Localization from Infinitesimal Kinetic Grading: Critical Scaling and Kibble-Zurek Universality

We study a one-dimensional lattice model with site-dependent nearest-neighbor hopping amplitudes that follow a power-law profile. The hopping variation is controlled by a grading exponent, $α$, which serves as the tuning parameter of the system. In the thermodynamic limit, the ground state becomes localized as $|α| \to 0$, signaling the presence of a critical point characterized by a diverging localization length. Using exact diagonalization, we perform finite-size scaling analysis and extract the associated critical exponent governing this divergence, revealing a universality class distinct from well-known Anderson, Aubry-Andre, and Stark localization. To further characterize the critical behavior, we analyze the inverse participation ratio, the energy gap between the ground and first excited states, and the fidelity susceptibility. We also investigate nonequilibrium dynamics by linearly ramping the hopping profile at various rates and tracking the evolution of the localization length and the inverse participation ratio. The Kibble-Zurek mechanism successfully captures the resulting dynamics using the critical exponents obtained from the static scaling analysis. Our results demonstrate a clean, disorder-free route to localization and provide a tunable platform relevant to photonic lattices and ultracold atom arrays with engineered hopping profiles.

💡 Research Summary

The authors investigate a one‑dimensional tight‑binding chain whose nearest‑neighbour hopping amplitudes follow a power‑law profile t_i ∝ i^α. The exponent α acts as a tuning parameter: α = 0 corresponds to a uniform lattice with completely delocalized Bloch states, while any infinitesimal non‑zero α (positive or negative) introduces a spatial bias in the kinetic term, breaking translational invariance without adding any on‑site disorder. The Hamiltonian possesses a sublattice (chiral) symmetry {S, Ĥ}=0, guaranteeing a symmetric spectrum about zero energy and a zero‑mode for odd‑length chains.

Using exact diagonalization for system sizes L = 500–2000, the authors compute several observables of the ground state: the localization length ξ (defined via the second moment of the probability distribution), the inverse participation ratio (IPR) χ = ∑_i p_i^2, the energy gap ΔE between the ground and first excited states, and the fidelity susceptibility η_Q. For finite L, ξ and χ display a flat region near α = 0 (extended behavior) followed by a rapid crossover to size‑independent values once |α| exceeds a size‑dependent threshold α_T(L). As L increases, α_T(L) drifts toward zero, indicating that in the thermodynamic limit the delocalized point collapses to the single value α_c ≈ 0.

Finite‑size scaling is performed with the standard ansatz X/L^κ = f