polymers Article
The Effect of Diluent Mixture with Upper Critical Solution Temperature on Membrane Formation Process, Microstructure, and Performance of PVDF Hollow Fiber Membrane by TIPS Process Zhenyu Cui 1, *, Shanshan Xu 1 , Jinyue Ding 1 , Jing Zhang 2 , Benqiao He 1 , Hao Wang 1 and Jianxin Li 1, * 1
2
*
State Key Laboratory of Separation Membranes and Membrane Processes/National Center for International Joint Research on Separation Membranes, School of Material Science and Engineering, Tianjin Polytechnic University, Tianjin 300387, China;
[email protected] (S.X.);
[email protected] (J.D.);
[email protected] (B.H.);
[email protected] (H.W.) School of Computer Science and Software Engineering, Tianjin Polytechnic University, Tianjin 300387, China;
[email protected] Correspondence:
[email protected] (Z.C.);
[email protected] (J.L.); Tel.: +86-022-83955055 (Z.C.); +86-022-83955798 (J.L.)
Received: 26 May 2018; Accepted: 27 June 2018; Published: 30 June 2018
Abstract: Thermally induced phase separation (TIPS) is a technique to prepare commercial membrane. However, the quick polymer crystallization during the quenching process will bring about a dense and thick skin layer and thus decrease permeability markedly. In this paper, a diluent mixture with upper critical solution temperature (UCST) was used to prepare polyvinylidene fluoride (PVDF) hollow fiber membrane. That is, the separation between diluent (propylene carbonate (PC)) and non-diluent (dioctyl terephthalate (DOTP)) occurred during the quenching process when the temperature of the dope was lower than 110 ◦ C. The effects of separation between PC and DOTP and the resulting coalescence of DOTP on the PVDF crystallization process, microstructure, and the permeability of the membranes were analyzed. The results showed that the suitable PC/DOTP weight ratio reduced the thickness of the skin layer near the outer surface markedly and resulted in a porous outer surface, and the microstructure evolution process was proposed. The maximum pure water flux for the prepared membrane is up to 128.5 L·m−2 ·h−1 even in a dry mode without using a hydrophilizing agent. The rejection rate of the carbonic particle is nearly 100%. This study presents a novel and simple way to fabricate the microporous membrane with the interconnected pore structure. Keywords: thermally induced phase separation; PVDF; microstructure; upper critical solution temperature; interconnection
1. Introduction Thermally induced phase separation (TIPS), based on the heat transfer, is one of the techniques for preparing commercial microporous membrane [1–3]. Compared with non-solvent induced phase (NIPS) technique, the membrane via the TIPS had a narrower distribution of pore size, high strength, and ease of regulation of the microstructure [4,5]. Many polymer membranes, such as polyolefin [6–8], polysulfone [9], poly(ethylene-co-vinylalcohol) (EVOH) [10], polyamide [11], polyphenylene sulfide (PPS) [12], poly(vinyl butyral) (PVB) [13], poly(methy methacrylate) (PMMA) [14], and especially polyvinylidene fluoride (PVDF) [15–17] and PVDF-based blends [18–20], have been prepared by this technique. Liquid–liquid (L–L) or solid–liquid (S–L) phase separation, relying on the interaction
Polymers 2018, 10, 719; doi:10.3390/polym10070719
www.mdpi.com/journal/polymers
Polymers 2018, 10, 719
2 of 17
between the polymer and diluent, occurred for the dope during the quenching process [21,22]. In detail, usually a strong interaction between the polymer and diluent brought about the S–L phase separation and formed the spherulites, while a weak interaction led to the L–L phase separation and formed a (closed or somewhat-opened) cellular or network structure [23,24]. During the past few years, some new developments in TIPS technology, such as the combination of TIPS and NIPS [25,26], modified TIPS [27], complex TIPS (c-TIPS) [28], low TIPS (L-TIPS) [29], and reverse TIPS (RTIPS) [30], have been achieved. Even the Janus membrane was developed via the nonsolvent thermally induced phase separation (NTIPS) [31]. The microstructure was regulated by changing the thermodynamics and dynamics of the phase separation for the dope, and the performance of the membrane was improved [24]. The network (formed via spinodal decomposition (SD)), somewhat-opened cellular (formed via nucleation growth (NG) in the metastable region with shorter coarsening time), and spherulitic structures are better than the closed cellular structures (formed via nucleation growth (NG) in the metastable region with longer coarsening time) for the membrane used in the separation field due to the well-interconnected microstructure [24]. Compared with the L–L phase separation, the temperature for the S–L phase separation is much lower, and the membrane formation process is relatively simple. This is because for the L–L phase separation, polymer-lean phase and polymer-rich phase were formed as the dope solution first entered the metastable region, and the coarsening process would go on until the temperature decreased to the crystallization temperature and polymer crystallization started [32]. Conversely, direct polymer crystallization occurred as the temperature of the dope decreased to the crystallization temperature for the S–L phase separation. This suggested that not only could the decomposition of the polymer be avoided, but also the microstructure regulation of the membrane became simple for the S–L phase separation. Up to now, besides a few single diluents, such as diphenyl ketone (DPK, resulting in a somewhat-closed cellular pore in a lower polymer content) and diphenyl carbonate (DPC, resulting in a somewhat-closed cellular pore) [33], most diluents, such as dibutyl phthalate (DBP) [1], glyceryl triacetate (GTA) [21], or propylene carbonate (PC) [27], had a strong interaction with PVDF and thus formed the spherulite. However, the rapid heat transfer between the outer dope and water bath facilitated polymer crystallization and thus resulted in a thick and dense skin layer for the resultant membrane due to the diluent evaporation, which was rejected by the polymer crystallization toward the outer surface of the dope [34]. Slowing down the temperature gradient, for example increasing the temperature of water bath [35] or adding additives, such as solvent (DMAC) into water bath or dope [27,30], or some pore former [36,37], can reduce the thickness of the skin layer and improve the pore interconnection. However, these methods either deteriorated the membrane’s mechanical strength markedly or produced a discarded solution containing multicomponents, which will increase the difficulty and expense of post treatment. The introduction of non-diluent into the polymer/diluent system was another common method to decrease the thickness and increase the interconnection of the skin layer by either weakening the interaction between the diluent and polymer or hindering the PVDF crystallization [34,38–40]. However, the thickness of the skin layer can become thinner only with the addition of enough non-diluent leading to the L–L phase separation [41]. As mentioned above, the temperature of the L–L phase separation is usually above the fusion temperature of the polymer [41] and resulted in a somewhat-closed cellular pore structure depending on the coarsening process of the polymer-lean phase [24]. Moreover, the diluent mixture used in the documents was miscible within the experimental temperature range. This suggested that the vacuum distillation had to be employed to separate the higher boiling point mixture of the diluent and non-diluent at a higher temperature after removing the extractant, and this will result in an expensive operating cost. If the diluent mixture with an upper critical solution temperature (UCST) was selected, the separation between the diluent and non-diluent occurred as the temperature decreased to a certain value. It can be conjectured that the separation between them not only omitted the later vacuum distillation operation (the two components of UCST system will automatically divide into two incompatible phases after the extractant was removed and the temperature decreased to the room temperature) and decreased
Polymers 2018, 10, 719
3 of 17
cost, but also restricted the polymer crystallization of the S–L phase separation and more importantly, influenced the coarsening process of the polymer-lean phase and polymer-rich phase for the L–L phase separation by penetrating the boundary between the polymer-lean phase and polymer-rich phase due to the resulting aggregation of each diluent (especially for the less component in the diluent mixture. The less and more component can be regarded as the dispersed phase and continuous phase, Polymers 2018, 10, x FOR PEER REVIEW 3 of 17 respectively. Compared with the continuous phase, the influence of aggregation for the dispersed phase onphase the coarsening process is more obviousthe due to the between distinctthesize change of the domain). for the L–L phase separation by penetrating boundary polymer-lean phase and polymer-rich phase due to the resulting aggregation of each diluent (especially for the less This suggested that the separation between dioctyl terephthalate (DOTP) and PC and the resulting component in the diluent mixture. The and more component can be regarded as the dispersed coalescence of DOTP may be favorable forless improving the interconnection of micropore (especially for phase and continuous phase, respectively. Compared with the continuous phase, the influence of the thick aggregation skin layer)for and decreasing filtration resistance. PVDF was insoluble in dioctyl terephthalate the dispersed phase on the coarsening process is more obvious due to the distinct at any temperature. meansThis that PC and DOTP can actbetween as diluent and non-diluent, respectively. size change of This the domain). suggested that the separation dioctyl terephthalate (DOTP) PCauthors and the resulting coalescence of DOTP may be only favorable for improving Moreover,and the found that PC and DOTP are miscible whenthe theinterconnection temperature is above of micropore for thewithout thick skinemulsification layer) and decreasing filtration resistance. was is to say, 110 ◦ C and becomes(especially two phases below 110 ◦ C (FigurePVDF 1). That insoluble in dioctyl terephthalate at any temperature. This means that PC and DOTP can act as the mixture of PC and DOTP composed the UCST system. However, up to now, no research about the diluent and non-diluent, respectively. Moreover, the authors found that PC and DOTP are only regulationmiscible effect when of a diluent mixtureis with membrane formation and membrane the temperature aboveUCST 110 °Con andthe becomes two phases withoutprocess emulsification below 110 °C (Figure 1). That is to say, the mixture of PC and DOTP composed the UCST system. microstructure had been reported. However, up to now, no research about the regulation effect of a diluent mixture with UCST on the In this paper, the diluent mixture of PC and DOTP with USCT was employed to prepare PVDF membrane formation process and membrane microstructure had been reported. hollow fiber In membrane via the S–L phase separation of the TIPS process. The aim is to investigate this paper, the diluent mixture of PC and DOTP with USCT was employed to prepare PVDF the effects of separation between PCphase and separation DOTP and the resulting coalescence of DOTP during hollow fiber membrane via the S–L of the TIPS process. The aim is to investigate the effects of separation between PC and formation DOTP and the resultingand coalescence of DOTP during of themembrane. the quenching process on the membrane process the microstructure quenching process on the membrane formation process and the microstructure of membrane. In In addition, crystallization behavior and the permeability of the resultant membrane were discussed addition, crystallization behavior and the permeability of the resultant membrane were discussed in in detail. detail.
