As shown in Fig. 12, the volume of the pure-water flux at steady state showed a linear dependency on the pressure applied across the membrane. This indicates that the transmembrane pressure is the only driving force for permeation. The pure-water fluxes were measured at 0.2e0.5 MPa of applied pressure. A significant difference in the permeabilities of the g-alumina nanofiltration membranes based on the different peptizing agents was observed. The permeability (24.8 L/(m2 h bar)) of A-membrane was three times more than that of N-membrane (7.4 L/(m2 h bar)) and was four times more than that reported in the literature. Fig. 13a shows the correlation between the MWCO and the pure water flux for the g-alumina membranes we prepared and that reported in literature [16e22]. It could be found that both A-membrane and Nmembrane showed higher flux than the g-alumina membranes reported in literature. The high permeability of the N-membrane could be explained by the asymmetric tubular supports we used, which showed much smaller resistance than the disk supports used in literature. The much higher permeability of A-membrane than that of N-membrane could be explained by the relatively high porosity, the uniform pore-size distribution and the small tortuosity. The contribution ratios of such factors to pure water flux were calculated according to Hagen-Poiseuille model [34]: where h is the liquid viscosity and B0 the total permeability which is
defined as