Elsevier

Journal of Membrane Science

Volume 618, 15 January 2021, 118708
Journal of Membrane Science

Blending modification to porous polyvinylidene fluoride (PVDF) membranes prepared via combined crystallisation and diffusion (CCD) technique

https://doi.org/10.1016/j.memsci.2020.118708Get rights and content

Highlights

  • Use of PMMA as a blending additive to reduce pore sizes of PVDF CCD membranes.

  • PMMA addition influences gelation thermodynamics and molecular diffusion kinetics.

  • DMSO solvent crystal growth during CCD impeded by PMMA addition; evidenced by SEM.

  • Molecular weight of PMMA influences degree of its enrichment on membrane surfaces.

  • Water permeance maintained for the blend membranes despite pore size reduction.

Abstract

PVDF membranes prepared by the combined crystallisation and diffusion (CCD) technique, a variation of ice templating, have shown permeation rates far superior to those made via conventional methods such as nonsolvent induced phase separation (NIPS) or thermally induced phase separation (TIPS). However, so far the smallest pore size achieved for PVDF membranes prepared by CCD is about 30 nm. In this study, the pore size of CCD PVDF membranes is reduced to 15 nm by blending poly(methyl methacrylate) (PMMA) into PVDF-dimethyl sulfoxide (DMSO) dope solutions without altering the operational conditions. Systemic study shows the pore size reduces when increasing PMMA concentration in the membranes; but after an initial drop compared to the pure PVDF membrane, the pure water flux doesn't decline with PMMA concentration. The effect of different molecular weights of PMMA (120 kDa and 35 kDa) was also investigated, where the results showed a smaller pore size for membranes prepared using the higher molecular weight PMMA. Scanning electron microscopy (SEM), cloud-point temperature measurement and Fourier transform infrared spectroscopy (FTIR) was used to characterise the membranes and help understand how PMMA affects membrane formation. From the evidence gathered, we attribute the effects of PMMA to the change in the number of nucleation centres, diffusional rates, as well as the gelation point based on a previously proposed CCD mechanism.

Introduction

Polyvinylidene fluoride (PVDF) is one of the most commonly used materials for making ultrafiltration and microfiltration membranes. This is because of its very desirable properties such as high chemical and thermal resistance, excellent mechanical strengths, and most importantly its ability to withstand chorine disinfection, making it one of the most suitable materials for use in wastewater treatment applications [1]. Despite the numerous advantages, PVDF membranes face several challenges such as low permeation fluxes and high fouling. Typical commercial ultrafiltration PVDF membranes have pure water permeation fluxes less than 150 LMH bar−1 with pore sizes around 0.03 μm [2]. Therefore, when using PVDF membranes to process large quantities of feed water, large membrane areas are required to compensate for low permeation fluxes. This leads to a more significant footprint requirement, therefore, leads to increased installation and maintenance costs.

Most of these commercial PVDF membranes that suffer from low permeation fluxes are prepared via either NIPS, TIPS, or the two combined [[3], [4], [5], [6], [7], [8]]. Also, many modifications to these fabrication techniques are suggested, which only marginally improve the permeation performance while maintaining the rejection properties [2]. A new method of fabrication called combined crystallisation and diffusion (CCD) was developed to address this problem [2]. The CCD technique not only produces the desired asymmetric structure, where the membrane has a tight separation layer and a much porous support layer characterised by large interconnected microchannels, but also has fewer influencing factors compared to the traditional techniques, making batch reproducibility much easier [2,9].