Figure 1. The images of the dioctyl terephthalate (DOTP)/ propylene carbonate (PC) mixture below
Figure 1. and Theabove images of the dioctyl terephthalate (DOTP)/ propylene carbonate (PC) mixture below 110 °C. ◦ and above 110 C. 2. Experimental Section
2. Experimental Section 2.1. Materials and Methods (Solvay 6020, Mw = 6.85 × 10 ) was supplied by Solvay Solexis, Brussels, Belgium. The 2.1. Materials PVDF and Methods 5
diluent mixture is composed of PC (analytical reagent) and DOTP (analytical reagent). Bore liquid is 5 ) was supplied DOTP. Anhydrous used to remove by DOTP and PC from the nascent Belgium. PVDF (Solvay 6020,ethanol Mw =was 6.85 × as 10extractant Solvay Solexis, Brussels, membrane. They were all purchased from Tianjin Fine Chemical Institute (Tianjin,Bore liquid The diluent mixture is composed of PC (analyticalGuangfu reagent) and DOTPResearch (analytical reagent). P.R. China). Deionized water was used as the quenching bath. Carbon ink was provided by Ostrich is DOTP.Ink Anhydrous ethanol was used as extractant to remove DOTP and PC from the nascent Co., Ltd. (Tianjin, China), and the average diameter size was about 0.16 μm [34].
membrane. They were all purchased from Tianjin Guangfu Fine Chemical Research Institute (Tianjin, P.R. China). Deionized water was used as the quenching bath. Carbon ink was provided by Ostrich Ink Co., Ltd. (Tianjin, China), and the average diameter size was about 0.16 µm [34].
Polymers 2018, 10, 719 Polymers 2018, 10, x FOR PEER REVIEW
4 of 17 4 of 17
2.2. 2.2.Phase PhaseDiagram DiagramofofPVDF/PC/DOTP PVDF/PC/DOTPSystem System Polarizing Polarizing optical optical microscopy microscopy (BX51C, (BX51C, Olympus, Olympus, Tokyo, Tokyo, Japan) Japan) and and differential differential scanning scanning calorimeter ) and the calorimeter(DSC, (DSC,204F1, 204F1,Selb, Selb,Germany) Germany)were wereused used to to determine determine the the cloud cloud point point (Tcloud ) cloud and the dynamics crystalline temperature (T c ), respectively. After the dope, composed of the measured dynamics crystalline temperature (Tc ), respectively. After the dope, composed of the measured PVDF, PVDF, PC, and DOTP, was quenched in nitrogen, liquid nitrogen, mg of solidified wasbetween placed PC, and DOTP, was quenched in liquid about 5about mg of5solidified sample sample was placed between twoslips cover slips to prevent evaporation, heated 180a°C a hot(LK-600 stage (LK-600 two cover to prevent evaporation, heated to 180 ◦to C on hotonstage THMS, THMS, Linkam −1, kept ◦ C·min Linkam Scientific Instruments Ltd, UK) Surrey, UK) at of the50rate of 50−1°C· minat 1802 °C forand 2 min, Scientific Instruments Ltd, Surrey, at the rate , kept 180 ◦ Catfor min, then −1 ◦ ◦ − 1 and thentocooled 50 rate °C at of 10. T°C· min . T cloud is the temperature corresponding to the cooled 50 C attothe of the 10 rate C·min is the temperature corresponding to the appearance cloud appearance turbidity under the optical polarizing optical microscopy. Likewise, 7 mg ofsample solidified of turbidityofunder the polarizing microscopy. Likewise, about 7 mgabout of solidified was −1, held for 3 min to ◦ ◦ − 1 sample was sealed in an aluminum pan, heated to 180 °C at the rate of 40 °C· min sealed in an aluminum pan, heated to 180 C at the rate of 40 C·min , held for 3 min to erase the ◦ Crate −110 erase the history, thermaland history, and then cooled to room temperature °C·onset min −1temperature . The onset thermal then cooled to room temperature at the rate ofat10the ·minof . The temperature of the exothermic peak was regarded as T c . Each temperature was tested three times, of the exothermic peak was regarded as Tc . Each temperature was tested three times, and the average and the average was calculated. was calculated.
2.3.Preparation PreparationofofHollow HollowFiber FiberMembrane Membrane 2.3. Self-madespinning spinningequipment equipment(Figure (Figure2)2)was wasemployed employedtotoprepare preparethe thePVDF PVDFhollow hollowfiber fiber Self-made membrane.The ThePVDF PVDFcontent contentwas was25% 25%and andthe thesum sumofofthe thePC PCand andDOTP DOTPwas was75%. 75%.The Theweight weightratio ratio membrane. PC/DOTP range 10/0–10/5. The measured PVDF, PC, DOTP and DOTP 1) were fed ofofPC/DOTP is is in in thethe range of of 10/0–10/5. The measured PVDF, PC, and (Table(Table 1) were fed into ◦ into the vessel and dissolved at to 180form C to form dope. wasthe fed into the twin-screw the vessel and dissolved at 180 °C dope. Then, theThen, dope the wasdope fed into twin-screw extruder, −1 under extruder,from extruded from the spinneret at 48−1 mL ·minthe the pump, metering pump, and on a extruded the spinneret at 48 mL·min under metering and wound onwound a take-up − 1 ◦ −1 take-upatwinder at the linear velocity 20 mafter ·minit was afterquenched it was quenched in abath water C) to winder the linear velocity of 20 m·of min in a water (20bath °C)(20 to form 1 . The diameters −1·. min form nascent membrane. The theliquid bore liquid is 24 mL of theand outer and nascent membrane. The flux offlux the of bore is 24 mL· min The−diameters of the outer inner inner the spinneret were2 4mm, andrespectively. 2 mm, respectively. Thewas air gap wasDOTP 0.5 cm. DOTP and the tube fortube the for spinneret were 4 and The air gap 0.5 cm. and the residual residual themembrane nascent membrane wereextracted further extracted by ethanol. The final membrane was PC withinPC thewithin nascent were further by ethanol. The final membrane was dried air at temperature room temperature and reserved without any post-treatment (for example, being wetted indried air atinroom and reserved without any post-treatment (for example, being wetted by by glycerol aqueous solution). glycerol aqueous solution).
Figure2.2.The Theflow flowchart chartfor forpreparing preparingthe thehollow hollowfiber fibermembrane. membrane. Figure Table1.1.Code Codeof ofmembrane, membrane,PC/DOTP PC/DOTPweight weightratio, ratio,and andspinning spinningand andbore boreliquid liquidtemperature. temperature. Table
Code Code M0 M0 M1 M1 M2 M2 M3 M3 M4 M4 M5 M5
PC/DOTP PC/DOTP (wt./wt.) (wt./wt.) 10/0 10/0 10/1 10/1 10/2 10/2 10/3 10/3 10/4 10/4 10/5 10/5
Spinning temperature liquid temperature Spinning temperatureBore Bore liquid temperature (°C)◦ (°C) ◦ ( C) ( C) 100 70 120100 90 70 130120 100 90 130 100 150 120 150 120 170170 130 130 180180 140 140
Polymers 2018, 10, 719
5 of 17
2.4. Characterization of Membrane The membrane was sputtered with gold after it was freeze-fractured in liquid nitrogen. The surface and cross-sectional images of the membrane were observed by a field-emitting scanning electron microscope (SEM, Sirion-100, FEI Co., Eindhoven, The Netherlands). DSC was employed to investigate the crystallization behavior of the membrane. Approximately 7 mg of the membrane was weighted and sealed in an aluminum pan. The temperature range was 40–210 ◦ C, and the heating rate was 10 ◦ C min−1 . The crystallinity (Xc ) was calculated using the following Equation (1). Each sample was tested three times, and the average was calculated. Xc =
∆H f ∆H of
(1)
where ∆Hf o , ∆Hf and Xc represented the PVDF standard fusion enthalpy (100% crystal, 104.7 J·g−1 [42]), the fusion enthalpy (J·g−1 ), and the crystallinity (%) of the membrane, respectively. The crystal structure of the membrane was characterized by X-ray diffraction (XRD D8DIS, COVEX, Karlsruhe, Germany, Cu Kα radiation, 40 kV, 30 mA). The scanning speed was 4◦ ·min−1 . The step size was 0.05◦ . FTIR (Nicolet, Madison, WI, USA) was used to quantify the content of β-phase. The wave-length range was 700–3500 cm−1 , and the resolution was 2 cm−1 . The fraction of β-phase in the membranes was calculated using the following equation [43]: F (β) =
Aβ 1.26Aα + Aβ
(2)
where Aα and Aβ are the absorbencies of the α- and β-phase at 764 and 840 cm−1 , respectively. F(β) is the fraction of β-phase. Porosity was determined by the weighing method and calculated using the following equation [20]: Mb /Pb ε (%) = × 100 (3) M p /Pb + Mb /Pb where ε is the porosity of the membrane. Mp is the mass of dry membrane. Mb is the mass of the membrane after absorbing n-butanol. ρp and ρb are the density of the membrane and the n-butanol, respectively. Each temperature was tested three times, and the average was calculated. An outside-in operating mode was used to determine the pure water flux (PWF) of the membrane on self-made membrane evaluation equipment. The membrane was pre-pressed at 0.15 MPa for 30 min, and then, the operating pressure decreased to 0.1 MPa, and the permeated water was collected. The PWF was calculated as follows [44]: Jw =
V A×t
(4)
where Jw is the PWF (L·m2 ·h−1 ). A is the effective outer surface area of the hollow fiber membrane (m2 ). V is the volume of permeation (L), and t is the running time (h). Each sample was tested three times, and the average was calculated. The solute rejection was measured at 0.1 MPa in the same apparatus using a suspension system of carbonic ink (0.5 g·L−1 ), and the turbidity of the permeation was measured by turbidimeter (HACH, 2100Q, Loveland, CO, USA). Each temperature was tested three times, and the average was calculated.
