which were then connected with the vertical cracks initiated at the surface of the YSZ layer, resulting in partial spalling of the YSZ layer. It is considered that the main cause of the horizontal cracks near the interface is the thermal stress resulting from the difference in the thermal expansion property between the YSZ and NiCrAlY layers w1x, and the formation of these vertical cracks can be explained by thermal shock due to a high cooling rate of the tests w2,9x. In the FGM TBC specimens, the cracks were formed in much less quantity compared to the duplex TBCS with the aid of the accommodation of thermal stress by the functionally graded structure, and in different location as well; horizontal cracks inside the FGM TBCs were mainly observed inside the ceramic-rich area of the FGM layer, as indicated by arrows in Fig. 7a. In the case of the FGM TBCs, it would, therefore, appear that the ceramicrich area of the FGM layer constitutes a weak point in respect to resistance against thermal shock. This can be explained by the fact that this ceramic-rich area lacks metal deposits, which can accommodate the thermal strain developed by a thermal load, even though the thermal expansion mismatch between the YSZ and FGM layers is relatively small. 5. Conclusions In the present study, the fabrication of FGM TBCs was attempted using a d-gun spray technique. In a new

approach to preparing an excellent FGM TBC, the ‘shotcontrol method’ is proposed in this study and has a definite advantage of spraying both ceramic and metal powders with their optimum spraying parameters. The FGM layer of the FGM TBCs prepared by using this method exhibited a finely mixed microstructure of metals and ceramics with a desired compositional gradient in the thickness direction. The micro-hardness measurements and XRD analyses revealed that the ceramics and metals mixed in the FGM layer maintained their individual properties without severe oxidation or phase transformation. The expected improvement of the thermal shock resistance due to the realization of a functionally graded layer between the ceramic and the metal coating layers in TBC systems was also proven, as evidenced by thermal shock tests using a burner rig tester. Acknowledgments This work was supported financially by the Agency for Defense Development (ADD) under the contract No. UD980014AD. References w1x R.A. Miller, C.E. Lowell, Thin Solid Films 95 (1982) 265–273. w2x A. Kawasaki, R. Watanabe, M. Yuki, Y. Nakanishi, H. Onabe, Mater. Trans. JIM 37 (1996) 788–795. w3x J. Musil, J. Fiala, Surf. Coat. Technol. 52 (1992) 211–220. w4x Y. Shinohara, Y. Imai, S. Ikeno, I. Shiota, T. Fukushima, ISIJ Int. 32 (1992) 893–901. w5x A.S. Demirkiran, E. Avci, Surf. Coat. Technol. 116y119 (1999) 292–295. w6x R.C. Tucker Jr., J. Vac. Sci. Technol. 11 (1974) 725–734. w7x H.W. Jin, Y.M. Rhyim, M.C. Kim, C.G. Park, J. Kor. Inst. Met. Mater. 36 (1999) 707–714. w8x J.H. Kim, M.C. Kim, C.G. Park, Met. Mater. Int. 7 (6) (2001) 557–563. w9x D. Zhu, R.A. Miller, Mater. Sci. Eng. A 245 (1998) 212–223.