UCLA

Na2IrO3 and other “Kitaev candidate” materials are of much scientific and technological interest. In such materials, proposed anisotropic, Ising-like interactions promote magnetic frustration, potentially leading to a quantum spin liquid (QSL) ground state with excitations (non-Abelian anyons) that could enable topological quantum computing – a form of quantum computing particularly robust to decoherence. Unfortunately, experimental work indicates that Na2IrO3 and other Kitaev candidates do not exhibit a QSL ground state. While there are several proposals for manipulating Kitaev candidates into a QSL state, there is no consensus over whether the effective-spin interactions of these materials are proximate to the values necessary for such a state. Here, we use density functional theory (DFT) and mathematical techniques to investigate the electronic properties and effective-spin interactions of Kitaev candidate materials. Our approach for determining spin interaction terms, Compressive Sensing Spin Dynamics (CSSD), involves two main steps: 1) by appending the Kohn-Sham energy functional with a Lagrangian term that allows for quickly finding the lowest energy solution of a material given fixed magnetic moments, we perform many calculations of the material with the spins slightly perturbed from the equilibrium state; 2) we perform compressive sensing on these data (on the fixed spins and magnetic fields required to stabilize the desired spin arrangement) to yield the interaction terms with far fewer data than otherwise required. Performing this procedure on the Kitaev candidates Na2IrO3 and α-RuCl3 indicates neither is close to the Kitaev QSL regime. We further investigate manipulated versions of Kitaev candidate materials and find all have problems: epitaxial and single layers of Na2IrO3 are conductors (implying Na2IrO3 cannot be cleanly separated into individual layers without changing the band structure), while straining Na2IrO3 in-plane introduces different problems if under tension (the Ir-O-Ir angles increase, leading to less destructive interference of undesirable interactions) and compression (the Ir-Ir distances decrease, leading to larger isotropic Heisenberg interaction, which competes with the Kitaev interaction). We additionally use CSSD to evaluate the interaction terms of idealized versions of Na2IrO3 and α-RuCl3 (90� Ir-O-Ir and Ru-Cl-Ru angles), and our results indicate that these, too, are far from the Kitaev QSL state. Finally, we investigate if YbBr3, as a lanthanide material, may circumvent the problems with 4d and 5d candidates, due to smaller f-f orbital overlap compared to d-d overlap. While the results with YbBr3 are somewhat ambiguous, we conclude f-orbital Kitaev candidate materials warrant further investigation.