Chemical potentials from structure factors: II. Charged multi-component mixtures

What the math gives to ML

The paper provides a constructive separation of long-wavelength composition fluctuations from forbidden macroscopic charge fluctuations in charged mixtures. Its transferable asset is a covariance and structure-factor representation in which neutral composition modes have finite small-wavenumber susceptibility, while the charge mode is suppressed by electroneutrality and Coulomb correlations, typically with a charge structure factor proportional to k squared. This can become a physics-informed regularizer and hard projection for neural fields or generative models that predict multicomponent concentration fields, preventing unphysical global charge drift while preserving chemically meaningful neutral fluctuations.