Level statistics in the fractal phase of generalized Rosenzweig--Porter models
summary
The gist
Level statistics in the fractal phase of generalized Rosenzweig–Porter models explores how level statistics behave in systems exhibiting an integrability-to-chaos crossover through generalized…
In short
The study investigates level statistics in generalized Rosenzweig-Porter models, which describe a quantum many-body system's transition between chaotic and localized states via a fractal phase. Using free probability and replica methods, the authors found universal scaling forms for full counting statistics around the Thouless energy scale ($E_T$). Numerical tests confirm this universality across different model variations.
Key concepts
- Rosenzweig--Porter (RP) Ensemble
- This is a minimal mathematical model used to describe how quantum systems move from being fully chaotic to being localized. It features a 'fractal phase' where the energy levels have intermediate properties between completely random and completely ordered states.
- Thouless Energy ($E_T$)
- This specific energy scale marks the crossover point in the system. It represents the characteristic width of the spectrum that quantum evolution can explore in a single time step, defining the boundary between different spectral behaviors.
- Full Counting Statistics (FCS)
- This statistical tool measures how often eigenvalues fall within a specific energy interval. The paper calculates this using advanced methods like free probability theory and replica techniques to find universal scaling laws near the crossover point.
Terminology used across episodes
This episode discusses
- Level statistics in the fractal phase of generalized Rosenzweig--Porter models · Paper Radio
- Random matrix theory of integrability-to-chaos transition
- Localization transition on the Random Regular Graph as an unstable tricritical point in a log-normal Rosenzweig-Porter random matrix ensemble
- Free Probability approach to spectral and operator statistics in Rosenzweig-Porter random matrix ensembles
- Fading ergodicity and quantum dynamics in random matrix ensembles
- On a Rosenzweig-Porter-type model
- Refinements of the Eigenstate Thermalization Hypothesis under Local Rotational Invariance via Free Probability
- Quantum Signatures of Chaos from Free Probability
- Multifractal Dynamics of the QREM
- A Software Tool for "Gluing" Distributions
The paper
Level statistics in the fractal phase of generalized Rosenzweig--Porter models · Read on arXiv
King’s College London · Sorbonne Université · Institut für Physik Universität Oldenburg
The Rosenzweig--Porter (RP) random matrix ensemble has emerged as a minimal model for the integrability-to-chaos crossover in quantum many-body systems. Its phase diagram features a region with fractal eigenstates, exhibiting intermediate spectral and localization properties between the fully localized and fully delocalized regimes. In this work, we explore several generalizations of the RP model and determine their level statistics at the scale of the Thouless energy E T, which characterizes the crossover. Using tools from free probability theory and the replica method, we compute the full counting statistics in the limit of large system size, and show that it takes a simple, universal scaling form around E T, shared across all variations of the model. We validate our analytical predictions using exact numerical diagonalization of large samples, and large-deviation algorithms that resolve the full counting statistics down to probabilities as low as 10-40. We also contrast our predictions with measurements on the quantum random energy model, which is the simplest model displaying many-body localization.
Transcript
Introduction to the show: ident: Condensed Matter. The day in condensed matter: materials, phases and the measurements under them.
Kai: We're looking at the paper: "Level statistics in the fractal phase of generalized Rosenzweig--Porter models".
Mira: Here's the gist of the paper. The paper explores the level statistics in the fractal phase of generalized Rosenzweig–Porter models to understand how quantum many-body systems transition between chaotic and localized regimes through a crossover scale characterized by the Thouless energy The authors compute full counting statistics for these models using free probability theory and the replica method, revealing universal scaling forms across different generalizations of the model Model Framework and Crossover Scale The Rosenzweig–Porter (RP) random matrix ensemble serves as a minimal model for the integrability-to-chaos crossover in quantum many-body systems, featuring a fractal phase with intermediate spectral and localization properties The generalized RP (GRP) model is defined by the sum of a diagonal random matrix A with i.i.d.
Kai: entries and a matrix M drawn from either rotationally invariant or Wigner classes The crossover scale is characterized by the Thouless energy ET, which defines the width in the spectrum accessible in time of O(1) by quantum evolution This energy scale is defined as ET ∼ Xj(̸=i) ⟨hij 2⟩ ∼ σ ∼ N(1-γ) Analytical Derivation of Full Counting Statistics The full counting statistics, specifically the cumulant generating function (CGF) of the number IN
α, β: of eigenvalues in an interval
α, β: , is computed using a replica path-integral formulation The CGF is obtained by evaluating a saddle-point calculation in the large-N limit For the intermediate phase 1 < γ < 2, the action of the path integral is expanded up to order σ, leading to a form that coincides with that of the standard GRP model at leading order in σ Universality and Scaling Functions The key finding is that for an interval of size O(ET), the CGF takes a simple scaling form F
-E,E: (s) = pa(0)ET G(s, E/ET) + O(σ 2) This scaling function G is universal across all variations of the model up to non-universal rescaling Furthermore, the level compressibility χT (y) for energies of order ET follows the universal scaling function χT (E/ET) given in Eq.
