**4. Investigations of thermal barrier properties of ceramic coatings with the use of heating**

## **4.1 Investigations of thermal barrier properties of ceramic coatings with the use pulse of heating**

Thermal barrier coating application efficiency depends on ceramic layer thermal conductivity, which determines the cooled blade temperature drop and corresponding increase in its service life. To measure thermal conductivity of a TBC ceramic layer, a laser flash method is used (Siegwart et al., 2006) (Parker et al., 1961). The method is based on irradiating the surface of a flat sample surface with an energy pulse, followed by recording a temperature rise on its backside. Thermal diffusivity and heat capacity are determined experimentally using the pulse method of heating. Then, thermal conductivity *()* is calculated from these characteristics:

$$
\lambda = \mathbf{a} \cdot \mathbf{b} \cdot \mathbf{C}\_{\mathbf{p}} \tag{1}
$$

in W/mK, where *a* is thermal diffusivity (cm2/s); is density (g/cm3); *Cp* is heat capacity (J/gK).

Measuring each thermophysical characteristic is an independent task. The most developed method is that of thermal diffusivity calculation, because the main formula for thermal diffusivity includes only one experimentally measured parameter. It is a period for the temperature to reach half of its maximum level:

$$a = 0.1388 \ (\\$2/\tau\_{1/2})\tag{2}$$

in cm2/s, where is sample thickness, and 1/2is the time required for the temperature of the sample backside to reach the level equal to one-half of the maximum temperature. Coefficient 0.1388 corresponds to an ideal case when the following conditions are met: instantaneous and uniform heat pulse, heat pulse absorption in a thin surface layer, and no heat losses. For experimental thermal diffusivity determination, one should know neither absolute temperatures nor parameters of a heat flow affecting a sample. Measuring heat capacity by the flash technique, especially for coated samples, is a much more complicated task. Analysis of thermal diffusivity - and thermal conductivitjrof cerarrricxoaflngs are discussed elsewhere (Pawlowski et al., 1984). For thermophysical studies of ceramics condensates, the TC-3000H unit manufactured by the Sinku-Riko Company was used. A ruby laser with a wave length of 6.943 m was used as an energy source, and as a temperature pickup on the backside of the sample, either a thermocouple (Pt-PtRo) or an infrared sensor was used (Maesono, 1983). The tested sample is essentially a flat disc 10 mm in diameter and 0.8 to 2 mm thick. When thermal diffusivity is studied in this unit, two types of experimental errors are possible. The first type of errors results from some lack of information on the parameter values used in the design formulas. They are due to the available accuracy of sample thickness and time of т1/2 measurements, exactness of the temperature rise assessment, and of catching the moment of the sample irradiation start. These errors are covered in detail in (Cape & Lehman, 1963). On the basis of the results reported in the literature, one can deduce that, with the modern data collection systems used, the contribution of this type error does not exceed 0.5%. The second type of errors is due to the difference between the experimental conditions and

Investigations of Thermal Barrier Coatings for Turbine Parts 153

Fig. 23. Sketch of specimen: *1* – specimen half with coating, *2* - specimen half without coating, *3* – coating, *4, 5* – thermocouples, *6* – axis of specimen (flame), *7* – burner

of a cold wall 600 °C. The results of investigations are presented on Fig. 24 and 25.

rig and use the small-size specimens whose surfaces during tests are accessible for inspecting the thermal state both by the contact and contactless methods. This rig in particular is usable effectively for conducting the comparison thermal barrier propertiesand thermocycles tests of various coatings. The rate of change of the temperature in a thermocycle reaches 100 °C/s. For performing these investigations, a test rig has been developed with gas-flame heating of model specimens. The gas generator is a water electrolysis device equipped with a control system; it provides the variable flammable gas flow. Hydrogen has a high combustion temperature and this fact ensures high-speed heating of the specimens. This test rig has a system for providing enrichment of the flammable gas with different fuels. This makes it possible to attain the required gas composition. While testing, the burner is installed fixed, however the attachment allows its position to be adjusted. The hollow specimen is of an axisymmetrical form. Before the test, the burner is installed in a way ensuring coincidence of the specimen axis with the flame torch axis in the process of heating. While investigating the efficiency of influence of ceramic coating on specimen temperature state, the unit with specimens was fixed. A special specimen construction was developed for these tests. The hollow specimen was cut longitudinally in two equal portions. The ceramic coating under investigation was applied on one half of the specimen, the other half remained uncoated. The thermocouples ХА by diameter of 0,2 mm weld on an external surface of halfs of a compound specimen (Fig. 23) and are connected to recording computer system. The half of a specimen is protected from products of combustion by a coating it is warmed up less, than unprotected half. The difference of temperatures *t* of protected wall with coating and unprotected wall characterizes the efficiency and thermal conducting of the thermal barrier coatings. Heat insulating material was placed between them to exclude the influence of heat transfer through the contacting edges of the specimen halves. While heating, temperature was measured on the outer (opposite to flame torch) specimen side. Conditions for heating the inner surfaces of both specimen halves by flame were the same, but with a difference in heat protection efficiency the outer surface of the specimen with a TBC had a lower temperature than the surface without a TBC. The after of lighting of a combustible gas the heating of an internal walls of both halfs of model begins. The difference of temperatures on lateral side grows until the heat transfer from a hot surface of a wall to cold surface is less, than a heat-conducting from an external surface in an environment. At absence of the organized cooling lateral side of a wall the maximal difference of temperatures *t*max outside of both of halfs corresponds to a gradient of temperatures on TBC under these conditions. In experiment *t*max it is reached at temperature

assumptions in the mathematical model used for calculating thermal diffusivity and heat capacity. These errors are related to the finite pulse duration and its spatial inhomogeneity, to heat losses (due to irradiation, mainly), and to violation of pulse absorption conditions in the thin surface layer. These errors may be avoided by using certain corrections (Clark & Taylor, 1975). For the TC-3000H unit, pulse duration and spatial inhomogeneity errors determined according to the Sinku-Riko Company recommendations are unessential (less than 1%). Heat losses in the experiment result in a quick temperature rise to its maximum and then a sharply defined smooth temperature decrease. The main cause that gives rise to measurement errors is radiation heat exchange, whose effect rises simultaneously with a temperature rise. The errors caused by radiation may account for 30%. To meet the requirements of pulse absorption in the thin surface layer, the ceramic samples, which are partially transparent, were coated with a thin layer (10 to 12 m) of the NiAl intermetallic compound (20% Al). This layer ensured steady surface optical parameters of the samples as well.
