Thermophysical properties of polymer composite materials based on TO-29-2 epoxy binder alloyed with AlN

. The results of an experimental study of the volumetric, caloric and transport properties of polymer composite materials based on the TO-29-2 grade epoxy binder alloyed with aluminum nitride (AlN) in a wide range of superconducting power equipment operating temperatures are presented. The influence of the inorganic component on the thermophysical properties of polymer composites has been established.


Introduction
A significant factor limiting the widespread use of modern power and energy plants based on high-temperature superconductors (HTSC) is the use of complex shaped dielectric products [1].One of the possible ways to solve this problem can be the creation of polymer composite materials (PCM) by vacuum infusion [2].
An experimental study of the thermophysical properties complex (linear thermal expansion coefficient, heat capacity, thermal conductivity, and thermal diffusivity) of a promising PCM based on the TO-29-2 grade epoxy binder alloyed with aluminum nitride (AlN), which is used in the development and creation of modern power and energy plants based on high-temperature superconducting devices, has been carried out.

Experimental details
PCM samples based on the epoxy binder of TO-29-2, as well as TO-29-2 containing 5, 10, 15, 20, and 50 wt.%AlN were studied (table 1).The average size of the AlN inclusions is 4-20 microns.To study the thermophysical properties, PCM samples were made in the form of cylinders of various sizes (from one initial billet for each of the studied materials), depending on the measurement method.It should be noted that the thermophysical properties of TO-29-2 + 50% AlN composite were studied for two samples obtained from two different billets.
The linear thermal expansion coefficient (LTEC, α) was investigated by the dilatometric method on a DIL-402C setup [3] in the temperature range of 100-300 K at 5 K/min heating-cooling rate.Samples had a length of 16-25 mm and a diameter of 5-6 mm were mounted on fused quartz supports and clamped between the sample holder and the pushrod with a force of 45 cN, which was maintained constant during the entire experiment.The elongation was measured by a linear variable differential transformer sensor (LVDT) with a resolution of 1 nm, and the temperature was measured by a thermocouple (type E), the bead of which was located in the immediate vicinity of the side surface of the sample (~3 mm).The specific isobaric heat capacity (cp) was investigated by the method of differential scanning calorimetry on a DSC 404 F1 Pegasus experimental setup [4].The mass of the samples was 25-35 mg.The samples under investigation were placed in platinum crucibles with corundum inserts and covered with platinum lids.The measurements of cp were made in the temperature range from 78-185 K to 300-500 K at a heating rate of 2 K/min in a dynamic (20 ml/min) atmosphere.Sapphire with a mass of 85.28 mg was used as a calibration sample.
The measurements of the thermal diffusivity (a) at room temperature were carried out by the laser flash technique using LFA-427 apparatus.A detailed description of the measurement method and the experimental setup is described in [5].Samples had a diameter of 12.6 mm and a thickness of 1.2-2.4mm.To increase the absorption coefficient of the radiation, the samples external surfaces were covered evenly by thin layer (up to 5 μm) of graphite using a special graphite spray.
The thermal conductivity (λ) was carried out by the transient plane source method on a Hot Disk TPS 2500S installation [6,7] in the temperature range from 78 to 300 K. Two plane-parallel cylindrical samples with carefully polished ends for each PCM with a diameter of 20 mm and a thickness of 15 mm were fabricated for λ measurements.Transiently heated plane sensor (30 µm thick) of the Hot Disk setup was clamped in the center between two disks of the investigated material.The samples container was placed in a Lauda Kryomat RUL80 cryostat (200-300 K) or Dewar vessel with liquid nitrogen (78 K) for cooling to a required temperature, which was controlled by a PT100 sensor.
The experiments were carried out in dry air for a and λ, helium (99.995 vol.%) for α, and argon (99.998 vol.%) for cp.Immediately before and after the experiments the samples mass and geometrical dimensions were measured, from which the PCM density at room temperature (ρ0) was calculated.The mass of samples was weighed on AND GH-252 analytical balances with an error of no more than 0.3 mg.The diameter was measured with a Kraftool electronic caliper with an error of no more than 0.03 mm.The samples thickness was determined with a Tesa Digico 10 electronic indicator at five points with an error of no more than 2 μm.
The estimated errors of the obtained data, confirmed by measurements of reference materials, were 2, 2-3, 2-3 and 5% depending on the temperature for α, cp, a, and λ, respectively.

Thermal expansion
Figure 1 shows the LTEC raw data on the one of the studied samples, smoothed by the Savitsky-Golay method.It can be seen that the results of successive experiments demonstrate good reproducibility.The behavior of the expansion coefficient is monotonous, without abrupt changes or jumps.The LTEC concentration dependence is shown in figure 4. It can be seen from the graph that the dependence is nonmonotonic.The addition of aluminum nitride in an amount of up to 10% significantly increases the thermal expansion.However, at a high content of the AlN component, the LTEC values are lower than those of the TO-29-2 sample without AlN.Deviations from the values obtained on the sample without doped component are given in table 4.

