Editing Precision Thermocouple Based Temperature Measurement System (section)

=== 5.3. Physical Interpretation of the Anomalous Behaviour Beyond ΔT = 14°C ===

The reproducible reversal in the Vout versus ΔT trend beyond 14°C is worth examining carefully. Up to ΔT = 14°C, the voltage climbs steadily toward less negative values in the manner expected of a well-behaved thermoelectric response. At ΔT = 16°C, it drops back down instead of continuing upward, and this happened the same way every time the measurement was repeated. That rules out a contact issue or an instrument fluctuation. Something in the pellet's response changes once the temperature difference crosses that threshold.

The most straightforward explanation is grain boundary Schottky barrier breakdown. Within the linear regime, the electrostatic barriers at ZnO grain boundaries are stable, and the thermoelectric voltage builds up coherently along the pellet. As ΔT climbs beyond 14°C, the electric field across individual grains grows to the point where it begins to overwhelm those barriers. Carriers start crossing grain boundaries through tunnelling or avalanche-type mechanisms, which disrupts the orderly voltage accumulation and pulls the net measured output back down (Özgür et al., 2005; Rowe, 2006). This is closely related to the varistor-like switching behaviour that ZnO grain boundaries are well known for, where each active grain boundary interface sustains a charge barrier before breakdown occurs.

Resistive self-heating adds to this. At higher heater input voltages, more power is dissipated within the pellet, creating an internal temperature gradient that partially works against the externally applied ΔT. The net thermal driving force is reduced, and with it the thermoelectric voltage. In undoped ZnO, which already carries high resistivity, this effect is more significant than it would be in a more conductive sample, and it tends to concentrate at the silver paste contacts where current density peaks (Goldsmid, 2010). Between these two mechanisms: barrier breakdown and self-heating, the reversal at ΔT = 16°C has a clear physical basis, and restricting the primary measurements to ΔT ≤ 14°C was the appropriate response to it.