Theoretical prediction of decomposition paths for Ca(BH$_4$)$_2$ and Mg(BH$_4$)$_2$

Experimental and theoretical studies on Ca(BH$_4$)$_2$ indicate that the decomposition pathway of this compound is not a simple one-step reaction to final products, but instead may be a multi-step decomposition path (though the reaction pathway is currently unknown). We have studied the decomposition pathways of both Ca- and Mg-borohydride using Density Functional Theory (DFT) as well as a Monte Carlo-based crystal structure prediction method called PEGS. We find that a recently experimentally proposed CaB$_2$H$_2$ intermediate is extremely high in energy and hence very unlikely to form. We then symmetrically studied the low-energy structures of CaB$_2$H$_4$ and CaB$_2$H$_6$ stoichiometries by PEGS+DFT simulations. Based on all these possible reaction paths, the Ca(BH$_4$)$_2$ decomposition convex hull using our calculated reaction enthalpies finds a new compound, CaB$_2$H$_6$, which forms a nearly degenerate pathway to the previously-proposed CaB$_{12}$H$_{12}$ phase. Similar calculations for the Mg system show that the MgB$_2$H$_6$ predicted structure does not form a stable intermediate in the decomposition of Mg(BH$_4$)$_2$.