REMOVING THE DIVERGENCE AT THE KONDO TEMPERATURE BY MEANS OF SELF-CONSISTENT PERTURBATION THEORY

R. D. Mattuck, Lars Kai Hansen, C.y. Cheung1971 · DOI

TL;DR

This paper tackles a long-standing issue in condensed matter physics: a spurious divergence in theoretical predictions for the Kondo effect at a specific temperature ($T_K$). By employing a self-consistent approach to "dress" electron propagators, the authors demonstrate that this divergence is an artifact of earlier methods and that perturbation theory remains valid down to very low temperatures. This work resolves a theoretical puzzle and provides a more robust framework for understanding the Kondo effect.

What Problem Does This Paper Address?

The Kondo effect describes the phenomenon where magnetic impurities in a metal lead to a peculiar increase in electrical resistance as temperature decreases. A key theoretical challenge in understanding this effect has been the appearance of a divergence in the calculated scattering amplitude (a measure of how electrons interact with the impurity) at a characteristic temperature, $T_K$. This divergence implied a breakdown of the theoretical framework (perturbation theory) and suggested a transition to a bound state at $T_K$. However, other theoretical approaches, like those using equation-of-motion methods, did not show this divergence, creating a discrepancy. Furthermore, ground-state calculations indicated a bound state only at absolute zero ($T=0$), not at $T_K$. This paper aims to reconcile these findings by investigating the origin of the divergence and proposing a method to remove it.

Methods

The authors employ a self-consistent perturbation theory approach. The core idea is to go beyond using "bare" electron propagators (which represent electrons in their simplest, unperturbed state) and instead use "clothed" propagators. These clothed propagators account for the electron's interactions with the impurity and other electrons in a self-consistent manner. This is achieved by iteratively calculating and updating the electron's self-energy and the scattering amplitude until a stable solution is reached. Specifically, they extend Abrikosov's "parquet diagram" summation by including non-parquet diagrams through this self-consistent clothing of the s-electron propagators. They also explore the effect of clothing the d-electron propagator and introducing a finite width for the d-electron energy level.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • The $T_K$ divergence is an artifact: The paper convincingly shows that the divergence at the Kondo temperature ($T_K$) is not a physical phenomenon but rather a consequence of using bare electron propagators in the calculation.
  • Self-consistent clothing is key: By self-consistently "dressing" the electron propagators, the authors effectively sum over a broader set of diagrams, leading to a stable and non-divergent scattering amplitude at $T_K$.
  • Perturbation theory is valid down to $T \approx 0$: With the corrected approach, perturbation theory remains a valid tool for describing the Kondo effect at temperatures just above absolute zero, resolving a major theoretical hurdle.
  • Bound state at $T=0$ persists: While the divergence at $T_K$ is removed, the theory still predicts a divergence at $T=0$, consistent with the existence of a bound state at absolute zero, as suggested by other studies.
  • Resistance matches Hamann's form: The calculated electrical resistance closely resembles the form derived by Hamann, suggesting the physical outcome of the Kondo effect is captured.
  • Finite d-level width is important for low-T behavior: Introducing a finite width to the d-electron energy level is shown to be crucial for accurately describing the resistance at very low temperatures, bringing the theoretical predictions closer to experimental observations.

Key Results

  • The spurious divergence of the s-d scattering amplitude at $T_K$ is removed by using self-consistently clothed electron propagators.
  • Perturbation theory is shown to be valid for temperatures down to just above absolute zero.
  • A divergence at $T=0$ is still present, indicating a bound state at absolute zero.
  • The calculated electrical resistance takes the form derived by Hamann.
  • Giving the d-electron energy level a finite width improves the agreement of the resistance curve with experimental data at low temperatures, particularly by avoiding a negative infinite slope at $T=0$.

Limitations and Future Work

The authors acknowledge that their approximation of treating the d-electron spectral density as a delta function is not valid at very low temperatures ($T < W$, where $W$ is the d-level width). This leads to a disagreement with experimental results in that regime, where the calculated resistance goes to zero while experiments show a finite or even positive slope. They suggest that a more complete solution is required to accurately describe this low-temperature behavior. Future work could involve a more rigorous treatment of the d-electron's finite width and its impact on the self-consistent equations.

Summary updated

Abstract

Previous applications of perturbation theory to the Kondo problem have shown a divergence at the Kondo temperature, TK. We show here that this divergence is spurious and comes from the use of bare electron propagators in the calculation. When the propagators are self-consistently clothed, there is no divergence at TK, and perturbation theory is valid all the way down to just above zero temperature. However, there is a divergence at T = 0, indicating a bound state there. Our resistance has the Hamann form.
BibTeX
@inproceedings{mattuck1971removing,
  author = {R. D. Mattuck and Lars Kai Hansen and C.Y. Cheung},
  title = {{REMOVING THE DIVERGENCE AT THE KONDO TEMPERATURE BY MEANS OF SELF-CONSISTENT PERTURBATION THEORY}},
  booktitle = {Le},
  year = {1971},
  doi = {10.1051/jphyscol:19711151},
}
Figure