Alain Vidonne

Analysis of the salt retention of a titania membrane using the “DSPM” model: effect of pH, salt concentration and nature

Abstract Retention measurements with single salt solutions of KCl, LiCl, K2SO4, MgCl2 and MgSO4 were carried out as a function of the permeate flux for a commercial titania membrane close to the nanofiltration (NF) range. The effect of both pH and salt concentration was studied. The membrane shows amphoteric behavior with an isoelectric point (iep) at pH 6.2 (in presence of an indifferent electrolyte: KCl or LiCl). The obtained results agree qualitatively with the Donnan exclusion principle, characteristic of electrically charged membranes: a higher co-ion valence leads to a higher retention, a higher counter-ion valence leads to lower retention and retention decreases with increasing concentration. The analysis of the retention data by the Donnan steric partitioning pore model (DSPM) allowed to evaluate the effective volume charge of the membrane. It was shown that the membrane volume charge depends not only on pH, but also salt and its concentration. At low pH values (when the membrane is positively charged), the membrane charge is higher for magnesium salts than for potassium salts and lower for sulfate salts than for chloride salts. Also, in the high pH range (when the membrane is negatively charged), the membrane charge is higher in absolute value for sulfate salts than for chloride salts and lower (in absolute value) for magnesium salts than for potassium salts. Moreover, it was shown that the membrane charge does not increase with concentration for sulfate salts unlike chloride salts, when the membrane is positively charged. Also, the membrane charge does not increase (in absolute value) with concentration for magnesium salts unlike potassium salts, when the membrane is negatively charged. These results have been attributed to specific adsorption of magnesium and sulfate ions on the membrane material.

A comparison of membrane charge of a low nanofiltration ceramic membrane determined from ionic retention and tangential streaming potential measurements

Abstract The retention of neutral solutes and ionic solutions on a NF ceramic membrane was measured. The Donnan Steric Pore Model (DSPM) was fitted to these observations yielding effective pore radius (rp) and thickness/porosity ratio (Δx/Ak) from neutral solutes data and the effective volume charge (X) from electrolyte data. Effective volume charges were also directly determined using tangential streaming potential measurements. The DSPM gives X in good agreement with the values obtained by streaming potential. However, it overestimates Δx/Ak compared to the value deduced from neutral solute retention.

Theoretical study of the electrokinetic and electrochemical behaviors of two-layer composite membranes

Abstract A theoretical study concerning the effect of structure (porosity, pore radius and layer thickness) and surface characteristics (zeta potential) of two-layer composite membranes on global streaming potential (SPg), membrane potential (Emg) and membrane conductivity (λg) is presented. To this end, each layer of the composite membrane (composed of a support layer and a filtering layer) was modeled as a bundle of identical capillary tubes with connections between pores of the two layers (the pores in the filtering layer being smaller than those of the support layer). The global parameters SPg, Emg and λg were calculated by using the theory of thermodynamics of irreversible processes and a space charge model. SPg, Emg and λg were expressed as a function of the individual parameters of each layer SP(i), Em(i) and λ(i), respectively, the length fraction of the support layer, the porosity and pore radius ratios. It was shown that the electrokinetic (streaming potential and membrane conductivity) and electrochemical (membrane potential) behaviors of such composite membranes vary between that of single layers. For streaming potential, the results indicate that the contribution of the filtering layer to the global streaming potential is very little influenced by zeta potentials of both types of pores. It appears that the individual streaming potential of the filtering layer greatly dominates the global streaming potential. This is due to the fact that the streaming potential of the filtering layer is weighted by the pore radius ratio which is a predominant parameter in determining the global streaming potential. In contrast to the streaming potential, the contribution of the filtering layer to the global membrane potential (Emg) or membrane conductivity (λg) depends more or less on the zeta potentials of both kinds of pores and the corresponding electrokinetic radii as well. As to the membrane potential, the contribution of the filtering layer to Emg is all the more sensitive to the zeta potentials than the electrokinetic radii are small. The filtering layer greatly dominates the global membrane potential when its pores are narrow (with regard to the Debye length) and strongly charged. For the electrolyte conductivity inside pores, the smaller pores (inside the filtering layer) have an effect all the more dominant on the apparent membrane conductivity than their zeta potential is low and that of larger pores (inside the support layer) is high.

Hydraulic resistance measurements combined with electrical or diffusional resistance measurements for determination of pore size in MF membranes

Abstract An easy operating, low-cost and fast method for determining the mean pore size of microfiltration (MF) membranes is offered by combining hydraulic resistance measurements with either electrical resistance measurements (ERM) or diffusional resistance measurements (DRM). ERM and DRM allow determination of the structural parameter e/τ 2 which represents the ratio of membrane porosity to tortuosity factor squared. This parameter being determined, it is then possible to evaluate the mean pore radius from water flux measurements. The mean pore radius can be then determined from water flux measurements by taking into account the tortuosity of pores. This method was applied to MF ceramic membranes with two different pore sizes. Results show that e/τ 2 values obtained by diffusion and ERM are very close. The mean pore radius determined from this method was found to be in very good agreement with that afforded by mercury intrusion porosimetry (MIP) for both membranes.