Seebeck Coefficient Metrology: Do Contemporary Protocols Measure Up?

Abstract

Comparative measurements of the Seebeck coefficient are challenging due to the diversity of instrumentation and measurement protocols. With the implementation of standardized measurement protocols and the use of Standard Reference Materials (SRMs^®), for example, the recently certified National Institute of Standards and Technology (NIST) SRM^® 3451 ‘‘Low Temperature Seebeck Coefficient Standard (10–390 K)’’, researchers can reliably analyze and compare data, both intra- and inter-laboratory, thereby accelerating the development of more efficient thermoelectric materials and devices. We present a comparative overview of commonly adopted Seebeck coefficient measurement practices. First, we examine the influence of asynchronous temporal and spatial measurement of electric potential and temperature. Temporal asynchronicity introduces error in the absolute Seebeck coefficient of the order of ≈10%, whereas spatial asynchronicity introduces error of the order of a few percent. Second, we examine the influence of poor thermal contact between the measurement probes and the sample. This is especially critical at high temperature, wherein the prevalent mode of measuring surface temperature is facilitated by pressure contact. Each topic will include the comparison of data measured using different measurement techniques and using different probe arrangements. We demonstrate that the probe arrangement is the primary limit to high accuracy, wherein the Seebeck coefficients measured by the 2-probe arrangement and those measured by the 4-probe arrangement diverge with the increase in temperature, approaching ≈14% at 900 K. Using these analyses, we provide recommended measurement protocols to guide members of the thermoelectric materials community in performing more accurate measurements and in evaluating more comprehensive uncertainty limits.