Comprehensive first-principles analysis of phonon thermal conductivity and electron-phonon coupling in different metals

Abstract

Separating electron and phonon components in thermal conductivity is imperative for understanding thermal transport in metals and highly desirable in many applications. In this work, we predict the mode-dependent electron and phonon thermal conductivities of 18 different metals at room temperature from first principles. Our first-principles predictions, in general, agree well with those available experimental data. For phonon thermal conductivity, we find that it is in the range of 2--18 W/mK, which accounts for 1%--40% of the total thermal conductivity. It is also found that the phonon thermal conductivities in transition metals and transition-intermetallic compounds (TICs) are non-negligible compared to noble metals due to the high phonon group velocities of the former. We further show that the electron-phonon coupling effect on phonon thermal conductivity in transition metals and intermetallic compounds is stronger than that of nobles, which is attributed to the larger electron-phonon coupling constant with a high electron density of states within the Fermi window and high phonon frequency in the former. For electron thermal conductivity, we observe that the transition metals and TICs have lower electron thermal conductivities compared to noble metals, which is mainly due to the weak electron-phonon coupling in noble metals. It is found that the Lorenz number of transition metals and TICs hold larger deviations from the Sommerfeld value ${L}_{0}=2.44\ifmmode\times\else\texttimes\fi{}{10}^{\ensuremath{-}8}\phantom{\rule{0.16em}{0ex}}\mathrm{W}\phantom{\rule{0.16em}{0ex}}\mathrm{\ensuremath{\Omega}}\phantom{\rule{0.16em}{0ex}}{\mathrm{K}}^{\ensuremath{-}2}$. We also find the mean free paths extracted at 50% accumulation function for phonons (within 10 nm) are generally smaller than those of electrons (5--25 nm).