We find that the initial, solvent-cast state of nanoparticles (NPs) in a polymer matrix temporally evolves during thermal annealing such that, at steady state, NPs maximize their distance from each other subject to mass balance constraints. The observed timescales for this unexpected structural reorganization, as probed by small-angle X-ray scattering, are temperature-dependent and can be prohibitively large, especially at temperatures around and below 1.2T g. X-ray photon correlation spectroscopy measurements during reorganization reveal that the collective NP dynamics slow down with annealing at constant temperature; this is accompanied by changes in the low-frequency regime in macroscopic viscoelastic measurements in equilibrated materials. By ruling out other potential sources for these effects (i.e., electrostatic interactions, adsorbed layers), we attribute these results to a long-ranged repulsive force between the NPs caused by fluctuations in the polymer phase, i.e., the “anti-Casimir” effect proposed by Obhukhov and Semenov [Long-range interactions in polymer melts: The anti-Casimir effect038305Phys Rev Lett200595 Thus, our results highlight the important role of long-term, slow NP reorganization on the structure and, subsequently, the properties of polymer nanocomposites (PNCs), even in the case of nominally miscible polymer nanoparticle hybrids.

In this work, we construct distinct first-principlesbased machine-learning models of CO 2 , reproducing the potential energy surface of the PBE-D3, BLYP-D3, SCAN, and SCAN-rvv10 approximations of density functional theory. We employ the Deep Potential methodology to develop the models and consequently achieve a significant computational efficiency over ab initio molecular dynamics (AIMD) that allows for larger system sizes and time scales to be explored. Although our models are trained only with liquid-phase configurations, they are able to simulate a stable interfacial system and predict vapor−liquid equilibrium properties, in good agreement with results from the literature. Because of the computational efficiency of the models, we are also able to obtain transport properties, such as viscosity and diffusion coefficients. We find that the SCAN-based model presents a temperature shift in the position of the critical point, while the SCAN-rvv10-based model shows improvement but still exhibits a temperature shift that remains approximately constant for all properties investigated in this work. We find that the BLYP-D3-based model generally performs better for the liquid phase and vapor−liquid equilibrium properties, but the PBE-D3-based model is better suited for predicting transport properties.