No Arabic abstract
We show that the off-diagonal exchange anisotropy drives Mott insulators with strong spin-orbit coupling to a classical spin liquid regime, characterized by an infinite number of ground states and Ising variables living on closed or open strings. Depending on the sign of the anisotropy, quantum fluctuations either fail to lift the degeneracy down to very low temperatures, or select non-collinear magnetic states with unconventional spin correlations. The results apply to all 2D and 3D tri-coordinated materials with bond-directional anisotropy, and provide a consistent interpretation of the suppression of the x-ray magnetic circular dichroism signal reported recently for $beta$-Li$_2$IrO$_3$ under pressure.
We report a new classical spin liquid in which the collective flux degrees of freedom break the translation symmetry of the honeycomb lattice. This exotic phase exists in frustrated spin-orbit magnets where a dominant off-diagonal exchange, the so-called $Gamma$ term, results in a macroscopic ground-state degeneracy at the classical level. We demonstrate that the system undergoes a phase transition driven by thermal order-by-disorder at a critical temperature $T_c approx 0.04 |Gamma|$. At first sight, this transition reduces an emergent spherical spin-symmetry to a cubic one: spins point predominantly toward the cubic axes at $T < T_c$. However, this seems to simply restore the cubic symmetry of the $Gamma$ model, and the non-coplanar spins remain disordered below $T_c$. We show that the phase transition actually corresponds to plaquette ordering of hexagonal fluxes and the cubic symmetry is indeed broken, a scenario that is further confirmed by our extensive Monte Carlo simulations.
Iridate oxides on a honeycomb lattice are considered promising candidates for realization of quantum spin liquid states. We investigate the magnetic couplings in a structural model for a honeycomb iridate K$_2$IrO$_3$, with $C_3$ point group symmetry at the Ir sites, which is an end member of the recently synthesized iridate family K$_x$Ir$_y$O$_2$. Using textit{ab-initio} quantum chemical methods, we elucidate the subtle relationship between the real space symmetry and magnetic anisotropy and show that the higher point group symmetry leads to high frustration with strong magnetic anisotropy driven by the unusually large off-diagonal exchange couplings ($Gamma$s) as opposed to other spin-liquid candidates considered so far. Consequently, large quantum fluctuations imply lack of magnetic ordering consistent with the experiments. Exact diagonalization calculations for the fully anisotropic $K$-$J$-$Gamma$ Hamiltonian reveal the importance of the off-diagonal anisotropic exchange couplings in stabilizing a spin liquid state and highlight an alternative route to stabilize spin liquid states for ferromagnetic $K$.
We report the structural transformation of hexagonal Ba3YIr2O9 to a cubic double perovskite form (stable in ambient conditions) under an applied pressure of 8GPa at 1273K. While the ambient pressure (AP) synthesized sample undergoes long-range magnetic ordering at 4K, the high pressure(HP) synthesized sample does not order down to 2K as evidenced from our susceptibility, heat capacity and nuclear magnetic resonance (NMR) measurements. Further, for the HP sample, our heat capacity data have the form gamma*T+beta*T3 in the temperature (T) range of 2-10K with the Sommerfeld coefficient gamma=10mJ/mol-Ir K2. The 89Y NMR shift has no T-dependence in the range of 4-120K and its spin-lattice relaxation rate varies linearly with T in the range of 8-45K (above which it is T-independent). Resistance measurements of both the samples confirm that they are semiconducting. Our data provide evidence for the formation of a 5d based, gapless, quantum spin-liquid (QSL) in the cubic (HP) phase of Ba3YIr2O9. In this picture, the T term in the heat capacity and the linear variation of 89Y 1/T1 arises from excitations out of a spinon Fermi surface. Our findings lend credence to the theoretical suggestion [G. Chen, R. Pereira, and L. Balents, Phys. Rev. B 82, 174440 (2010)] that strong spin-orbit coupling can enhance quantum fluctuations and lead to a QSL state in the double perovskite lattice.
In systems where electrons form both dispersive bands and small local spins, we show that changes of the spin configuration can tune the bands through a Lifshitz transition, resulting in a continuous metal-insulator transition associated with a progressive change of the Fermi surface topology. In contrast to a Mott-Hubbard and Slater pictures, this spin-driven Lifshitz transition appears in systems with small electron-electron correlation and large hybridization. We show that this situation is realized in 5$d$ distorted perovskites with an half-filled $t_{2g}$ bands such as NaOsO$_3$, where the strong $p-d$ hybridization reduces the local moment, and spin-orbit coupling causes a large renormalization of the electronic mobility. This weakens the role of electronic correlations and drives the system towards an itinerant magnetic regime which enables spin-fluctuations.
Motivated by recent experiments on $alpha$-RuCl$_3$, we investigate a possible quantum spin liquid ground state of the honeycomb-lattice spin model with bond-dependent interactions. We consider the $K-Gamma$ model, where $K$ and $Gamma$ represent the Kitaev and symmetric-anisotropic interactions between spin-1/2 moments on the honeycomb lattice. Using the infinite density matrix renormalization group (iDMRG), we provide compelling evidence for the existence of quantum spin liquid phases in an extended region of the phase diagram. In particular, we use transfer matrix spectra to show the evolution of two-particle excitations with well-defined two-dimensional dispersion, which is a strong signature of quantum spin liquid. These results are compared with predictions from Majorana mean-field theory and used to infer the quasiparticle excitation spectra. Further, we compute the dynamical structure factor using finite size cluster computations and show that the results resemble the scattering continuum seen in neutron scattering experiments on $alpha$-RuCl$_3$. We discuss these results in light of recent and future experiments.