1. Introduction
Over the last two decades, membrane filtration products - microfiltration (MF) and ultrafiltration (UF) - have proliferated globally to remove suspended solids, colloids, pathogens and emulsions[1-3]. Unlike conventional methods such as gravity based flotation and clarification, MF and UF method do not require chemical coagulation because the effective desired pore size could be controlled[4].
UF and MF membranes have been typically synthesized by phase inversion method due to cost effect mass production[5]. If the target materials could be dissolved in organic solvents, the pore is simply formed as mixing the poor solvents. This method is named as non-solvent induced phase inversion (NIPS). In case of some polymers such as polyethylene or polypropylene, it would be difficult to find proper organic solvent to dissolve them at room temperature. At this time, those polymers could be dissolved in diluents at high temperatures. Based on the thermodynamic, the phase separation takes place when the temperature is cooled until binomial points. This method is called as thermally induces phase separation (TIPS)[6,7]. Current commercialized MF and UF membranes are released using TIPS or NIPS. However, since both methods have been well known relatively, the technical barrier is relatively low. A novel technique to achieve a competitive price must be developed with well controllable pore size[8].
As one approach, the amount of used materials must be minimized. The volume portion of solvents or diluents for NIPS and TIPS is around 80%[9-12]. They occupy around 50% in the production cost except for incidental expenses. Even though those solvent or diluent could be recyclable, those are wasted in most cases. Also, it takes at least 1 week to remove remained solvents and diluents or use pore-forming chemical agents such as glycerin. In addition, the coagulation liquids are also the additional cost. The additional process could also increase the production cost.
This study introduces a novel method, named as a solution spreading phase inversion method, which produces a porous membrane in several seconds without any additional next process. The solution spreading is based on the solvent spreading on water. The solvent spreading must stratify two conditions; 1) the solvent should be buoyant in the subphase (water) without sinking and dissolving, 2) the solvent must spread on the surface of water. The solvent spreading is easily observed when the polymeric solutions were dropped on water. However, the film is typically not uniform because of the uncontrollable combination phenomenon among gravity, surface tension, evaporation, or shrinkage, etc. To put the handing condition, the suitable solvent is primary searched. However, it would be difficult to find the suitable solvent with relatively cost-effective solvent removal and Hansen solubility parameter matching. More important things, the film must have pore structures to show the ability of membrane. However, until recently, there is no study about a solution spreading phase inversion method to fabricate the membrane. Therefore, this study exhibits that the suggested method is very unique to fabricate the MF membrane with satisfying all the needs.
2. Experimental
As the model polymer, polysulfone (PSf) was purchased from BASF (Germany). The solvent, N,N-dimethyformaide (DMF), dichloromethane (DCM), and chloroform (CF) were purchased from Duck San Pure Chem. (Korea). As pore-forming agents, dimethylsulfoxide (DMSO) and blenched polyethyleneimine (PEI) were purchased from Sigma-Aldrich (USA). The polyester mesh (350 mesh) was used as the support layer. The images of samples were obtained using an optical microscopy (OM; Olympus, Japan), and a mini-SEM (COXEM, Korea) after Pt coating. The flux of water was measured using a home-made coupon cell (effective area: 18 cm2) with distilled water.
The mesh was deposited on the glass. The water was filled in the all area to occupy all the holes. After PSf solutions was spread on the mesh as Fig. 1. The mesh was transferred into a convection oven at 60°C for 5 minutes to remove water.
3. Results and Discussion
PSf polymer has the Hansen solubility parameter of 21.2 MPa1/2. DMF (24.9 MPa1/2), DCM (20.2 MPa1/2) and CF (19.0 MPa1/2) as the proper solvent to dissolve PSf were selected. Three solvents are used to spread on water (not shown data). The solution spreading ability was simply observed as one solution was very gently dropped on water. When the mesh filled by water was used, the spreading phenomena were different. When the DMF solution was dropped on water-filled mesh, the solution disappeared and did not show clearly spreading. Unlike free subphase water, the mesh could influence an additional effect during the solution spreading because of physical interactions between solution and mesh. However, although PSf solutions with DMF could be spread on the mesh, PSf film did not be fabricated as shown in Fig. 2(a).
