Computationally driven discovery of new borides in the ternary Li-Ni-B system

The crystal structure prediction of multinary inorganic materials is a challenging task given the diversity of potential structures. Experimental validation of the computational predictions is a vital step for accelerated materials discovery, however solid state chemistry synthetic routes lack the predictability of the synthesis outcome. Yet, theoretical predictions of structure and thermodynamic stability of the new ternary phases can yield the desired roadmap for the targeted synthesis. On the other hand, efficient synthesis allows for the fast screening of compositional space, providing timely feedback for the iterative improvement of theoretical predictions.
Here we report on two novel lithium nickel boride polymorphs RT-LiNiB and HT-LiNiB with layered crystal structures.¹ This family of compounds was theoretically predicted by using the adaptive genetic algorithm (AGA)² and subsequently synthesized by a hydride route with LiH as the lithium source. Unique among the known ternary transition-metal borides, the LiNiB structures feature Li layers alternating with almost planar [NiB] layers. A comprehensive study using a combination of single-crystal/synchrotron powder X-ray diffraction, solid-state NMR spectroscopy, scanning transmission electron microscopy, quantum-chemical calculations, and magnetism revealed on the intrinsic features of these polymorphic compounds. The unique layered structures of LiNiB compounds make them ultimate precursors for preparation of two-dimensional transition-metal borides, MBenes.