Entropy-Driven Phase Separation in Multicomponent Fluid Mixtures

Povzetek

Phase separation in multicomponent fluid mixtures is conventionally attributed to unfavorable enthalpic interactions
between chemically dissimilar species, but the combinatorial entropy of mixing itself becomes an increasingly
important, and increasingly counter-intuitive, driver of demixing as the number of components grows and species
differ in molecular size. We use lattice-based Monte Carlo simulations within a generalized Flory-Huggins
framework to study phase behaviour in mixtures of 3 to 6 components with size asymmetry ratios (large-to-small
molecular volume) from 1 to 8, running 60 independent composition/temperature combinations per mixture size,
240 simulations in total, each sampled for 2×10^6 Monte Carlo sweeps after equilibration. For size-symmetric
mixtures (ratio = 1), demixing required unfavorable interaction parameters (chi > 2.1/z) in every case tested,
consistent with classical regular-solution predictions, and no entropy-driven phase separation was observed. Once
size asymmetry exceeded a ratio of 3, however, we observed spontaneous demixing at interaction parameters as
low as chi = 0.3/z, and in 14 of 60 highly asymmetric configurations demixing occurred even at chi = 0 (athermal
mixing), driven purely by the loss of combinatorial entropy available to the larger species when confined among
smaller ones. The critical temperature for entropy-driven demixing scaled approximately linearly with
size-asymmetry ratio above the threshold of 3, rising from 0.31 (in reduced units) at ratio 3 to 0.89 at ratio 8, an
increase of 187%. Increasing the number of components from 3 to 6 broadened the two-phase region's
compositional extent by 24% on average at fixed size asymmetry and interaction strength, and shifted the critical
composition toward the species with the largest size disparity in 52 of 60 six-component configurations.
Order-parameter fluctuation analysis confirmed second-order-like critical behaviour near the entropy-driven phase
boundary, with the fluctuation susceptibility exponent estimated at 1.24 ± 0.09, within range of mean-field
expectations for this lattice model. These results demonstrate that combinatorial entropy alone, independent of any
enthalpic penalty, can drive phase separation in size-asymmetric multicomponent mixtures once asymmetry
exceeds a well-defined threshold, with direct relevance to polymer blends, colloidal suspensions, and biomolecular
condensate formation.