When DCM and CF were used for PSf solutions, the solution spreading was observed. The boiling points of DCM and CF were 39.6 and 61.2°C, respectively. The evaporation kinetic depends on the environmental conditions (e.g., temperature, wind, and viscosity) but spreading thickness is around several millimeter (which is approximately calculated by the thickness of PSf films) [13,14]. Therefore, the spreading solution was visually solidified under several seconds.
The fast evaporation rate could influence complicated solidification and phase separation mechanism. Under high humidity conditions, the fast evaporation could condense water droplet on the polymer solution that Breath-Fig. structure could take place. However, OM images show that CF solution (Fig. 2(c)) seem to have porous structure compared to DCM (Fig. 2(b)). The relative humidity in laboratory was around 62%. Also, the liquid water was filled in the mesh. Thus, this porous structure could not result from the water condensation.
To understand the pore formation mechanism, one hypothesis is suggested; if an additive is miscible and dissolve with PSf as well as water, it could act as the water absorbent during evaporation and solidification. If the additives could absorb the water molecules, the phase separation could occur and form the pores such as a non-solvent phase separation mechanism[15]. To prove this assumption, PEI was chosen.
Figs. 3 and 4 exhibit OM and SEM images at various PEI concentrations. When PEI was not added into PSf solutions, pores of PSf films cannot be observed. However, the pore is formed when PEI was added into PSf solutions. The result might provide that the assumption could be reasonable. Furthermore, the pore size could be facilely controlled by only concentration of water-soluble contents without complicated operating factors (e.g., temperature, humidity, or flow rate, etc.) and long-term solvent exchange time in the case of a non-solvent exchange membrane[16].
To confirm our assumption, a water-soluble solvent, DMSO, was used. As shown in Fig. 5, the pore is formed like PEI. The pore size also increases with increasing in the concentration of DMSO. There are various water soluble organic solvents in the range of PSf’s Hensen solubility parameter. Although this study did not show the results using the other organic solvents effects, tetrahydrofuran, DMF, and 1-methyl-2-pyrrolidinone could also form the pore similar to DMSO trends. Although the pore characterization had not been systematically examined, OM and SEM images provide the possibility about controlling the pore size.
Fig. 6 shows the cross-section images. Without PEI addition, the PSf (Fig. 6(a)) film has the thickness of below 1 μm. When PEI was added within PSf solutions, the thickness is around 2 μm because the addition of PEI corresponds to increase of concentration. The thickness of porous PSf film increases with increasing solution concentrations. Thus, the film thickness could be relatively simply controlled by the concentration of polymer solutions.
The flux is a critical important factor to characterize the porous membrane. The PSf film without PEI does not show any water flux. However, when PEI is added, the flux (Fig. 7) increases with increasing in the concentration similar to Fig. 4. Thus, the flux as well as pore size is closely related to the concentration of PEI. At the same composition ratio (1 : 1 w/w) between PSf and PEI, the flux does not have dominant dependence on the thickness in the error range (not shown data).
The solution spreading phase inversion method is unique to control the pore size as well as thickness. This method could hold a great potential with respect to economic feasibility; no additional post-treatment, very fast fabrication process, and reasonable reproducibility without high-cost environment control process. More interesting advantage is that this method can fabricate the three-dimensional structure as Fig. 8 in which the mesh provides the three-dimensional architecture. The density of DCM is heavier than water so the film could compress the water. Thus, the level of PSf solutions at the hole of meshes is lower than that of the mesh surface. Thus, the PSf film was solidified on the curvature of the interface and surface zone. The level dependent results in the three dimensional structure of PSf films. Due to no chemical adhesion force between PSf film and mesh, the porous PSf film can be easily detached. After that, the porous PSf film has the original structure of mesh. Although this study used only mesh, any type of supports could be used to design the architecture.
4. Conclusion
This study focused on the fabrication of porous polysulfone membrane using the solution spreading phase inversion method. As the immiscible organic solvent was used for the solution spreading on water, the film could be easily formed. The pore size can be controlled by the concentration of water-miscible matters in organic solution. The thickness of porous PSf films is controlled by the total solution concentration. The spreading methods are interesting to fabricate three-dimensional architecture. Although this study could not show more detailed fabrication mechanism as well as characterization, this method could hold intensive potentials with respect to economic feasibility as well as novel membrane designs.