CCD technique involves unidirectional cooling of a cast polymer solution to freeze the solvent into crystals which then serve as pore templates after leaching of the solvent crystals. To achieve this, a polymer solution is spread via a casting knife on a casting plate and then placed on a cooling plate maintained at −30 °C. After solidification of the cast film, it is placed in an ice-cold water bath to leach the solvent crystals out. The result is a porous membrane, ready for filtration applications. Factors influencing the permeation and rejection properties of the CCD membranes include solvent choice, cooling temperature, cooling rate, and polymer dope concentration. Dimethyl sulfoxide (DMSO) is the ideal solvent choice as it freezes at a relatively higher temperature at 19 °C and therefore enables solvent crystallisation to precede phase separation via polymer precipitation [10]. The cooling rate and the cooling temperature influences the extent of solvent crystallisation. It has been observed that faster cooling rates produce membranes with smaller pore sizes. The effect of cooling rate and temperature on the pore sizes is elaborately detailed in the previous work on CCD [9].

Previously for PVDF membranes prepared by CCD technique, the smallest mean flow pore size achieved was 29 ± 3 nm with a pure water flux (PWF) as high as 570 ± 37 LMH.bar−1 [2]. To decrease the pore size further, one could either apply an even faster cooling rate or further increase the dope concentration or both. For faster cooling rates, one could either improve heat conduction or use a lower temperature, but these have design challenges or intensified energy requirements associated with them. Increasing the polymer dope concentration further is also challenging as dissolution would be a lot slower, and a higher dissolution temperature might be needed, both of which are undesirable from a production standpoint. Additionally, the PWF would be considerably affected in membranes prepared with higher concentration dopes owing to the reduced membrane porosity.

This work focuses on reducing the pore size of the membranes by neither changing the cooling rates, nor by increasing the polymer dope concentration, but rather by using an additive to the dope solution, namely poly(methyl methacrylate) (PMMA). PMMA was chosen as the additive as it is one of the few polymers that is thermodynamically compatible with PVDF over a wide range of blend compositions, is water-insoluble and hydrophilic [[11], [12], [13]]. Alternatively, polyethylene glycol (PEG) could also be chosen; however, PMMA, unlike PEG, is not water-soluble and hence would not get washed away during operation and therefore maintain the membrane porosity [14]. We will show that PMMA can change the thermodynamic and kinetic properties of the polymer solution during the CCD process, and limit solvent crystal growth, resulting in smaller solvent crystals and therefore smaller pore sizes in the final membrane.

Section snippets

Materials

Commercial PVDF (Kynar®K-761, MW = 440 kDa, ρ = 1.79 g cm−3) was purchased from Elf Atochem. Two different molecular weights (MW) of PMMA (120 kDa, 35 kDa) were used, where 120 kDa PMMA was obtained from Merck Sigma Aldrich while 35 kDa PMMA was obtained from VWR, UK. DMSO (HPLC grade), Isobutanol (HPLC grade), n-hexane (GPR), and ethanol (Analar) were purchased from VWR, UK. DMSO was used as a solvent for PVDF in preparing the dope solutions. Isobutanol was used to prepare isobutanol-water

SEM

Fig. 1 shows the membrane morphology of PVDF-PMMA blend membranes (18PVDF-2PMMA-120kDa and 18PVDF-2PMMA-35kDa) and pure PVDF membrane (20PVDF-0PMMA). The separation side surface images for 20PVDF-0PMMA (Fig. 1-1a) and 18PVDF-2PMMA-35kDa (Fig. 1-2a) look very similar, while that for 18PVDF-2PMMA-120kDa (Fig. 1-3a) shows a much smaller pore size. For support side surface images, the differences between pure PVDF membrane sample (Fig. 1–1b) and the blend membranes' samples (Fig. 1, Fig. 2, Fig. 3b

Increase in PMMA and a decrease in pore size

The SEM images and the mean flow pore size results indicate that increasing the proportion of PMMA, produces membranes characterised by narrower micro-channels in the support layer and smaller pore sizes in the separation layer.