Polymers 2018, 10, 719
6 of 17
3. Results and Discussion Polymers 10,Diagram x FOR PEER REVIEW 3.1. The2018, Phase of the Dope
6 of 17
For the PVDF/PC PVDF/PCsystem, system,neither neither coarsening process of the polymer-lean For the thethe coarsening process of the polymer-lean phasephase (the (the appearance and growth of liquid droplets), nor the nucleation and growth of the PVDF appearance and growth of liquid droplets), nor the nucleation and growth of the PVDF crystallization, crystallization, can can be be observed observed clearly clearly under under the the polar polar optical optical microscope microscope (Figure (Figure 33 (M0)) (M0)) during during the the ◦ cooling Only muddy muddy was was found found when when the the temperature temperature decreased decreased to to about about 50 50 °C, C, and and the cooling process. process. Only the ◦ C. This suggested that the phase muddy temperature further decreased to 38.1 muddy became becameobvious obviousasasthe the temperature further decreased to 38.1 °C. This suggested that the separation indeed occurred for the system, and the temperature of solidification crystallization phase separation indeed occurred for the system, and the temperature of and solidification and for PVDF is very which can be ascribed better compatibility between the carbonyl groupthe in crystallization forlow, PVDF is very low, whichtocan be ascribed to better compatibility between PC and fluorine atom in PVDF. Likewise, for the PVDF/PC/DOTP system, neither the coarsening carbonyl group in PC and fluorine atom in PVDF. Likewise, for the PVDF/PC/DOTP system, neither process of polymer-lean phase nor the PVDF crystallization cancrystallization be observed during the cooling process the coarsening process of polymer-lean phase nor the PVDF can be observed during ◦ C, and the (Figure 3 (M4)). The liquid appeared when the temperature decreased to about 113.8 the cooling process (Figure 3 (M4)). The liquid appeared when the temperature decreased to about ◦ C. It should be pointed liquid sizeand became larger.size In particular, the temperature decreased to about 52 113.8 °C, the liquid became larger. In particular, the temperature decreased to about 52 °C. out that the observed the polar optical microscope not the liquid droplets by It should be liquid pointed out that under the liquid observed under the polarisoptical microscope is not caused the liquid the polymer-lean phase coarsening process of the L–L phase separation. It should be ascribed to droplets caused by the polymer-lean phase coarsening process of the L–L phase separation. It should the separation between PC and DOTPPC andand theDOTP aggregation DOTP, which coveredwhich up the PVDF be ascribed to the separation between and theofaggregation of DOTP, covered crystallization and made it invisible under the polar optical above phenomena up the PVDF crystallization and made it invisible under the microscope. polar optical The microscope. The above suggested only the S–L phase occurred within the whole experimental process and not the phenomena suggested only separation the S–L phase separation occurred within the whole experimental L–L phase Although Tc increased with the of with the PC/DOTP weight in the process andseparation. not the L–L phase separation. Although Tcincrease increased the increase of theratio PC/DOTP phase diagram (Figure 4), the values of T for the dopes are all very low and only a little higher c weight ratio in the phase diagram (Figure 4), the values of Tc for the dopes are all very low and than only that of the PVDF/PC system. meant that the addition of DOTP hadaddition less influence on Thad the c of less a little higher than that of theThis PVDF/PC system. This meant that the of DOTP PVDF/PC system. However, thesystem. result isHowever, reasonable, theisreasons are asand follows. No turbidity influence on Tc of the PVDF/PC theand result reasonable, the reasons are as ◦ C, suggesting that the and crystallization can be observed when the temperature is higher than 110 follows. No turbidity and crystallization can be observed when the temperature is higher than 110 ◦ C. The mass transfer between PC and DOTP should be faster T the system is the lower thanthe 110system c for °C, suggesting that Tc for is lower than 110 °C. The mass transfer between PC and −1). than PVDFbecrystallization of the slowerbecause coolingofrate (10 ◦ C·cooling min−1 ).rate This that DOTPthe should faster than the because PVDF crystallization the slower (10meant °C·min ◦ as the temperature decreased to 110 C and the separation between PVDF and PC occurred, PVDF This meant that as the temperature decreased to 110 °C and the separation between PVDF and PC will dissolve in PC until the temperature decreased tofurther the crystallization for the occurred, PVDF will dissolve in PC untilfurther the temperature decreased totemperature the crystallization PVDF/PC system. This is true of the PVDF crystallization within PC regardless of the addition of temperature for the PVDF/PC system. This is true of the PVDF crystallization within PC regardless DOTP. Finally, the difference of T is not obvious for the different PC/DOTP weight ratios. The slight of the addition of DOTP. Finally, cthe difference of Tc is not obvious for the different PC/DOTP weight increase of Tslight the addition DOTP mainly ascribed the effect of PVDFtocontent in PC. c withincrease ratios. The of Tc of with the can addition ofbe DOTP can to mainly be ascribed the effect of That is, the higher the PC/DOTP weight ratio, the higher the PVDF content in the PVDF/PC system, PVDF content in PC. That is, the higher the PC/DOTP weight ratio, the higher the PVDF content in which in turn resulted in an easy nucleation the PVDF/PC system, which in turn resultedof inPVDF. an easy nucleation of PVDF.
Figure 3. 3. The The representative representative polarizing polarizing microscope microscope images images of of the the dope dope with withdifferent differentthe thePC/DOTP PC/DOTP Figure weight ratios ratios (M0) (M0) and and (M4). (M4). weight
Polymers 2018, 10, 719 Polymers 2018, 10, x FOR PEER REVIEW
7 of 17 7 of 17
Figure 4. Phase diagram of the PVDF/PC/DOTP system. PVDF/PC/DOTP system.
