Nonequilibrium Statistics of Biased Kondo Resonance

Basic Information

• Paper ID: 2503.14400
• Title: Nonequilibrium Statistics of Biased Kondo Resonance
• Author: Jong E. Han (State University of New York at Buffalo)
• Classification: cond-mat.str-el (Strongly Correlated Electron Systems)
• Publication Date: October 14, 2025
• Paper Link: https://arxiv.org/abs/2503.14400

Abstract

This paper formulates a numerical renormalization group (NRG) method for nonequilibrium steady states by converting finite-lattice many-body eigenstates into scattering states. By extending the full density matrix NRG to the biased Anderson impurity model using the primitive orbital basis as the Hamiltonian representation, the paper enables detailed investigation of sub-Kondo spectral evolution in the zero-temperature limit, confirming the dual-resonance structure under bias at the Kondo energy scale $T_K$. The distributions reveal unique multiscale spectral features below the Kondo scale ($\omega \lesssim T_K$) and near the bias ($\omega \gtrsim V$), leading to a nonequilibrium local temperature $T_{\rm loc}$ at the Kondo quantum dot that scales as $k_BT_{\rm loc} \approx V$ when $V \gg T_K$. In the low-temperature limit ($T \ll T_K$), the current-voltage relationship deviates from the unitary limit and reaches a current saturation regime when the bias exceeds the Kondo scale ($V/2 \gtrsim T_K$).

Research Background and Motivation

Problem Description

1. Core Challenge: Solving nonequilibrium statistical mechanics in biased quantum systems has been a tremendous challenge in recent decades, particularly for strongly correlated quantum impurity models.
2. Significance:
   • The Kondo problem serves as a testing ground for strongly correlated nonequilibrium techniques, crucial for understanding nonequilibrium behavior in quantum many-body systems
   • Quantum impurity models play a key role in solving the condensed matter limit through dynamical mean-field theory
3. Limitations of Existing Methods:
   • The theoretical rigor and numerical efficiency of nonequilibrium impurity solvers remain unsatisfactory
   • Computational difficulties limit understanding of how the Kondo singlet decomposes into sub-Kondo regions with dual peaks under bias
   • Nonequilibrium statistical understanding of spectra lags behind spectral features in the model
4. Research Motivation:
   • Construct an accessible steady-state nonequilibrium extension based on the mature NRG method
   • Provide clear understanding of Kondo resonance evolution under bias
   • Develop a general method applicable to complex impurity interactions

Core Contributions

1. Methodological Innovation: Proposes a novel nonequilibrium NRG method that handles steady-state nonequilibrium problems by converting finite-lattice eigenstates into scattering states
2. Technical Simplification: Uses primitive orbital basis representation of the Hamiltonian, simplifying implementation and making it more suitable for extension to generalized problems
3. Physical Discoveries:
   • Confirms the splitting behavior of the Kondo peak at $V/2 \approx T_K$
   • Discovers multiscale distribution features and population inversion phenomena
   • Reveals the scaling relationship of nonequilibrium temperature: $T_{\rm loc} \approx V$
4. Transport Properties: Verifies current saturation in the low-temperature limit for the region $T_K \lesssim V/2 \lesssim 3T_K$

Methodology Details

Theoretical Framework

Anderson Impurity Model

The Hamiltonian is defined as: $\hat{H} = \sum_\sigma \epsilon_d n_{d\sigma} + U\left(n_{d\uparrow} - \frac{1}{2}\right)\left(n_{d\downarrow} - \frac{1}{2}\right) + \sum_{\alpha k\sigma} \epsilon_k c^\dagger_{\alpha k\sigma} c_{\alpha k\sigma} - \frac{t}{\sqrt{\Omega}} \sum_{\alpha k\sigma} (d^\dagger_\sigma c_{\alpha k\sigma} + c^\dagger_{\alpha k\sigma} d_\sigma)$

where $\alpha = L,R$ denotes source and drain reservoirs, $U$ is the Coulomb repulsion parameter, and $\epsilon_d$ is the impurity level energy.

