Exact bounds for dynamical critical exponents of transverse-field Ising chains with a correlated disorder

Basic Information

• Paper ID: 2007.07439
• Title: Exact bounds for dynamical critical exponents of transverse-field Ising chains with a correlated disorder
• Authors: Tatsuhiko Shirai (Waseda University), Shu Tanaka (Keio University)
• Classification: quant-ph cond-mat.dis-nn
• Publication Date: April 7, 2021 (arXiv preprint)
• Paper Link: https://arxiv.org/abs/2007.07439

Abstract

This study investigates the dynamical critical exponents of disordered Ising chains under a transverse field to examine the effects of correlated disorder on quantum phase transitions. Correlated disorder refers to the case where the transverse field strength at a site depends on the nearest-neighbor coupling strength connected to that site, which yields qualitatively different results compared to uncorrelated disorder. In the uncorrelated disorder case (where the transverse field is uniform or randomly chosen independently of the nearest-neighbor coupling strength), the dynamical critical exponent is infinite. In contrast, in the presence of correlated disorder, the authors analytically prove that the dynamical critical exponent is finite and depends on the tuning process of the transverse field strength.

Research Background and Motivation

Problem Definition

1. Core Problem: Investigate the effects of correlated disorder on the properties of quantum phase transitions (QPT) in transverse-field Ising chains, particularly the behavior of the dynamical critical exponent z.
2. Problem Significance:
   • Quantum phase transitions are important phenomena in condensed matter physics and statistical physics, occurring at absolute zero temperature
   • The dynamical critical exponent z determines the relationship between the characteristic energy scale and length scale near the critical point: Δ ∼ ξ^(-z)
   • Understanding the effects of disorder on quantum phase transitions is crucial for applications such as quantum adiabatic computation
3. Limitations of Existing Methods:
   • In the uncorrelated disorder case, due to the existence of Griffiths-McCoy singularities, the dynamical critical exponent z = ∞
   • This divergent z value limits the performance of quantum adiabatic algorithms
   • Lack of exact analytical results for the correlated disorder case
4. Research Motivation: Explore suppressing the z value through introducing correlated disorder, reducing it from infinity to a finite value, thereby improving quantum computational performance.

Core Contributions

1. First Analytical Proof: In the correlated disorder case, the dynamical critical exponent z is finite, forming a sharp contrast with z = ∞ in the uncorrelated disorder case.
2. Exact Bound Results:
   • Weak disorder case: z = 1 (same as the disorder-free case)
   • Strong disorder case: max(D(1/2 + |s-1/2|) + 1/2, 1) ≤ z ≤ D + 1
3. Parameter Dependence Analysis: Demonstrates that z depends on the tuning parameter s and disorder strength D, providing a theoretical basis for controlling the z value.
4. Numerical Verification: Theoretical predictions are verified through exact diagonalization methods, particularly for boundary behavior in the strong disorder case.

Methodology Details

Task Definition

Study the quantum phase transition properties of a one-dimensional N-spin Ising chain under a transverse field, with Hamiltonian:

H = -∑[i=1 to N-1] J_i σ_i^z σ_{i+1}^z - ∑[i=1 to N] Γ_i σ_i^x

where correlated disorder is defined through the following relationship:

ln(Γ_1/Γ) = (1-s)ln(J_1)
ln(Γ_i/Γ) = s ln(J_{i-1}) + (1-s)ln(J_i)  (2 ≤ i ≤ N-1)
ln(Γ_N/Γ) = s ln(J_{N-1})

Model Architecture

1. Disorder Distribution Model

• Weak Disorder: Uniform distribution π_w(J_i) = (1-J^(0))^(-1) for J^(0) < J_i ≤ 1
• Strong Disorder: Power-law distribution π_s(J_i) = (1/D)J_i^(-1+1/D) for 0 < J_i ≤ 1

2. Fermionization Mapping

Map the spin system to a free fermionic system via Jordan-Wigner transformation:

H = ∑∑ [c_i† [A]_{ij} c_i + [B]_{ij}/2 (c_i† c_j† + c_j c_i)] + const.

3. Eigenvalue Problem

The energy gap problem reduces to a matrix eigenvalue problem:

M φ_n = Λ_n² φ_n

where M = (A-B)(A+B), and the energy gap Δ = Λ_1.

Technical Innovations

1. Correlated Disorder Design

Unlike conventional uncorrelated disorder, the transverse field strength Γ_i depends on the adjacent coupling strengths J_ and J_i, with parameter s controlling this dependence.

