First-principles study on lithium borohydrideLiBH4

Abstract

First-principles calculations have been performed on lithium borohydride $\mathrm{Li}\mathrm{B}{\mathrm{H}}_{4}$ using the ultrasoft pseudopotential method, which is a potential candidate for hydrogen storage materials due to its extremely large gravimetric capacity of $18\phantom{\rule{0.3em}{0ex}}\text{mass}\phantom{\rule{0.3em}{0ex}}%$ hydrogen. We focus on an orthorhombic phase observed at ambient conditions and predict its fundamental properties; the structural properties, electronic properties, dielectric properties, vibrational properties, and the heat of formation. The calculation gives a nearly ideal tetrahedral shape for $\mathrm{B}{\mathrm{H}}_{4}$ complexes, although the recent experiment suggests that their configuration is strongly distorted [J-Ph. Souli\'e et al., J. Alloys Compd. 346, 200 (2002)]. Analyses for the electronic structure and the Born effective charge tensors indicate that Li atoms are ionized as ${\mathrm{Li}}^{+}$ cations. The internal bonding of ${[\mathrm{B}{\mathrm{H}}_{4}]}^{\ensuremath{-}}$ anions is primarily covalent. The high-frequency dielectric permittivity tensor ${\ensuremath{\epsilon}}_{\ensuremath{\infty}}$ is predicted as almost isotropic, but the static dielectric permittivity tensor ${\ensuremath{\epsilon}}_{0}$ as considerably anisotropic. The $\mathrm{\ensuremath{\Gamma}}$-phonon eigenmodes can be classified into three groups, namely, the librational modes involving the displacements of ${\mathrm{Li}}^{+}$ cations (less than $500\phantom{\rule{0.3em}{0ex}}{\mathrm{cm}}^{\ensuremath{-}1}$), and the internal B-H bending and stretching modes of ${[\mathrm{B}{\mathrm{H}}_{4}]}^{\ensuremath{-}}$ anions (around 1100 and $2300\phantom{\rule{0.3em}{0ex}}{\mathrm{cm}}^{\ensuremath{-}1}$, respectively). The molecular approximation fairly reproduces the phonon frequencies in the latter two groups, implying the strong internal bonding of $\mathrm{B}{\mathrm{H}}_{4}$ complexes. The librational modes have significant contributions to the large anisotropies of ${\ensuremath{\epsilon}}_{0}$. The agreement of the heat of formation with the experimental value is reasonably good.