3.2. 3.2.Morphology Morphologyofofthe theMembrane Membrane Figure Figure 55 showed showed the the cross-sectional cross-sectional images images of of the the membrane. membrane. The Thespherulites spherulitescan canbe be observed observed from all samples, and this further proved that the S–L phase separation occurred for the from samples, and this further proved that the S–L phase separation occurred for the systemsystem during during the cooling This demonstrated that the of quantity of the non-diluent in is this is the cooling process.process. This demonstrated that the quantity the non-diluent in this paper notpaper enough not enough to result in a L–L phase separation even if the addition of DOTP will weaken the to result in a L–L phase separation even if the addition of DOTP will weaken the interaction between interaction PVDF and PC. The asymmetric gradient structure can be found the That whole PVDF and between PC. The asymmetric gradient structure can be found from the whole crossfrom section. is, cross there existed a skin layer The nearsize the of outer surface. The size of gradually the spherulite there section. existed aThat skinis, layer near the outer surface. the spherulite increased from increased gradually from the surface outer surface to thetoinner surface to no skin layer the the outer surface to the inner and finally, no skin layerand nearfinally, the inner surface. This near gradient inner surface. This gradient structuretocould mainly beofascribed to the difference ofwhich the temperature structure could mainly be ascribed the difference the temperature gradient, in turn led gradient, which in turn led to the different crystallization for PVDF at different positions of dope. to the different crystallization rate for PVDF at differentrate positions of dope. In detail, the difference In detail, the difference between the dope and the water bath is much bigger than that between the between dope and the water bath is much bigger than that between the dope and the bore liquid. dope and the Therefore, thedown heat transfer rate from slowed sequence from thesurface, outer Therefore, thebore heatliquid. transfer rate slowed in sequence thedown outerin surface to the inner surface to the inner surface, or it corresponded to the decrease of the nucleation rate and facilitated or it corresponded to the decrease of the nucleation rate and facilitated the crystal growth rate from the growthtorate outer surface inner surface. Finally, thetofaster heatskin transfer thecrystal outer surface thefrom innerthe surface. Finally, to thethe faster heat transfer rate led a dense layerrate or a led to a dense layer orwhile a spherulite of a heat smalltransfer size, while slower transfer of rate resulted spherulite of askin small size, the slower rate the resulted in heat a spherulite a big size orina aporous spherulite a big size or a porous skin of layer. The schematic diagram of theis gradient skin of layer. The schematic diagram the gradient structure formation shown instructure Figure 6. formation is shown in Figure 6. Although there existed the transfer between andsolubility the water Although there existed the mass transfer between PC and themass water bath due to the PC water of bath due to the water solubility of PC,and no this finger-like pores can be found, andTIPS, this the suggested that PC, no finger-like pores can be found, suggested that compared with influence of compared with the influence of isNIPS on thewhich bulk membrane is negligible, is NIPS on the bulkTIPS, membrane structure negligible, is different structure from the reports by Jungwhich [45] and different the reports by Jung [45] [36].the This may be ascribed fact that unlikewith the Lee [36]. from This may be ascribed to the factand thatLee unlike PolarClean solvent, to PCthe cannot be soluble − 1 PolarClean solvent, PC cannot be soluble with water at any proportion (the solubility is about 240 water at any proportion (the solubility is about 240 g·L at room temperature) and thus cannot bring g· L−1 atanroom temperature) and ofthus bring with about instantaneous of NIPS . about instantaneous demixing NIPScannot . Compared thean pristine membranedemixing (M0), the spherulite Compared with the for pristine membrane using (M0), the PC/DOTP spherulite size became bigger for the be membranes size became bigger the membranes mixture. This can mainly explained using the analysis PC/DOTP mixture.3.1, This can showed mainly that be explained the in analysis in Section from the in Section which the PVDFfrom content PC increased with3.1, the which higher showed thatweight the PVDF in PC increased withfacilitated the higher ratio.led The PC/DOTP ratio.content The higher polymer content thePC/DOTP growth ofweight crystal and to ahigher larger polymer content facilitatednot theonly growth of crystal and led a larger spherulite not spherulite size. Moreover, the space between the to spherulites increased size. with Moreover, the addition of only the space between the spherulites increased with the addition of DOTP, but also the spherulites DOTP, but also the spherulites became irregular and some cavities on the surface of the spherulite became irregular and some on the surfaceofofthe thecenter spherulite can be seen (as shown in the can be seen (as shown in thecavities higher magnification cross-section in Figure 5). These can higher magnification of the center cross-section in Figure 5). These can mainly be ascribed to the mainly be ascribed to the separation between PC and DOTP and the resulting DOTP aggregation. The separation between PC and DOTP the resulting DOTP aggregation. The S–L For phase S–L phase separation process can beand depicted as “nucleation-growth” (NG) mode. theseparation PVDF/PC process can depicted the as “nucleation-growth” (NG) mode. For as thethat PVDF/PC of this system of thisbe manuscript, membrane formation process is the same reportedsystem in the literature. manuscript, the membrane formation process is the same as that reported in the literature. However, However, the separation between PC and DOTP occurred as the temperature decreased to 110 ◦ C the separation between PC and DOTP occurred asagglomeration the temperature 110 °C the for the PVDF/PC/DOTP system, and the resulting anddecreased growth oftoDOTP arefor bound PVDF/PC/DOTP system, and the resulting and growth and of DOTP boundwill to to influence the PVDF crystallization process. agglomeration In detail, the agglomeration growthare of DOTP influence the PVDF crystallization process. In detail, the agglomeration and growth of DOTP will occupy the space, which not only enlarged the space between the spherulites, but also hindered the occupy space, notcrystallization only enlargedand the resulted space between the spherulites, butwith alsocavities hindered further the growth of which polymer in an irregular spherulite onthe the further growth of polymer crystallization and resulted in an irregular spherulite with cavities on the
Polymers 2018, 10, 719
8 of 17
Polymers 2018, 10, x FOR PEER REVIEW
8 of 17
spherulitesurface surfaceafter afterthe theaggregated aggregatedDOTP DOTPwas wasextracted. extracted.The Thehigher higherthe thePC/DOTP PC/DOTPweight weightratio, ratio, spherulite the bigger the size of cavity. Finally, the interconnection of the pores was improved. the bigger the size of cavity. Finally, the interconnection of the pores was improved.
Polymers 2018, 10, x FOR PEER REVIEW
8 of 17
spherulite surface after the aggregated DOTP was extracted. The higher the PC/DOTP weight ratio, the bigger the size of cavity. Finally, the interconnection of the pores was improved.
Figure 5. SEM images of the membrane. The whole cross section (MX1), magnification of the cross Figure 5. SEM images of the membrane. The whole cross section (MX1 ), magnification of the cross section near the center (MX 2), cross section near the inner surface (MX 3), and cross section near the Figure 5. center SEM images membrane. The whole section (MX1),(MX magnification of thesection cross section near the (MXof2 ),the cross section near thecross inner surface near the 3 ), and cross outer surface (MX 4). the X represents the sample code from 0 to 5. section near center (MX 2), cross section near the inner surface (MX 3), and cross section near the outer surface (MX ). X represents the sample code from 0 to 5. 4 outer surface (MX4). X represents the sample code from 0 to 5.
Figure 6. The schematic diagram of the gradient structure formation.
Figure 6. The schematic diagram of the gradient structure formation.
Figure 6. The schematic diagram of the gradient structure formation.
Polymers 2018, 10, 719
Polymers 2018, 10, x FOR PEER REVIEW
9 of 17
9 of 17
Withthe theaddition additionofofDOTP, DOTP,four fourdifferent differentstructures structuresfor forthe theskin skinlayer layernear nearthe theouter outersurface surface With (Figure 7) can be seen as follows: (1) the thick and dense skin layer when the PC/DOTP weight (Figure 7) can be seen as follows: (1) the thick and dense skin layer when the PC/DOTP weightratio ratiois loose skin skin layer layer with with many manysmall smallsize sizeand andopened openedcellular cellular isno nomore morethan than10/1 10/1 (M0 (M0 and and M1); M1); (2) (2) the the loose pores when the PC/DOTP weight ratio is in the range of 10/2–10/3 (M2 and M3); (3) the loose cellular pores when the PC/DOTP weight ratio is in the range of 10/2–10/3 (M2 and M3); (3) the loose cellular pores with thin skin layer for the 10/4 PC/DOTP weight ratio (M4); and (4) the loose skin layer pores with thin skin layer for the 10/4 PC/DOTP weight ratio (M4); and (4) the loose skin layer with with big some big isolated pores as the PC/DOTP ratio further 10/5 (M5). The some isolated cellularcellular pores as the PC/DOTP weight weight ratio further reachedreached 10/5 (M5). The change change of structure for the skin layer near the outer surface can be ascribed to the comprehensive of structure for the skin layer near the outer surface can be ascribed to the comprehensive effects of effects of three factors: effect of separation and and DOTP the aggregation of DOTP, three factors: the effect the of separation betweenbetween PC and PC DOTP theand aggregation of DOTP, the the diffusion between PC and the water bath, and the evaporation difference of PC and DOTP caused diffusion between PC and the water bath, and the evaporation difference of PC and DOTP caused by by elevated the elevated spinning temperature. In detail, as reported Ji et [35],the thethick thickand anddense denseskin skin the spinning temperature. In detail, as reported by by Ji et al.al. [35], layerfor forthe thepristine pristinePVDF PVDFwas wasmainly mainlycaused causedby bythe theevaporation evaporationofofthe thediluent diluentnear nearthe theouter outersurface surface layer at the higher spinning temperature due to the exclusion by the direct polymer crystallization. When at the higher spinning temperature due to the exclusion by the direct polymer crystallization. When the PC/DOTP weight ratio was no more than 10/1 (M1), the influences for both the separation between the PC/DOTP weight ratio was no more than 10/1 (M1), the influences for both the separation PC and DOTP andDOTP the effect resulting penetration the growth of of the crystallization by between PC and andfor thethe effect for the resultingofpenetration the PVDF growth of the PVDF the aggregation of DOTP was not obvious because of the lower DOTP content. Therefore, the structure crystallization by the aggregation of DOTP was not obvious because of the lower DOTP content. of M1 is similar to that of of M1 M0.isWhen the ratio is in the range 10/2–10/4, one Therefore, the structure similar toPC/DOTP that of M0.weight When the PC/DOTP weightofratio is in the for range thing, the elevated spinning temperature not only promoted the mass transfer diffusion between PC of 10/2–10/4, for one thing, the elevated spinning temperature not only promoted the mass transfer and the water bath, but also facilitated the evaporation rate of the diluent. However, the boiling point diffusion between PC and the water bath, but also facilitated the evaporation rate of the diluent. ◦ C) is much higher than that of PC (242 ◦ C), and this suggested that the evaporation of DOTP (400 However, the boiling point of DOTP (400 °C) is much higher than that of PC (242 °C), and this rate of PCthat is much faster than that of PC DOTP. Therefore, the PC/DOTP weight ratio forthe thePC/DOTP skin layer suggested the evaporation rate of is much faster than that of DOTP. Therefore, near the outer the increase DOTP. The PC/DOTP weight led weight ratio for surface the skinincreased layer nearwith the outer surface of increased withhigher the increase of DOTP. Theratio higher to the occurrence of the L–L phase separation and formed the cellular structure for the skin layer PC/DOTP weight ratio led to the occurrence of the L–L phase separation and formed the cellular near the for outer For another, thesurface. influence the separation between PCseparation and DOTPbetween and the structure thesurface. skin layer near the outer Forofanother, the influence of the resulting aggregation of DOTP aggregation became obvious due to the increase of due DOTP content, which in turn PC and DOTP and the resulting of DOTP became obvious to the increase of DOTP hindered the coarsening process of the polymer-rich phase and enhanced the penetrating effect content, which in turn hindered the coarsening process of the polymer-rich phase and enhanced theto the boundary between polymer-rich phase the polymer-lean phase. the small size penetrating effect to thethe boundary between the and polymer-rich phase and the Finally, polymer-lean phase. and opened cellular for thecellular skin layer near surface (M3) even thin (M3) skin layer Finally, the small size pores and opened pores forthe theouter skin layer near theand outer surface and (M4) was formed. As the PC/DOTP weight ratio further increased to 10/5, the ever higher spinning even thin skin layer (M4) was formed. As the PC/DOTP weight ratio further increased to 10/5, the temperature caused temperature the increasedcaused evaporation of PC and much faster diffusion between and the ever higher spinning the increased evaporation of PC and much fasterPC diffusion water bath, suggesting further increase the PC/DOTP ratio. Finally, coarsening between PC and the waterthe bath, suggesting theoffurther increase ofweight the PC/DOTP weightthe ratio. Finally, process of the polymer-lean phase developed more perfectly and formed a bigger size, creating more the coarsening process of the polymer-lean phase developed more perfectly and formed a bigger size, somewhat-closed and isolated cellular porescellular (M5). The formation process diagram for the skin layer creating more somewhat-closed and isolated pores (M5). The formation process diagram for near the outer surface is proposed in Figure 8. the skin layer near the outer surface is proposed in Figure 8.
