Simultaneous structure and carrier tuning of dimorphic clathrateBa8Ga16Sn30

Abstract

We report structural, transport, and thermal properties of carrier-tuned ${\text{Ba}}_{8}{\text{Ga}}_{16}{\text{Sn}}_{30}$ single crystals with the type-1 clathrate structure ($\ensuremath{\beta}$ phase), demonstrating that ${\text{Ba}}_{8}{\text{Ga}}_{16}{\text{Sn}}_{30}$ is a unique thermoelectric clathrate material wherein both the structure type and the carrier type are tunable. The results are compared with the properties of the better known type-8 structure ($\ensuremath{\alpha}$ phase) and of ${A}_{8}{\text{Ga}}_{16}{\text{Ge}}_{30}$ $(A=\text{Sr},\text{Eu})$. Differential thermal analysis and powder x-ray diffraction show that both phases are stable up to their virtually identical melting point of $520\ifmmode\pm\else\textpm\fi{}3\text{ }\ifmmode^\circ\else\textdegree\fi{}\text{C}$. Refinements of single-crystal x-ray diffraction data indicate that the Ba(2) guest ion in the tetrakaidecahedron occupies the off-center $24k$ sites which are $0.43--0.44\text{ }\text{\AA{}}$ away from the centered $6d$ site. The temperature-linear coefficient of the specific heat is $29\text{ }\text{mJ}/\text{mol}\text{ }{\text{K}}^{2}$ for both $n$- and $p$-type carriers in the $\ensuremath{\beta}$ phase, four times larger than that for the $\ensuremath{\alpha}$ phase, suggesting contributions from tunneling of the Ba(2) guest ions between off-center minima. Analysis of specific heat with a soft-potential model (SPM) gives a characteristic energy of 20 K for the Ba(2) vibration, significantly lower than 50 K for the $\ensuremath{\alpha}$ phase and in fact the lowest among type-1 clathrates. The lattice thermal conductivities ${\ensuremath{\kappa}}_{L}$ for the $\ensuremath{\beta}$ phase with both charge carrier types are very similar and show a glasslike temperature dependence. This behavior in ${\ensuremath{\kappa}}_{L}(T)$ is also described using SPM analysis, where it is found that the coupling strength between guest modes and acoustic phonons for the $\ensuremath{\beta}$ phase is significantly larger than that for ${\text{Sr}}_{8}{\text{Ga}}_{16}{\text{Ge}}_{30}$.