JPS63115042A - Method for measuring thermal diffusivity - Google Patents

Abstract

PURPOSE:To permit such measurement which minimizes thermal influence of a coating material by using a platinum wire the surface of which is coated with aluminum nitride as a probe. CONSTITUTION:The aluminum nitride has high heat conductivity and can transmit calorific value quickly to a semiconductor melt to be measured by energizing a fine platinum wire even if the fine platinum wire is coated with the aluminum nitride. For example, the surface of the fine platinum wire having 1,400mum diameter is coated with the aluminum nitride to 30mum thickness and the thermal diffusivity of molten indium antimonide is measured by a nonstationary fine wire method using such fine wire as the probe in common use as a heating element. The thermal diffusivity can then be measured at 1% error as compared to the theoretical value of the case in which there is no coating material. The measurement error by the presence of the coating material is, therefore, restricted within a permissible range.

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to a technique for measuring thermophysical properties of a melt. More specifically, the present invention relates to a method for measuring thermophysical properties of a melt using an unsteady thin wire method.

(Prior Art) In growing a perfect semiconductor single crystal, it is important to simulate the flow of melt and heat during the crystal growth process. For this purpose, it is necessary to accurately measure the thermophysical constants of the semiconductor melt. Among several thermophysical constants, the value of thermal diffusivity is a physical constant that should be accurately measured prior to simulation.

Until now, the thermal diffusivity of semiconductor melt has been determined by Glaz
ov), “Liquid Sem1conduct” by Chizhevskaya and Glagoleva et al.
ors”) (PlenumPress
), New York, 1969), p. 30-33.
As reported in , measurements have been performed using a steady-state method using concentric cylinders. On the other hand, in the measurement of thermal diffusivity in relatively low-temperature liquids, long plates and long heights are used to measure the thermal diffusivity of relatively low-temperature liquids.
ysics) B14. p., 1435.1981, measurement by the unsteady thin wire method is recommended as it guarantees excellent measurement accuracy.

In measuring the thermal diffusivity by the unsteady thin wire method when the surface of the thin wire is covered with a coating material, there is a relationship as shown in the following equation between the temperature rise ΔT detected by the thin wire and the measurement time t.

ΔT = q/4n people (Int+A+1/1(B1nt
+C)) (1) Here, q is the amount of heat generated from the thin wire, λ is the thermal conductivity, t is the time during which electricity is applied to the thin wire, and A is a constant related to the thermal diffusivity of the melt and the diameter of the thin wire. B and C are constants related to the thermal conductivity and thermal diffusivity of the sample melt, the thin wire and the coating material, as well as the diameter and thickness of the thin wire and the coating material. In this equation, if the thermal influence of the covering material can be ignored, the third term in parentheses can be ignored. In this case, the thermal diffusivity k can be determined from the slope q/4nλ of a graph in which ΔT is plotted on the horizontal axis and Int is plotted on the horizontal axis. However, the diffusivity is =A/p-C, where p is the density and C9 is the specific heat.

(Problems to be Solved by the Invention) When applying the measurement of thermal diffusivity by the unsteady thin wire method to a semiconductor melt, the following two problems arise. That is, the semiconductor melt has electrical conductivity and corrosivity. For this reason, the current may be short-circuited through the semiconductor melt during measurement, and the platinum wire, which has excellent properties as a probe in this method, may react with the semiconductor melt and be corroded.

For this reason, it is necessary to coat the surface of the thin platinum wire serving as the probe with a substance that has insulating properties and does not react with the semiconductor melt. However, when the surface of the platinum thin wire is coated with an insulating and non-reactive substance, each of these substances exhibits thermal behavior, so the third value in parentheses in equation (1)
term cannot be ignored, the measured value of thermal diffusivity deviates from the theoretical value, and its accurate measurement becomes difficult.

An object of the present invention is to provide a measurement method that minimizes the thermal influence of a coating material in measuring the thermal diffusivity of a semiconductor melt by an unsteady thin wire method using a thin platinum wire as a probe.

(Means for Solving the Problems) The gist of the present invention is to use a platinum wire whose surface is coated with aluminum nitride as a probe in measuring the thermal diffusivity of a semiconductor melt by the unsteady thin wire method.

