Realization of ITS-90 from 273.15 K through 1234.93 K: One Company’s Approach
Xumo Li,Steve Iman,M. Hirst,Mingjian Zhao
2003-01-01
Abstract:All of the fixed points of the International Temperature Scale of 1990 (ITS-90) in the range from 273.15 K to 1234.93 K have been realized in specially-designed, permanently-sealed cells. The purity of all of the metals used was 99.9999% or greater. Many improvements in design and techniques used were made in order to get the highest possible accuracy. Since it is important that all of the fixed points be traceable to NIST (National Institute of Standards and Technology), these fixed points were compared with NIST data by using two NIST-calibrated standard platinum resistance thermometers (SPRTs). The differences between the fixed point values and those measured at NIST were within 0.5 mK for the freezing points (FP) of tin and zinc, and within 1 mK for the FP of aluminum. The expanded uncertainties (coverage factor K=2) at the fixed points were 0.4 mK for the melting point (MP) of gallium, 1.0 mK for the FP of indium, tin, and of zinc, 4.0 mK for the FP of aluminum and 10 mK for the FP of silver. Two SPRTs were used to interpolate between the fixed points. A 25-ohm SPRT was used up to 933.473 K and a 0.25-ohm SPRT up to 1234.93 K. Introduction The International Temperature Scale of 1990 (ITS-90) became the official international temperature scale on January 1, 1990, superseding the International Practical Temperature Scale of 1968 (IPTS-68). During the last couple of years the efforts of our laboratory have been concentrated on establishing the new international temperature scale between 273.15 K and 1234.93 K. In order to realize all of the defining fixed points through this temperature range it was necessary to develop and manufacture high-quality furnaces and fixed-point cells including the triple point of water, the melting point of gallium, and the freezing points of indium, tin, zinc, aluminum, and silver. These fixed-point cells are of the sealed cell design to aid in the ease of use and convenience of intercomparison. In addition, two highly-stable SPRTs of the authors’ design were calibrated by NIST and were used as the standard interpolation instruments in realizing the ITS-90 temperature scale. Standard Platinum Resistance Thermometer According to the ITS-90, the temperature between 13.8033 K and 961.78°C is defined by means of a platinum resistance thermometer calibrated at specified sets of defining fixed points. The uncertainties of the realization and dissemination of the ITS-90, therefore, depend closely on the stability of an SPRT. Special attention was paid in developing, testing and improving SPRTs. Three types of SPRTs have been developed and tested. The main specifications are shown in Table 1. The detailed test data are listed in Table 2 and Table 3. A 25-ohm SPRT (S/N 92017) and a 0.25-ohm SPRT (S/N 92256) were calibrated by NIST and used throughout these tests to compare our fixed-point data with NIST. The data given in the NIST Reports of Calibration for these SPRTs are listed in Table 4. The resistance ratios W(t) at the exact temperatures of the freezing points of indium, tin, zinc, aluminum and silver were calculated by us Table 1: Main specifications of three types of SPRTs. Model 5681 5684 5685 Temperature Range –189°C to 661°C 0°C to 1070°C 0°C to 1070°C† Nominal Resistance at T. P. Water 25Ω 0.25Ω 2.5Ω Sensor Support Quartz cross Quartz strip with notches Quartz cross Diameter of Sensor Pt wire 0.07 mm 0.4 mm 0.2 mm Protecting Sheath Quartz Diameter: 7 mm Length: 520 mm Quartz Diameter: 7 mm Length: 680 mm Quartz Diameter: 7 mm Length: 680 mm Typical stability of Rtp 1 mK/100h at 661°C 1 mK/100h at 1085°C 1 mK/100h at 1085°C The official maximum temperature of an SPRT as a defining interpolation instrument of the ITS-90 is 961.78°C, but these types of SPRTs were found to be stable up to at least 1070°C. The annealing temperature during the stability test was 1085°C. The lower temperature limit of these types of SPRTs can be as low as –189°C. In general, it is suggested that a 25-ohm SPRT be used in the range below 0°C. Table 2: Results of stability tests with 2.5-ohm SPRTs. Rtp (ohms) Total Annealing time at 1085°C (hours) Notes S/N 2001 S/N 2002 S/N 2004 S/N 2006 2.475910 2.463342 2.469168 2.458491 25 2.475914 2.463354 2.469171 2.458499 50 2.475905 2.463357 2.469168 2.458504 75 2.475910 2.463355 2.469160 2.458514 After 1 thermal cycle† 2.475917 2.469162 2.458523 After 1 thermal cycle† Thermal cycle procedure: Put tested SPRT into a furnace at 962°C. Take the SPRT out of the furnace after twenty minutes and cool it in the air. Insert the SPRT into the furnace again, and after thirty minutes decrease the temperature of the furnace from 962°C to 480°C at a rate of 100°C per hour. Finally take the SPRT out of the furnace, cool it in the air and measure its Rtp. according to the coefficients given in the NIST Reports. These calculated values are listed in Table 4 as well. Defining Fixed Points According to the