149
path crosses the binodal line, in which demixing occurs instantaneously (Fig. 6.2a).
On the other hand, the compositions just below the top layer shown in Fig. 6.2b stay
in the region of one phase and are still miscible reflecting no demixing occurred
immediately after immersion. Demixing, in this case, will start after a long time
period when the compositions beneath the top layer will cross the binodal line.
Those two distinct demixing processes will give membranes with completely different morphologies (Strathmann et  al. 1975). Demixing rate measured as the time
difference between immersion of the casted solution into coagulation bath and that
solution turns to opaque determines the ultimate structure of the membrane.
Strathmann reported that slow demixing rates form sponge-like membrane morphologies, while instantaneous precipitation rates produce large finger-like voids
extended across the membrane thickness as illustrated insets in Fig. 6.2. Low water
fluxes and high salt rejections when tested as reverse osmosis membranes are the
characteristics of the former case and vice-versa for the latter case. Consequently,
the structure of a membrane can be tailored through variation of polymer types and
additives (i.e., pore former, inorganic nanoparticles), their concentrations, and temperature and content (nonsolvents or their mixtures with different miscibilities) of
the coagulation bath.
Recently, a few new works on the development of membrane adsorbers have
been reported in literature. For example, Yurekli has prepared polysulfone ultrafiltration membranes impregnated with different amount of NaX nanoparticles
(Yurekli 2016). The author calculated the change in membrane pore radius as a
function of NaX loading by filtration velocity method using scanning electron
microscopy analysis, porosity, and filtration results. Pore radius of the membrane
was reported to increase from 10 to 18 nm when the NaX loading was 10% w/w of
the casting solution. The large increment in the pore size was explained by the
instantaneous precipitation and agglomerations of nanoparticles occurred during
phase separation. However, rejection of bovine serum albumin was not significantly
affected when compared to the neat polysulfone membrane reported in that study.
In the study of Rahimpour et  al. (2012), functionalized multi-walled carbon
nanotube- incorporated polyethersulfone membranes have been prepared via
nonsolvent- induced phase separation method. Raising the content of functionalized
multi-walled carbon nanotubes to 1 wt.% has been reported to improve the membranes’ pore size, porosity, and surface roughness.. Besides, incorporation of functionalized multi-walled carbon nanotubes into the casting solution enhanced the
bovine serum albumin rejection of the polyethersulfone membrane. Several studies
consisting of polymers such as polyvinylidene difluoride, polyamide, polyethersulfone, and polysulfone blended with nanoparticles, including ZnO, Al 2 O 3 , TiO 2 , graphene oxide, silica, zeolite, multi-walled carbon nanotubes, SiO 2 –graphene oxide
nanohybrid, and TiO 2 -coated multi-walled carbon nanotubes, have been summarized by Esfahania et al. (2019) in terms of their functions on morphologies, hydrophilicities, and water permeabilities of the composite membranes.
6 Recovery of Heavy Metals by Membrane Adsorbers
Précédent

- 159/468

Suivant