When selecting a sensor, one should consider whether the environment in which it is to be used will cause a measurement error. As part of this series on thermocouples, this article discusses a few environmental effects that can alter the structure of the materials that make up a thermocouple and thereby affect their accuracy. This degradation of thermocouple accuracy is referred to as drift.
Effects of Temperature on Thermocouples
High temperatures can lead to thermocouple drift in a couple ways. High temperatures can cause metallurgical changes to the thermocouple materials through grain growth and alterations of the crystalline structure [1]. The effects of temperature-induced changes to the material can be amplified at the surfaces of thermocouple materials through oxidation or contamination with the environment [1]. These material changes tend to accumulate over time: the longer the exposure to the adverse environment, the greater the impact. These effects can be highly non-linear; for example, the drift effect in Chromel (used in Type K thermocouples) increases with temperature up to ~350-400°C and then decreases at higher temperatures [3].
Thermocouple drift can be significant with exposure to high temperatures, particularly if the wires are bare and exposed to air that can lead to oxidation. Since the properties of the two wires that make up the thermocouple are affected differently by temperature, drift isn’t generally linear with temperature. For example, one study showed that sheathed Type N thermocouples exhibited the largest drift (-10 to -13°C) when exposed to ~400°C, but the drift went back to nearly 0°C when they were subjected to temperatures around ~800°C. Exposure time also affected those results, which included test durations ranging from 300-1200 hours [4]. Another factor that affects drift is the type of sheath used on the thermocouples; for example, two Type K thermocouples with different sheathing exhibited drifts of approximately -20°C and +5°C after 1000 hours at 1200°C [5].
The impact of temperature driven thermocouple drift varies significantly with the thermocouple type. Thermocouples that include nickel alloys (Types K and N) are particularly susceptible, however studies have been conducted to characterize the drift of almost all types. One exception appears to be Type T thermocouples, presumably because their temperature range (Tmax = 370°C) is low enough that they are unlikely to be subjected to sufficiently high temperatures to produce the metallurgical changes that cause substantial drift (typically in the range of 500-1000°C).
Effects of Magnetic Fields on Thermocouples
When thermocouples are used in an environment with magnetic flux, users should be aware of how the magnetic fields can interact with the materials that make up the thermocouple. The impact of magnetic fields can include permanent changes to the material properties of the wire as well as temporary effects, such as induction heating of the thermocouple resulting from the magnetic field. Unsurprisingly, thermocouples that include a ferrous material, such as Type J thermocouples that are comprised of iron and constantan, tend to be more affected than thermocouple types that don’t include iron [6]. Dynamic (oscillating) magnetic fields can affect thermocouples, including non-ferrous types, presumably due to voltage generated by the magnetic interactions with wires. One study found that a high magnetic field induced up to ~7°C measurement error on a Type T thermocouple. This error was substantially reduced to less than 2°C when the test was repeated with a twisted wire thermocouple [7].
Effects of Radiation on Thermocouples
The accuracy of thermocouples used in nuclear reactors will change over time due to neutrons causing atomic displacement and transmutation of the materials that make up the thermocouples. Reported errors in thermocouples can be up to ~15% of the absolute temperature, particularly type W (tungsten/rhenium) and platinum-based devices, while nickel-based thermocouples, such as N type, are less affected by radiation [8].
Discussion
Clearly, thermocouples can drift when exposed to adverse environments like extreme temperature, magnetic fields, or radiation. The drift is due to changes to the thermocouple wires, not just the junction where they meet. Therefore, one method for reducing these effects is to minimize the amount of wire exposed to them. If a thermocouple is to be used in a high temperature chamber, as much of the thermocouple wire as possible should be routed outside the chamber to reduce the length of wire that sees the high temperatures.
From an electronics cooling perspective, it would be rare for thermocouples to be exposed to temperatures high enough to cause drift, simply because the electronics would fail before the thermocouple would exhibit substantial drift. Similarly, most electronics are not exposed to sufficiently high magnetic flux or irradiation to create a significant issue. Unusual situations can happen though, so it is best to at least be aware of whether a specific environment could introduce measurement error. If thermocouples are used in a potentially damaging environment, such as high temperatures, it may be necessary to replace or recalibrate them on a regular basis to ensure that they are accurate.
References
[1] Michele Scervini, “Drift: A Short Explanation”, last update August 31, 2009, https://www.msm.cam.ac.uk/utc/thermocouple/ pages/Drift.html
[2] Robert Torgerson, “What to Know about Aging and Drift in Type K Thermocouples”, https://blog.wika.us/knowhow/ aging-and-drift-in-type-k-thermocouples/?doing_wp_cron=1731072533.6360089778900146484375#:~:text=Drift%20is%20 generally%20a
[3] Michele Scervini, “Type K Thermocouple: Bare Wire Configuration”, https://www.msm.cam.ac.uk/utc/thermocouple/pages/ DriftInTypeKBareWiresThermocouples.html
[4] Abdelaziz, Yasser et al., “Characterizing Drift Behavior in Type K and N Thermocouples After High Temperature Thermal Exposures”. Journal of Advanced Research in Fluid Mechanics and Thermal Sciences. 97. 62-74 (2022).
[5] Michele Scervini, “Type K Thermocouple: MIMS Configuration”, Drift in Type K MIMS thermocouples
[6] Samo Beguš, et al., “Magnetic effects on thermocouples”, Measurement Science and Technology, (25) 2014, DOI: 10.1088/0957- 0233/25/3/035006
[7] Shir, F., Mavriplis, C., & Bennett, L. H. (2005). “Effect of Magnetic Field Dynamics on the Copper‐Constantan Thermocouple Performance”. Instrumentation Science & Technology, 33(6), 661–671. https://doi.org/10.1080/10739140500311239
[8] M. Scervini and C. Rae, “Low drift type N thermocouples for nuclear applications,” 2013 3rd International Conference on Advancements in Nuclear Instrumentation, Measurement Methods and their Applications (ANIMMA), Marseille, France, 2013, pp. 1-7, doi: 10.1109/ANIMMA.2013.6727899.
