Understanding EMF Drift in High-Temperature Thermocouples

Type K drifts −2 to −5 °C in 500 h at 1000 °C; Type N holds under ±1 °C.

What is EMF Drift?

EMF drift is the gradual change in a thermocouple’s output voltage at a given temperature over time. It is the single largest source of measurement uncertainty in long-duration industrial temperature measurement.

A Type K thermocouple operating at 1000 °C in air will drift by approximately −2 to −5 °C after 500 hours. In reducing atmospheres, the drift can be −10 °C or more over the same period.

Four Mechanisms of Drift

1. Selective Oxidation (Green Rot — Type K)

The most well-known drift mechanism. At 800–1050 °C in low-oxygen atmospheres, chromium in the Chromel positive leg oxidises preferentially, forming Cr₂O₃. This depletes chromium from the alloy matrix, changing its Seebeck coefficient. The result is a negative EMF drift (the thermocouple reads progressively lower than the true temperature).

Prevention: Use Type N instead. The silicon in Nicrosil forms a protective SiO₂ layer that suppresses chromium oxidation. Type N drift at 1000 °C: < ±1 °C after 500 hours.

2. Contamination from the Environment

Sulphur, carbon, and metal vapours (zinc, lead) diffuse into the thermocouple wire at high temperature, forming intermetallic compounds that alter the Seebeck coefficient. This is particularly severe in petrochemical reformer tubes and heat treatment furnaces with carbon-rich atmospheres.

Prevention: Use mineral-insulated (MI) cable with an appropriate sheath material (SS316 or INCL600). The compacted MgO insulation acts as a diffusion barrier against contaminants.

3. Cold Working and Recrystallisation

Mechanical deformation during installation, or vibration in service, introduces dislocations into the metal lattice. As the wire heats up, these dislocations anneal out and the grain structure recrystallises, changing the Seebeck coefficient. This is most pronounced in Type K and least pronounced in Type N.

Prevention: Use fully annealed wire. For vibration-prone installations, use stranded wire (which accommodates flexing without work-hardening). Specify homogeneity testing per ASTM E1350 for critical applications.

4. Thermal Ageing (Grain Growth)

Prolonged exposure to high temperature causes grain growth in the metal. Larger grains change the electron scattering behaviour, which alters the Seebeck coefficient. This effect is slow (hundreds to thousands of hours) but irreversible.

Prevention: Select the appropriate alloy for the temperature. Type N was specifically designed to resist thermal ageing above 1000 °C. For temperatures above 1200 °C, consider noble-metal thermocouples (Type R, S or B).

Quantifying Drift in Your Application

If your process requires ±2 °C accuracy over 6 months of continuous operation at 1000 °C:

• Type K: Will exceed tolerance within 200–500 hours → Not suitable
• Type N: Will maintain tolerance for 2000+ hours → Suitable
• Type R (Pt/Pt-Rh): Will maintain tolerance for 5000+ hours → Overkill for most applications, 10× cost

Our Recommendation

Specify in this order:

1. Set the accuracy target over time — for example ±2 °C held for six months at 1000 °C.
2. Match the alloy to the mechanism — Type N where oxidation or thermal ageing governs; MI cable with an SS316 or Inconel 600 sheath where contamination governs.
3. Lock the batch — Type N, Class 1 tolerance, single furnace batch.
4. Verify homogeneity — request testing per ASTM E1350.

For long-duration, high-temperature processes (power generation turbines, petrochemical reformers, aerospace heat treatment), specify Type N, Class 1 tolerance, from a single furnace batch, with homogeneity testing per ASTM E1350. The incremental cost over Type K is repaid many times over by avoiding unplanned shutdowns for sensor replacement.

Concerned about EMF drift in your process? Send us your operating temperature, atmosphere and required accuracy — we recommend the optimal alloy and provide drift data from our test records.

Bottom line: Drift is the alloy reacting to its environment over time, not a defect that can be inspected out. Where ±2 °C must hold for months above 1000 °C, specify Type N in Class 1 tolerance from a single furnace batch with homogeneity testing per ASTM E1350.