(322f) Targeted Separation Scheme of Polyurethane Depolymerization Products | AIChE

(322f) Targeted Separation Scheme of Polyurethane Depolymerization Products

Authors 

Long, T. E., Virginia Polytechnic Institute and State University
Green, M. D., Arizona State University
In the United States, end of life polyurethane foams such as mattresses or cushions go predominantly to landfills. Mechanical recycling can prolong the inevitable fate of these materials, but not indefinitely. An emerging option is chemical recycling: breaking down these foams into smaller re-polymerizable feedstocks with a goal of creating a polyurethane circular economy. Chemical recycling creates a chemical mixture, requiring separation mechanisms to produce contaminant-free feedstocks. The proposed separation methods utilize membrane technology through a variety of processes such as pervaporation and filtration because these processes require lower energy. Pervaporation is a chemical potential driven process which uses dense polydimethylsiloxane (PDMS) membranes to remove any volatile components such as methanol, ethanol, butanediol, and ethylene glycol early in the process. Diffusion experiments show that all volatile components mentioned have diffusion coefficients on the same magnitude as ethanol, which is industrially separated through pervaporation. Pervaporation experiments provide permeability of each volatile component. Sorption coefficients represent the thermodynamic interactions as particles move through the membrane and are dependent on the diffusion coefficients and permeability. The proposed microfiltration process is pressure driven that uses porous bisphenol A polysulfone (BPA-PS) and 4,4′-dihydroxydiphenylmethane polysulfone (BPF-PS) membranes to remove larger components or oligomers such as polypropylene glycol (PPG). Microfiltration membrane process conditions are integral to the overall performance of membranes. Polymer composition, molecular weight, concentration, thickness, and operating pressures influence water permeance through the membrane. Higher molecular weight (~40 kDa) polymers at 15 wt% loading in N-methyl pyrrolidone (NMP) show the best performance for pure water.

This study investigates the separation of degradation products from both model and conventional polyurethanes. The model compounds considered are a poly(methylene diphenyl diisocyanate)-co-(1,4-butanediol) (pMDI-BDO) and a poly(MDI)-co-(bis(2-hydroxyethyl) terephthalate) (pMDI-BHET). The conventional polyurethanes consist of copolymers of either MDI or 2,4-toluene diiosocyantate (TDI) with either poly(propylene glycol) (PPG) or poly(ethylene glycol) (PEG). Methanolysis of the MDI-BDO model compound yields a mixture of MDI-dimethyl carbamate (MDI-DC), BDO, a potassium tert-butoxide (KOtBu) base, methanol (MeOH), and water, from washing. Methanolysis of the MDI-BHET model compound can be controlled to give a mixture of MDI-DC, BHET, terephthalic acid (TPA), ethylene glycol (EG), KOtBu, MeOH, and water. Glycolysis of conventional polyurethanes gives a mixture of insoluble carbamate solids: MDI-DC or TDI-dimethyl carbamate (TDI-DC); water and MeOH soluble polyols: PPG or PEG; partially soluble diamines: 4,4-methylenedianiline (MDA) or 2,4-diaminetolune (DAT); and relatively volatile solvents and catalysts, 1,4-diazabicyclo[2.2.2]octane (DABCO), EG, MeOH, and water.

Leveraging the distinct physical and chemical properties of each component in these mixtures allows for a targeted separation strategy using the above membrane systems. Dimethyl carbamates insolubly in water and methanol allows them to be efficiently removed with microfiltration membranes. For the remaining components, ultrafiltration membranes are suited for capturing the slightly soluble molecules: the diamines and BHET. Ultrafiltration may also remove any soluble polyols. Liquid-liquid extraction is ideal for separating out the acids and bases: TPA and KOtBu. Pervaporation effectively targets the removal of the less volatile liquid components like BDO and EG. Finally, pervaporation can also be applied to the final separation of the water:MeOH mixture. To evaluate the effectiveness of microfiltration membranes, binary mixtures of each component with either MeOH or water were tested. Similarly, for pervaporation, binary mixtures of each liquid component (ex. BDO and EG) and MeOH or water were tested with PDMS membranes. Microfiltration effectively removes MDI-DC from both MeOH and water mixtures. BDO and EG permeate through microfiltration membranes, and are effectively captured by subsequent pervaporation, removing them from both MeOH and water.