Nonequilibrium Density Matrix

The steady-state nonequilibrium density matrix is constructed as: $\hat{\rho}_{\rm neq} = \frac{1}{Z_{\rm neq}} e^{-\beta(\hat{H} - V\hat{Y})}$

where Hershfield's Y operator is defined as: $\hat{Y} = \frac{1}{2}\sum_{k\sigma}(\psi^\dagger_{Lk\sigma}\psi_{Lk\sigma} - \psi^\dagger_{Rk\sigma}\psi_{Rk\sigma})$

Scattering State Construction

The key innovation lies in constructing the Y operator through: $\hat{Y} = \frac{i\eta}{-\mathcal{L} + i\eta}\hat{Y}_0$

where $\mathcal{L}$ is the Liouvillian operator and $\hat{Y}_0$ is the initial Y operator when the impurity is decoupled. In the energy eigenstate representation: $\langle\alpha|\hat{Y}|\beta\rangle = \frac{i\eta}{-E_\alpha + E_\beta + i\eta}\langle\alpha|\hat{Y}_0|\beta\rangle$

Numerical Implementation

Wilson Chain Representation

• Wilson chain (WC) simulates the reservoir with chain length $N_{\rm WC} = 100$ and discretization parameter $\Lambda = 1.2$
• Environmental orbital treatment: considers free energy changes of environmental orbitals under bias
• Density matrix weight peaks at Wilson chain length $n = M$, satisfying $D_M \approx V/2$

Full Density Matrix NRG (FDM-NRG)

• Retained state number: $N_{\rm kept} = 2000$
• z-averaging uses 8 z-values to reduce Wilson chain discreteness
• Hilbert space dimension per NRG iteration: $16 \times N_{\rm kept}$

Experimental Setup

Model Parameters

• Strong Kondo Regime: $U/\Gamma = 10$, $\Gamma = 0.1$, $T_K = 0.0044$
• Particle-Hole Symmetry: $\epsilon_d = 0$
• Particle-Hole Asymmetry: $\epsilon_d = U/4$
• Temperature: $T = 0.1T_K$ (low-temperature limit)

Computational Parameters

• Discretization Parameter: $\Lambda = 4$
• Spectral Broadening Parameter: $\delta = 0.2 \times \max(T_K, T, V/2)$
• Infinitesimal Parameter: $\eta = 0.1 \times \max(T_K, T, V/2)$

Evaluation Metrics

• Spectral Function: $\rho_\Sigma(\omega)$ (computed using self-energy technique)
• Distribution Function: $f(\omega) = G^<_d(\omega)/[2\pi i\rho(\omega)]$
• Nonequilibrium Temperature: $T_{\rm loc}^2 = \frac{6}{\pi^2}\int_{-\infty}^{\infty} \omega[f(\omega) - \Theta(-\omega)]d\omega$
• Current: Calculated via graphical techniques

Experimental Results

Spectral Evolution Analysis

Non-interacting Verification

• Validates method correctness in the non-interacting resonant level model
• Small bias: current follows $I = G_0 V$ with $G_0 = 1/(2\pi)$ as unitary quantum conductance
• Large bias: conforms to exact relation $I = (\Gamma/\pi)\tan^{-1}(V/2\Gamma)$

Kondo Resonance Splitting

1. Zero Bias: Spectral function satisfies unitary limit at $\rho_\Sigma(0) \approx 1/(\pi\Gamma)$
2. Critical Bias: When $V/2 \approx T_K$, Kondo resonance begins showing peak splitting
3. High Bias: Peak positions align at $\pm V/2$, but linewidth far exceeds $T_K$, should be viewed as weakly correlated

Nonequilibrium Statistical Features

Distribution Function Evolution

• Low Bias: Distribution function approaches Fermi-Dirac function
• Moderate Bias: Develops large, broad peaks at $|ω| \gtrsim V$
• Population Inversion: Exhibits $f'(0) < 0$ between split Kondo peaks ($-V/2 \lesssim \omega \lesssim V/2$)

Nonequilibrium Temperature Scaling

Key finding: In the high-bias limit $T_{\rm loc} \approx V \quad \text{for} \quad V/2 \gg T_K$

This indicates that local electron temperature is proportional to the applied bias.

