**3. Thermal properties**

## **3.1 Thermal conductivity of gadolinium zirconate-based thermal barrier coatings**

It can be said that the thermal conductivity is the most important property of the TBC. Therefore, mechanisms of heat transfer of the TBC material have a great significance. The phonons are primarily responsible for thermal conduction of ceramic materials. So thermal resistivity of the ceramic materials depends on the scattering of phonons. There are a lot of ways to scattering of phonons. One of them is to disturb the lattice vibration with additional scattering centers such as crystallographically changing to a lower symmetry, presence of different atoms inside a unit cell, porosities, grain boundaries, and impurities. As the number of these scattering centers increases, the thermal conductivity decreases [16]. Therefore, Gd2Zr2O7 having pyrochlore and disordered fluorite structure has low thermal conductivity due to the oxygen vacancies in the unit cell.

**Table 1** summarizes the lowest and highest thermal conductivity values of the GZ- and YSZ-based TBCs at specific temperature ranges produced by different techniques and having different coating designs. In addition, the thermal conductivity data of YSZ-based TBCs were given in **Table 1**. As understood from the data in the table, the thermal conductivity of the coating was influenced by the production method. This difference is entirely related to the different characteristics of microstructural morphology of the production processes. The lowest thermal conductivity values were obtained in the TBCs produced by APS due to its porous and lamellar morphology containing a lot of defects such as unmelted particles and cracks at splat boundaries. They acted as phonon scattering centers. On the other hand, thermal conductivity values of the TBCs produced by SPS and EB-PVD technique were higher than the TBCs produced by APS. This was because columnar morphology of the TBCs is produced by SPS and EB-PVD. Moreover, it is seen that the GZ-based TBCs had lower thermal than the YSZ-based TBCs produced by the same method and having the same design. This situation could be explained by oxygen vacancies acting as phonon scattering centers in the unit cell of GZ-based TBCs.

MLed and FGed GZ/CYSZ TBCs consisting 2, 4, 8, and 12 layers were produced by APS technique, and their thermal conductivity was analyzed by laser flash method [8]. Thermal conductivity values of the single-layered GZ and MLed and FGed GZ/CYSZ TBCs were lower than single-layered YSZ TBC (0.91–1.79 W/mK<sup>−</sup><sup>1</sup> at 25–1105°C). Results showed that the lowest thermal conductivity values at 1105°C were achieved for ML12 and FG12 coatings having the highest number of layers. From this point of view, it was concluded that thermal conductivity value decreased by increasing the number of layers. This situation was attributed to the interfaces of the layered periodicity and increasing porosity level by increasing number of layers. Because these porosities and interfaces between GZ and CYSZ phases in the MLed


#### **Table 1.**

*Thermal conductivity of GZ- and YSZ-based materials produced via different processes and having different designs.*

and FGed designs acted as phonon scattering centers and reduced phonon conduction. Moreover, hemispherical reflection, that is, reduced phonon conduction, was thought to be active in the MLed and FGed designs. A similar phenomenon regarding the increasing hemispherical reflection by increasing the number of layers was reported in other studies [36].

Moskal et al. [31] produced single-layered GZ and YSZ TBCs by APS processes and compared their thermal conductivity at the temperatures between 25 and 1100°C. According to the results, GZ-based TBC had lower thermal conductivity (0.59–1.50 W/mK<sup>−</sup><sup>1</sup> ) than YSZ-based TBC (0.8–2.25 W/mK<sup>−</sup><sup>1</sup> ). Moreover, differences between the thermal conductivity values of the GZ and YSZ were higher at higher temperatures (900–1100°C). In different studies [29, 30], Mahade et al. produced single-layered YSZ, double-layered GZ/YSZ, and triple-layered GZ dense/GZ/YSZ TBCs by SPS process. The thickness of the layers was different in the studies. As a result of thermal conductivity measurements, they showed that single-layered YSZ-based TBC had higher thermal conductivity than that of doublelayered and triple-layered GZ-based systems. This situation was attributed to both oxygen vacancies in the crystal structure and larger difference between atomic weights of the cations of the GZ-based TBC. In another study [35], single-layered GZ and multilayered GZ/YSZ + Yb2O3 + Gd2O3 TBCs were produced by EB-PVD process and analyzed their thermal conductivities at temperatures between 25 and

