AbstractPolyamide (PA) thinfilm composite (TFC) reverse osmosis (RO) membranes with high permselectivityand excellent mechanical/chemical durability were prepared using porouspolyethylene (PE) supports. Although the uniform pore structure and highsurface porosity of the PE support are beneficial for enhancing membranepermselectivity, its intrinsic hydrophobicity makes the formation of a PA selectivelayer challenging. The oxygen plasma pretreatment on the PE support, combinedwith the use of a sodium dodecyl sulfate surfactant during interfacialpolymerization, enabled the production of a PA layer on top of the support byimproving its wettability.
The systematic optimization of the membranefabrication parameters (e.g., plasmapretreatment, monomer and SDS compositions and post-heat treatment) led to theformation of a highly permselective PA layer. The fabricated PE-supported TFC (PE-TFC)membrane showed ~30% higher water flux with ~0.4% enhancement in NaCl rejectioncompared to a commercial RO membrane.
- Thesis Statement
- Structure and Outline
- Voice and Grammar
In addition, the PE-TFC membraneexhibited mechanical properties and organic solvent resistance superior to the commercialmembrane, which is attributed to the excellent mechanical and chemical stabilityof the PE material. The proposed strategy could expand the application of ROmembranes to mechanically and chemically harsh operating environments. Keywords: reverseosmosis; thin film composite membrane; polyamide; interfacial polymerization; polyethylenesupport 1.
IntroductionOverthe past few decades, polyamide (PA) thin film composite (TFC) membranes,consisting of a topmost PA selective layer on a porous polymer support (~100 mmin thickness), have been extensively used as commercialized reverse osmosis(RO) membranes for desalination and water treatment owing to their highpermselectivities 1-3. PA TFC membranes are fabricated byforming a PA selective layer viainterfacial polymerization (IP) of m-phenylenediamine(MPD) and trimesoyl chloride (TMC) dissolved in two immiscible solvents on aporous support 4,5. For the TFC membrane, the support determines its chemicaland mechanical stability while the PA selective layer mainly controls its separationperformance 6.
Although each membrane component can be independently chosen, thestructure and chemistry of the support have a critical influence on the structureand thus separation performance of the formed PA layer, thus posing constraintson the selection of the support material 7,8. For example, it is known that asupport having moderate hydrophobicity is favorable for fabricating a highly permselectivePA layer 9, which is consistent with the fact that commercial RO membranesare fabricated mostly using polysulfone (PSF) or polyethersulfone (PES) as a supportmaterial 10. However, PSF or PES-supported TFC membranes have a technicallimitation when applied to harsh environments due to their relatively lowmechanical strength and poor chemical stability (or organic solvent resistance)11,12. To improve the mechanical and chemical durability of TFC membranes, various polymer materialsincluding polypropylene (PP) 13-15,polyacrylonitrile 16,17, polyvinylidene fluoride 12,18, poly(tetrafluoroethylene)19, polyimide 20 and sulfonated polyphenylsulfone 21 have been exploredas supports.Amongthem, PP, one of common polyolefins, has been considered a strong candidate forsuch supports owing to its high mechanical strength and excellent chemicaldurability. Although its strong hydrophobicity, one of the obstacles infabricating TFC membranes, was resolved by adopting hydrophilic modificationssuch as plasma or chemical pretreatments 13-15,its irregular slit-like pore shape(an aspect ratio of ~2.
8), unevenlydistributed surface pores and low surface porosity (~11%) stood in the way of attemptsto fabricate a defect-less and permeable PA layer. In fact, it has beenreported that the fabricated PP-supported TFC membrane had a very lowwater flux and NaCl rejection (~87%) far below RO performance level due to itsunfavorable support pore structure 13.Polyethylene(PE) is another class of polyolefins and exhibits remarkably high mechanical and chemical durability,like PP.
PE has been manufactured into a porous membrane by sequentialprocesses consisting of melt extrusion, mechanical stretching and diluentextraction, unlike PP membranes whose pore structures are formed by the simplemechanical stretching of an extruded film 22. As a result, the porous PE membrane has higher porositywith a more regular pore shape than the PP membrane and thus has been successfullycommercialized as a lithium ion batteryseparator 23,24. This PE membrane also has beneficial structural features asa TFC membrane support together with its inherently strong mechanical andchemical stability. Its relatively uniform pore structure and high surfaceporosity could facilitate the formation of a uniform and highly cross-linked PAselective layer via the IP process 25.In particular, its high surfaceporosity could improve the mass transport at the PA layer-support interface, therebyenhancing the membrane water flux 8. In addition, its highly interconnected, openpore structure could further improve membrane permeation 26-28.
