Tail states and unusual localization transition in low-dimensional Anderson model with power-law hopping
arXiv:2103.06001 · doi:10.1016/j.aop.2021.168524
Abstract
We study deterministic power-law quantum hopping model with an amplitude and local Gaussian disorder in low dimensions under the condition . We demonstrate unusual combination of exponentially decreasing density of the "tail states" and localization-delocalization transition (as function of disorder strength ) pertinent to a small (vanishing in thermodynamic limit) fraction of eigenstates. At sub-critical disorder delocalized eigenstates with energies near the bare band edge co-exist with a strongly localized eigenstates in the same energy window. At higher disorder all eigenstates are localized. In a broad range of parameters density of states decays into the tail region as simple exponential, , while characteristic energy varies smoothly across edge localization transition. We develop simple analytic theory which describes dependence on power-law exponent , dimensionality and disorder strength , and compare its predictions with exact diagonalization results. At low energies within the bare "conduction band", all eigenstates are localized due to strong quantum interference at ; however localization length grows fast with energy decrease, contrary to the case of usual Schrodinger equation with local disorder.
References in corpus (9)
- Correlation-induced localization
- Critical transport in weakly disordered semiconductors and semimetals
- Statistics of pre-localized states in disordered conductors
- Duality in power-law localization in disordered one-dimensional systems
- Unconventional localisation transition in high dimensions
- Localization properties of a one-dimensional tight-binding model with non-random long-range inter-site interactions
- Monitoring the localization-delocalization transition within a 1D model with non-random long-range interaction
- Non-ergodic extended phase of the Quantum Random Energy model
- Cavity approach to the first eigenvalue problem in a family of symmetric random sparse matrices