The Platinum Resistance Thermometer Range of the International Temperature Scale of 1990

Abstract

Within the International Temperature Scale of 1990 (ITS-90) the platinum resistance thermometer (PRT) is used to realize the scale from approximately 13,8 K to 1 235 K. Such a temperature range is wider than the corresponding range in the International Practical Temperature Scale of 1968 (IPTS-68) because the PRT is used up to the freezing point of silver (1 234,93 K). In this way, the ITS-90 can be realized with much more precision than the IPTS-68, particularly between 901 K and 1 235 K where the standard Pt-10% Rh vs Pt thermocouple was previously used.This paper describes some important steps in the construction of the PRT reference function and the criteria for the selection of the PRT interpolating equations of the ITS-90. In contrast to previous international scales, the PRT range of the ITS-90 is based on two reference functions, one from 13,8 K to 273,16 K and the other from 273,15 K to 1 234,93 K. The two reference functions were obtained from two real PRTs.A set of interpolating equations is used to account for the differences of other real PRTs from those on which the reference functions are based. They provide flexibility of calibration and high precision as expressed, for instance, by non-uniqueness and sub-range inconsistency not exceeding 0,5 mK over the range from 13,8 K to 693 K. Such good properties are the consequence of the choice of suitable forms for the interpolating equations and of fine adjustments in the values assigned to some defining fixed points.

An evolutionary approach to updating the international temperature scale

“…As for the matter of the interpolation, Working Group 3 (WG3) of the CCT published "The Platinum Resistance Thermometer Range of the International Temperature Scale of 1990" [26] in 1991 to explain the rationale and process behind the construction of the interpolation function. ITS-90 partitions the PRT range into a low-temperature portion (13.81 K T 90 273.16 K) and a high-temperature range (273.15 K T 90 1234.93K).…”

“…In this way, the reference function accurately expresses the relation of the resistance of a representative PRT to the thermodynamic temperature." [26] Ultimately, the reference functions were based on two real PRTs, a 25 : The construction of the high-temperature reference function was described in detail by Jung [27]. Unfortunately, this publication is not widely available so I will elaborate the key details.…”

The Discontinuity in the First Derivative of the ITS-90 at the Triple Point of Water

“…The process was an iterative one, continuing into the 1989 meeting of the CCT which finalized the scale, and is discussed by Crovini et al [2]. In particular, an adjustment of −0.4 mK (a minor amount in the context of the thermodynamic uncertainty) was made to the assigned temperature of the triple point of mercury to ensure that the sub-range to the argon point is consistent with subranges extending to lower temperatures.…”

“…In particular, an adjustment of −0.4 mK (a minor amount in the context of the thermodynamic uncertainty) was made to the assigned temperature of the triple point of mercury to ensure that the sub-range to the argon point is consistent with subranges extending to lower temperatures. Figure 1.7 of the Supplementary Information for the ITS-90 [3], shows that an inconsistency or error in the mercury point exerts considerable leverage at temperatures down to 84 K. Similar consistency tests were carried out for SPRT sub-ranges above T TPW , and minor adjustments were made to some of the assigned temperatures [2].…”

“…The ITS-90 reference and deviation functions for representing the calibrations of standard platinum resistance thermometers (SPRTs) are analytic and continuous, except at the junction between the two main ranges at the triple point of water (T TPW , 273.16 K), and also at the freezing point of aluminum, 660.323 • C. At T TPW the reference functions are forced to be continuous as far as the second derivative [1,2], but the mathematics do not include any constraints on even the first derivative of the deviation functions at that point. At 660.323 • C the first derivative is continuous, but the second is not.…”