Transport Properties

Current-Voltage Relationship

1. Low-Temperature Limit ($T = 0.1T_K$):
   • Satisfies unitary limit for small bias: $G_0 = (2\pi)^{-1}$
   • Begins deviating from unitary limit when $V/2 \gtrsim T_K$
   • Current saturation appears in region $T_K \lesssim V/2 \lesssim 3T_K$
2. High Temperature ($T = T_K$):
   • No current saturation observed
   • Demonstrates that current saturation originates from strong correlation effects

Related Work

Nonequilibrium Quantum Impurity Methods

The paper systematically reviews various nonequilibrium techniques for treating the Kondo problem:

• Analytical Theory: Bethe ansatz methods
• Graphical Methods: Landauer formula, perturbation theory
• Numerical Methods: Real-time diagrammatic simulations, imaginary-time simulations, tensor network methods
• Renormalization Group: Real-time RG, perturbative RG, functional RG

Comparison with Existing NRG Methods

• SNRG Method: Steady-state NRG proposed by Anders using non-interacting scattering states to construct $\hat{Y}_0$
• Present Method: Uses primitive Hamiltonian basis, more direct implementation, suitable for extension to complex impurity interactions

Conclusions and Discussion

Main Conclusions

1. Method Validation: Successfully combines Hershfield scattering state theory with NRG method, providing a reliable numerical tool for nonequilibrium strongly correlated systems
2. Physical Picture: Confirms Kondo peak splitting at $V/2 \approx T_K$, revealing complex multiscale structure in the sub-Kondo region
3. Nonequilibrium Features: Discovers population inversion and linear scaling relationship of nonequilibrium temperature, providing new insights into understanding strongly correlated nonequilibrium states
4. Transport Phenomena: Verifies current saturation phenomenon, confirming theoretical predictions

Limitations

1. Computational Cost: Rapid Hilbert space growth due to handling two Wilson chains limits the number of retained states
2. Parameter Dependence: Method is sensitive to choice of infinitesimal parameter $\eta$, requiring careful tuning
3. Temperature Constraints: Primarily applicable to low-temperature regime $T \ll T_K$

Future Directions

1. Extended Applications: Applicable to complex impurity interaction models
2. Lattice Theory: Application to nonequilibrium lattice models through dynamical mean-field theory
3. Experimental Comparison: Detailed comparison with quantum dot transport experiments

In-Depth Evaluation

Strengths

1. Theoretical Rigor: Based on Hershfield's rigorous theoretical framework with clear mathematical derivations
2. Methodological Innovation: Cleverly converts finite-lattice eigenstates into scattering states, avoiding complexity of explicit scattering state construction
3. Numerical Reliability: Verified through non-interacting limit, ensuring method correctness
4. Physical Insights: Provides deep understanding of nonequilibrium evolution of Kondo resonance

Weaknesses

1. Computational Complexity: Significantly increased computational cost compared to equilibrium NRG
2. Parameter Tuning: Choice of $\eta$ parameter requires experience, potentially affecting result universality
3. Experimental Verification: Lacks direct comparison with actual quantum dot experiments

Impact

1. Methodological Contribution: Provides new numerical tool for nonequilibrium strongly correlated systems
2. Theoretical Value: Deepens understanding of nonequilibrium behavior of Kondo effect
3. Application Prospects: Extensible to more complex impurity models and lattice systems

Applicable Scenarios

• Theoretical analysis of quantum dot transport experiments
• Nonequilibrium research in strongly correlated electron systems
• Impurity solver for dynamical mean-field theory
• Theoretical modeling of molecular junction electronics

References

The paper cites important literature in the field, including:

• Original Kondo work and Hewson's classical textbook
• Hershfield's nonequilibrium steady-state theory
• Wilson's numerical renormalization group method
• Anders' steady-state NRG method
• Various nonequilibrium quantum impurity theories and numerical methods

This paper makes important contributions to the field of nonequilibrium quantum many-body theory, providing both a rigorous and practical numerical method for studying strongly correlated nonequilibrium systems, with particular value for understanding the nonequilibrium behavior of the Kondo effect.