Theory of Anderson localization on the hyperbolic plane
Alexander Altland, Tobias Micklitz, Devasheesh Sharma, Maksimilian Usoltcev, Carolin Wille · 2026
Read Theory of Anderson localization on the hyperbolic plane by Alexander Altland, Tobias Micklitz, Devasheesh Sharma, Maksimilian Usoltcev, Carolin Wille (2026) 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-2604-24917.txt.
Introduction
Altland and four colleagues develop a unified theory of Anderson localization on the hyperbolic plane. This plane is locally flat at short range, but vast at large scales. It offers a paradigm for disorder-driven phase transitions, notably the Anderson transition. Building on work that treated short and large scales separately, they build a framework bridging low- and high-dimensional trapping rules. Their main result is a two-part flow in a plane spanned by scale-changing bend and flow. A long key line divides metal and non-metal phases.
Catalog record
| CREATOR | Alexander Altland, Tobias Micklitz, Devasheesh Sharma, Maksimilian Usoltcev, Carolin Wille |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 45901 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2604.24917
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Theory of Anderson localization on the hyperbolic plane Alexander Altland, Tobias Micklitz, Devasheesh Sharma, Maksimilian Usoltcev, Carolin Wille Abstract The two-dimensional hyperbolic plane, ℍ 2 \mathbb{H}^{2} , is an unusual system in that dimensionality changes with scale: locally two-dimensional and planar at short distances, but effectively infinite-dimensional at large scales, it provides an interesting paradigm for the study of (quantum) phase transitions, notably the disorder-driven Anderson transition. Generalizing previous work, which treated short and large distance scales separately, we develop a unified framework interpolating between the principles of low- and high-dimensional Anderson localization. As a main result, we derive a two-parameter flow in a plane spanned by scale-dependent curvature (setting the system’s effective dimensionality) and conductivity, with an extended critical line separating metallic and insulating phases. Introduction:— Physical dimensionality is one of the key parameters governing the long-distance and critical behavior of complex systems. Unlike symmetries or emergent collective phenomena, it is typically treated as an immutable attribute, organizing criticality into well-defined universality classes. The hyperbolic plane, \mathbb{H}^{2} , provides a notable exception to this paradigm 28 ; 22 ; 20 ; 4 ; 9 ; 7 ; 18 ; 8 ; 10 . As a two-dimensional surface of constant negative curvature with radius L , it exhibits a scale-dependent effective dimensionality: at short distances, r\ll L , the geometry is locally flat with d_{\mathrm{eff}}=2 , while at large distances, r\gg L , exponential volume growth renders the space effectively infinite-dimensional, d_{\mathrm{eff}}\to\infty . In this way, \mathbb{H}^{2} provides a setting in which regimes associated with different dimensionalities coexist. This crossover is clearly reflected in quantum transport, where, depending on spatial dimension d , the localization of quantum states by disorder is governed by different principles. For d\leq 2 , the high return probability of semiclassical trajectories leads to constructive interference and — in generic, non-interacting systems — to localization for arbitrarily weak disorder. For d>2 , returns become irrelevant, and localization requires strong disorder to overcome escape into the exponentially large volume. While conventional systems realize either one of these regimes, the hyperbolic plane exhibits both within a single setting. Previous work has addressed these limits separately, using weak-disorder perturbation theory 10 and lattice exact diagonalization 18 ; 8 . The aim of the present work is to develop a unified theory and to analyze the resulting dimensional crossover. Central to this construction is a scale-dependent change in the semiclassical point-to-point transmission probability — from two-dimensional diffusive behavior at short distances to effectively ballistic propagation at large scales (see Fig. 1 ) — which mediates an interpolation between the two localization mechanisms. We will address this problem using a renormalization group approach generalized to curved backgrounds, complemented by a transfer-matrix analysis for lattice discretizations of \mathbb{H}^{2} . The combination of these concepts will lead to a two-parameter flow in a phase plane spanned by curvature and conductivity, with an extended critical line separating metallic and insulating regimes, the main result of this work. Figure 1: A number of diffusive Brownian motion processes simulated on \mathbb{H}^{2} . The dashed curves indicate geodesic circles of radii r=(1,2,5)L . Note the crossover from diffusive to ballistic behavior, as trajectories escape to infinity. Definition of the problem:— The hyperbolic plane may be realized as the upper sheet of the two-sheeted hyperboloid x_{1}^{2}+x_{2}^{2}-x_{3}^{2}=-L^{2} , x_{3}>0 , in ambient Minkowski space \mathbb{R}^{2,1} with metric ds^{2}_{\rm amb}=dx_{1}^{2}+dx_{2}^{2}-dx_{3}^{2} . The radius L determines the negative scalar curvature, R=-2L^{-2} . In polar coordinates, x=(r,\phi) the metric tensor takes the form ds^{2}=dr^{2}+L^{2}\sinh^{2}\left(r/L\right)d\phi^{2} , where \phi\in[0,2\pi] is an angle and r the geodesic distance from the origin. We are interested in the physics of quantum particles of characteristic energy E=k_{0}^{2}/2m governed by the random Hamiltonian H=-\frac{1}{2m}\Delta+V(x) . Here, \Delta=g^{-1/2}\partial_{\mu}\!