*State of the Art of Gadolinium Zirconate Based Thermal Barrier Coatings: Design, Processing… DOI: http://dx.doi.org/10.5772/intechopen.85451*

1316°C. They observed slight reduction in the thermal conductivity of the coatings thanks to multilayered design (1.13–1.42 and 1.10–1.22 W/mK<sup>−</sup><sup>1</sup> ) for single and multilayered systems, respectively. This decrease was tried to be explained by increasing porosity and hemispherical reflection of the multilayered design. In the case of sintering (fully dense material), GZ had lower thermal conductivity than YSZ (1.6–2.3 W/mK<sup>−</sup><sup>1</sup> at 700°C for GZ and YSZ, respectively) due to the higher concentration of oxygen vacancies of GZ and atomic weight difference between Gd and ZrO2 [17]. These oxygen vacancies and atomic weight differences gave rise to an effective phonon scattering.

## **3.2 Mechanical properties of gadolinium zirconate-based thermal barrier coatings**

**Table 2** shows some of the mechanical properties of YSZ- and GZ-based TBCs produced by APS and sintering techniques. The bonding strength of a coating to a substrate or cohesion strength of the coating has been determined by using adhesion test. This test has been carried out according to the ASTM C-633 standard test method. The bonding strengths of GZ/CYSZ multilayered and functionally graded TBCs that were produced by APS were determined by using ASTM C-633 test method [8]. The bonding strength values of the single-layered GZ and MLed and FGed GZ/CYSZ TBCs having 2, 4, 8, and 12 layers changed between 8.87 and 12.1 MPa. This changing in the bonding strength was attributed to the changing porosity level of the coatings. As the porosity value increased, the adhesion strength of the coating decreased. On the other hand, bonding strength values of the GZ-based TBCs (8.87–12.1 MPa) were comparable to single-layered YSZ-based TBC (10.1 MPa) produced by the same technique [37]. Fracture type of the singlelayered GZ coating was fully adhesive (fracture at the bond coat/ceramic top coat interface), but complex adhesive/cohesive fracture was seen (fracture both bond coat/ceramic top coat interface and within the ceramic top coat layers) in the MLed and FGed coatings. No fracture was observed at the bond coat/substrate interface.

Carpio et al. [23] analyzed the hardness of double-layered and functionally graded GZ/YSZ TBCs produced by APS. The hardness values of the TBCs were 4.0 and 4.1 for double-layered and functionally graded GZ/YSZ, respectively. On the


#### **Table 2.**

*Mechanical properties of GZ- and YSZ-based materials produced via different processes and having different designs.*

other hand, hardness value of the sintered (fully dense) GZ was changing between 6.0 and 10.0 GPa [3, 17]. The reason of this increase in the hardness was related with the reduction of porosity and cracks after sintering process. It is suggested to investigate the mechanical properties of the GZ-based TBCs produced by EB-PVD technique in future works.

## **3.3 Thermal cycling and thermal shock performance of gadolinium zirconate-based thermal barrier coatings**

Thermal cycling performance test has been generally conducted by heating the TBCs to the exact temperatures in exact time, holding them at that temperature in exact time and cooling them to a certain temperature. Therefore, the failure mechanism in the TBC that is exposed to the thermal cycle test is associated with the stress produced by growth and shrinkage of TGO during heating and cooling. As a result, complex failure morphologies such as cracks propagating within the TBC, at TBC/TGO and TGO/bond coat interface, come into existence. On the other hand, thermal shock test has been carried out by sudden heating and cooling of the TBCs. Thermal shock resistance of the TBC material depends on the its thermal expansion coefficient, elastic module, fracture resistance, and phase transformation phenomenon.

