Abstract
The "band structure" of a disordered stripe array is computed and compared, at a qualitative level, to angle-resolved photoemission experiments on cuprate high-temperature superconductors. The low-energy states are found to be strongly localized transverse to the stripe direction, so the electron dynamics is strictly one dimensional (along the stripe). Despite this, aspects of the two dimensional band-structure Fermi surface are still vividly apparent.
Introduction
A central paradox in the study of cuprate high-temperature superconductors is the coexistence of one-dimensional (1D) "stripes" and a two-dimensional (2D) Fermi surface. While evidence suggests that local, self-organized 1D structures dominate electron dynamics, Angle-Resolved Photoemission Spectroscopy (ARPES) typically reveals a "nodal region"—spectral weight along the Brillouin zone diagonal—that resembles a standard 2D metal.
Standard models of perfectly ordered 1D stripes predict a spectral gap in these nodal regions, directly contradicting experimental observations. This paper addresses whether introducing disorder into stripe patterns can resolve this tension, explaining how strictly 1D electron dynamics can manifest as a 2D-like spectral map.
Methods
The authors utilize a non-interacting tight-binding model on a square lattice. Electrons interact with a static, staggered field $m(x,y)$ representing antiferromagnetic order, where stripes function as anti-phase domain walls. Through exact numerical diagonalization on large systems (typically $320 \times 320$ sites), the energy eigenstates and wavefunctions are determined.
The single-particle spectral function $A(\mathbf{k}, \omega)$ is computed and integrated over a narrow energy window below the chemical potential to simulate ARPES intensity maps. The study compares three primary scenarios: isolated stripes, perfectly ordered arrays, and disordered arrays representing snapshots of fluctuating systems.
Key Results
The investigation reveals that 1D dynamics do not necessarily imply a 1D spectral appearance. Although electrons are confined to move along 1D stripes, their wavefunctions possess a finite transverse width, allowing them to "map out" the underlying 2D Fermi surface in momentum space.
A critical finding is the role of "bi-stripes"—anomalously close pairs of stripes occurring in disordered configurations. These bi-stripes create low-energy states that fill the spectral gaps seen in ordered models. As shown in the disordered model results, these arrays produce a broad region of spectral weight near the $(\pi, 0)$ point and signatures in the nodal region, qualitatively matching the "smearing" observed in experimental ARPES data for materials like $La_{2-x}Sr_xCuO_4$ (LSCO).
Participation ratio calculations confirm that low-energy states remain exponentially localized transverse to the stripes while remaining extended along them. This confirms that the underlying dynamics are 1D, even when the spectral map suggests a 2D nature.
Discussion and Limitations
The model demonstrates that disorder acts as a bridge between stripe theory and experimental data. However, it treats stripes as static snapshots, neglecting full dynamical fluctuations. Furthermore, the results are derived from a non-interacting model; while the authors suggest that 1D electron interactions (Luttinger liquid theory) would preserve these features, a full many-body treatment remains for future work. The simplicity of the tight-binding setup also omits realistic parameters like second-neighbor hopping ($t'$). Despite these limitations, the study suggests that nodal spectral weight intensity could serve as an experimental metric for stripe disorder in cuprates.
Summary updated