Vol.3 No.1 2010

36/110

Research paper : Improving the reliability of temperature measurements up to 1550 ℃ (M. Arai et al.)−33−Synthesiology - English edition Vol.3 No.1 (2010) is changed as shown in Fig. 10. The thermocouple where the Seebeck coefficient differs according to the position of the wire is generally called the “inhomogeneous thermocouple.” In actual temperature measurement, the emf change generated when the position where the temperature gradient falls is changed by altering the insertion length of the thermocouple is often called “inhomogeneous.” In the inhomogeneous thermocouple, the emf is not determined by the temperature values of the measuring and reference junctions only, as it is dependent on the temperature distribution of the furnace. Therefore, the inhomogeneity may give erroneous standard if the calibrated transfer standard is used in different temperature gradient. The Seebeck coefficient S of the general thermocouple is temperature dependent, and a similar way of thinking is available.4.2.2 Development of evaluation method of drift and inhomogeneityAs mentioned earlier, the drift and inhomogeneity of the thermocouple are extremely important factors when evaluating the stability of the transfer standard. Particularly, since the drift and inhomogeneity increase in the high temperature range, they become major components of the uncertainty of thermocouple calibration.To evaluate the stability of the thermocouple, at first, we studied the drift and inhomogeneity of the Pt/Pd thermocouple fabricated according to the paper[11] of the joint research of the National Institute of Standards and Technology (NIST) (USA), and the Istituto di Metrologia “Gustavo Colonnetti” (IMGC) (Italy). Following this paper, the wire annealing was performed for 10 hours at 1200 °C by direct current application to the thermocouple wires. After assembling the thermocouple, the annealing in a furnace was performed for 3 hours at 1100 °C , and then 10 hours at 450 °C.The Cu fixed-point device was used to maintain the measuring junction of the thermocouple at a constant temperature, and the drift and inhomogeneity of the thermocouple was measured simultaneously using the stability and the uniformity of the freezing temperature of the fixed-point device[12]. The measuring junction of the thermocouple was inserted to the position 1 cm above the deepest point of the measurement well of the Cu fixed-point device, the copper was repeatedly frozen and melted, and the emf changes were measured. The temperature control of the Cu fixed-point device was programmed to repeat the melting and freezing constantly, and the measuring junction of the thermocouple was always exposed to the Cu fixed-point temperature. Figure 11 shows the result of the measurement of in-situ drift for approximately 500 hours by fixing the position of the measuring junction. The emf of the Pt/Pd thermocouple changed markedly in the first 50 hours after starting the exposure, and showed almost constant value after 100 hours.Figure 12 is a plot of the emf change when the thermocouple was moved up and down when the freezing of copper was in progress during the drift measurement. The position of the measuring junction when conducting the drift measurement was set as “0 cm”. The 0 hour to 505 hours in the figure is the time lapse from the start of exposure. The data at the start of exposure (0 h) was the measurement of the emf change when the thermocouple was pushed in at the rate of 0.5 cm/min, and the other data were the measurements taken when the thermocouple was pulled up at the rate of 0.5 cm/min. The act of “changing the insertion length of the thermocouple” to obtain the data for Fig. 12 was the same as Fig. 11 Drift of the Pt/Pd thermocouple when exposed to the Cu fixed point.Time/ hEmf Change / µV00200300400500100132450 mKPt/PdFig. 12 Inhomogeneity of the Pt/Pd thermocouple when exposed to the Cu fixed point.Position of measuring junction / cm00481216246Emf change / µVInhomogeneity25h13h7h1.5h0h505h241h79h49h50 mK

※このページを正しく表示するにはFlashPlayer9以上が必要です