As already mentioned before (introduction part), GZ has very advantageous properties (low thermal conductivity, high phase stability and higher CMAS, and hot corrosion resistance than YSZ) as TBC material, but its thermal cycle and thermal shock performance is poor. Therefore, works on GZ-based TBC are focused on improving its thermal cycling and thermal shock performance. Thermal cycling performance of the MLed and FGed GZ/CYSZ TBCs was tested by using gas burner method (oxy-propane flame) [8]. TBCs were heated to 1250°C for 1 min and cooled below 150°C by using air jet. According to the results, single-layered GZ fully spalled after 165 cycles, and there was a small edge spallation in the MLed GZ/ CYSZ TBCs after 300 cycles. However, there was no visible spallation in the FGed GZ/CYSZ TBCs even after 300 cycles. The microstructural characterizations of MLed and FGed GZ/CYSZ TBCs after thermal cycling test revealed that TGO layer played an active role on the spallation of single-layered GZ. In the case of MLed GZ/CYSZ TBCs, horizontal cracks propagating inside the ceramic coating layers were observed. On the other hand, there was no microcrack or spallation in the FGed GZ/CYSZ TBCs. It was concluded that an improvement in the thermal cycling performance of the single-layered GZ TBC took place thanks to MLed and FGed designs. Furthermore, GZ/CYSZ TBCs having FGed design had superior thermal cycling performance than that of MLed design. As a continuation of the previous work, in another study [14], these TBCs (MLed and FGed GZ/CYSZ coatings) were subjected to the thermal shock test. During the test, the TBCs were directly placed to the furnace at 1250°C and hold for 5 min. Then they were dropped into the cold water. As seen in **Figure 7**, single-layered GZ and YSZ (sample codes: GZ1 and YSZ1, respectively) TBCs were subjected to the thermal shock test as reference samples as well as MLed and FGed TBCs, and the number of cycles to failure for each coating was written on the images. The failure or spallation took place on the coating of single-layered GZ1 after only ten cycles. On the other hand, a significant improvement happened in the thermal shock lifetime of the MLed and FGed TBCs.

This improvement in thermal shock lifetime of TBCs having MLed and FGed designs was ascribed to the decreasing high thermal mismatch stress generated due to the sharp differences in the CTE between bonding layer and CYSZ and GZ phases. As mentioned before CYSZ had higher CTE (13 × 10<sup>−</sup><sup>6</sup> K<sup>−</sup><sup>1</sup> ) than GZ (10.4 × 10<sup>−</sup><sup>6</sup> K<sup>−</sup><sup>1</sup> ). Therefore, large thermal expansion mismatch among ceramic

### *State of the Art of Gadolinium Zirconate Based Thermal Barrier Coatings: Design, Processing… DOI: http://dx.doi.org/10.5772/intechopen.85451*

layers, metallic bond coat, and substrate was balanced thanks to gradual compositional variation of CYSZ layers having in GZ layers. Furthermore, microstructural evaluations were carried out on GZ/CYSZ MLed and FGed TBCs exposed to thermal shock test. Microstructures of the GZ-based TBCs after thermal test showed the evidence of the thermal stress on the ceramic layer. The vertical, horizontal, and propagating cracks were formed on the ceramic top layers. Propagating cracks are located around the vertical cracks. The shape of the vertical cracks was looking like capillary vessels. Other than this, there were penetrating cracks forming in the TGO layer and penetrating into the bond coat. A similar microstructure regarding the shape of the cracks of thermally cycled TBCs was reported in other studies [23, 38]. As a result, the improvement in thermal shock life of GZ-based TBCs was directly reducing the thermal mismatch among the layers thanks to CYSZ. Furthermore, from the microstructures, it was assumed that energy of thermal fatigue cracks absorbed at the interface between GZ and CYSZ layers or splats causing improvement in thermal shock life.

In a study [22], double-layered GZ/YSZ TBCs were produced by APS technique so as to have different porosity values. Then thermal cycling behavior of the TBCs were tested at 1400°C by using a gas flame (for heating) and air jet (for cooling). Results showed that thermal cycling lifetime of the single-layered GZ was significantly improved thanks to a double-layered design. Moreover, GZ/YSZ TBCs having the highest porosity level (%32) withstood up to 1627 thermal cycles. Carpio et al.

