High-throughput screening for efficient two-dimensional thermoelectric materials

A high-throughput ab-initio framework is used to identify two-dimensional materials suitable for thermoelectric applications. Out of 290 experimentally exfoliable materials, a total of 150 monolayers are selected through band gap screening. For these, we compute the thermoelectric transport coefficients and power factor scaled by the relaxation time, S 2 σ / τ , finding that most of them exceed 5 × 10 11 W/mKs. Further screening is then applied by using the average atomic weight criterion. It is found that the p -type monolayers with high S 2 σ / τ are mostly tetradymites, such as Sb 2 SeTe 2 , Bi 2 STe 2 , and Bi 2 SeTe 2 , except for the less known Sb 2 SnTe 4 . In contrast, n -type monolayers with high S 2 σ / τ are found over a wider range of structures, although typically their S 2 σ / τ is lower than that of the p -type monolayers. The five best-performing p -type and n -type monolayers are selected for studying their electronic and phonon transport properties. The electronic relaxation time is computed from the electron-phonon coupling matrix, while the lattice thermal conductivity is calculated from the ab-initio force constants up to the third order. Sb 2 SeTe 2 is predicted to provide a high power factor of 26 mW/mK 2 , while GeTe and Sb 2 SnTe 4 show at 300 K ultra low lattice thermal conductivities of 0.34 and 0.46 W/mK, respectively. Overall the conducted large-scale screening exercise allows us to identify both p -type and n -type monolayers with excellent thermoelectric performance in terms of the figure of merit.