Disordered Electron Systems

• Disordered electron systems are materials where intrinsic randomness in atomic positions induces localization, quantum criticality, and unconventional transport phenomena.
• They are modeled using Hamiltonians with random onsite potentials and hopping terms, such as the Anderson and Anderson–Hubbard models, to capture complex electron dynamics.
• Experimental realizations in amorphous solids, doped semiconductors, and 2D electron gases validate theories predicting variable-range hopping, glassy behavior, and novel quantum phase transitions.

Disordered electron systems comprise materials and models in which the quantum coherence of electrons is strongly influenced by spatial randomness in atomic positions, site energies, coupling strengths, or other microstructural aspects. This randomness can induce localization, nontrivial quantum criticality, glassiness, or anomalous transport, with behavior that deviates dramatically from the standard Fermi-liquid paradigm. Disordered electronic systems are experimentally realized in amorphous solids, granular metals, oxide films, doped semiconductors, transition metal glasses, two-dimensional electron gases, and artificially engineered lattices. Both non-interacting and interacting cases display rich phenomena, including Anderson localization, glassy nonequilibrium dynamics, variable-range hopping, pronounced magnetoresistive effects, and correlated quantum phase transitions.

1. Fundamental Models and Hamiltonians

Disordered electron systems are typically modeled by lattice or continuum Hamiltonians that incorporate random variables to represent disorder. The canonical tight-binding Hamiltonian includes both random onsite potentials and random hopping amplitudes: $$H = \sum_{i\ne j} t_{ij} c_i^\dagger c_j + \sum_i V_i c_i^\dagger c_i,$$ where $V_i$ are random onsite potentials (Anderson disorder), and $t_{ij}$ may also be random (off-diagonal disorder), with microscopic origin in compositional randomness, lattice deformations, or electron–phonon interaction (Chen et al., 2024, Slavin et al., 14 Jun 2025). In models of correlated electrons, interaction terms ($U$ for onsite repulsion, $V$ for nearest neighbor, etc.) are added: $$H = -t \sum_{\langle i j \rangle} c_{i \sigma}^\dagger c_{j \sigma} + \sum_i V_i n_i + U \sum_i n_{i \uparrow} n_{i \downarrow},$$ as in the Anderson–Hubbard model (Liu et al., 2022, Chakraborty et al., 2011).

For electron glass or localized phases, long-range Coulomb interaction and fluctuating disorder are dominant (Delahaye et al., 2021). Real systems extend these models with multi-orbital, spin–orbit, and magnetic effects, and may exhibit chiral, orthogonal, unitary, or symplectic symmetries (as captured in the Cartan classification) (Markos et al., 2012, Dell'Anna, 2016).

2. Electron Localization and Quantum Criticality

Disorder disrupts electronic wavefunctions, leading to spatial localization (Anderson localization) and new scaling phenomena. Key diagnostics include the localization length $\xi$ (or $\lambda$) and the local density of states (LDOS). In one dimension, all states are localized for arbitrarily weak disorder, $\xi\sim O(1/W^2)$ (Chen et al., 2024, Rivas et al., 2023), while in higher dimensions a mobility edge and critical disorder $W_c$ separate localized and metallic regimes, with $\xi(W)\sim |W-W_c|^{-\nu}$ and $\nu\approx1.57$ in three dimensions.

At quantum critical points—for example, the chiral quantum critical point at $E=0$ in 2D systems with off-diagonal disorder—scaling behavior becomes logarithmic: the Lyapunov exponent $\gamma(E)\sim 1/|\ln(E_0/|E|)|^\kappa$ shows non-universality depending on symmetry class and disorder type (Markos et al., 2012). Chiral symmetry (spectral pairing $E\leftrightarrow -E$) allows for extended critical states strictly at the band center, violating the tenets of conventional scaling theory.

Modern approaches use statistical field theory (non-linear sigma model, NLSM; Finkel'stein model) to capture diffusive modes and interactions, yielding RG equations for disorder strength $\rho$, (triplet/singlet/Cooper) Landau parameters $\gamma$, and dynamical exponents $z_{en/m}$ (Finkel'stein et al., 2023, Dell'Anna, 2016). The interplay of disorder and interactions produces novel quantum phase transitions, including interaction-induced metallicity in systems near ferromagnetic quantum criticality wherein both disorder and residual interactions flow logarithmically to zero, yielding a perfectly conducting metal (Nosov et al., 2020).

3. Glassy Dynamics and Non-equilibrium Effects

In strongly disordered insulating films (granular Al, discontinuous Au, amorphous NbSi, InOx), slow, glassy dynamics of charge carriers can be directly observed by gate-voltage protocols (Delahaye et al., 2021). The electron glass regime is characterized by two hallmarks:

• Glassy relaxation: Conductance $G(t)$ exhibits a slow, logarithmic time dependence after a thermal quench or gate voltage perturbation, $\Delta G(t)\simeq -A \ln(t/\tau_0)$, indicating a scale-free distribution of relaxation times.
• Memory dips (MD): After equilibration at voltage $V_{g,eq}$, sweeping the gate reveals a conductance minimum centered on $V_{g,eq}$, analogous to the thermal memory in structural glasses.

These effects persist up to room temperature, with MD width increasing with temperature and amplitude decreasing as $T^{-\alpha}$ or $\exp(-E_a/k_BT)$, with effective barriers up to hundreds meV. The Coulomb glass model—localized electrons interacting via long-range repulsion and disorder—predicts a soft gap in the DOS, $N(E)\propto |E-E_F|^{d-1}$, and a hierarchy of metastable many-electron configurations separated by large barriers (Delahaye et al., 2021). No sharp glass transition is observed up to 300 K.

