Abstract
This research extends the classical theory of Large Deviations into the quantum realm, specifically targeting lattice systems such as quantum spins and Fermi gases. By moving beyond traditional cluster expansions and utilizing a functional-analytic approach involving matrix inequalities and analyticity estimates, this work establishes the existence of a logarithmic moment generating function in the "one-phase region." Specifically, results are proven for high-temperature regimes in any dimension and for one-dimensional systems at any temperature. These findings provide a rigorous foundation for calculating upper bounds on the probability of rare macroscopic fluctuations in quantum systems.
Introduction
Large Deviation Theory (LDT) is a cornerstone of classical statistical mechanics, providing the mathematical tools necessary to understand the probability of rare events—specifically, how likely a macroscopic average is to deviate significantly from its expected value. While LDT is well-established for classical systems, its application to quantum mechanics has historically been hindered by the fundamental property of non-commutativity.
In classical physics, observables are commuting variables, allowing for a direct relationship between probability distributions and thermodynamic pressures. In quantum systems, however, macroscopic observables (such as total magnetization) generally do not commute with the system's Hamiltonian. This breakdown of commutativity renders standard classical mathematical techniques inapplicable, previously limiting quantum LDT results to highly specific cases like ideal gases or extremely dilute systems.
Methodology
To address the challenges posed by non-commutativity, this paper adopts a functional-analytic approach rather than the mathematically cumbersome cluster expansion techniques often used in the field. The methodology is built upon three primary pillars:
- Matrix Inequalities: The authors derive a specific inequality (Lemma 3.6) to bound the difference between the traces of non-commuting operators. This is crucial for managing the "imaginary time" evolution of quantum observables.
- Analyticity Estimates: Utilizing established results from Ruelle (for high temperatures) and Araki (for 1D systems), the authors demonstrate that the dynamics of the system are analytic within a specific complex strip, ensuring the system remains "well-behaved."
- Subadditivity Arguments: To prove that macroscopic limits exist as the system size approaches infinity (the thermodynamic limit), the authors employ subadditivity. This shows that the properties of a large system can be bounded by the sum of its parts, with controllable error terms arising from the boundaries.
The One-Phase Region and Quantum Equilibrium
The study focuses on the "one-phase region," a regime where the system is disordered and lacks phase transitions. The mathematical advantage of this region is that correlations between different parts of the system decay rapidly, making the derivation of macroscopic limits tractable. Furthermore, the authors utilize the Gibbs-KMS (Kubo-Martin-Schwinger) condition, a sophisticated mathematical definition of thermal equilibrium for infinite quantum systems, to ensure their results remain valid as the number of particles goes to infinity.
Key Results
The central result of this research is the rigorous proof that the logarithmic moment generating function $f(\alpha)$ exists and is finite in the following regimes:
- High Temperature: In any dimension, provided the inverse temperature $\beta$ is sufficiently small and the deviation $\alpha$ is relatively small.
- One Dimension: At any temperature, for any value of $\alpha$, provided the interactions have a finite range.
By establishing the existence of this function, the authors provide Large Deviation Principle (LDP) upper bounds. This allows for the rigorous calculation of the maximum probability that a quantum system will fluctuate away from its average state. In specific cases where observables do commute with the energy (such as particle density), the authors obtain a full Large Deviation Principle, matching the strength of classical statistical mechanics results.
Limitations and Future Research
While this work represents a significant step forward, several open questions remain:
- The Smoothness Gap: A full LDP (including lower bounds) requires proving that $f(\alpha)$ is $C^1$ differentiable. While conjectured to be true in 1D, this remains unproven in the current work.
- Phase Transitions: The results are explicitly restricted to the one-phase region. Understanding large deviations near quantum phase transitions, where systems become highly ordered and correlations do not decay quickly, remains a major challenge.
- Bosonic Systems: This framework is currently limited to spins and fermions. Extending these results to bosonic systems is technically more difficult because bosonic operators are "unbounded," requiring a shift toward $W^*$-algebraic mathematical frameworks.
Summary updated