arXiv:2610.33393
Controlling transport with structured long-range hopping
Boris Gurevich, Michael H. Kolodrubetz
Abstract
The combined effects of disorder and long-range hopping play an important role in determining localization and transport in low-dimensional quantum systems. We study transport in a one-dimensional disordered chain with nearest-neighbor hopping supplemented by a long-range hopping term. The long-range coupling is characterized by site-dependent amplitudes that may be either uniform or random. We investigate the interplay between diagonal disorder and long-range hopping, characterizing the resulting transport efficiency. We find that nearest-neighbor hopping in the presence of uniform long-range hopping amplitudes strongly suppresses delocalization over a range of weak disorder. The disorder-enhanced transport regime, established in previous studies, can be understood here as the recovery of delocalization from this strongly suppressed regime. An analytical solution in the Anderson eigenstate basis reveals the microscopic origin of this behavior. In particular, we show that the low-disorder regime of strongly suppressed delocalization can be eliminated by introducing randomness into the long-range hopping amplitudes, providing an additional mechanism for enhancing transport and qualitatively changing the response of the system to disorder.
AI Summary
The paper asks how the spatial profile of cavity-mediated, long-range hopping changes localization and disorder-enhanced transport in a one-dimensional Anderson chain. It studies nearest-neighbor hopping supplemented by a rank-one term $H_L=-\gamma\sum_{ij}\eta_i\eta_j|i\rangle\langle j|$, comparing uniform, random, and sinusoidal coupling profiles. For uniform $\eta_j$, a weak-to-moderate-disorder regime has strongly suppressed eigenstate spatial variance; the previously observed disorder-enhanced transport regime is interpreted as recovery from this suppression. Random amplitudes, especially zero-mean $\eta_j\in[-1,1]$, largely remove the suppression, whereas sinusoidal profiles retain it. The explanation uses an exact rank-one secular equation in the Anderson eigenbasis: uniform and sinusoidal profiles concentrate coupling weight on selected low-disorder momentum states, suppressing mixing among other Anderson orbitals. Random profiles distribute that weight more broadly. Neglecting interference and statistically factorizing orbital-mixing weights from spatial profiles gives $\langle\sigma_T^2\rangle\simeq P_A\langle\sigma_A^2\rangle+(1-P_A)N^2/24$, where $P_A$ is the full eigenstates’ inverse participation ratio in the Anderson basis. Numerical comparisons use chains of $N=1000$–$3000$, generally nine disorder realizations. Single-excitation drain-transfer calculations, averaged over 100 realizations, show corresponding qualitative transport trends. The variance approximation fails at very weak disorder for nonuniform profiles, where interference matters. An explicit analytical disorder dependence of $P_A$ remains unavailable. Transport is estimated from transient transfer times rather than a full pumped steady state. Experimental implementation, accompanying nearest-neighbor hopping disorder, and many-excitation dynamics remain open extensions.
Error found: The weak-disorder level-spacing argument neglects the clean ring’s ±k degeneracies, whose generic first-order disorder splitting scales as W/√N rather than the asserted W, undermining the quantitative condition inferred from that estimate. The general-observable expression and appendix transpose the eigenvector coefficients relative to their definition, although correcting these indices leaves the final statistically factorized formulas unchanged.
Written by gpt-6.1-sol from the paper's full text. Automated: it can be wrong.
- Comments
- 15 pages, 7 figures