The proposed polymer-solvent binary phase diagram for CCD method, as illustrated in Fig. 7, can be used to explain this trend [9]. The schematic phase diagram shows that when the homogenous polymer solution is cooled way below the freezing point (Tm) of the solvent, even

Comparison with industry standards

PVDF membranes is one of the most popular membranes used in water based filtration applications and therefore have many commercial producers. The pure water fluxes (PWF) along with the pore sizes of the PVDF membranes produced by the industry leaders are compared to our current work and presented in Table 7. In terms of PWF, one can see that CCD based membranes with similar pore sizes clearly outperform the commercial industry standards.

Conclusion

In this study, PMMA was added as an additive to PVDF-DMSO dope solutions to decrease the pore size of PVDF based CCD membranes. Addition of PMMA to pristine PVDF does decrease the mean flow pore size and the PWF, however, further increase in the proportion of PMMA keeps the PWF levels maintained, while further reducing the pore size. Furthermore, of the two molecular weight sizes of PMMA chosen (120 kDa and 35 kDa) in preparation of the blend membranes, the ones prepared using higher molecular

CRediT authorship contribution statement

Vatsal Shah: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing - original draft, Writing - review & editing, Software, Data curation, Visualization. Bo Wang: Conceptualization, Methodology, Writing - review & editing. Kang Li: Conceptualization, Methodology, Resources, Writing - review & editing, Supervision, Project administration, Funding acquisition.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors gratefully acknowledge the research funding provided by EPSRC in the United Kingdom (Grant no EP/M01486X/1).

References (50)

  • J.I. Calvo et al.

    Comparison of pore size distributions from dextran retention tests and liquid-liquid displacement porosimetry

    Microporous Mesoporous Mater.

    (2017)
  • R.I. Peinador et al.

    Characterisation of polymeric UF membranes by liquid–liquid displacement porosimetry

    J. Membr. Sci.

    (2010)
  • M.B. Tanis-Kanbur et al.

    Membrane characterization via evapoporometry (EP) and liquid-liquid displacement porosimetry (LLDP) techniques

    J. Membr. Sci.

    (2019)
  • S. Rajabzadeh et al.

    Preparation of PVDF/PMMA blend hollow fiber membrane via thermally induced phase separation (TIPS) method

    Separ. Purif. Technol.

    (2009)
  • Y.P. Mahant et al.

    Electrospun poly (vinylidene fluoride)/poly (methyl methacrylate) composite nanofibers polymer electrolyte for batteries

    Procedia Materials Science

    (2015)
  • H.P. Zhang et al.

    A novel sandwiched membrane as polymer electrolyte for lithium ion battery

    Electrochem. Commun.

    (2007)
  • M. Peyravi et al.

    Tailoring the surface properties of PES ultrafiltration membranes to reduce the fouling resistance using synthesized hydrophilic copolymer

    Microporous Mesoporous Mater.

    (2012)
  • A. Rahimpour

    UV photo-grafting of hydrophilic monomers onto the surface of nano-porous PES membranes for improving surface properties

    Desalination

    (2011)
  • S. Belfer et al.

    Surface characterization by FTIR-ATR spectroscopy of polyethersulfone membranes-unmodified, modified and protein fouled

    J. Membr. Sci.

    (2000)
  • J.L. Garcia et al.

    Mechanical and wear properties of polymethylmethacrylate and polyvinylidene fluoride blends

    Polymer

    (1998)
  • T. Lindvig et al.

    A Flory–Huggins model based on the Hansen solubility parameters

    Fluid Phase Equil.

    (2002)
  • A. Bottino et al.

    The formation of microporous polyvinylidene difluoride membranes by phase separation

    J. Membr. Sci.

    (1991)
  • V.Y. Goldstein et al.

    Heterogeneous nucleation in crystallization: impact of impurities and local melt inhomogeneity

    Mater. Lett.

    (2018)
  • B. Wang et al.

    Crystal nuclei templated nanostructured membranes prepared by solvent crystallization and polymer migration

    Nat. Commun.

    (2016)
  • J. Lee et al.

    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)
  • Cited by (24)

    View all citing articles on Scopus
    View full text