Editing Precision Thermocouple Based Temperature Measurement System (section)

== 6. Conclusion ==

This study set out to measure the Seebeck coefficient of an undoped ZnO pellet at room temperature using a Keysight B2901A Source Measure Unit with silver paste contacts. Four independent measurement runs were carried out, each covering a temperature difference range of 1°C to 14°C, and the Seebeck coefficient was extracted from the slope of the Vout versus ΔT relationship in each case.

The clearest finding across all four runs was the consistently negative output voltage — every reading, at every temperature difference, in every run, came out negative. That alone is a meaningful result. It confirms that the undoped ZnO pellet behaves as an n-type semiconductor, with electrons as the dominant charge carriers — behaviour that is well documented in the literature and understood to arise from native point defects, chiefly oxygen vacancies and zinc interstitials, that introduce free electrons into the ZnO lattice (Özgür et al., 2005; Look, 2001). Averaging the slopes extracted from the four linear fits gave a final Seebeck coefficient of S = −0.934 ± 0.094 μV/K, where the spread in values across the runs reflects primarily the variability introduced by reapplying the silver paste contacts before each run.

This value is considerably smaller in magnitude than what the literature reports for undoped ZnO pellets, where values of −350 to −430 μV/K are typical. That gap is large, but it has a clear explanation — and more than one factor feeds into it. The thermocouples were not placed directly on the pellet faces, which means the recorded ΔT was larger than the real temperature difference across the pellet, pulling the extracted S downward. The connecting wires contribute their own thermoelectric voltages to the circuit, and without correcting for these, the measured slope cannot be taken as the pellet's Seebeck coefficient alone. The narrow temperature differences used in some runs pushed the signal into the sub-microvolt range, where noise and drift become competitive. And the pellet itself sintered in ambient air, carrying high resistivity from grain boundary Schottky barriers — limited how much of the thermoelectric voltage could be extracted at the terminals (Rawat & Paul, 2016; Goldsmid, 2010; Rowe, 2006; Snyder & Toberer, 2008). Crucially, the literature values of −350 to −430 μV/K are not room-temperature figures — they come from measurements at 600 K to 1273 K under inert atmospheres, which is a fundamentally different regime from what was used here. The comparison is therefore not a straightforward one.

A reproducible anomaly was also identified beyond ΔT = 14°C, where the output voltage reversed its upward trend rather than continuing to climb. This happened consistently every time the measurement was repeated past that threshold, pointing to a genuine physical change in the pellet's response rather than a measurement glitch. Grain boundary Schottky barrier breakdown and resistive self-heating within the pellet are the most likely causes, both of which are characteristic of polycrystalline undoped ZnO sintered in air at elevated thermal gradients (Özgür et al., 2005; Rowe, 2006). All primary measurements were kept within ΔT ≤ 14°C on this basis.

Overall, the study demonstrates that the undoped ZnO pellet exhibits thermoelectric behaviour that is physically consistent and interpretable. The n-type character of the material was confirmed unambiguously, a Seebeck coefficient was successfully extracted within a well-defined linear measurement window, and the limitations of the two-probe room-temperature setup were identified and accounted for. The results, while not matching literature values in magnitude, are defensible and informative within the context of the experimental conditions employed.