Figure 7. The representative magnification images of the skin layer near the outer surface. Figure 7. The representative magnification images of the skin layer near the outer surface.
Polymers 2018, 10, 719
Polymers 2018, 10, x FOR PEER REVIEW
10 of 17
10 of 17
Figure 8. The formation process diagram for the skin layer near the outer surface. Figure 8. The formation process diagram for the skin layer near the outer surface.
Although the inner surface of all samples is not dense (Figure 9), compared with the pristine Although the inner surface ofmembranes all samplesprepared is not dense 9), compared the is pristine membrane, the inner surface for the with(Figure the PC/DOTP diluent with mixture more membrane, for the prepared with the the dope PC/DOTP mixture porous. This the can inner mainlysurface be ascribed tomembranes the mass transfer between and thediluent bore liquid (theis more porous. This can mainly be ascribed to the mass transfer between the dope and the bore liquid influence of TIPS on the inner surface is the same considering the same temperature difference (the influence of TIPS on bore the inner surface is the sameofconsidering the same temperature difference between the dope and the liquid). For the sample M0, the interfacial temperature between the between the dope and the bore liquid). For the sample of M0, the interfacial temperature between dope and the bore liquid is less than 110 °C, because the temperature for both of them is less than 110 ◦ C, because the temperature for both of them is less theThis dope and the liquid is lessoccurred than 110between °C. meant thatbore no mass transfer them. Therefore, only the lower temperature ◦ C. This meant that no mass transfer occurred between them. Therefore, only the lower than 110resulted gradient in a slower PVDF crystallization and formed a handful of pores within the inner temperature gradient in a slower PVDF crystallization and spinning formed a temperature handful of pores within surface. However, for resulted the PC/DOTP diluent mixture, the elevated led to the the inner surface. However, for the PC/DOTP diluent mixture, the elevated spinning temperature increase of interfacial temperature between the dope and the bore liquid. The mass transfer between led tomade the increase ofsurface interfacial temperature thetemperature dope and the bore liquid. The mass transfer them the inner porous when thebetween interfacial was over 110 °C [35]. Likewise, ◦ C [35]. between them made the inner surface porous when the interfacial temperature was over 110 compared with M0, the outer surface for the membranes prepared by the PC/DOTP diluent mixture Likewise, compared the outer surfaceasforthe thePC/DOTP membranes prepared the PC/DOTP diluent were porous, and thewith poreM0, number increased weight ratioby increased from 10/0 to mixture were porous, and the pore number increased as the PC/DOTP weight ratio increased 10/4 (Figure 10). When the PC/DOTP weight ratio further increased to 10/5, the number of poresfrom on 10/0 to 10/4 (Figure 10). When PC/DOTP weightfor ratio increased to 10/5, the PC/DOTP number of the outer surface decreased. The the porous outer surface thefurther membranes prepared by the pores on the outer decreased. The outer surface for PC theand membranes prepared by the diluent mixture wassurface also mainly ascribed toporous the separation between DOTP and the resulting PC/DOTP diluent mixture was also mainly ascribed to the separation between PC and DOTP and aggregation of DOTP. This is because according to the mechanism of NIPS and TIPS, the elevated the resulting aggregation DOTP. This is because according theand mechanism NIPS and TIPS, spinning temperature couldoffacilitate double-diffusion betweentoPC the waterof bath and form a the elevated spinning temperature could facilitate double-diffusion between PC and the water bath dense and thick skin layer by the instantaneous phase separation of NIPS. In addition, it can promote andevaporation form a dense anddiluent thick skin layer by the instantaneous phase separation of also NIPS. In addition, the of the by the increased temperature gradient of TIPS and form a dense it can promote the evaporation of the diluent by the increased temperature gradient of TIPS and11. also and thick skin layer. The formation process diagram for the outer surface is proposed in Figure formThe a dense and thick skin layer. The formation process diagram for the outer surface is proposed relationship between the PC/DOTP weight ratio and porosity is shown in Figure 12. It canin Figure 11. be seen that porosity increased from 57.9% to 72.2% when the PC/DOTP weight ratio increased from relationship between thetoPC/DOTP weight ratio and shown in Figure It can −1) is larger 10/0 toThe 10/5. This can be ascribed the fact that the density ofporosity PC (1.2 is g·L than12. that of be seen that porosity increased from 57.9% to 72.2% when the PC/DOTP weight ratio increased from DOTP (0.984 g·L−1). That is, the greater the addition of DOTP, the smaller the density or the higher 10/0 toof 10/5. This canmixture. be ascribed to the fact that theporosity density is ofachieved PC (1.2 gafter ·L− 1 )the is larger than of that volume the diluent Therefore, the higher extraction theof − 1 DOTP (0.984 g · L ). That is, the greater the addition of DOTP, the smaller the density or the higher diluent mixture.
Polymers 2018, 10, 719
11 of 17
Polymers 2018, 10, x FOR PEER REVIEW
11 of 17
volume of the diluent mixture. Therefore, the higher porosity is achieved after the extraction of the diluent mixture. Polymers 2018, 10, x FOR PEER REVIEW 11 of 17 Polymers 2018, 10, x FOR PEER REVIEW
Figure 9. The inner surface image of the membrane. Figure 9. The inner surface image of the membrane. Figure 9. The inner surface image of the membrane. Figure 9. The inner surface image of the membrane.
Figure 10. The outer surface image of the membrane. Figure 10. The outer surface image of the membrane.
Figure10. 10.The Theouter outersurface surfaceimage imageofofthe themembrane. membrane. Figure
Figure 11. The schematic diagram of the membrane for the porous outer skin layer. Figure 11. The schematic diagram of the membrane for the porous outer skin layer.
Figure Figure11. 11.The Theschematic schematicdiagram diagramofofthe themembrane membranefor forthe theporous porousouter outerskin skinlayer. layer.
11 of 17
Polymers 2018, 10, x FOR PEER REVIEW
12 of 17
2018, PolymersPolymers 2018, 10, 71910, x FOR PEER REVIEW
12 of12 17of 17
Figure 12. relationship The relationship between porosityand andthe the PC/DOTP PC/DOTP ratio. Figure 12. between porosity Figure 12. The The relationship between porosity and the PC/DOTPweight weight ratio.
3.3. Crystallization Behaviors of the Membrane 3.3. Crystallization Behaviors of the Membrane There existed three crystal forms (α, β, and γ) for PVDF, and the α-phase is more common [46].
