Perm-selectivity optimization of MWNT composite membranes for wastewater treatment by an ab-initio modelling

Functionalization of membranes with nanomaterials, ranging from zeolites to carbon nanotubes (CNTs), is one current way to address the main challenges related to wastewater treatment i.e. increasing flux while maintaining selectivity and reducing fouling.1 Modelling and ab-initio methodologies can play an important role in the optimization of the functionalized membrane structure thanks to their independence from fitting or adjustable parameters. In this work, permeability and rejection of CNT- composite membranes were optimized using ab-initio. Two MWNT-membranes were modelled, a polyester thin film composite (TFC) membrane2 and a chitosan porous membrane,3 with the nanotubes aligned perpendicularly to the membrane surface. Permeability and selectivity towards rac-fluoxitine, glucose, ethanol and water have been investigated. The permeability of the CNT-composite membranes was modelled using the experimental water flow rate of the virgin membranes2,3 and a slip-modified Hagen-Poiseuille equation to describe the water flow in the MWNTs inner core, therefore including the flow enhancement. The MWNTs outer size was hence used to calculate the fraction of the membrane area covered by MWNTs. Both the inner and outer diameters were changed to evaluate their effects on permeability. The selectivity of the MWNTs was evaluated through a bespoke algorithm free from empirical or fitting parameters and based on a topological analysis of the target compounds.4 The algorithm evaluates the steric hindrance offered by functional groups (FGs) anchored on the CNT inlet. The FGs and target compound geometries were evaluated by a quantum mechanics approach and the algorithm provides information on the compounds rejection by a molecular sieving mechanism. This methodology is not limited to CNTs but can be applied to the functionalization of nano-pores with regular contour. Monomers belonging to the new type of 'polymers of intrinsic microporosity5' were selected among various types of FGs after an accurate evaluation of their arrangement on the CNT inlet. The selectivity-permeability optimization revealed that the enhanced CNTs fluid dynamics properties depend not only on a good alignment of the MWNTs but also on their inner and outer diameters. In fact, for the chitosan membrane it was shown that the addition of nanotubes with specific dimensions leads to a marked reduction in the membrane permeability with respect to the virgin one. Instead, for the TFC membrane, the optimization showed that a low density of MWNTs with internal diameter of 4.44 nm yielded an enhancement in permeability of more than 500 times compared to the un-functionalized membrane. A high rejection of target solutes is achieved by an accurate design of the FG-CNTs. As a result, functionalized CNT-polymer membrane can result in a more efficient membrane for wastewater treatment, with increased permeability and optimal rejection toward solutes with very low molecular weight. Finally, this study allows also identifying which membranes are more appropriate to be modified with functionalized CNTs