Peierls Distortion and Low-Dimensional Solids

Introduction

A uniform one-dimensional chain with one electron per site seems metallic in elementary tight binding: one spin-degenerate band is half filled. Yet a chain can lower its total energy by alternating short and long bonds. This dimerisation doubles its repeat distance, changes the reciprocal lattice and opens a gap at the Fermi level. The idea is known as a Peierls distortion. It connects a structural change that chemists can recognise to a band instability that physicists calculate.

Core explanation

Take a uniform chain with lattice spacing a and one relevant electron per site. Its spin-degenerate band holds two electrons per site when full, so one electron half-fills it. In a simple one-dimensional picture the occupied wavevectors reach k F = ±π/(2a) . If the chain dimerises, alternating links become short and long. The new repeat length is 2a and its reciprocal period becomes G′ = 2π/(2a) = π/a. The new first-zone boundaries are ±π/(2a), exactly the old Fermi points. The distortion's wavevector G′ = 2k F couples states at opposite Fermi points and can open a gap where occupied and empty states previously met.

Why might a crystal choose a nonuniform arrangement? Splitting the states near the Fermi energy lowers the energy of the newly occupied bonding-like states while raising the corresponding empty states. This can lower total electronic energy. However, moving atoms costs elastic energy and may change other bonds. A stable distortion occurs when the net electronic gain exceeds the lattice cost under the model's conditions. It is not enough merely to draw an alternating structure; one must compare total energies. In a real material, transverse coupling between chains, finite temperature, disorder, electron repulsion and phonon dynamics can alter whether and how an ordered distortion appears.

In tight-binding language, alternation creates two hopping integrals, t short and t long. The old one-site unit cell becomes a two-site unit cell, giving two subbands. At half filling, the lower subband can be full and the upper empty, separated by a gap whose magnitude depends on the hopping difference in the elementary model. If t short = t long, the gap closes back to the uniform-chain crossing. This mechanism resembles the nearly-free-electron boundary gap but begins from local orbitals and changing bonds rather than weak scattering of free waves.

Trans-polyacetylene provides a chemically clear example. An ideal equal-bond π chain would have one p z electron per carbon and a half-filled π band, but the real chain has bond-length alternation and a semiconducting gap. The alternating pattern is often sketched as single and double bonds, although the measured structure contains shorter and longer bonds rather than perfectly localised textbook bond orders. A primary ACS account of conducting polymers connects half filling, k F = π/(2a), unit-cell doubling and the resulting gap. A primary Physical Review Letters proof in a specified one-dimensional model makes clear that rigorous conclusions depend on model assumptions.

Peierls-like effects also appear in low-dimensional inorganic solids and charge-density-wave systems. The observed transition may involve more than simple dimerisation: an incommensurate modulation, magnetism or correlations can coexist. In two or three dimensions, nesting only some parts of a Fermi surface may open a partial gap while leaving other carriers. Do not infer a fully insulating state from the word “Peierls” without looking at the filling and whether the entire Fermi surface is removed.

Step-by-step reasoning

1. Determine band filling and locate Fermi wavevectors. 2. Propose a lattice modulation with wavevector connecting occupied and empty near-Fermi states. 3. Determine the new unit cell and folded Brillouin-zone boundary. 4. Find whether the modulation couples degenerate states and opens a gap. 5. Compare electronic energy gain with elastic cost and other physical effects.

Visual explanation

Draw equal-spaced sites above a half-filled cosine band whose Fermi points lie at ±π/(2a). Below, draw alternating short and long links labelled a−δ and a+δ, with a two-site repeat of 2a. Fold the band into the new smaller zone and show its crossing at the Fermi point splitting into a filled lower branch and empty upper branch.

Real-world analogy

Imagine a row of identical seats with people filling exactly half of a possible paired arrangement. If the seats reorganise into close pairs, each person can occupy a lower-energy seat in a pair while the costly partner stays empty. Rearranging the seats takes effort, so the change happens only if the energy saved by the occupants outweighs the furniture-moving cost.

Real-world example

Polyacetylene's alternating bond lengths change its electronic spectrum compared with a hypothetical equal-bond chain. Chemical doping then introduces charge carriers and can greatly increase conductivity, but this is not the same as saying the undoped alternating chain is metallic. The example illustrates how chemical structure, periodicity and carrier concentration work together.

Why?

Why does doubling a half-filled chain's period target the Fermi level? A period 2a gives reciprocal vector π/a, equal to the separation 2k F between the two Fermi points. The distortion can therefore scatter one Fermi-state wave into the other and lift their degeneracy.

Common misconception

"All one-dimensional materials must be insulating because of Peierls distortion." The textbook instability concerns idealised systems and conditions. Real materials can have finite temperature, transverse coupling, incomplete nesting, different filling, competing interactions or disorder, and may remain conducting over relevant ranges.

Worked example

Question: An ideal one-dimensional chain has one spin-degenerate orbital and one electron per atom, with atom spacing a = 0.30 nm. Determine the original k F, the reciprocal period after dimerisation, and whether it equals 2k F.

Reasoning: At half filling, k F = π/(2a) = π/(0.60 nm) ≈ 5.24 nm⁻¹. Dimerisation doubles the repeat to 2a = 0.60 nm, so the new reciprocal period is 2π/(2a) = π/a ≈ 10.47 nm⁻¹. Twice the original k F is also 10.47 nm⁻¹. The new periodic potential can therefore connect the two old Fermi points and open a gap at the folded zone edge under the simple model.

Answer: k F ≈ 5.24 nm⁻¹ and G′ ≈ 10.47 nm⁻¹ = 2k F.

Quick check

1. What happens to the first-zone width when a one-dimensional lattice period doubles? Answer: It halves, because the reciprocal period changes from 2π/a to π/a.

Exam focus

Derive k F and G′ rather than only repeating “alternating bonds open a gap.” Distinguish a gap from a guaranteed thermodynamic transition by mentioning elastic cost. Use polyacetylene as an example with approximate shorter and longer bonds, and state that real low-dimensional materials can have competing physics.

Advanced insight

The mathematical tendency toward a Peierls modulation is related to enhanced electronic response at q = 2k F in an ideal one-dimensional metal. Coupling to a lattice displacement at that wavevector can lower energy. In strongly interacting one-dimensional systems, a Mott gap or spin ordering can compete with or modify the Peierls state; a single-particle band picture is no longer a complete theory.

Summary

A half-filled one-dimensional band has Fermi points at ±π/(2a). Dimerisation to period 2a introduces a reciprocal vector 2k F, folds those points to the new zone boundary and can open a gap. Stability follows from electronic energy gain versus lattice cost, and real materials may include transverse coupling and correlation effects.

Practice questions

1. What is the new repeat distance after simple dimerisation of a chain of period a? Answer: 2a. 2. Where are the Fermi points of an ideal half-filled spin-degenerate one-dimensional band? Answer: At ±π/(2a). 3. Why does alternating hopping split the old uniform band? Answer: It creates a two-site cell and couples states related by the new reciprocal vector, opening a boundary gap. 4. Why can a proposed distortion fail to occur in a real sample? Answer: Elastic cost, finite-temperature effects, transverse coupling, filling or competing interactions can outweigh the electronic gain.