In a paper published in PRX Energy, Samuel Poncé and Xavier Gonze revisit a quiet approximation built into many first-principles materials calculations: the Born–Oppenheimer separation of electronic and nuclear motion. They set out a more exact formulation of total energy and identify a small electron–phonon term that can sit beyond the main electronic and vibrational contributions. In calculations for diamond and gallium nitride (GaN), this leading term is small in absolute energy, but the authors describe it as non-negligible.
The Born–Oppenheimer approach treats nuclei as fixed while calculating the electronic state, then brings nuclear motion back in afterward. This two-step procedure is approximate: it decouples the nuclear kinetic operator from the electronic Hamiltonian and nuclear potential. The authors distinguish different forms of this decoupling and formulate total energy in a basis of Born–Oppenheimer electronic wave functions.
They then organize contributions beyond the main electronic and vibrational terms in a perturbative expansion based on nuclear mass. Some higher-order terms can be grouped as an electron–phonon contribution to total energy, which first appears at fourth order in powers of the inverse fourth root of a typical nuclear mass. The electronic inertial-mass contribution appears at sixth order. This accounting helps clarify which effects belong in each part of the total-energy calculation.
The researchers implemented a calculation of the lowest-order electron–phonon term. They report values of 3.8 millielectronvolts per atom for diamond and 9.7 millielectronvolts per atom for GaN. Their summary compares these terms with vibrational energy: about 2% in diamond and up to 18% in GaN. The comparison shows why a correction that is modest in absolute terms may still deserve attention, particularly when energy differences between candidate states are small.
That precision matters in problems such as predicting which crystal polymorph is stable or describing a material’s magnetic energy landscape, both examples the authors identify as sensitive to small energy differences. The work’s broader motivation is that first-principles simulations help guide searches for materials relevant to batteries, solar cells and energy-efficient electronics. The reported results are calculations for diamond and GaN, rather than demonstrations of improved devices; the paper’s case is that a more complete energy accounting can make close comparisons more reliable.
The authors also draw a boundary between this total-energy term and a different quantity often discussed in electron–phonon physics. They say that summing Allen–Heine–Cardona zero-point renormalizations of electronic eigenvalues over occupied states does not give the electron–phonon contribution to total energy; that sum belongs to the phononic contribution. Keeping the categories distinct helps prevent different physical contributions from being conflated. The paper also corrects a recent proposal in the literature about this accounting, though the supplied text does not specify further details.
For diamond, Poncé and Gonze additionally estimate the electronic inertial-mass and quasiharmonic contributions. They report that all computed terms are size consistent. Together, the formulation and calculations offer a framework for including small corrections when they could affect energy comparisons. How much such corrections change the stability ranking of particular materials remains a separate question from the values reported here.
Source: Samuel Poncé and Xavier Gonze, In Search of the Electron-Phonon Contribution to Total Energy, PRX Energy 5, 043004 (2026), DOI 10.1103/zx5f-cw7t.