Interplay of Power-Law correlated Disorder and Long-Range Hopping in One Dimension: Mobility Edges, Criticality, and ML-Based Phase Identification
Mohammad Pouranvari · 2025
Read Interplay of Power-Law correlated Disorder and Long-Range Hopping in One Dimension: Mobility Edges, Criticality, and ML-Based Phase Identification by Mohammad Pouranvari (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2511-18341.txt.
Introduction
Mohammad Pouranvari maps the phases of a one dimensional model where electrons hop between sites with disordered energies. The site energies are correlated over distance with a power law, and hopping strength also falls as a power law, so two exponents control the whole family of models.
Using large exact studies, Pouranvari combines spectral data and localization checks to chart where states are localized, extended, or critical, finding robust mobility edges and mixed regimes. A scaling analysis extracts critical exponents and transition lines. A guided autoencoder trained on raw eigenstate features matches the same phase map, backing the physics based picture.
Catalog record
| CREATOR | Mohammad Pouranvari |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2511-18341
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Interplay of Power-Law correlated Disorder and Long-Range Hopping in One Dimension: Mobility Edges, Criticality, and ML-Based Phase Identification Mohammad Pouranvari Abstract We investigate a one-dimensional tight-binding model in which onsite potentials { ε i } \{\varepsilon_{i}\} exhibit power-law spatial correlations (with exponent α \alpha ) and the hopping amplitudes decay as t i j ∼ | i − j | − β t_{ij}\sim|i-j|^{-\beta} . This two-parameter family interpolates continuously between short-range Anderson-like disorder, correlated disorder with conventional hopping, and long-range hopping models with nontrivial delocalization tendencies. Using large-scale exact diagonalization, we construct a comprehensive phase map in the ( α , β ) (\alpha,\beta) plane by combining spectral statistics, density-of-states analysis, and energy-resolved localization indicators such as the participation ratio, single-particle entanglement entropy, level-spacing ratio r r , and the ratio of the geometric to arithmetic density of states. From these observables we define phase-indicator functions that compactly quantify localization behaviour across the spectrum. Our analysis reveals robust mobility edges and multiple regimes of spectral coexistence between localized, extended, resonant, and critical states. Finite-size scaling, implemented via an explicit smoothness-based cost function, enables extraction of critical exponents and delineation of transition lines across the ( α , β ) (\alpha,\beta) parameter space. To validate and complement these physics-based diagnostics, we employ a supervised autoencoder that learns high-level representations of eigenstate structure directly from raw features and reliably reproduces the phase classification defined by the indicator functions. Together, these approaches provide a coherent and internally consistent picture of the spectral transitions driven by correlated disorder and long-range hopping, establishing a unified framework for characterizing mobility edges in long-range one-dimensional systems. I Introduction The interplay of disorder and quantum interference in low-dimensional systems is a paradigmatic problem in condensed-matter physics Santos and Rigolin 2005 ; Toldin et al. 2013 ; Walschaers et al. 2016 ; Dobrinevski et al. 2011 ; Mishra et al. 2016 . In one dimension, the conventional result is simple and robust: for short-range models with uncorrelated disorder, all single-particle eigenstates are exponentially localized in the thermodynamic limit Anderson 1958 ; Abrahams et al. 1979 . Departures from this simple picture arise, however, when one of the basic assumptions is relaxed — for example, when the potential landscape contains long-range correlations de Moura and Lyra 1998 ; Izrailev and Krokhin 1999 or when the kinetic energy is long-ranged Levitov 1990 ; Mirlin and Fyodorov 1996 . Each of these ingredients alone can produce atypical spectral and transport behaviour in one dimension; when combined the resulting phenomenology becomes rich and — in many respects — only partially charted. On one hand, tight-binding models with long-range hopping, where the hopping amplitude between sites i and j decays as a power law |i-j|^{-\beta} , can host a localization-delocalization transition even in 1D. For sufficiently slow decay ( \beta\leq 1 ), the hopping is non-local enough to overcome the localizing effects of disorder, leading to extended eigenstates. Such models have been instrumental in understanding phenomena ranging from transport in disordered systems to the nature of eigenstates in