There existed three crystal forms (α, β, β, and and γ) γ) for for PVDF, PVDF, and the α-phase α-phase is is more common common [46]. [46]. XRD was used to detect the crystal form of the as-prepared membranes. As shown in Figure 13, the used form as-prepared membranes. As shown in Figure 13, the XRD was used to to detect detectthe thecrystal crystal formof ofthe the as-prepared Figure characteristic diffraction peaks at 18.8° and 20.5° correspondedmembranes. to (020) plate ofAs theshown α-phaseinand also 13, ◦ and ◦ corresponded characteristic diffraction peaks at 18.8° and 20.5° corresponded to (020) plate of the α-phase and the characteristic diffraction peaks at 18.8 20.5 to (020) plate of the α-phase and corresponded to (110) and (002) plates of the β-phase, respectively [47]. This demonstrated that thealso corresponded to (110) and (002) plates of the β-phase, respectively [47]. This demonstrated that also corresponded to (110) and (002) plates of the β-phase, respectively [47]. This demonstrated as-prepared membranes showed the complex α and β phases. The values of F(β), derived from the the as-prepared membranes showed the complex α F(β), derived from the FTIR measurement (Figure 14)showed are presented in and Tableβ2.phases. of values F(β) is inofthe range 48%–57%, that the as-prepared membranes the complex αThe andvalue βThe phases. The values ofofF(β), derived andFTIR this means that the amounts of14) α-phase and β-phase are roughly equivalent the FTIR measurement (Figure 14)(Figure are presented in Table 2. in The value F(β)value is in or the range of from the measurement are presented Table 2.ofThe of that F(β) is PC/DOTP in 48%–57%, the range weight ratio has less influence on the content of β-phase. This is probably ascribed thePC/DOTP small and this means that the amounts of α-phase and β-phase are equivalent or equivalent thattothe of 48–57%, and this means that the amounts of α-phase androughly β-phase are roughly or that difference in crystallization temperature (Figure 4), because the content of β-phase was mainly weight ratio has less influence on the contentonofthe β-phase. is probably to the small the PC/DOTP weight ratio has less influence contentThis of β-phase. Thisascribed is probably ascribed influenced by the crystallization temperature [24]. The fusion enthalpy and the melting peak (Tmp, difference crystallization temperature (Figure 4),(Figure because4),the content β-phase was mainly to the smallindifference in crystallization temperature because theof content of β-phase was reflecting the thickness of the lamella) obtained from the DSC curves of the membranes (Figure 15) p influenced by the crystallization temperature [24]. The [24]. fusionThe enthalpy the melting peak (Tm , mainly influenced by the crystallization temperature fusion and enthalpy and the melting and the calculated crystallinity are listed in Table 3. With the increase of the PC/DOTP weight ratio, pthe reflecting thickness of the lamella) obtained from the DSC curves of the membranes (Figure 15) peak (T , reflecting the thickness of the lamella) obtained from the DSC curves of the membranes m the values for both Xc and Tmp first increased and then decreased, reaching the maximum as the and the calculated crystallinity are listed in Table 3. With the increase of the PC/DOTP weight ratio, (Figure 15) and the calculated crystallinity are listed in Table 3. With the increase of the PC/DOTP PVDF/DOTP weight ratio became 10/3. This can be ascribed to the comprehensive effects of the the values Xc for and(1) Tthe mpXfirst and then decreased, reaching the maximum as the weight ratio,for theboth values both Tm p first then decreased, reaching the maximum following two factors: PVDF content in increased PC and (2)and the separation between PC and DOTP and c andincreased PVDF/DOTP weight ratio of became 10/3. This becan ascribed to thetocomprehensive of the aggregation DOTP. The increased PVDF content helped to Xc,effects while the as the resulting PVDF/DOTP weight ratio became 10/3.can This be ascribed theincreasing comprehensive effects of separation between PC and DOTP and the resulting aggregation of DOTP hindered the following two factors: (1) the PVDF content in PC and (2) the separation between PC and DOTP and the following two factors: (1) the PVDF content in PC and (2) the separation between PC and DOTP crystallization (reducing c) and made the spherulites irregular (reducing PVDF microcrystal size the resulting aggregation of DOTP. The increased PVDF content helped totoincreasing X and resulting aggregation ofXDOTP. The increased PVDF content helpedthe increasing Xcc,, while or the thickness of the lamella). Therefore, the two opposing factors suggested the maximum for both separation between betweenPCPC DOTP and the resulting aggregation of DOTP hindered the separation andand DOTP and the resulting aggregation of DOTP hindered the crystallization Xc and Tmp and in a moderate PC/DOTP weight ratio. crystallization (reducing Xc)spherulites and made the spherulites irregular (reducing the PVDF size (reducing Xc ) and made the irregular (reducing the PVDF microcrystal sizemicrocrystal or the thickness pboth or the the lamella). thicknessTherefore, of the lamella). Therefore, two opposing suggested maximum of the two opposingthe factors suggestedfactors the maximum forthe both Xc and Tfor and m Xc aand Tmp andPC/DOTP in a moderate PC/DOTP in moderate weight ratio. weight ratio.
Figure 13. XRD patterns of the membrane.
Figure 13. XRD patterns of the membrane.
Polymers 2018, 10, 719 Polymers 2018, 10, x FOR PEER REVIEW Polymers 2018, 10, x FOR PEER REVIEW
13 of 17 13 of 17 13 of 17
Figure 14. FTIR of the membrane. Figure FTIR spectra spectra Figure 14. 14. FTIR spectra of of the the membrane. membrane. Table2.2.AAαα,, A Aβ, and Table andF(β) F(β)values valuesof ofthe themembranes. membranes. Table 2. Aα, Aββ,, and F(β) values of the membranes. Samples M0 Samples Samples M0
Aα 1963.2 Aα 1963.2 A2537.5 α Aβ Aβ A 2537.5 F(β) (%) F(β)β(%) 51.0 F(β) (%) 51.0
M1 M2 M2M2 M1 1398.9 1484.6 1398.9 1484.6 1963.2 1484.6 2334.2 1398.9 2138.6 2537.52334.2 2334.2 2138.6 2138.6 57.0 53.453.4 51.0 57.0 57.0 53.4 M0 M1
M3 M3 M3 1614.0 1614.0 1614.0 2064.8 2064.8 2064.8 50.3 50.3 50.3
M4 M5 M5 M5 1753.0 1361.4 1753.0 1361.4 1753.0 2085.2 1361.4 2064.5 2085.2 2064.5 2064.5 2085.2 48.0 54.6 48.048.0 54.6 54.6 M4 M4
Figure 15. DSC the membrane. Figure DSC curves curves of of Figure 15. 15. DSC curves of the the membrane. membrane. Table 3. The values of melting peak temperature, fusion, and crystallinity. Table 3. The values of melting peak temperature, fusion, and crystallinity. Table crystallinity. Samples
Samples Samples M0 Tmp (°C)
Tmp (°C) Tm ΔHf (J170.0 g−1) f (J g−1) 71.45 ∆Hf (J g−1 ) ΔH Xc (%) Xc (%) Xc (%) 68.24 p (◦ C)
M0 M0 M1 170.0 170.0 170.5 71.45 71.45 72.17 68.24 68.24 68.93
M1 M2 M3 M1 M2M2 M3 M3 170.5 170.6 170.8 170.5 170.6 170.8 170.8 72.17 170.6 72.98 75.52 72.1772.98 72.98 75.52 75.52 68.93 69.70 72.13 68.9369.70 69.70 72.13 72.13
M4 M5 M4 M4 M5 170.6 170.2 170.6 170.2 69.05170.6 63.43 69.0569.05 63.43 65.95 60.58 65.9565.95 60.58
M5 170.2 63.43 60.58
3.4. Permeability Properties of the Membrane 3.4. Permeability of the the Membrane Membrane 3.4. Permeability Properties Properties of Figure 16 indicates the PWF of the membranes. It can be seen that the value of the PWF is near Figure 16 16 indicates indicates the the PWF PWF of of the the membranes. membranes. It seen that the value of the PWF Figure It can can be befrom seen 23.2 that L· the ofthe themaximum PWF is is near near −2·h−1 to zero for the pristine membrane. However, the PWF increased mvalue of −2 −1 − 2 − 1 zero for the pristine membrane. However, the PWF increased from 23.2 L· m · h to the maximum of · zero for the pristine membrane. However, the PWF increased from 23.2 L · m h to the maximum −2 −1 128.5 L·m −2·h −1 when the PC/DOTP weight ratio increased from 10/1 (M1) to 10/4 (M4) and then − 2 − 1 128.5 L· m · h when the PC/DOTP weight ratio increased from 10/1 (M1) to 10/4 (M4) and then of 128.5 L·to m97.2 ·h L mwhen the PC/DOTP ratio increased fromfeatures 10/1 (M1) to 10/4 and −2·h−1 with decreased the furtherweight increase to 10/5 (M5). Two should be of(M4) concern. −2 ·h−the 1 with decreased to 97.2 L97.2 m−2L ·h−1 with further increase to 10/5 (M5). Two features should beshould of concern. then decreased to m the further increase to 10/5 (M5). Two features be of One is that the as-prepared membranes using the PC/DOTP diluent mixture had a lower filtration One is that the as-prepared membranes using the PC/DOTP diluent mixture had a lower filtration concern. One is if that as-prepared PC/DOTP mixture hadand a lower resistance, even thethe membrane wasmembranes reserved in using a dry the mode without diluent any post treatment the resistance,resistance, even if the membrane was reserved in a dry mode without any postany treatment and the filtration even if the membrane was reserved in a dry mode without post treatment measurement of the PWF was directly conducted without being wetted by a hydrophilic reagent (for measurement of the PWF directly conducted without being wetted by awetted hydrophilic reagent (for and the measurement of was the PWF directly conducted without being by a hydrophilic example, the PWF is near zero for was the dry PVDF membrane without hydrophilization by ethanol example, the PWF is near zero for the dry PVDF membrane without hydrophilization by ethanol [34]). The other is that the PWF is relatively high despite the cellular pores (M3, M4, and M5), and [34]). The other is that the PWF is relatively high despite the cellular pores (M3, M4, and M5), and
Polymers 2018, 10, 719
14 of 17
reagent2018, (for 10, example, theREVIEW PWF is Polymers x FOR PEER
near zero for the dry PVDF membrane without hydrophilization 14 ofby 17 ethanol [34]). The other is that the PWF is relatively high despite the cellular pores (M3, M4, and M5), and suggested this suggested the cellular pores in this work completely closed. mainly this that that the cellular pores in this work are are not not completely closed. ThisThis cancan mainly be be ascribed to the separation between DOTP resulting aggregation of DOTP, which ascribed to the separation between PC PC andand DOTP andand thethe resulting aggregation of DOTP, which in in turn hindered the coarsening process of the polymer-lean phase for the skin layer near the outer turn hindered the coarsening process of the polymer-lean phase for the skin layer near the PWF was influenced by by the interconnection of pores and surface to form a closed cellular pore. The PWF porosity. As shown in Figures 5–7, for the pristine membrane, the filtration resistance is tremendous porosity. because of of little littlepores poreson onthe theinner innerand andouter outer surfaces and thick dense layer formed because surfaces and thethe thick andand dense skinskin layer formed via via S–L the S–L phase separation similar the reports byal. Ji [34] et al.and [34]Cha andetCha et al. [27]. However, the phase separation similar to thetoreports by Ji et al. [27]. However, for the for the membranes prepared by the PC/DOTP there only existed someon pores membranes prepared by the PC/DOTP diluent diluent mixture,mixture, there not onlynot existed some pores the on theand inner andsurfaces outer surfaces thespace largeramong space the among the spherulites, skin layer inner outer and theand larger spherulites, but also but the also skin the layer became becameand porous and interconnected the PC/DOTP weight ratio from increased from 10/1 porous interconnected when thewhen PC/DOTP weight ratio increased 10/1 to 10/4. Thattois,10/4. the That is, the interconnection for the membranes prepared by the diluent mixture with USCT is better. interconnection for the membranes prepared by the diluent mixture with USCT is better. In addition, In addition, also increased with the of DOTP. Thesethat all meant that theresistance filtration porosity alsoporosity increased with the addition of addition DOTP. These all meant the filtration resistance and decreased and a higher flux. with Compared with the larger, somewhat-closed, decreased resulted in resulted a higher in flux. Compared the larger, somewhat-closed, and isolated and isolated of M5, thinof skin of M4 is much more interconnected due lots of cellular porescellular of M5,pores the thin skinthe layer M4layer is much more interconnected due to lots of to smaller, smaller, cellular opened pores. cellularTherefore, pores. Therefore, PWF is higher. opened the PWFthe is higher.