Despite many potential advantages, the PEmembrane has not been explored as a support material for the RO membrane. Inthis work, for the first time, we demonstrate that the commercialized, porousPE membrane can be used as a support to fabricate highly performing,mechanically strong and chemically durable RO membranes via conventional IP. The fabrication parameters, including the supportpretreatment, monomer and additive compositions and post-heat treatment, were systematicallycontrolled to achieve high separation performance of the PE-supported TFC (PE-TFC)membrane. The fabricated PE-TFC membrane exhibited ~30% higher water flux with ~0.4%enhancement in NaCl rejection compared to a commercial RO membrane (SWC4+).
Inaddition, it was demonstrated that the PE-TFC membrane has superior mechanicalstrength and organic solvent resistance compared to the commercial counterpartby characterizing the mechanical and chemical stability of the membranes. Thestructures and physicochemical properties of the prepared TFC membranes werecomprehensively characterized using various analysis tools including scanningelectron microscopy (SEM), atomic force microscopy (AFM), Fourier transforminfrared spectroscopy (FT-IR), X-ray photoelectron spectroscopy (XPS) andcontact angle measurement and correlated to their separation performance tounderstand the membrane structure-property-performance relationship. 2. Materials andmethods2.1. MaterialsTrimesoyl chloride(TMC, 98.0%, Tokyo Chemical), m-phenylenediamine(MPD, 99.
0%, Tokyo Chemical), sodium dodecyl sulfate surfactant (SDS, 99.0%, Sigma-Aldrich), n-hexane (95.0%, Daejung Chemical), sodium chloride (NaCl, 99.0%, Samchun Chemical),dodecane (99.0%, Sigma-Aldrich), tetrahydrofuran (THF, 99.0%, Daejung Chemical)and toluene (99.5%, Daejung Chemical) were purchased and used without purification. De-ionized(DI) water (18.
2 M? cm) was suppliedfrom a Millipore Milli-Q purification system. A PE support (~20 mmin thickness) and a commercial RO membrane (SWC4+) were received from SKInnovation Co., Ltd. and Hydranautics/Nitto Denko, respectively. 2.2. MembranefabricationThe fabricationprocess of a PE-TFC membrane is depicted in Fig.
1. High hydrophobicity of the pristine PEsupport whose watercontact angle is ~120° preventsthe support pores from being impregnated with an MPD aqueous solution 25. Inaddition, the intrinsically weak chemical interaction between the PA layer andthe support due to the absence of polar functional groups on the pristine PE hampersthe formation of a TFC membrane 13,14. Thus, the PE support was pretreatedwith O2 plasma prior to IP to improve support hydrophilicity (waterwettability) and to enhance the PA-support interfacial adhesion by generating oxygen-containingpolar functional groups on the support 13,14. It has been reported that the O2plasma process can effectively modify porous polyolefin (PP and PE) membranesto increase their hydrophilicity 29,30. The support hydrophilicity was systematicallycontrolled by adjusting the plasma exposure time (0 ~ 300 s) at the fixedplasma power of 20 W and operating pressure of 0.09 kPa using an oxygen plasmatreatment system (UVFAB systems, CUTE-MPR) to find the optimum treatment conditionthat can result in the best separation performance of the prepared TFCmembrane.
We selected the plasma power of 20 W because the power higher than 20W tended to significantly deform the PE support structure. Unfortunately, the plasmatreatment alone was found unable to completely wet the support pores with theMPD solution, failing to form a uniform PA layer, presumably because the support pores cannot be uniformlymodified with a one-sided plasma dose. Hence, a SDS surfactant was used duringthe IP process to ensure the complete wetting of the support with the MPDsolution by reducing the support-water interfacial tension. This was evidencedby the observation that theplasma-treated PE support can be completely soaked into the MPD solution onlywith the use of SDS (see Supplementary Material S1) and has a lower contactangle with a SDS/water mixture than that with pure water (see SupplementaryMaterial S2).
The plasma-treated PEsupport was mounted on a glass plate with a silicone frame. Next, an MPD (0.5 ~7.0 wt.%) aqueous solution containing SDS (0.03 ~ 0.2 wt.%) was poured on the plasma-treated PE support.
After 10 min, the MPD solution was decanted, and the excess MPD solution was carefullyremoved with a rubber roller. Subsequently, a TMC (0.05 ~ 0.4 wt.
