Phase separation in the binary-alloy problem: The one-dimensional spinless Falicov-Kimball model

J. K. Freericks, C. Gruber, Nicolas MacrisPhysical 1996 · DOI

TL;DR

This paper investigates the ground states of the one-dimensional spinless Falicov-Kimball model, a simplified model for binary alloys. They discover that for certain electron and ion densities, the system undergoes phase separation into distinct configurations, deviating from the expected Peierls-like insulating state. This suggests a more complex ground state behavior than previously assumed, particularly at small coupling strengths.

What Problem Does This Paper Address?

The study of electronic properties in solids often relies on understanding how atoms arrange themselves to minimize the system's energy. In one-dimensional systems, a key concept is the Peierls instability, where a periodic distortion of the atomic lattice can open a gap in the electronic energy spectrum, leading to an insulating state and lowering the overall energy. This paper explores a similar phenomenon in the context of binary alloys, specifically using the Falicov-Kimball model. The core question is: under what conditions does the system favor a uniform, periodic arrangement of ions (leading to an insulating state) versus a phase-separated state where different regions of the material have distinct ionic arrangements? They focus on the "small-coupling limit" where the interaction between electrons and ions is weak, and investigate how electron and ion densities influence the ground state.

Methods

The authors employ nearly degenerate perturbation theory to analyze the ground-state energy of the Falicov-Kimball model. This approach is suitable for situations where the system's energy levels are close to each other. They expand the ground-state energy in a series of powers of the coupling strength ($U$), focusing on the leading corrections, particularly the $U^2 \ln U$ term. This term is crucial for understanding the stability of different configurations. They also utilize Rayleigh-Schroedinger perturbation theory and Wigner-Brillouin perturbation theory to handle degeneracies and singularities that arise in the calculations. The analysis involves calculating the structure factor of the ion configuration, which quantifies how ordered the ions are, and maximizing this factor to find the lowest energy state.

Datasets

Not discussed in the paper.

Key Insights and Ideas

  • Peierls Analogy Breaks Down: The conventional wisdom that a periodic ion arrangement (like in the Peierls instability) always leads to the lowest energy state (an insulator) is challenged. For certain densities, phase separation becomes more favorable.
  • Phase Separation is Common: The paper rigorously demonstrates that for small coupling ($U$), phase separation is a common ground state, occurring when ion densities deviate significantly from 1/2.
  • Density Dictates Behavior: The electron and ion densities are critical parameters. Specifically, the ratio of ion density to electron density plays a key role in determining whether the system remains in a homogeneous, periodic phase or separates.
  • Structure Factor as a Key: The maximization of the square of the ion structure factor ($|W(2\pi\rho_e, \{\omega_i\})|^2$) emerges as the central criterion for selecting the ground state in the small-$U$ limit.
  • Mixture of States: The phase-separated ground state is not simply a disordered arrangement but a mixture of distinct, well-defined phases: a metallic phase (like the empty lattice) and an insulating phase (a period-q ionic configuration).

Key Results

  • For rational electron densities $\rho_e = p/q$ and ion densities $\rho_i$ close to 1/2, the ground state is a period-q ionic configuration, analogous to the Peierls distortion.
  • However, when $\rho_i$ deviates significantly from 1/2 (specifically, outside the range $[p_c, 1-p_c]$ where $p_c \approx 0.371$), the system becomes unstable and phase separates.
  • The phase-separated ground state is a mixture of two phases: one with a period-q ionic configuration and another that is either an empty or full lattice. Both phases have the same electron density to leading order.
  • The stability of the pure period-q phase is determined by a condition (Eq. 24) related to the ion density and the electron density. If this condition is not met, phase separation occurs.
  • The analysis shows that for small $U$, the ground state is determined by maximizing the square of the structure factor.

Limitations and Future Work

  • Small Coupling Limit: The rigorous results presented are valid only for sufficiently small coupling ($U$). The authors acknowledge that for finite $U$, the phase separation might disappear, and numerical evidence suggests this is true for some densities.
  • Generalization: The study is restricted to the spinless Falicov-Kimball model. The authors suggest that the general ideas might extend to more complex models (e.g., tertiary alloys, Holstein model), but the analysis would become significantly more complicated.
  • Phase Transitions with U: The paper strongly suggests that the Falicov-Kimball model exhibits phase transitions as a function of $U$. For large $U$, the ground state is known to be either a segregated phase or the most homogeneous phase. The authors conjecture that in the neutral case, the Peierls-type state might not undergo phase transitions for $0 < U < \infty$, but they are unable to prove this.
  • Experimental Verification: The authors are not aware of any experiments that have directly observed this predicted phase separation. They suggest that it might be observable in quasi-one-dimensional metals and insulators.

Summary updated

Abstract

The ground states of the one-dimensional Falicov-Kimball model are investigated in the small-coupling limit, using nearly degenerate perturbation theory. For rational electron and ion densities, respectively, equal to p/q, ${\mathit{p}}_{\mathit{i}}$/q, with p relatively prime to q and ${\mathit{p}}_{\mathit{i}}$/q close enough to 1/2, we find that in the ground state the ion configuration has a period q. The situation is analogous to the Peierls instability, where the usual arguments predict a period-q state that produces a gap at the Fermi level and is insulating. However for ${\mathit{p}}_{\mathit{i}}$/q far enough from 1/2, this phase becomes unstable against phase separation. The ground state is a mixture of a period-q ionic configuration and an empty (or full) configuration, where both configurations have the same electron density to leading order. Combining these results with those previously obtained for strong coupling, it follows that a phase transition occurs in the ground state, as a function of the coupling, for ion densities far enough from 1/2. \textcopyright{} 1996 The American Physical Society.
BibTeX
@inproceedings{freericks1996phase,
  author = {J. K. Freericks and C. Gruber and Nicolas Macris},
  title = {{Phase separation in the binary-alloy problem: The one-dimensional spinless Falicov-Kimball model}},
  booktitle = {Physical},
  year = {1996},
  doi = {10.1103/physrevb.53.16189},
}

Loading related content...

Figure