ITS-90, the calibration of an SPRT from 273.16 K to 1234.93 K requires the use of defined fixed points. Actually, the fixed points are the most important part of the ITS-90. All of the defining fixed points in this range were realized in specially-designed, permanently-sealed cells manufactured at Hart Scientific. Table 3: Results of stability tests with 25-ohm SPRTs. Rtp (ohms) Total Annealing time at 660°C (hours) Notes S/N 1008 S/N 1009 S/N 1010 S/N 1013 25.57653 25.70259 25.57474 25.53706 25 25.57641 25.70255 25.57452 25.53702 50 25.57639 25.70248 25.57450 25.53696 75 25.57644 25.70247 25.57449 100 25.57641 25.70256 25.57443 25.53962 After one thermal cycle† 25.57641 25.70258 25.57443 25.53942 After one thermal cycle† Thermal cycle procedure: Put a tested SPRT into a furnace at 660°C. Take the SPRT out of the furnace after twenty minutes and cool it in the air. Insert the SPRT into the furnace again, and after thirty minutes decrease the temperature of the furnace from 660°C to 480°C at a rate of 100°C per hour. Finally take the SPRT out of the furnace, cool it in the air and measure its Rtp. Table 4: The NIST calibration data of two SPRTs used for comparison. SPRT S/N: 92017 92256 NIST Test No: 253699 255156 Date of Report 3 May 1994 13 April 1995 For zero-power For 1.0 mA For zero-power For 14.14 mA a7 or a6 –1.5526301E-04 –1.5988934E-04 3.6806657E-05 3.2513557E-05 b7 or b6 –1.5007187E-05 –1.7190604E-05 7.1081548E-05 6.1878236E-06 c7 or c6 2.2808655E-06 2.8945480E-06 –2.1652828E-06 –1.8774523E-06 d 5.1265611E-06 –2.6454136E-06 W(29.7646 °C) 1.11812035 1.11811977 W(156.5985 °C) 1.60970212 1.60969863 W(231.928 °C) 1.89264875 1.89264332 1.89283467 1.89283031 W(419.527 °C) 2.56864562 2.56863636 2.56898418 2.56897629 W(660.323 °C) 3.37558565 3.37557056 3.37610714 3.37609560 W(961.78 °C) 4.28654566 4.28652538 Fixed-Point Cell The construction of all metal fixed-point cells is basically the same. Some of the materials used in construction are different, however. The particular metals used for each of the different type of cells are all 99.9999% pure or better. The gallium cell differs slightly in that a PTFE (Teflon) crucible is used instead of graphite. All cells are constructed using a quartz shell and reentrant well. Each cell is connected to a vacuum system after assembling. The cell is then evacuated at a selected temperature for more than 100 hours and during that period the cell is repeatedly purged with 99.999% pure argon. Once it has been determined that the cell is free from contaminants, it is filled with pure argon and sealed at a known atmospheric pressure exactly at the temperature of the fixed point. The sealed cell solves most problems associated with open cell designs. Information on each cell is outlined in Table 5. The actual temperature of the sensor of an SPRT calibrated at a fixed point might be different from the value assigned to the fixed point by the ITS-90 because of small pressure differences. Realization of the Fixed Points Realization of a fixed point to a high level of accuracy requires a strict envelopment, the temperatures of which must be very uniform, stable and controlled. The equipment used to realize these fixed points is listed in Table 6. The three-zone furnace used for the freezing points of indium, tin, zinc and aluminum is shown in Figure 1 and the heat-pipe furnace for the freezing point of silver and aluminum is shown in Figure 2. A typical axial temperature distribution in the furnace at the freezing point of zinc is shown in Figure 3. All resistance measurements were made using a Guildline DC automatic bridge, model 6675, S/N 60956. The resistance of the 25-ohm SPRT was measured using DC currents of 1 mA and 1.41 mA. The resistance of the 0.25-ohm SPRT was measured using DC currents of 14.14 mA and 20.0 mA. These currents allowed the measurement results to be extrapolated to values at zero power. The reTable 5: Fixed point cell Thermometric Fixed Point Purity of Material, % Quantity, kg Immersion Depth, ‡cm H2O TP † 0.76 23 Ga MP 99.99999 0.45 15 In FP 99.9999+ 0.97 18 Sn FP 99.9999+ 0.96 18 Zn FP 99.9999+ 0.95 18 Al FP 99.9999+ 0.35 18 Ag FP 99.9999+ 1.35 18 Water from a purification system, with a resistivity of about 18 megohm-cm, was distilled into the cell. Immersion depth is calculated with the SPRT in its fully-immersed position in the cell. The distance from the midpoint of the sensor of the SPRT to the tip of the SPRT quartz sheath is about 1.5 cm. The actual distance from the upper surface of the metal to the bottom of the central well is 19.5 cm for the metal freezing points, and 16.5 cm for the melting point of gallium. sistance at the triple point of water was measured immediately after each measurement at the melting or freezing points described in this paper. The triple point of water is the most important fixed point of the ITS-90. The stability of all SPRTs should be checked at the triple point of water periodically. According to the ITS-90, the temperatures are interpolated in terms of the ratio of the resistanc