Bridging molecular dynamics and process engineering to predict the chemical recyclability of polyurethane foams

Abstract

Polyurethanes are versatile materials, but their thermoset nature limits mechanical recycling as they do not melt under heating. However, chemical recycling by depolymerization does not yet provide complete circularity, given the highly diverse monomers and additives. Especially for rigid foams, with short and multifunctional polyols, the separation of depolymerization products is challenging. In this paper, we propose a strategy to screen the separation of these products via liquid-liquid extraction by comparing molecular dynamics simulations with experimental results performed on a synthesized model rigid foam depolymerised by hydrolysis. Bifunctional ethyleneoxide polyols with molecular weights of 200, 400, 600 and 1000 are used together with 4,4’-methylene dianiline (MDA). The tested solvent systems are acetonitrile/n-hexane, water/dichloroethane, water/ethyl acetate and water/noctanol. Coarse-grained (CG) molecular dynamics simulations were used to model the large-scale morphological organization during solution-processing. Our results show a clear separation between the polyols and the depolymerised aromatic components, except for the acetonitrile/n-hexane solvent system, where all solutes dissolve in acetonitrile. Polyols preferentially migrate to the aqueous phase, while aromatic amines such as MDA favour the organic phase. The simulations reveal that larger polyols tend to accumulate at the interface, particularly in the water/dichloroethane and water/ ethyl acetate systems. Experimental partition coefficients, consistent with simulation predictions within error margins reported for the Martini 3 forcefield, indicate a strong correlation between polyol molecular weight and phase behaviour. This separation is further influenced by solvent polarity, with water/n-octanol systems exhibiting 2-3 times more MDA aggregation than the other systems, as predicted by the analysis of radial distribution functions and density profiles. The combined approach of molecular dynamics and experimental validation offers a promising strategy for optimizing liquid-liquid extraction in polyurethane depolymerization, guiding future design-for-circularity processes.