Measurement of the Inhomogeneity in Type B and Land–Jewell Noble-Metal Thermocouples

Abstract

Inhomogeneity is the largest contributor to uncertainty in temperature measurements made with thermocouples, and the knowledge of inhomogeneity is essential if low-uncertainty measurements are required. Inhomogeneity is a particular problem for long-term applications at temperatures near or above 1500 $^{\circ }\hbox {C}$, where pairs of alloyed noble-metal thermocouples must be used and the alloy components and potential contaminants become very mobile and cause large deviations in the Seebeck coefficient. While changes in inhomogeneity are a known and well-studied problem in noble-metal alloys at temperatures below 1100 $^{\circ }\hbox {C}$, the effects are not well quantified at higher temperatures. This paper reports the first detailed measurements of inhomogeneity in a number of Type B and Land–Jewell thermocouples exposed to either short-term calibration up to 1600 $^{\circ }\hbox {C}$ or long-term in situ measurements for a period of approximately 3000 h at 1600 $^{\circ }\hbox {C}$. The inhomogeneity is measured in a high-resolution scanner operating over the range from 600 $^{\circ }\hbox {C}$ to 900 $^{\circ }\hbox {C}$. The results show that drifts of between 0.2 % and 0.6 % can be expected for reversible crystallographic and oxidation effects, whereas drift caused by irreversible contamination effects can be expected to be between 0.6 % and 1.1 %. It is also shown that the deviations in emfs caused by irreversible homogeneities in these thermocouples scale approximately linearly with temperature. This scalability allows uncertainties assessed at one temperature, to be extrapolated to other temperatures. Additionally it is shown that a preconditioning anneal at 1100 $^{\circ }\hbox {C}$ should be applied both before and after calibration to remove undesirable crystallographic and rhodium-oxidation effects.