Figure 16. 16. Pure Pure water water flux flux of of the the membranes membranes (0.1 (0.1 MPa MPa operating operating pressure). pressure). Figure
thethe particle rejection of the membrane (M4). It can The carbon carbon ink ink suspension suspensionwas wasselected selectedtototest test particle rejection of the membrane (M4). It be seen fromfrom Figure 17 that the permeating solution is transparent compared withwith the the feedfeed andand the can be seen Figure 17 that the permeating solution is transparent compared turbidity is nearly zero.zero. ThisThis indicated that the of theofcarbon black black particles is nearly 100%, the turbidity is nearly indicated thatrejection the rejection the carbon particles is nearly which indicated a bettera rejection capability. 100%, which indicated better rejection capability. Polymers 2018, 10, x FOR PEER REVIEW
15 of 17
Figure 17. The feed (left) and permeating solution (right) obtained from M4.
Figure 17. The feed (left) and permeating solution (right) obtained from M4. 4. Conclusions
The diluent mixture system with UCST was firstly introduced to prepare the PVDF hollow fiber membrane via the S–L phase separation of the TIPS process. The separation between PC and DOTP and the resulting aggregation of DOTP not only influenced the polymer crystallization process of the bulk dope, the coarsening process polymer-rich phase, and the polymer-lean phase for the dope near the outer surface, but can also penetrate the boundary between the polymer-rich phase and the polymer-lean phase, which in turn led to a membrane with a porous outer surface and decreased the thickness of skin layer near the outer surface. Finally, a membrane with an interconnected structure −2
−1
Polymers 2018, 10, 719
15 of 17
4. Conclusions The diluent mixture system with UCST was firstly introduced to prepare the PVDF hollow fiber membrane via the S–L phase separation of the TIPS process. The separation between PC and DOTP and the resulting aggregation of DOTP not only influenced the polymer crystallization process of the bulk dope, the coarsening process polymer-rich phase, and the polymer-lean phase for the dope near the outer surface, but can also penetrate the boundary between the polymer-rich phase and the polymer-lean phase, which in turn led to a membrane with a porous outer surface and decreased the thickness of skin layer near the outer surface. Finally, a membrane with an interconnected structure was obtained. The maximum PWF of the dry membrane reached 128.5 L m−2 ·h−1 without being wetted by a hydrophilizing agent in advance. The rejection rate of the carbonic particle was nearly 100%. In summary, our study provides a new avenue to tailor the micropore structure of a hollow fiber membrane with improved interconnection during the TIPS process. Author Contributions: Z.C. conceived of and designed the experiments; S.X., J.D., and H.W. performed the experiments; J.Z. and B.H. analyzed the data; Z.C. and J.L. contributed materials/analysis tools and wrote the paper. Funding: This work was supported by the National Nature Foundation of China (No. 21576209), Applied Basic Research and Advanced Technology Programs of Science and Technology Commission Foundation of Tianjin (No. 16PTSYJC00100), and the Program for Changjiang Scholars and Innovative Research Team in University (PCSIRT) of the Ministry of Education of China (No. IRT-17R80). Acknowledgments: The authors thank Tiantian Li (Ningbo Institute of Industrial Technology, Chinese Academy of Sciences) for depicting the membrane formation mechanism. Conflicts of Interest: The authors declare no conflict of interest.
References 1.
2. 3. 4. 5. 6. 7.
8.
9. 10.
11.
Ji, G.L.; Zhu, B.K.; Cui, Z.Y.; Zhang, C.F.; Xu, Y.Y. PVDF porous matrix with controlled microstructure prepared by TIPS process as polymer electrolyte for lithium ion battery. Polymer 2007, 48, 6415–6425. [CrossRef] Lloyd, D.R.; Kinzer, K.E.; Tseng, H.S. Microporous membrane formation via thermally induced phase separation. I. Solid–liquid phase separation. J. Membr. Sci. 1990, 52, 239–261. [CrossRef] Tsujimoto, T.; Hosoda, N.; Uyama, H. Fabrication of Porous Poly(3-hydroxybutyrate-co-3-hydroxyhexanoate) Monoliths via Thermally Induced Phase Separation. Polymers 2016, 8, 66. [CrossRef] Gu, M.H.; Zhang, J.; Wang, X.L.; Ma, W.Z. Crystallization behavior of PVDF in PVDF–DMP system via thermally induced phase separation. J. Appl. Polym. Sci. 2006, 102, 3714–3719. [CrossRef] Cui, Z.Y.; Du, C.H.; Xu, Y.Y.; Ji, G.L.; Zhu, B.K. Preparation of porous PVDF membrane via thermally induced phase separation using sulfolane. J. Appl. Polym. Sci. 2007, 108, 272–280. [CrossRef] Wang, Y.J.; Zhao, Z.P.; Xi, Z.Y.; Yan, S.Y. Microporous polypropylene membrane prepared via TIPS using environment-friendly binary diluents and its VMD performance. J. Membr. Sci. 2018, 548, 332–344. [CrossRef] Park, M.J.; Kim, C.K. Fabrication of polyethylene microporous membranes using triethylolpropane tris(2-ethylhexanoate) as a novel diluent by a thermally induced phase separation process. J. Membr. Sci. 2014, 449, 127–135. [CrossRef] Tang, N.; Jia, Q.; Zhang, H.J.; Li, J.J.; Cao, S. Preparation and morphological characterization of narrow pore size distributed polypropylene hydrophobic membranes for vacuum membrane distillation via thermally induced phase separation. Desalination 2010, 256, 27–36. [CrossRef] Liu, Z.; Cui, Z.Y.; Zhang, Y.W.; Qin, S.H.; Yan, F.; Li, J.X. Fabrication of polysulfone membrane via thermally induced phase separation process. Mater. Lett. 2017, 195, 190–193. [CrossRef] Matsuyama, H.; Iwatani, T.; Kitamura, Y.; Tearamoto, M.; Sugoh, N. Formation of porous poly(ethylene-covinyl alcohol) membrane via thermally induced phase separation. J. Appl. Polym. Sci. 2015, 79, 2449–2455. [CrossRef] Jeon, S.; Karkhanechi, H.; Fang, L.F.; Cheng, L.; Ono, T. Novel preparation and fundamental characterization of polyamide 6 self-supporting hollow fiber membranes via thermally induced phase separation (TIPS). J. Membr. Sci. 2018, 546, 1–14. [CrossRef]
Polymers 2018, 10, 719
12. 13. 14. 15.
16. 17. 18.
19.
20.
21.
22. 23. 24. 25. 26. 27. 28. 29. 30.
31.
32. 33.