**Figure 7.**

*Macroscopic image showing remaining top coats on the substrate after thermal shock test and number of cycles to failure.*

produced MLed and FGed GZ/YSZ TBCs by APS technique [23]. The GZ/YSZ TBCs having two and five layers were exposed to thermal cycling test in the furnace. The TBCs were heated to 1050°C and then cooled by air. The thermal cycling life of the MLed GZ/YSZ TBC was about 1000 cycles. On the other hand, FGed GZ/ YSZ TBC had resistance up to about 2300 cycles. This situation was attributed to the homogenous distribution of thermal stress in the layers of FGed coating. Mahade et al. subjected the specimens of double- and triple-layered (GZ dense/GZ/ YSZ) TBCs that were produced by SPS technique to the thermal cycling test. The thermal cycling tests were carried out in the furnace (heating the specimens up to 1100 and 1200°C for 1 hour and cooling to 100°C). Researchers concluded that an improvement in the thermal cycling lifetime took place thanks to GZ dense/GZ/ YSZ triple-layer approach. This was due to the lower porosity content of triple-layer approach. Thanks to dense GZ layer, less oxygen has reached to the bond layer. As a result, the thickness of TGO had remained below the critical value causing failure in the coating. In another study by Zhang et al. [6], double-layered Yb2O3-doped GZ (GYbZ)/YSZ TBCs are produced by EB-PVD technique. They produced two sets of (GYbZ)/YSZ TBCs. One of them had normal (sharp) interface between GYbZ and YSZ phases. On the other hand, another set of the TBCs had a uniform interface with a fluent transition (gradient coating) region (nearly 10 μm). Results of thermal cycling test, which was carried out by heating the surface of the TBCs up to 1350°C by a flame and cooling by compressed air, showed that TBC having a gradient transition region had higher lifetime (1346 cycles) than that of TBC having sharp interface (942 cycles). This beneficial effect of the gradient transition region was attributed to the gradually compositional change reducing the thermal mismatch at the interface between ceramic layers.

Consequently, it is possible to prolong the thermal cycling and thermal shock lifetime of the GZ by using a second material (such as YSZ and CYSZ) in multilayered and functionally graded designs. The stress generated by thermal mismatch has been reduced and reaction tendency of the GZ with TGO layer prevented thanks to multilayered and functionally graded designs.

### **4. Attack of harmful foreign substances**

### **4.1 CMAS and hot corrosion resistance of gadolinium zirconate-based thermal barrier coatings**

The jet engine fuel especially used in military aircrafts has low quality, and it contains appreciable levels of impurities such as Na2SO4 + V2O5. They cause severe damage on the ceramic top layer at high temperatures [14, 26, 39]. On the other hand, calcium-magnesium-alumino-silicate (CMAS) is known as airborne silicate particles (sand, ash, dust, etc.). The CMAS entering to the turbine melts on the hot TBC surfaces and penetrates into the ceramic top coat. This situation leads to premature failure of the coating [12, 14]. There are a lot of studies in the literature proving that GZ-based TBCs have excellent CMAS and hot corrosion resistance.

In the studies that were carried out by Habibi et al. [21, 24], YSZ- and GZ-based TBCs were produced by APS technique. To evaluate hot corrosion behavior of the TBCs, hot corrosion powders (Na2SO4 + V2O5) were applied (20 mg/cm2 ) on the TBCs, and they were cyclically exposed to the 1050°C. The failure of the YSZ-based top coat started after 24 hours and spallation occurred. The detailed characterizations revealed that YVO4-type crystals formed as result of the reactions between yttria (Y2O3) and V2O5 or NaVO3. This formation of the YVO4 caused the transformation of ZrO2 from tetragonal to monoclinic due to the leaching of the Y2O3 from

*State of the Art of Gadolinium Zirconate Based Thermal Barrier Coatings: Design, Processing… DOI: http://dx.doi.org/10.5772/intechopen.85451*

the YSZ. On the other hand, failure in the GZ-based coating started after 36 hours. This situation showed that GZ-based coating had much better hot corrosion resistance than that of YSZ. In the case of GZ-based coating, molten Na2SO4 + V2O5 mixture reacts with the bulk Gd2Zr2O7 layer to form GdVO4. However, Na2SO4 + V2O5 mixture attacks to the Y2O3 that is stabilizer of YSZ. Liu et al. [26] investigating hot corrosion behavior of sintered GZ in the presence of V2O5 explained the reaction mechanism. According to this, at 700°C molten V2O5 started to react with Gd2Zr2O7 to form ZrV2O7 and GdVO4 (see Eq. (1)). In the case of 750–850°C, the final reaction products of GdVO4 and m-ZrO2 formed depleting V2O5 (see Eq. (2)):

