Résumé
IntroductionMembrane distillation (MD) is a promising process for water desalination, which is not limited byosmotic pressure - contrary to reverse osmosis - and hence by salt concentrations. In MD, pure vaporwater permeates through the membrane thanks to a transmembrane water partial pressure differencewhile liquid salty water remains on the membrane feed side. MD process requires thus membraneswith high porosity, micrometer pore size, low thickness to maximize permeate fluxeswhereas itrequires membranes with high hydrophobicity (or even better super-hydrophobicity) and importantliquid entry pressure in order to prevent wetting of the membrane during MD operation. However, thereis still today a lack of specific polymeric membranes designed for MD.The aim of this work was to develop macro-porous highly hydrophobic membranes specificallyprepared for MD by Non-solvent Induced Phase Separation (NIPS).Material and Methods20%wt. PVDF was initially dissolved in solvent N,N-dimethylacetamid (DMAc). This dope solution wasthen casted on a glass support and pure water was used as non-solvent in two successive NIPSsteps. First, the casted dope solution was exposed to water vapor at fixed temperature and humidity(Vapor Induced Phase Separation step or VIPS step) during a fixed duration. Casted doped solutionwas then immersed into liquid water bath during 4 hours at ambient temperature (Liquid InducedPhase Separation step or LIPS step). Influence of the VIPS step parameters (duration, temperatureand humidity) was investigated and the obtained membranes were thoroughly characterized byscanning electron spectroscopy observations, liquid drop contact angle, liquidentry pressure (LEP),porosity and roughness measurements.ResultsSuccession of VIPS and LIPS steps allowed successfully obtaining macro-porous membranes withhigh porosity (70 – 80%). Figure 1a presents the influence of the VIPS duration (at 20°C and 99% RH)on (i) the liquid drop contact angle and (ii) the Liquid Entry Pressure. Duration of the VIPS step is acrucial parameter in the control of the morphology and properties of final membranes. For high VIPSdurations (more than 2 minutes), VIPS step is predominant and controls the membrane morphology:an interconnected nodule structure is obtained. In such case, membrane surfacehad a near super-hydrophobicity (with liquid drop contact angle around 125°C) while the LEP is verylow. These two properties are directly related to the nodule structure of the membrane surface. On thecontrary, for shorter VIPS durations (less than 1 minute), influence of the VIPS step is weak since thewater intake is too low during this first VIPS step. In this case, the membrane morphologies were verysimilar to those of membranes obtained by LIPS step only :a three-layermorphology with a very thin dense layer, a finger-like macro-void layer and a macro-porous nodulesublayer. These membranes exhibited higher LEP thanks to the thin dense upper layer whilemembrane surface was more hydrophilic. LIPS step seems dominant and influence of the VIPS stepmay be neglected.For VIPS duration between 1 and 2 minutes, both VIPS and LIPS affect the membrane morphologysimultaneously. When the VIPS duration was increased from 1 to 2 minutes, typical LIPS processfinger-like macro-void layer was reducing while the typical VIPSprocess nodule structure wasincreasing. In relation to the change of structure, the liquid drop contact angle increased and the LEPdecreased when the VIPS duration was increasing. The time limit between the dominance range of the VIPS and LIPS steps was shown to be strongly related to the operating conditions of the VIPS step(RH, temperature…)ConclusionIn order to obtain both high surface hydrophobicity and important liquid entry pressure, an optimummust be researched for intermediate VIPS duration when both VIPS and LIPS affect the finalmembrane morphology. More studies of the influence of the operating parameters during the VIPSstep at intermediate VIPS durations are currently underway.