Vol.3 No.1 2010

38/110

Research paper : Improving the reliability of temperature measurements up to 1550 ℃ (M. Arai et al.)−35−Synthesiology - English edition Vol.3 No.1 (2010) 100 hours after exposure to 1085 °C. Largest change was seen in emf at 100 hours at approximately 4 µV (corresponds to approximately 0.2 K) when the final heat treatment temperature was 730 °C. Here, it should be noted that the thermocouples that underwent 100 hours final heat treatment at 850 °C or 1030 °C showed very small drift and inhomogeneity. The emf change was within 0.5 µV (corresponds to 24 mK) over 150 hours, and this showed that the drift and inhomogeneity of the Pt/Pd thermocouple can be kept low by conducting appropriate heat treatment[12].As a result of conducting similar measurement of the drift and inhomogeneity for the exposure to 962 °C using the Ag fixed-point device, it was found that the effect of the final heat treatment at 850 °C was effective for different lots of wire[15]. Also, as a result of similar measurement for the exposure to 1324 °C using the Co-C eutectic point, it was found that the final heat treatment at 1030 °C was effective in reducing the drift at the Co-C eutectic point[16]. In the exposure to temperature lower than the Ag fixed point, the drift and inhomogeneity of the Pt/Pd thermocouple decreased rapidly[17]. The above results mean that the drift and inhomogeneity of the Pt/Pd thermocouple at temperatures up to 1330 °C can be kept extremely small by selecting the appropriate wires, the fabrication method, and the heat treatment method[16]. We were able to find a method for reducing and stabilizing the drift and inhomogeneity at each temperature of the Ag fixed point, Cu fixed point, and Co-C eutectic point.4.2.4 Comparison between Pt/Pd thermocouple and type R thermocoupleIn case of the type R thermocouple, the emf tended to decrease gradually with exposure to 1085 °C, and drifted approximately 0.2 °C after 300 hours exposure[9][14]. In contrast, as shown in Fig. 15, in the Pt/Pd thermocouple with appropriate final heat treatment, the drift was held within 0.03 °C even after 150 hours exposure to 1085 °C. As in the drift, the inhomogeneity of the type R thermocouple continued to change gradually while the Pt/Pd thermocouple with appropriate final heat treatment showed almost constant value within 0.04 °C at approximately 150 hours, as shown in Fig. 16. Due to these clear differences, we selected the Pt/Pd thermocouple as the transfer standard for Ag fixed point, Cu fixed point, and Co-C eutectic point. On the other hand, since the Pt/Pd thermocouple used palladium as its wire, it could not be used at the Pd fixed-point temperature because it would melt. Therefore, we decided to use the type R thermocouple, which had been used widely for a long period of time in Japan, as the transfer standard for the Pd fixed point.4.3 Element 3: Evaluation of the uncertainly of fixed-point calibrationThere are several sources of uncertainty when providing the temperature standard at the temperature fixed point as the national measurement standard using the thermocouple as the transfer standard. As an outline, the sources are: “the uncertainty inherent in the fixed-point device,” “the uncertainty inherent in the measurement system (voltmeter, reference junction device, etc.) when calibrating the thermocouple,” and “the uncertainty arising from the drift and inhomogeneity of the thermocouple itself that is being calibrated.” Table 1 shows the calibration uncertainty of the thermocouple at the Ag fixed point, Cu fixed point, Co-C eutectic point, and Pd fixed point. The expanded uncertainties (level of confidence of approximately 95 %) for the Ag fixed Fig. 15 Drift of the Pt/Pd thermocouples annealed at different temperatures as final heat treatment.Fig. 16 Inhomogeneity of the Pt/Pd thermocouples annealed at different temperatures as final heat treatment.450 ℃550 ℃600 ℃660 ℃730 ℃800 ℃850 ℃900 ℃960 ℃1030 ℃1080 ℃0050100150200250Time / h-112345Emf change / µV50 mk450 ℃550 ℃600 ℃660 ℃730 ℃800 ℃850 ℃900 ℃960 ℃1030 ℃1080 ℃0050100150200250Time / h-112350 mk(Emf8 cm-Emf0 cm) / µV

※このページを正しく表示するにはFlashPlayer9以上が必要です