No abstract

No abstract

Membrane-based organic solvent separations are rapidly emerging as a promising class of technologies for enhancing the energy efficiency of existing separation and purification systems. Polymeric membranes have shown promise in the fractionation or splitting of complex mixtures of organic molecules such as crude oil. Determining the separation performance of a polymer membrane when challenged with a complex mixture has thus far occurred in an ad hoc manner, and methods to predict the performance based on mixture composition and polymer chemistry are unavailable thus far. Here, we combine physics-informed machine learning algorithms and mass transport simulations to create an integrated predictive model for the separation of complex mixtures containing up to 400 components via any arbitrary linear polymer membrane. We experimentally demonstrate the effectiveness of the model by predicting the separation of two crude oils within 6-7 % of the measurements. Integration of ML predictors of key sorption and transport properties of molecules with transport simulators constitutes a radical departure from traditional approaches and will open new avenues for the rapid screening of polymer membranes prior to physical experimentation for the separation of complex liquid mixtures.

Membrane-based organic solvent separations are rapidly emerging as a promising class of technologies for enhancing the energy efficiency of existing separation and purification systems. Polymeric membranes have shown promise in the fractionation or splitting of complex mixtures of organic molecules such as crude oil. Determining the separation performance of a polymer membrane when challenged with a complex mixture has thus far occurred in an ad hoc manner, and methods to predict the performance based on mixture composition and polymer chemistry are unavailable. Here, we combine physics-informed machine learning algorithms (ML) and mass transport simulations to create an integrated predictive model for the separation of complex mixtures containing up to 400 components via any arbitrary linear polymer membrane. We experimentally demonstrate the effectiveness of the model by predicting the separation of two crude oils within 6-7% of the measurements. Integration of ML predictors of diffusion and sorption properties of molecules with transport simulators enables for the rapid screening of polymer membranes prior to physical experimentation for the separation of complex liquid mixtures.

Existing polymeric membranes struggle to separate small molecule solvents in the liquid phase due to low selectivity from solvent-induced plasticization and dilation. Mixed matrix membranes (MMMs) can potentially alleviate this issue via diffusion-based separations within rigid framework materials. Previous work from our lab and others has shown that organic solvent reverse osmosis membranes have different responses to transmembrane pressure depending on whether the material is a rigid structure (e.g., a carbon, zeolite, or metal-organic framework) or a swollen polymer. This work combines two Maxwell−Stefan transport models, representing the flexible polymer phase and a rigid microporous filler, with the Maxwell model to predict mixed matrix membrane solvent separation performance as a function of pressure and membrane material properties. The model demonstrates that for every filler perm-selectivity, there is a filler permeability that provides the largest separation factor in the final MMM. This optimum permeability increases with the filler's perm-selectivity. Dual-layer UiO-66/Matrimid hollow fiber MMMs were created to evaluate the model's prediction on the influence of transmembrane pressure on the separation of toluene and mesitylene as a test case. The UiO-66/Matrimid membrane demonstrated a predicted decline in permeance as pressure was increased. The separation factors increased as higher pressures increased the driving force for separation, consistent with the model. UiO-66 was shown to have superior selectivity to Matrimid in toluene/mesitylene; however, we conclude that ultraselective materials are ultimately needed to enable the mixed matrix membrane concept for the most challenging solvent− solvent separations, and open questions remain about polymer−filler pairings for organic solvent reverse osmosis.