$$\text{\textbullet V}\_2\text{O}\_5\text{\textdegree\_{(liquid)}} + \text{Gd}\_2\text{Zr}\_2\text{O}\_{7\text{(solid)}} \rightarrow \text{2ZrV}\_2\text{O}\_{7\text{(solid)}} + \text{2Gd}\text{VO}\_4\text{\textdegree\_{(solid)}}\tag{1}$$

$$\text{V}\_2\text{O}\_5\text{(liquid)} + \text{Gd}\_2\text{Zr}\_2\text{O}\_7\text{(solid)} \rightarrow 2\text{Gd}\text{VO}\_4\text{(solid)} + 2\text{ m} - \text{ZrO}\_2\text{(solid)}\tag{2}$$

Suspension plasma-sprayed (SPSed) GZ/YSZ TBCs having multilayered designs were subjected to the hot corrosion test [25]. V2O5 and Na2SO4 salts were applied to the specimen surface at a concentration of 4 mg/cm<sup>2</sup> . The test was conducted at 900°C for 8 hours. Microstructures revealed that GZ-based coatings had lower reactivity with the corrosive salts and the formation of gadolinium vanadate (GdVO4).

Another promising feature of the GZ is its high resistance to CMAS attack. In contrast to conventional YSZ, GZ is highly effective in resistance to the penetration of molten Ca-Mg-Al-silicate (CMAS) glass deposit. This higher resistance of the APSed GZ-based TBC was proven for prolonged durations [40]. The CMAS sand was applied to the surface of the TBCs at a concentration of 35 mg/cm<sup>2</sup> . Then TBCs were heated to 1200°C in air for 24 and 168 hours in a furnace. Results showed that CMAS melt was fully penetrated inside of the YSZ. The penetration depth of YSZ-based TBC was approximately 200 μm. However, CMAS melt was arrested on the surface of GZ-based coating (at a penetration depth 20 μm). This resistance of the GZ-based TBC was attributed to the formation of a crystalline sealing layer. The chemical composition of the sealing layer was Ca2Gd8(SiO4)6O2 (apatite phase), and it formed as a result of the high-temperature chemical interactions between the GZ and CMAS melt. Similar results were found for GZ-based TBC that was produced by EB-PVD technique [20]. The mechanism of the CMAS resistance of GZ-based TBC was explained via formation of a highly stable apatite phase. This phase sealed the flow channels of GZ-based SPSed TBC.

In a unique study [14], GZ- and YSZ-based TBCs were produced by APS technique, and they were exposed to CMAS and hot corrosion (CMAS + hot corrosion) test. This test was executed in the same experiment at once. Moreover, a defocused CO2 laser beam was used as heat source, and TBCs were simultaneously cooled from the back surface of the substrate to create a thermal gradient and harsh environment. Hot corrosion and CMAS powders were applied on the surface of the TBCs together at a concentration of 30 mg/cm2 , and they were heated by a laser beam to 1250°C and held at that temperature for 1 hour. The rodlike structures formed on both YSZ- and GZ-based coatings. The chemical composition of the rodlike structures was in the form of YVO4 and GdVO4 for YSZ and GZ, respectively. The micrographs of the YSZ-based TBC clearly demonstrate that CMAS + hot corrosion products penetrated through microcracks and pores. However, there was a continuous reaction layer formed at the boundary between CMAS and GZ-based TBC. The thickness of reaction layer was approximately 6 μm. The lamellar microstructural morphology of APSed TBCs turned into a denser structure seeming to extend crystals into the CMAS. This situation was consistent with the similar studies that investigated interactions between GZ-based TBCs and different CMAS sands [20, 23, 40].

Further penetration of the CMAS products into the GZ based was prevented thanks to this reaction (sealing) layer. The formation and sealing mechanism of the reaction layer were defined as follows:


Gledhill et al. [2] applied the lignite fly ash on the surface of APSed YSZand GZ-based TBCs at 1200°C. They showed that the molten lignite fly ash completely penetrated (throughout entire thickness) inside the YSZ-based TBC. Unlike this, the molten lignite fly penetrated only approximately 25% of the GZ-based TBC thickness thanks to a crystalline reaction layer arresting molten fly ash on the TBC.
