Vol.3 No.1 2010

37/110

Research paper : Improving the reliability of temperature measurements up to 1550 ℃ (M. Arai et al.)−34−Synthesiology - English edition Vol.3 No.1 (2010) the act of “changing the temperature distribution along the wire of the thermocouple” while maintaining the temperature of the measuring junction, and the emf change reflected the inhomogeneity of the thermocouple. To quantitatively assess the inhomogeneity of the thermocouple, the emf change from 0 cm to upper 8 cm was defined as the “inhomogeneity” of the thermocouple, as shown in the example for 7 hours in Fig. 12.The stability of the thermocouples at a temperature fixed point could be evaluated almost automatically in large amounts using this method, and the stable development and evaluation of the transfer standard that will be explained in the next section could be done efficiently. 4.2.3 Fabrication of the stable thermocoupleThe Pt/Pd thermocouple is composed of the platinum and palladium wires that are pure metals, and it is reported that the emf change due to inhomogeneity is mostly related to the palladium wire[13]. To investigate the effect of the palladium wire from different lots, four Pt/Pd thermocouples were fabricated using the palladium wires from four different lots with the same nominal purity of 99.99 %[14]. Here, the four fabricated thermocouples will be called TC-a, TC-b, TC-c, and TC-d. Of the palladium wires from four different lots, the wires of TC-a and TC-b were made by the same manufacturing procedure by the same manufacturer. The wires of TC-c and TC-d were each purchased from manufacturers different from that of TC-a and TC-b. The electric current was applied to heat the wires for 10 hours at 1200 °C. After assembly, the thermocouple was annealed in a furnace for 3 hours at 1100 °C , and finally 10 hours at 450 °C.Figures 13 and 14 show the results of the measurements of the drift and the inhomogeneity of these thermocouples, respectively. It can be seen that the changes of the drift and inhomogeneity differ by different lots even if the palladium wires were made by the same manufacturing procedure by the same manufacturer, as seen in TC-a and TC-b. On the other hand, TC-d showed small drift and inhomogeneity. This indicates that the drift and inhomogeneity can be greatly reduced by selecting an appropriate palladium wire lot. However, it is not easy to obtain a steady supply of the wire used in TC-d. Therefore, we investigated a method for reducing the drift and inhomogeneity using the wires that are relatively readily available, as used in TC-a, TC-b, and TC-c.Since the heat treatment in the furnace was the final procedure in the heating history of thermocouples before they were actually used, we thought this was most deeply involved in the property of the thermocouples. Therefore, after 3 hours heat treatment at 1100 °C in the furnace to remove the wire strain, we fabricated a total of 11 thermocouples with different final heat treatments by conducting the final heat treatment at temperature points in the range from 450 °C to 1080 °C. For the palladium wire of the thermocouple, the wires from the same lot as TC-a that showed the largest drift were used. Also, since the emf of the Pt/Pd thermocouple in Fig. 11 stabilized at 100 hours, we selected 100 hours as the length of the final heat treatment.Figure 15 shows the drift of the emf when the 11 Pt/Pd thermocouples fabricated as above were exposed to 1085 °C using the Cu fixed-point device, and Fig. 16 shows the inhomogeneity of the same thermocouples. The temperatures in the figure are final heat treatment temperatures. For reference, the result for the thermocouple annealed for 10 hours at 450 °C was plotted as black dots. In all thermocouples, the emf stabilized at approximately Fig. 13 Drift of the Pt/Pd thermocouples using palladium wires from different lots.Fig. 14 Inhomogeneity of the Pt/Pd thermocouples using palladium wires from different lots.Time / h00408012021-143Emf change / µV50 mKTC-cTC-dTC-bTC-aTime / h04080120-0.50.0-1.00.51.01.52.02.5(Emf8 cm-Emf0 cm) / µV50 mKTC-aTC-bTC-dTC-c

※このページを正しく表示するにはFlashPlayer9以上が必要です