Learning electron scattering parameters by analyzing thermoelectric data

The theoretical prediction of thermoelectric (TE) transport properties has made systematic use of the constant relaxation time approximation (CRTA), thus ignoring effects associated with the specific scattering mechanisms at work. The idea of improving the design of TE materials through scattering engineering is very powerful. Here we investigate the role of the CRTA in the electronic properties of well-known TE materials, by considering scattering models with energy and temperature dependence. The introduction of a self-consistent fitting procedure allows the existing models to be finely tuned for the individual experimental sample and, the contribution of each mechanism to be singled out without relying on experimental parameters. Transport coefficients, scattering models, and the fitting procedure have been computed with the recently developed PAOFLOW software (http://aflowlib.org/src/paoflow/).