LAUSR

Abstract Polymeric hollow fiber (HF) membranes are the solution for future applications of membrane distillation (MD) at industrial scale, with main advantages the enclosure of large active surface in modules of small volume and the reduced vulnerability of HFs to temperature polarization by using flow alteration aids [1] , [2] . Improved performance that might pave the way towards commercialization appears currently possible upon incorporation of nanostructured carbon fillers which allow for the control of pore architecture, hydrophilicity and mechanical properties of polymeric HFs. However, the spinning of advanced, nanostructure carbon incorporating, HFs with much needed improved properties as regards MD process performance, might be excessive. It is also essential to mitigate problems of flow maldistribution and/or poor hydrodynamics through a comprehensive investigation of MD module configuration and optimization of the spacer design. Hence, the cost of module assembly and testing rises significantly, with the module price accounting for more than 50% of the system price (module and HFs) at large scales. In this work, we are attempting to address the first challenge, achieving to develop high quality, though inexpensive (ca. 3 €/g), MWCNTs at high yield via CVD. Furthermore, the benefit of using the produced MWCNTs as fillers of Polyvinylidenefluoride (PVDF) and Poly(vinylidenefluoride Hexafluoropropylene) (PVDF-HFP) HFs is demonstrated in a direct contact membrane distillation (DCMD) module of 0.13 m2 active area, where a one-fold enhancement of the water vapor flux, complemented by a remarkable increase of the salt rejection efficiency, are achieved for the MWCNTs modified HFs as compared to their pristine analogues. Furthermore, a model is developed based on Knudsen diffusion and its combinations with viscous flow and molecular diffusion, with the target to define the temperature polarization indices and the thermal efficiency of the MD process. It is concluded that the thermal efficiency dropped slightly due to the high thermal conductivity of MWCNTs. However, such a minor adverse effect on the heat transfer properties is tolerable and countervailed by the enhanced vapor flux and salt rejection efficiency. In conclusion, the low cost of the presented in this work MWCNT nanofillers, combined with the amplified flux and salt rejection properties of the MWCNT-modified PVDF and PVDF-HFP HFs, support the claim for a very promising nanofiller for upscaled production and incorporation into HF membranes for MD applications at industrial scale.