%) solution in n-hexane was immediately poured on the MPD-impregnatedsupport and allowed to react at room temperature for 60 s, which is sufficientto ensure the complete reaction and thus the maximized NaCl rejection (seeSupplementary Material S3). Then, the membrane was rinsed with pure n-hexane to terminate the IP reaction andremove the unreacted TMC. The prepared PE-supported TFC (PE-TFC) membrane wasdried at 70 °C for a certain time period (0 ~ 15min) and stored in DIwater prior to the use. Fig.1.
Schematic illustration of the fabrication of the PE-TFCmembrane via the IP process. 2.3. Membrane characterizationThe surface andcross-section images of the PE supports and the PE-TFC membranes were collectedwith scanningelectron microscopy (SEM, FEI Inspect F50) under an accelerating voltage of 5 kV. The surfacepore size and porosity of the PE support were estimated from its SEM surfaceimage. To quantify the surface porosity, the obtained SEM surface images of thesupport were converted to black-and-white formats using an imageJ program. Theaverage surface porosity of the support was calculated from the percentage ofthe area occupied by black dots corresponding to the pores (see SupplementaryMaterial S4). The overall porosity of the PE support was measured using the gravimetric method proposed by others25,31,32.
The PE support was soaked into a bath containing a wetting solvent(dodecane) for 12 hand then removed fromthe bath. The excess solvent on the PE support was immediately removed usingtissue paper. The average overall porosity (?,%) of the PE support was calculated by measuring the weights of the dry (mdry) and wet (mwet) PE support with the densities of the PE (?m = of 0.97 g cm-3)and wetting solvent (?w = 0.
75g cm-3) as given by, The arithmeticaverage (Ra) roughness of the preparedPA selective layers was estimated from the topographical surface images (5 mm´5 mm)obtained using atomic force microscopy (AFM, NanoScope 5, Veeco) in a tappingmode. At least three different regions were scanned to obtain the average roughnessvalue for each sample. The chemical properties of the PE supports and PE-TFC membranes were characterized usingFourier transform infrared spectroscopy (FT-IR, SpectrumTwo). The chemical structure near the PE support surface was analyzed usingX-ray photoelectron spectroscopy (XPS). XPS spectra were collected using a PHIX-tool system with monochromatized Al-Karadiation at 1.49 keV.
The water contact angles of the PE supports and PE-TFCmembranes were measured by the sessile drop method using a contact anglemeasurement system (Phoenix-300, SEO Corporation).The membranemechanical properties were characterized using a universal tensile testingmachine (UTM, H5KT, Tinius Olsen). A membrane sample (length ´width = 50 mm ´ 10 mm) was clamped at both ends with an initial gauge length of 30mm and then stretched with a constantcross head speed of 20 mm min-1 to obtain the strain-stress curve. Thekey mechanical properties of the membrane, tensile strength, elongation atbreak and Young’s modulus, were determined from the strain-stress curve. Atleast five measurements were carried out to obtain the average values.
2.4. Membrane performanceevaluationWater flux and NaClrejection of the prepared PE-TFC membranes were evaluated using a cross-flowfiltration apparatus, where the circular flat sheet membrane coupon of 4 cm indiameter (an effective membrane area of 14.
5 cm2 (A))was mounted (seeSupplementary Material S5). Separationperformancewas evaluatedusing a NaClaqueous solution (2,000 mg L-1) at an operating pressure of 15.5 bar,an operating temperature of 25.0 ± 0.5 °Cand a flow rate of 1 L min-1(a corresponding cross-flow velocity of52.
7 cm s-1). Performance datawere collected after 12h, whenthe membrane water flux reacheda steady state. The water flux (Jw,L m-2 h-1) was calculated by measuring the totalvolumeof the permeate (?V)through the predeterminedmembrane area for a certain time period(?t),as given by, NaCl rejection (R, %) was calculatedby measuring the NaCl salt concentrationsof the feed (Cf) andpermeate (Cp) solutionsusing a conductivity meter (Cond 730P,INOLAB), as given by, 2.5. Organic solventresistance assessmentOrganicsolvent resistance?r each sampleerty ne btained from the strain-stresscurve. of the membranes was assessed by monitoring performancechanges before and after exposure to two organic?lvent resistance assessmentrations of MPD andSDS structure solution systems, an THF (20wt.%) aqueous solution and pure toluene, according to the previous reported method33.
THF and toluene were selected as representative water-miscible andimmiscible organic solvents, respectively. The membranes were immersed in eachorganic solution system at 25 °C for10 min, and then thoroughly rinsed with pure n-hexane and water. The separation performance of the membrane beforeand after the solvent treatment was measured and compared to assess its organicsolvent resistance.