4. Transport Phenomena in Disordered Systems

Disordered electron systems display rich transport regimes, ranging from coherent metallic conduction to variable-range hopping (VRH) and nonlinear response. In the strongly localized regime, DC conductivity is suppressed; VRH dominates at low carrier density, with hopping rates determined by thermal or field activation: $$\sigma(E)\propto \exp[-(E_0/E)^{\nu}],\quad \nu=1/3 \text{ (1D VRH)},$$ where $E_0$ depends on disorder strength and localization length (Mozumdar et al., 22 Mar 2025). Under sufficiently strong electric field, thermal activation is supplanted by field-driven hops, modifying the scaling exponent to $\nu=1/2$ in the extreme regime.

Mean-field descriptions (coherent potential approximation, CPA) yield exact single-particle spectral properties but fail to capture quantum interference (weak localization, backscattering). Two-particle vertex corrections, resummed nonperturbatively via Bethe–Salpeter equations, restore manifest positivity and incorporate quantum-coherent backscattering ladders (Cooperons), giving physical conductivity down to the metal–insulator transition (Pokorny et al., 2012).

In quantum Monte Carlo studies of the Anderson–Hubbard model, disorder suppresses antiferromagnetic tendencies but can enhance low-temperature ferromagnetic correlations—independently of Anderson localization, Stoner instability, or local moment formation. Mechanistically, disorder disrupts kinetic exchange pathways, favoring local ferromagnetic alignment even when the moment size is reduced (Chakraborty et al., 2011). Spintronic applications exploit controlled disorder to tune susceptibilities.

5. Interplay of Disorder, Interactions, and Quantum Phases

In 2D and 3D electron liquids, the Finkel'stein NLSM formally unifies quantum interference and interaction effects, predicting non-monotonic $\rho(T)$ imposed by the RG flows in disorder–interaction space (Finkel'stein et al., 2023). In Si-MOSFETs, repulsive triplet interactions lead to flow toward a metallic ground state, consistent with observed resistance drop and universal scaling. Inclusion of Cooper-channel attraction describes disorder-suppression of superconductivity, with $T_c$ curbed by the resistance-dependent RG flow (Finkel'stein et al., 2023).

Magnetic quantum phase diagrams of dilute magnetic impurities reveal anomalous power-law susceptibility and a sharp disorder-driven Kondo–RKKY balance, set by the statistical cutoff of $|J_{RKKY}|/T_K$ across the system. The PM, Kondo, and spin-coupled regimes are set by impurity density, exchange coupling, and localization scale (Lee et al., 2012). Graphene (honeycomb lattice) exhibits a universal $J_c$ threshold for Kondo suppression, enhanced stability of spin-coupled phases, and disorder-driven paramagnetic regimes.

Electron–electron scattering in Galilean-invariant multi-band systems with realistic impurity mechanisms re-distributes energy among bands, modifying resistivity even though collisions conserve total current for parabolic bands. The temperature-dependent crossover between “bare” and “hydrodynamic” conductivity is governed by the difference in impurity scattering rates between subbands (Nagaev, 2021).

6. Modern Computational and Machine Learning Techniques

Large-scale quantum linear algebra—block-encoding sparse disordered Hamiltonians, quantum singular value transformation (QSVT), and amplitude estimation—enables exponential quantum speed-up in simulating disordered systems and critical scaling in high dimensions (Chen et al., 2024). Local density of states, transport properties, and localization length can be estimated in polynomial time in system size and dimension, leveraging quantum resources far beyond classical reach.

Machine learning approaches exploiting locality and group-theoretical descriptors for the Anderson–Hubbard model predict local observables (electron number, double-occupancy) with near–VMC accuracy. Neural networks trained on variational Monte Carlo datasets can provide correlation-aware local properties at linear cost, facilitating multi-scale modeling and embedding into kinetic Monte Carlo and phase field frameworks (Liu et al., 2022).

7. Phase Coexistence, Field-Induced Transitions, and Phonon Effects

In 2D low-density electron systems, short-range disorder can induce robust mesoscopic WC–liquid phase coexistence without density difference, in contrast to long-range disorder and clean microemulsion scenarios (Joy et al., 16 Feb 2025). Disorder stabilizes the Wigner crystal phase to higher densities, with mesoscopic domains mediated by local impurity fluctuations.

Field-driven superconductor–insulator transitions (SIT) in amorphous InOx thin films are captured by time-dependent Ginzburg–Landau theory supplemented by quantum tunneling–pair-breaking corrections (Maniv et al., 8 Oct 2025). Cooper-pair fluctuations condense in real-space puddles, with quantum-tunneled pairs breaking into fermions and producing activated inter-puddle transport. The critical field and scaling exponents extracted from experimental magnetoresistance are consistent with theory.

Coupling strongly disordered systems to classical (Einstein) phonons via quantum–classical hybrid methods (Lanczos/TEBD–Ehrenfest) destabilizes many-body localization, leading to phonon-induced delocalization and ergodic relaxation, even in regimes where static disorder would otherwise produce persistent memory (Menzler et al., 11 Dec 2025).

---

Disordered electron systems thus provide a versatile platform for exploring fundamental phenomena—localization, glassiness, critical scaling, variable-range hopping, unconventional magnetism, complex quantum phases, and the practical interplay between quantum computation and machine learning in condensed matter theory.