Trials of Pt-20%Rh versus Pt thermocouples between 157 °C and 962 °C

Abstract

For over a century the Type S (Pt-10%Rh) and R (Pt-13%Rh) thermocouples have been used as primary or secondary reference thermometers. However, the measurement uncertainties of both types at the silver point (~962 °C) are typically limited to about 0.5 °C, due to drift caused by crystallographic ordering between 200 °C and 500 °C and rhodium oxidation between 500 °C and 900 °C. Although both processes can be reversed using an 1100 °C anneal, regular annealing can be inconvenient or impracticable. This paper follows up a prior study indicating that an alternative noble-metal thermocouple comprised of Pt-20%Rh and Pt is relatively insensitive to both drift mechanisms. Four thermocouples, assembled using wire from four different manufacturers, were evaluated using a gradient-furnace and homogeneity scanner. Measurements were also made using a salt-bath (200 °C to 500 °C) and fixed points between the indium and silver points. The experiments indicate that individual thermocouples are stable at the silver point to within 0.18 °C for periods of up to 100 h and all four thermocouples have emf versus temperature characteristics within 0.3 °C of each other. This intrinsic stability and similarity, coupled with cost, assembly and use conditions identical to a Type R or S thermocouple make the Pt-20%Rh versus Pt thermocouple an attractive alternative. Recommended annealing procedures enabling the greatest stability are also given.

A critical review of the common thermocouple reference functions

“…The aim is to produce optimised and highly stable reference thermocouples with new reference functions, which are not constrained by historical measurements and a limited knowledge of the causes of drift. The uncertainties for a calibrated Pt-20% Rh vs Pt are close to that of a Pt-Pd thermocouple (∼0.1 • C) at temperatures less than 1100 • C [75]. Other benefits of the Pt-20% Rh vs Pt thermocouple are a near identical cost, construction and annealing treatment to that of the traditional types R and S.…”

“…Currently, the only practicable way to achieve reliable accuracy for base-metal thermocouples (close to that given by LoE and class specification) is through regular annealing and calibration [75]. A similar argument holds for noble-metal thermocouples.…”

“…In the last 20 years a significant advance has been made in understanding how noble-metal types B, R and S vary from the reference function, as-supplied, and after exposure to various temperatures [70][71][72][73]. From this knowledge new Pt-Rh alloy pairs [74], able to replace type B, and a new Pt-20% Rh vs Pt reference thermocouple, able to replace types R and S, are being developed [43,75,76]. The aim is to produce optimised and highly stable reference thermocouples with new reference functions, which are not constrained by historical measurements and a limited knowledge of the causes of drift.…”

“…Webster E 2020 Base-metal thermocouple tolerances and their utility in real-world measurements Meas.Sci. Technol.…”