power-law random banded matrices, which often exhibit multifractal properties at the transition point Mirlin et al. 1996 ; Pérez-González et al. 2019 ; Celardo et al. 2016 ; Levitov 1999 ; Nag and Garg 2019 ; Singh et al. 2017 ; Zhou and Bhatt 2003 ; Qi et al. 2021 ; Wang et al. 2023 . On the other hand, introducing spatial correlations into the disorder potential can also induce delocalization. Potentials with long-range correlations, such as those generated from fractional Brownian motion, can create extended states and even mobility edges separating localized and delocalized states within the energy spectrum de Moura and Lyra 1998 ; Izrailev and Krokhin 1999 . The intuition is that correlated disorder is ”smoother” than white-noise disorder, reducing the effectiveness of random scattering Keen and Goodwin 2015 ; Carpena et al. 2002 ; Neverov et al. 2022 ; Li et al. 2011 ; Pouranvari and Moghaddam 2025 . While these two mechanisms for delocalization have been studied extensively in isolation, their interplay remains a rich area of investigation. The central question we address in this work is: What is the fate of localization in a 1D system where both long-range hopping and long-range correlated disorder are present simultaneously? This scenario is not merely of academic interest but is relevant to various physical systems, including trapped ions, Rydberg atoms, and polar molecules, where both interactions and external potentials can exhibit long-range characteristics Defenu et al. 2021 ; Kuhl et al. 2000 ; Chávez et al. 2021 ; Bixon et al. 1999 ; Singh et al. 2017 ; Ito et al. 2003 . In this work we consider a family of one-dimensional tight-binding Hamiltonians in which (A) the on-site potentials \{\varepsilon_{i}\} are constructed by Fourier synthesis with random phases so that the power spectrum scales as S(k)\propto k^{-\alpha} ; the resulting sequence is mean-subtracted and variance-normalized and is approximately Gaussian for large system size. and (B) the hopping amplitude between sites decays algebraically as |i-j|^{-\beta} with exponent \beta . This two-parameter family interpolates between several limits of current interest: (i) the short-range, uncorrelated Anderson limit ( \beta\to\infty , small \alpha ), (ii) models with long-range correlated potentials but nearest-neighbour hopping ( \beta\gtrsim 1 , \alpha\gg 1 ), (iii) long-range-hopping chains that have been argued to admit nontrivial critical behaviour and delocalization tendencies for sufficiently small \beta , and (iv) long-range-hopping with strongly correlated disorder with the expectation (at first sight) of having extended states across all spectrum. Physically, the correlated potential models capture situations where the disorder arises from a rough substrate or from long-wavelength structural variations; long-range hopping arises naturally in systems with dipolar or Coulomb-mediated tunnelling or in effective single-band descriptions with nonlocal overlap integrals. A central question we address is whether, and under what conditions in the (\alpha,\beta) parameter space, there will be a phase transition between delocalized/localized states for the entire spectrum or more interestingly there will be a single-particle mobility edges — energy-dependent transitions between localized and extended eigenstates — appear and remain robust with system size. Mobility edges in one-dimensional deterministic quasiperiodic systems are known (for example, in generalizations of the Aubry–André model), and mobility edges have been reported in various models with correlated disorder or long-range hopping Aubry and André 1980 . What remains unsettled is a systematic characterization, using a compact and mutually consistent set of diagnostics, of how correlated disorder and algebraic hopping cooperate or compete to produce energy-selective localization. To address these questions we use a set of complementary spectral and eigenstate diagnostics. We first analyse level statistics through the spacing distribution P(s) and number variance \Sigma^{2}(L) , which distinguish Poisson (localized) from Wigner–Dyson (extended) behaviour. We then employ energy-resolved probes—the participation ratio, single-particle entanglement entropy, and the auxiliary measures S_{\mathrm{proxy}} and the \rho -ratio—to characterize eigenstate structure across the band. Finally, we define four phase-classification functions f_{1}\!-\!f_{4} , which quantify the fractions of energy bins labelled localized, extended, resonant, or critical, providing a compact summary of spectral mixing for constructing the (\alpha,\beta) phase diagram. A central emphasis of this work is that energy-resolved analysis is essential: mobility edges appear throughout broad regions of the (\alpha,\beta) parameter space, and global (spectrum-averaged) quantities