16 of 17
Fan, T.T.; Li, Z.H.; Cheng, B.W.; Li, J.X. Preparation, characterization of PPS micro-porous membranes and their excellent performance in vacuum membrane distillation. J. Membr. Sci. 2018, 556, 107–117. [CrossRef] Fu, X.; Matsuyama, H.; Teramoto, M.; Nagai, H. Preparation of hydrophilic poly(vinyl butyral) hollow fiber membrane via thermally induced phase separation. Sep. Purif. Technol. 2005, 45, 200–207. [CrossRef] Tsai, F.J.; Torkelson, J.M. Microporous poly(methy methacrylate) membranes: Effect of a low-viscosity solvent on the formation mechanism. Macromolecules 1990, 23, 4983–4989. [CrossRef] Lin, Y.K.; Tang, Y.H.; Ma, H.Y. Formation of a bicontinuous structure membrane of polyvinylidene fluoride in diphenyl carbonate diluent via thermally induced phase separation. J. Appl. Polym. Sci. 2009, 114, 1523–1528. [CrossRef] Cui, Z.L.; Hassankiadeh, N.T.; Lee, S.Y.; Lee, J.M. Poly(vinylidene fluoride) membrane preparation with an environmental diluent via thermally induced phase separation. J. Membr. Sci. 2013, 444, 223–236. [CrossRef] Wu, L.; Sun, J. An improved process for polyvinylidene fluoride membrane preparation by using a water soluble diluent via thermally induced phase separation technique. Mater. Des. 2015, 86, 204–214. [CrossRef] Ma, T.; Cui, Z.Y.; Wu, Y.; Qin, S.H.; Wang, H.; Yan, F. Preparation of PVDF based blend microporous membranes for lithium ion batteries by thermally induced phase separation: I. Effect of PMMA on the membrane formation process and the properties. J. Membr. Sci. 2013, 444, 213–222. [CrossRef] Cheng, Q.; Cui, Z.Y.; Li, J.B.; Qin, S.H.; Yan, F. Preparation and performance of polymer electrolyte based on poly(vinylidene fluoride)/polysulfone blend membrane via thermally induced phase separation process for lithium ion battery. J. Power Sources 2014, 266, 401–413. [CrossRef] Wu, Z.G.; Cui, Z.Y.; Li, T.Y.; Qin, S.H.; He, B.Q. Fabrication of PVDF-based blend membrane with a thin hydrophilic deposition layer and a network structure supporting layer via the thermally induced phase separation followed by non-solvent induced phase separation process. Appl. Surf. Sci. 2017, 419, 429–438. [CrossRef] Rajabzadeh, S.; Maruyama, T.; Ohmukai, Y.; Sotani, T.; Matsuyama, H. Preparation of PVDF/PMMA blend hollow fiber membrane via thermally induced phase separation (TIPS) method. Sep. Purif. Technol. 2009, 66, 76–83. [CrossRef] Lloyd, D.R.; Kim, S.S.; Kinzer, K.E. Microporous membrane formation via thermally induced phase separation. II. Liquid–liquid phase separation. J. Membr. Sci. 1991, 64, 1–11. [CrossRef] Sung, S.K.; Yeom, M.O.; Cho, I.S. Thermally-induced phase separation mechanism study for membrane formation, Membrane Formation and Modification. ACS Symp. 1999, 744, 42–64. Liu, M.; Liu, S.H.; Xu, Z.L.; Wei, Y.M.; Yang, H. Formation of microporous polymeric membranes via thermally induced phase separation: A review. Front. Chem. Sci. Eng. 2015, 12, 1–19. [CrossRef] Matsuyama, H.; Takida, Y.; Maki, T.; Teramoto, M. Preparation of porous membrane by combined use of thermally induced phase separation and immersion precipitation. Polymer 2002, 43, 5243–5248. [CrossRef] Li, X.F.; Wang, Y.G.; Lu, X.L.; Xiao, C.F. Morphology changes of polyvinylidene ouoride membrane under different phase separation mechanisms. J. Membr. Sci. 2008, 320, 477–482. [CrossRef] Cha, B.J.; Yang, J.M. Preparation of poly(vinylidene fluoride) hollow ber membranes for microfiltration using modified TIPS process. J. Membr. Sci. 2007, 291, 191–198. [CrossRef] Zhu, Z.X.; Meng, G.Z. c-TIPS method for membrane production. Membr. Sci. Technol. 2010, 30, 1–6. Lu, Z.P.; Xiaolong, L.; Wu, C.; Jia, Y.; Wang, X.; Chen, H.Y. Preparation of polyvinylidene fluoride hollow fiber porous membrane via a low thermally induced phase separation. Membr. Sci. Technol. 2012, 32, 12–22. Liu, M.; Wei, Y.M.; Xu, Z.L.; Guo, R.Q.; Zhao, L.B. Preparation and characterization of polyethersulfone microporous membrane via thermally induced phase separation with low critical solution tempera; ture system. J. Membr. Sci. 2013, 437, 169–178. [CrossRef] Liu, Y.F.; Xiao, T.H.; Bao, C.H.; Fu, Y.H.; Yang, X. Fabrication of Novel Janus Membrane by Nonsolvent Thermally Induced Phase Separation (NTIPS) for Enhanced Performance in Membrane Distillation. J. Membr. Sci. 2018. [CrossRef] Kim, J.F.; Kim, J.H.; Lee, Y.M. Thermally Induced Phase Separation and Electrospinning Methods for Emerging Membrane Applications: A Review. AIChE J. 2016, 62, 461–490. [CrossRef] Tang, Y.H.; Lin, Y.K.; Zhou, B.; Wang, X.L. PVDF membranes prepared via thermally induced (liquid-liquid) phase separation and theirapplication in municipal sewage and industry wastewater for water recycling. Desalin. Water Treat. 2016, 57, 1–19. [CrossRef]
Polymers 2018, 10, 719
34. 35. 36.
37.
38. 39.
40.
41. 42. 43. 44. 45.
46. 47.
17 of 17
Ji, G.L.; Zhu, L.P.; Zhu, B.K.; Zhang, C.F.; Xu, Y.Y. Structure formation and characterization of PVDF hollow fiber membrane prepared via TIPS with diluent mixture. J. Membr. Sci. 2008, 319, 264–270. [CrossRef] Liu, L.Q. Studies on Preparation of iPP Hollow Fiber Microporous Membrane via Thermally Induced Phase Separation. Master Thesis, Tianjin University, Tianjin, China, 2007. Lee, J.; Park, B.; Kim, J.; Park, S.B. Effect of PVP, lithium chloride, and glycerol additives on PVDF dual-layer hollow fiber membranes fabricated using simultaneous spinning of TIPS and NIPS. Macromol Res. 2015, 23, 291–299. [CrossRef] Hassankiadeh, N.T.; Cui, Z.; Kim, J.H.; Shin, D.W.; Lee, S.Y. Microporous poly(vinylidene fluoride) hollow fiber membranes fabricated with PolarClean as watersoluble green diluent and additives. J. Membr. Sci. 2015, 479, 204–212. [CrossRef] Gu, M.; Zhang, J.; Wang, X.; Tao, H.; Ge, L. Formation of poly(vinylidene fluoride) (PVDF) membranes via thermally induced phase separation. Desalination 2006, 192, 160–167. [CrossRef] Li, X.; Xu, G.; Lu, X.; Xiao, C. Effects of mixed diluent compositions on poly(vinylidene fluoride) membrane morphology in a thermally induced phase-separation process. J. Appl. Polym. Sci. 2008, 107, 3630–3637. [CrossRef] Zhou, Q.H.; Wang, Z.; Shen, H.J.; Zhu, Z.Y.; Liu, L.P.; Yang, L.B.; Cheng, L.N. Morphology and performance of PVDF TIPS microfiltration hollow fiber membranes prepared from PVDF/DBP/DOP systems for industrial application. J. Chem. Technol. Biotechnol. 2016, 91, 1697–1708. [CrossRef] Zhang, H.; Lu, X.; Liu, Z.; Ma, Z.; Wu, S. The unidirectional regulatory role of coagulation bath temperature on cross-section radius of the PVDF hollow-fiber membrane. J. Membr. Sci. 2018, 550, 9–17. [CrossRef] Cao, J.H.; Zhu, B.K.; Xu, Y.Y. Structure and ionic conductivity of porous polymer electrolytes based on PVDF-HFP copolymer membranes. J. Membr. Sci. 2006, 281, 446–453. [CrossRef] Gregorio, R.; Cestari, M. Effect of crysallization Temperature on the cyystalline Phase Content and Morphology of poly(vinylidene fluoride). J. Polym. Sci. Part B 1994, 32, 859–870. [CrossRef] Santos, T.D.; Kelvin, A.; Poletto, P.; Meireles, C.S.; Grisa, M.C. Effect of cellulose fibers on morphology and pure water permeation of PSf membranes. Desalin. Water Treat. 2011, 27, 72–75. [CrossRef] Jung, J.T.; Kim, J.F.; Wang, H.H.; Nicolo, E.D.; Drioli, E.; Lee, Y.M. Understanding the non-solvent induced phase separation (NIPS) effect during the fabrication of microporous PVDF membranes via thermally induced phase separation (TIPS). J. Membr. Sci. 2016, 514, 250–263. [CrossRef] Gregorio, R. Determination of the α, β, and γ crystalline phases of poly(vinylidene fluoride) films prepared at different conditions. J. Appl. Polym. Sci. 2006, 100, 3272–3279. [CrossRef] Gregorio, R.; Ueno, E.M. Effect of crystalline phase, orientation and temperature on the dielectric properties of poly(vinylidene fluoride) (PVDF). J. Mater. Sci. 1999, 34, 4489–4500. [CrossRef] © 2018 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).