(Function) Aluminum nitride has high thermal conductivity, and even if the thin platinum wire is coated with aluminum nitride, the amount of heat generated by energizing the thin platinum wire can be immediately transmitted to the semiconductor melt to be measured. Therefore, it is possible to suppress measurement errors due to the presence of the covering material within an allowable range.

FIG. 2 is an example of measurement of thermal diffusivity by the unsteady thin wire method carried out by the present inventors, and shows measurement errors due to the presence of the covering material. Here, the diameter is 1400pm
The surface of the thin platinum wire was coated with aluminum nitride having a thickness of 1100p, and the thermal diffusivity of molten silicon was measured. Electricity was applied to the thin platinum wire that served as the probe, and the temperature of the molten silicon rose due to the amount of heat generated, and the rise in temperature of the thin platinum wire was detected as the electrical resistance of the thin platinum wire. This figure shows the relationship between the temperature rise read from the change in electrical resistance of the thin platinum wire and time expressed in logarithm. The straight line shows the theoretical rise in temperature in case 4 without the covering material, but according to experiments, the temperature rise deviates from the ideal state due to the presence of the covering material (in 3 in the figure). According to the principle of this measurement method, the slope of this graph corresponds to the thermal diffusivity, but deviation from the straight line causes errors in measurement.

As shown in Figure 1, if we plot the relationship between the size of the error and the thickness of aluminum nitride, which is the coating material, we can see that the error within the allowable range for the value used in the simulation is 2.
．． If the thickness is 5%, the thickness of aluminum nitride that satisfies this requirement is 1100p at most.

(Example) The present invention will be described in more detail below using Examples.

[Example IJ The surface of a thin platinum wire having a diameter of 1400 pm was coated with aluminum nitride having a thickness of 30 pm using the CVD method.

Using this thin wire as a heating element and probe, the thermal diffusivity of molten indium antimonide (InSb) was measured by the unsteady thin wire method and was found to be 12.3 W/m-.
As shown in Figure 1, compared to the theoretical value without coating material,
Thermal diffusivity could be measured with an error of 1%.

"Example 2" The surface of a thin platinum wire with a diameter of 1400 pm was coated with a thickness of 110 pm.
Coated with 0p aluminum nitride. When the thermal diffusivity of molten indium antimonide (InSb) was measured by the unsteady thin wire method using this thin wire as a heating element and probe, as shown in Figure 1, the thermal diffusivity of molten indium antimonide (InSb) was 2. Thermal diffusivity could be measured with an error of .5%.

"Example 3" The surface of a thin platinum wire with a diameter of 1400 pm was coated with a thickness of 30 pm.
coated with m aluminum nitride. Using this thin wire as a heating element and probe, molten silicon (
When the thermal diffusivity of Si) was measured, as shown in FIG. 1, the thermal diffusivity could be measured with an error of 0.5% compared to the theoretical value without the coating material.

"Example 4" The surface of a thin platinum wire with a diameter of 1400 pm was coated with a thickness of 110 pm.
Coated with 0p aluminum nitride. When the thermal diffusivity of molten silicon (Si) was measured by the unsteady thin wire method using this thin wire as a heating element and a probe, it was found that the thermal diffusivity of molten silicon (Si) was 1.5 ° The thermal diffusivity could be measured with an error of %.

(Effects of the Invention) As described above, by using the present invention, it is possible to measure the thermal diffusivity of a semiconductor melt having corrosivity and electrical conductivity.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing the thickness of the covering material aluminum nitride covering the surface of the thin platinum wire and the error of the measured value of thermal diffusivity from the theoretical value. 1 is when indium antimonite is measured, and 2 is when silicon is measured. FIG. 2 is a diagram showing the relationship between the temperature rise and the energization time of the thin wire when the surface of the thin wire is covered with a coating material. 3 is when there is a covering material, 4 is when there is no covering material.

Claims (1)

[Claims] A thermal diffusivity measuring method characterized by using a platinum wire whose surface is coated with aluminum nitride as a probe in measuring the thermal diffusivity of a semiconductor melt by an unsteady thin wire method.