Heat capacity
Figure 5 presents the heat capacity measurement results of TO-29-2 composite for the unannealed and annealed sample.At the beginning, heating was carried out in the range of 78-300 K (curve 1) in a low-temperature experiment, after which heating was made in the range of 318-475 K (curve 2) in a high-temperature experiment.Curve 2 has a maximum above 330 K associated with the phase transformation occurring in the studied PCM.After that, low-temperature measurements (curves 3, 4) from 185 to 375 K, as well as high-temperature measurements (curves 5, 6) from 325 to 500 K were carried out.As can be seen from the graph, after the first high-temperature experiment and the phase transition occurred, the sample was annealed, while the heat capacity of the annealed sample (curves 3-6) began to lie below the unannealed one (curves 1, 2).Wherein, the difference in cp at 300 K is 10%.The sample mass after the first high-temperature experiment decreased by 0.75%, and after subsequent heating did not change.Apparently, the sample properties changed after such annealing, which led to a significant decrease of the cp data.Figure 5 shows that the results of low-temperature and high-temperature experiments separately from each other have good reproducibility within the estimated measurement errors.Figure 6 shows the cp experimental results for samples containing 10, 20, and 50 wt.%AlN.The data of two heatings are presented for all samples.The first heating and the second one were carried out in the range of 185-300 K and 185-375 K, respectively.Both heatings for each composite are reproduced among themselves within the estimated measurement errors.As can be seen from the figure 6, a phase transformation occurs above 315 K in the studied PCM, which represents itself as a sharp maximum at cp (T).The heat capacity has not been determined above the phase transition starting temperature until its ending, because it represents the phase intrinsic heat capacity and the phase transformation heat.It can also be seen from the figure 6 that an increase in the AlN content in TO-29-2 composite leads to cp decrease.A similar situation was observed in experiments with the composite samples containing 5 and 15 wt.%AlN.The heat capacity experimental points were approximated by the least-squares method with polynomials of different degree.Table 5 shows the investigated composites heat capacity recommended data.Figure 8 shows the heat capacity concentration dependences of studied composites depending on the content of AlN cp (X) at 185 and 300 K, as well as calculation results according to the additivity rule.The heat capacity additive values were calculated from cp of TO-29-2 composite (present study) and AlN [8].Figures 7 and 8 illustrate that with increasing X the cp of PCM decreases.It can be seen from the figure 8 the composites cp experimental data lie above the additive one.Wherein, for TO-29-2 composites with the content of 5 and 10 wt.% AlN, the difference between the experimental results and the additive values does not exceed the measurement error limits.Starting from 15 wt.%AlN content, the difference between the experimental and calculated heat capacity data exceeds the measurement error (2-3%), and the maximum deviation (15%) at 300 K is achieved for a composite with a content of 50 wt.%AlN.

Thermal diffusivity and thermal conductivity
The thermal diffusivity measurement points at room temperature obtained by the laser flash technique on the LFA-427 apparatus are shows in figure 9.It is clearly seen from the thermal diffusivity concentration dependence a (X) that with increasing X the a of PCM increases.
Using the a measured values, cp recommended data and ρ0, the λ of PCM at room temperature was calculated according to the relation λ = a ρ0 cp.The measured values on a and calculation results on λ are presented in table 6. Taking into account the uncertainties in the values of a, ρ0, and cp, the error in calculating λ is estimated at 4.5% at room temperature.The thermal conductivity experimental results obtained by the transient plane source method on the Hot Disk TPS 2500S setup for the samples No. 1, No. 5, and No. 6 are presented in figure 10. 2-3 experimental points at each temperature were obtained.Results obtained for each composite are in good agreement among themselves within the measurement error.The thermal conductivity temperature dependence λ (T) of presented composites increases monotonically with increasing temperature.Figure 10 shows that an increase in the AlN content in composite leads to λ increase.A similar situation was observed in experiments with other composite samples.
The thermal conductivity experimental points were approximated by the least-squares method with polynomials of different degree.Table 7 shows the thermal conductivity recommended data.A comparison of investigated composite material thermal conductivity experimental data is presented in figure 11.It can be seen that with an increase in AlN content to 50 wt.%,a noticeable increase in thermal conductivity is observed.Compared to TO-29-2 composite without AlN content, the thermal conductivity of a composite with 50 wt.%AlN is higher by 48-57%.Figure 12 presents the thermal conductivity concentration dependences (obtained on the Hot Disk and LFA-427) at 300 K.It can be seen from the graph that with an increase in the AlN content, the λ of PCM increases, wherein λ (X) of composites undergoes a jump in the concentration ranges of 15-20 wt.% (Hot Disk) and of 10-20 wt.% (LFA-427), and it begins to rise more sharply above 20 wt.%.Comparison of PCM thermal conductivity concentration dependences at room temperature, obtained on the Hot Disk and LFA-427 installations, showed the same character of the change in λ (X).The mean deviation of the absolute values was 9.5%.This value is within the limits of the total measurement errors (taking into account additional errors in the thermal conductivity calculation due to the heat capacity and density, as well as possible differences in the structure of the samples), which confirms the reliability of the results obtained.

Conclusion
For the first time, experimental data on the volumetric, caloric and transport properties of polymer composite materials based on the TO-29-2 epoxy binder containing 0-50 wt.% AlN were obtained in a wide range of negative and positive temperatures which are used in the creation of superconducting power equipment.The ambiguous influence of the aluminum nitride on the thermophysical properties of polymer composites has been established.The approximation equations and the tables of thermophysical properties recommended values for the entire measurement interval were compiled based on the experimental information obtained, and their errors were estimated.

Fig. 1 .
Fig. 1.Thermal expansion of the sample No. 3. 1 -1-st heating; 2 -2-nd heating; 3 -3-rd heating.Similar results were obtained for other samples.To obtain smoothed dependences of the LTEC (figure2), the experimental data were approximated by the least-squares method with polynomials of different degree, and the values of relative elongation  (figure3) were obtained by integrating these equations with the condition that the

Table 1 .
Grades of the samples.

Table 4 .
LTEC and relative elongation values comparison.

Table 5 .
Heat capacity recommended values.

Table 6 .
Transport properties at T = 300 K on the LFA-427.