Mira: (85) Numerical Validation and QREM Comparison Numerical tests using exact diagonalization confirm the analytical predictions, showing that level compressibility collapses onto the universal prediction for energies of order ET However, when testing the Quantum Random Energy Model (QREM), the level compressibility exhibits a distinct crossover function that differs quantitatively from the GRP model's universal scaling function This discrepancy is attributed to the QREM eigenstates being genuinely multifractal rather than simply fractal The results suggest that the universality established in sections 3, 4, and 5 does not extend directly to the QREM because its eigenstates possess a more complex multifractal structure Nevertheless, a well-defined scaling regime still emerges in the intermediate non-ergodic extended phase of the QREM This suggests that random-matrix models may capture fundamental aspects of spectral statistics in this crossover regime The paper motivates future investigations into models displaying multifractality, such as random Cantor sets and correlations between matrix elements.
Kai: Appendix Results In Appendix A, the CGF for energies E ∼ O(1) is derived using a saddle-point calculation that generalizes the method to access energies outside the O(σ) scale This derivation yields a complex expression for the CGF that depends explicitly on pa(a), but still results in a scaling form when appropriately rescaled The full counting statistics for independent energy levels are also computed analytically, showing convergence to the Poisson rate function in the limit N → ∞ Numerical Results Summary Numerical simulations show that for medium-sized matrices, probabilities as low as 10−40 can be reached using large-deviation algorithms The empirical rate function for the QREM shows a logarithmic divergence as a function of N at k = 0, which is consistent with theoretical predictions The level compressibility in the QREM also exhibits a collapse onto the universal scaling function in the intermediate phase, although its specific scaling function differs from that of GRP models The study confirms that this universality is restricted to random-matrix ensembles whose eigenstates are characterized by a single fractal dimension The analysis also reveals that the crossover between RMT and Poisson statistics in the QREM is substantially broader than predicted by Eq.
Paper summary: Kai: So to wrap up, this paper on "Level statistics in the fractal phase of generalized Rosenzweig--Porter models" shows that even when you look at more complex generalizations of the RP model, there's still a simple universal scaling form for the full counting statistics around the Thouless energy ET >
Mira: The core claim is that this scaling function G is universal across different types of matrix constructions, which is pretty compelling because it suggests a fundamental mathematical structure governing how systems transition between integrable and chaotic regimes >
Lev: From an engineering standpoint, if we can trust this universal scaling form for the level compressibility, it gives us a baseline to test our error correction strategies when dealing with these many-body localization phenomena >
Kai: But we also have to keep in mind what the paper flags as a limitation. They admit that this universality doesn't extend directly to the QREM because its eigenstates have a more complex multifractal structure >
Mira: So, the implication is that while random matrix models capture some essential spectral behavior in this crossover, we still need to look at systems with genuinely multifractal states if we want a complete picture of what happens in real interacting systems >
Lev: It’s motivating for us to look into models that show multifractality more closely, like correlations between matrix elements, because the RP models give us a good starting point but stop short where the complexity gets too high for this simple scaling to hold >
Conclusion: Kai: So, this paper looks at level statistics in generalized Rosenzweig--Porter models, specifically focusing on that fractal phase where things get weird between chaotic and localized states.
Mira: Right, it’s about how you use these random matrix ensembles to map out that crossover scale using the Thouless energy ET.
Lev: So what they did was take these generalized RP models, which have this fractal phase, and they used free probability theory and the replica method to compute full counting statistics.
Kai: That means they figured out the distribution of eigenvalues in a certain window around that crossover energy, but they found a universal scaling form across all those different ways you can generalize the model.
Mira: Exactly, so it’s not just about one specific matrix construction; the scaling function for the cumulant generating function stays consistent.
Lev: But they did have to be careful with their assumptions; they looked at cases where gamma was between one and two, which is pretty specific. And they noted that this universality doesn't quite hold for the Quantum Random Energy Model because its states are genuinely multifractal.
Kai: So, the main result is that even in these complex systems, there’s a simple way to describe what’s happening around that energy scale ET, but the QREM shows us where those simple models stop working.
Mira: It suggests random matrix models are good at capturing the fundamental spectral structure in this transition zone, but you need to account for more complex multifractality when dealing with systems like the QREM.
Lev: So what this means for hardware is that if we use a simpler RP model, we get a reliable prediction near ET, but if our real system has that extra complexity—that multifractality—we’ll have to look at different scaling functions.
Kai: It opens the door for us to look at new models, like random Cantor sets or correlations between matrix elements, because these are where we need to go next.
More episodes
- 2610.32678-Probing the limits of materials design via a comprehensive crystal database of 14.6 million structures with 140 properties
- 2610.33595-Native-oxide-driven layer asymmetry, two-step switching and exchange bias in Co/Ir/Co trilayers
- 2503.19572-Spin models from nonlinear cellular automata
- 2606.23350-Irreversibility Enhances Quantum-Enhanced Markov-Chain Monte Carlo
- 2608.02744-Two-Parameter Ansatz for the Violation of Eigenstate Thermalization
- 2210.00806-Low- and high-energy localization landscapes for tight-binding Hamiltonians in two-dimensional lattices
- 2409.09625-Selective switch for band-edge alignments in pentagonal CdSeTe monolayer: Atom-valley locking
- 2606.20168-Norms, overlaps and Yangian descendants for the Haldane-Shastry spin chain
- 2512.19503-Strain-induced splitting of the CCDW-NCCDW phase transition in 1T-TaS 2
- 2307.12638-Strong localization of microwaves beyond two dimensions in aperiodic Vogel spirals