Potential energy contribution to the thermopower of correlated electrons

Abstract

Certain classes of strongly correlated systems promise high thermopower efficiency, but a full understanding of correlation effects on the Seebeck coefficient is lacking. This is partly due to limitations of Boltzmanntype approaches. One needs a formula for the thermopower that allows separate investigations of the kinetic and potential energy contributions to the evolution with temperature and doping of the thermopower. Here we address this issue by deriving for Hubbard-like interactions a formula for the thermopower that separates the potential from the kinetic energy contribution and facilitates a better understanding of correlation effects on the Seebeck coefficient. As an example, the thermopower of the one-band Hubbard model is calculated from dynamical mean-field. For interactions in both the intermediate and strong correlation limit, the contributions from kinetic and potential energy nearly cancel.

Magnon heat transport in a two-dimensional Mott insulator

“…For convenience we use κ for the DC limit κ(ω = 0) throughout the paper. To understand the temperature dependence, we separate out the kinetic (K) and potential contributions (P) to c v and κ [47,48] in Fig. 2(b) and (d), by splitting the total energy H and defining the hopping energy as the kinetic energy H K and the electron-electron interaction as the potential energy H P .…”