Information scrambling and butterfly velocity in quantum spin glass chains

Abstract

We make lattice generalization of two well-known zero-dimensional models of quantum spin glass, Sachdev-Ye (SY) and spherical quantum $p$-spin glass model, to one dimension for studying crossovers in non-local scrambling dynamics due to glass transition, complex dynamics, and quantum and thermal fluctuations in paramagnetic (PM) and spin glass (SG) phases. In the SY chain of quantum dots, each described by infinite-range random Heisenberg model of $N$ spin-$S$ $SU(M)$ spins, we obtain the quantum Lyapunov exponent $\lambda_\mathrm{L}$ and butterfly velocity $v_B$ as a function of temperature $T$ and the quantum parameter $S$ across the PM-SG phase boundary using a bosonic spinon representation in the large $N,M$ limit. In particular, we extract asymptotic $T$ and $S$ dependence, e.g., power laws, for $\lambda_\mathrm{L}$ and $v_B$ in different regions deep inside the phases and near the replica symmetry breaking SG transition. We find the chaos to be non-maximal almost over the entire phase diagram. Very similar results for chaos indicators are found for the $p$-spin glass chain as a function of temperature and a suitable quantum parameter $\Gamma$, with some important qualitative differences. In particular, $\lambda_\mathrm{L}$ and $v_B$ exhibit a maximum, coinciding with onset of complex glassy relaxation, above the glass transition as a function of $T$ and $\Gamma$ in the PM phase of the $p$-spin glass model. In contrast, the maximum is only observed as a function of $S$, but not with temperature, in the PM phase of SY model. The maximum originates from enhanced chaos due to maximal complexity in the glassy landscape. Thus, the results in the SY model indicate very different evolution of glassy complexity with quantum and thermal fluctuations.