\left(\sqrt{g}\,g^{\mu\nu}\partial_{\nu}\right) is the Laplace operator on the plane, which in the chosen coordinates assumes the form \displaystyle\Delta \displaystyle=\partial_{r}^{2}+\coth\left(\frac{r}{L}\right)\frac{1}{L}\partial_{r}+\frac{1}{L^{2}\sinh^{2}\left(\frac{r}{L}\right)}\partial_{\phi}^{2}, (1) with g=\det(g_{\mu\nu})=L^{2}\sinh^{2}(r/L) . We assume the random potential V to be \delta -correlated with a variance \left\langle V(x)V(y)\right\rangle=\left(W/k_{0}\right)^{2}|g(x)|^{-1/2}\delta(x-y) setting the disorder strength, W . Before turning to the quantum dynamics of this problem, consider the classical limit, whose dynamics is governed by the ‘heat operator’ (\partial_{t}-D\Delta) with a constant D (we hesitate to call it diffusion operator, as the dynamics is diffusive only locally). The matrix elements \Delta^{-1}(x,y)\equiv\braket{x|\Delta^{-1}|y} are measures for (the negative of) the time integrated probability of a particle to propagate from y to x , and they are known in terms of eigenfunction decompositions (cf. the Supplementary Material) with the simple result \Delta^{-1}(x,y)=\frac{1}{2\pi}\ln\tanh(d(x,y)/2L) , where d(x,y) is the geodesic distance between the observation points. Specifically, for d(x,y)=r\ll L small compared to the curvature, \Delta^{-1}(r)\approx\frac{1}{2\pi}\ln(r/2L) as for a diffusive Euclidean plane. Conversely, for r\gg L , \Delta^{-1}(r)\approx-\frac{1}{\pi}\exp(-r/L) . This exponential suppression indicates directed isotropic decay towards any of the points on the circle of radius r , and metric circumference \sim\exp(r/L) . Indeed, the analysis of the full space-time dependent heat equation shows that in this regime \langle r\rangle\sim 2Dt/L grows ballistically, consistent with the escape dynamics of the trajectory simulations in Fig. 1 . Effective field theory:— Turning to the quantitative analysis, we consider the uniquely determined extension of the effective field theory of Anderson localization 11 ; 2 to a background geometry with a Riemannian metric 1 , g_{\mu\nu} . This is a nonlinear \sigma -model defined by the functional integral Z=\int DQ\,e^{-S[Q]} , with the action 26 ; 1 , \displaystyle S[Q] \displaystyle=-\frac{\sigma}{16}\int d^{2}x\sqrt{g}\,\mathrm{str}\left(Q\Delta Q\right). (2) Here, Q=T\tau_{3}T^{-1}=\{Q^{\alpha\beta}(x)\} where the index \alpha=(a,c,r) comprises an advanced/retarded index a=\pm 1 (with \tau_{3} being a Pauli matrix in this space), an index c=\pm 1 for time-reversal, and a ‘super-index’ r=\mathrm{b,f} distinguishing between commuting ( \mathrm{bb,ff} ) and anti-commuting ( \mathrm{bf,fb} ) matrix entries, with the supertrace \mathrm{str}(A)=\mathrm{tr}(A^{\mathrm{bb}})-\mathrm{tr}(A^{\mathrm{ff}}) . (Anticipating the need for non-perturbative constructions in the strong-disorder regime, we here work with the supersymmetry formalism, where Q are 8\times 8 supermatrices, time-reversal symmetry (symmetry class AI) requires 11 T to take values in the ortho-symplectic supergroup \mathrm{OSp}(2,2|4) , and Q in a coset space \mathrm{OSp}(2,2|4)/\mathrm{OSp}(2|2)\times\mathrm{OSp}(2|2) dividing out transformations commuting with \tau_{3} .) Finally, the integral is over the invariant area element, defined by \sqrt{g}drd\phi=L\sinh(r/L)drd\phi , approaching the planar element rdrd\phi for r\ll L and exponentially growing otherwise. Renormalization:— \sigma -models in two-dimensional space generically fluctuate strongly at large length scales, leading to an effective downward renormalization of the coupling constant, and eventually localization 12 . Our goal is to find out how the geometric distortion, formally encoded in the volume element and the Laplacian, interferes with this picture. In concrete terms, this requires an RG protocol, implemented, e.g., by mode renormalization: depending on their Laplace eigenvalues, one organizes the field space T\to T_{\textrm{f}}T_{\textrm{s}} into ‘fast’ and ‘slow’ modes, and proceeds by iterative integration over the fast T_{\textrm{f}} . While in translationally invariant settings the corresponding eigenfunctions are just Fourier modes, we here need to work with the eigenfunctions of the hyperbolic Laplacian. These functions provide a convenient, physical basis for the expansion of field fluctuations in a variant of the Fourier transform known as the Fourier-Helgason transform. Apart from this modification, the one-loop integration over fast field modes with eigenvalues in an interval [\Lambda e^{-\ell},\Lambda] below a UV cutoff is standard 11 ; 2 (see the Supplementary Material for details), and it