[Paper Review] Random Current Representation for Transverse Field Ising Models

Nicholas Crawford, Dmitry Ioffe|ArXiv.org|Dec 28, 2008

Markov Chains and Monte Carlo Methods10 references4 citations

TL;DR

This paper introduces a space-time random current representation (RCR) for the transverse field Ising model (TFIM), extending classical RCR techniques to quantum systems. It proves exponential decay of truncated two-point correlations at positive longitudinal magnetic fields (h > 0), establishing uniform bounds in inverse temperature β and system size N, and derives differential inequalities implying sharpness of quantum phase transitions, particularly in the ground state limit (β → ∞).

ABSTRACT

Recently, a random current representation for transverse field Ising models has been introduced in \cite{ILN}. This representation is a space-time version of the classical random current representation exploited by Aizenman et. al. %It is a space-time version of the classical random current representation \cite{Ai82, ABF, AF}. In this paper we formulate and prove corresponding space-time versions of the classical switching lemma and show how they generate various correlation inequalities. In particular we prove exponential decay of truncated two-point functions at positive magnetic fields in $\sfz$-direction and address the issue of the sharpness of phase transition.

Motivation & Objective

• To develop a stochastic geometric framework—space-time random current representation—for quantum spin systems, specifically the transverse field Ising model (TFIM).
• To extend classical correlation inequalities, such as the switching lemma, to the quantum setting via a space-time formulation.
• To establish uniform exponential decay of truncated two-point functions in the presence of a positive longitudinal magnetic field (h > 0), independent of β and system size N.
• To derive differential inequalities for the z-magnetization that imply sharpness of the quantum phase transition in the ground state (β = ∞).

Proposed method

• Formalizes the transverse field Ising Hamiltonian on a d-dimensional lattice torus Z^d with finite-range, translation-invariant interactions and parameters h (longitudinal field), λ (transverse field), ρ (interaction strength), and β (inverse temperature).
• Introduces a space-time random current representation by mapping the quantum partition function to a stochastic process involving Poisson processes and random currents on a space-time lattice.
• Defines pivotal and loop-pivotal events for currents to express truncated correlation functions as expectations over configurations with specific connectivity constraints.
• Applies a generalized switching lemma in the space-time setting to relate configurations with different boundary conditions and derive correlation bounds.
• Uses stochastic geometric arguments and path-counting techniques to bound the probability of rare events (e.g., absence of arrivals on certain sets), leading to exponential decay estimates.
• Derives differential inequalities for the z-magnetization M by analyzing its dependence on h, ρ, and λ, leveraging the uniformity of bounds in β and N.

Experimental results

Research questions

• RQ1 Can a space-time version of the random current representation be formulated for the transverse field Ising model to enable stochastic geometric analysis of quantum correlations?
• RQ2 Does the truncated two-point function of σ^z operators decay exponentially in space at positive h > 0, uniformly in β < ∞ and N?
• RQ3 Can differential inequalities for the z-magnetization be derived that imply the sharpness of the quantum phase transition in the ground state?
• RQ4 How does the random current representation facilitate the derivation of correlation bounds and phase transition properties in the limit β → ∞?
• RQ5 What is the role of the transverse field λ and longitudinal field h in controlling the decay rate and critical behavior of the system?

Key findings

• For every h > 0, λ ≥ 0, ρ ≥ 0, there exist positive constants c₁(h,λ,ρ) and c₂(h,λ,ρ) such that the truncated two-point functions satisfy |⟨σ^z_i;σ^z_j⟩| ≤ c₂ e^{-c₁|i−j|}, uniformly in β < ∞ and N.
• The truncated x-spin correlations also decay exponentially: 0 ≤ ⟨Σ^x_i;Σ^x_j⟩ ≤ c₂ e^{-c₁|i−j||, and the mixed zx-correlations are bounded below by −c₂ e^{-c₁|i−j|}.
• The z-magnetization M satisfies the differential inequality: M ≤ h ∂M/∂h + M³ + M²ρ ∂M/∂ρ − 2λM² ∂M/∂λ, uniformly in β < ∞ and N.
• Additional differential inequalities hold: −∂M/∂λ ≤ (M/(1−M²)) ∂M/∂h and ∂M/∂ρ ≤ J̄M ∂M/∂h, uniformly in β < ∞ and N.
• The exponential decay of correlations is uniform in β, so the bounds extend to the ground state (β = ∞), and the differential inequalities remain valid in this limit.
• The derived inequalities imply the sharpness of the phase transition in the z-magnetization as a function of λ or β, even in the quantum ground state, via methods analogous to those in classical statistical mechanics.