2. Exact Bound Method

• Lower Bound: Obtained through matrix inequalities M ≥ M_0 and Weyl's monotonicity theorem
• Upper Bound: Derived using the Rayleigh-Ritz variational principle and carefully constructed trial wavefunctions

3. Moment Estimation Technique

Determine z by analyzing the system size dependence of the m-th moment of the energy gap Δ^m_ave:

[Δ^m]_ave ∼ N^(-mz)

Experimental Setup

Dataset

• System Sizes: N = 10² to 10⁴
• Disorder Realizations: Each data point averaged over 1000 disorder realizations
• Parameter Range: s ∈ 0,1, D = 1 (main results)

Evaluation Metrics

• Dynamical critical exponent z determined from power-law scaling behavior of the energy gap
• Critical exponents ν and ν̄ for the correlation length ξ and typical correlation length ξ̄

Comparison Methods

• Comparison with uncorrelated disorder case (z = ∞)
• Comparison with disorder-free case (z = 1)
• Comparison of theoretical bounds with numerical results

Implementation Details

• Use exact diagonalization to compute eigenvalues of matrix M
• Critical point determination: Γ_c = 1 (via condition ΔJ = ΔΓ)
• Fitting range: N ≥ 10³ and 10² ≤ N ≤ 10³

Experimental Results

Main Results

1. Weak Disorder Case

• Theoretical Result: z = 1 (exact)
• Physical Meaning: Same as the disorder-free case; correlated disorder does not alter critical behavior

2. Strong Disorder Case

• Theoretical Bounds: max(D(1/2 + |s-1/2|) + 1/2, 1) ≤ z ≤ D + 1
• Numerical Results: At D = 1, z is minimized at s = 0.5, and z ≈ 1 + |s-1/2| for s ≤ 0.3 and s ≥ 0.7

3. Parameter Dependence

• z is symmetric at s and 1-s
• z reaches minimum value at s = 0.5
• Strong disorder bounds are well verified at extreme s values

Ablation Studies

• s Parameter Effect: Finite-size effects are largest at s = 0.5, requiring larger systems for study
• Bound Tightness: Numerical results approach theoretical lower bounds when s is close to 0 or 1
• Correlation Length: ν = ν̄ = 1, differing from the uncorrelated disorder case

Experimental Findings

1. Qualitative Change in Critical Behavior

Correlated disorder causes z to drop from ∞ to a finite value, representing a qualitative change.

2. Tuning Mechanism

Parameter s enables effective tuning of the z value, providing a control mechanism for practical applications.

3. Unification of Correlation Lengths

The average correlation length and typical correlation length coincide in the correlated disorder case, eliminating sample-to-sample fluctuations.

Related Work

Main Research Directions

1. Disorder-Free Transverse-Field Ising Chain: z = 1, completely solvable
2. Uncorrelated Disorder Case: z = ∞, caused by Griffiths-McCoy singularities
3. Strong Disorder Renormalization Group: Theoretical framework for studying disordered quantum spin chains

Relationship to Related Work

• Distinguished from Fisher et al.'s uncorrelated disorder research
• Extends numerical work by Hoyos et al., providing exact analytical results
• Resonates with Knysh et al.'s research on quantum annealing acceleration

Comparative Advantages

• First to provide exact analytical bounds for the correlated disorder case
• Systematic study of parameter dependence
• Connects theoretical analysis with numerical verification

Conclusions and Discussion

Main Conclusions

1. Protective Effect of Correlated Disorder: Correlated disorder effectively suppresses Griffiths-McCoy singularities, reducing z from ∞ to a finite value
2. Parameter Tunability: By tuning s and D, the specific value of z can be controlled
3. Application Prospects: Finite z values help enhance quantum adiabatic computational performance

Limitations

1. One-Dimensional Restriction: Results apply only to one-dimensional systems
2. Specific Disorder Form: The particular form of correlated disorder may affect universality
3. Exact Values in Strong Disorder: The exact value of z cannot be determined in certain parameter regions

Future Directions

1. Higher-Dimensional Generalization: Study correlated disorder effects in d ≥ 2 dimensions
2. Other Disorder Types: Explore effects of different correlated disorder forms
3. Experimental Implementation: Seek experimental schemes for realizing correlated disorder

In-Depth Evaluation

Strengths

1. Theoretical Rigor: Provides exact mathematical proofs and rigorous bound analysis
2. Methodological Innovation: Cleverly combines matrix inequalities and variational methods to obtain tight bounds
3. Practical Value: Provides theoretical guidance for parameter optimization in quantum computing
4. Completeness: Theoretical analysis and numerical verification mutually validate each other

Weaknesses

1. Bound Gap: Gap remains between upper and lower bounds in the strong disorder case
2. Finite-Size Effects: Finite-size effects are significant in certain parameter regions
3. Physical Mechanism: Insufficient explanation of the microscopic mechanism by which correlated disorder suppresses singularities

Impact

1. Theoretical Contribution: Provides new perspective for theory of disordered quantum systems
2. Application Potential: Important guidance for quantum annealing and adiabatic computation
3. Methodological Value: Bound analysis methods can be generalized to other disordered systems

Applicable Scenarios

1. Quantum Annealing Optimization: Design quantum annealing protocols with finite z values
2. Disordered System Design: Engineer quantum systems with specific critical properties
3. Theoretical Research: Serve as methodological foundation for studying other disordered quantum systems

References

The paper cites important works in the field, including:

• Sachdev's theoretical foundations of quantum phase transitions
• Fisher's pioneering work on random transverse-field Ising chains
• Numerical studies of correlated disorder by Hoyos et al.
• Theoretical and applied research on quantum adiabatic computation

This paper makes important contributions at the intersection of theoretical physics and quantum computing, revealing through rigorous mathematical analysis the protective effects of correlated disorder on quantum phase transitions, and providing valuable theoretical guidance for practical applications.