Elsevier

Chemical Engineering Science

Volume 123, 17 February 2015, Pages 283-291
Chemical Engineering Science

High performance PVDF-TiO2 membranes for water treatment

https://doi.org/10.1016/j.ces.2014.10.047Get rights and content

Highlights

  • TiO2 nanoparticles improve membrane structure and permeability.

  • TiO2 addition and UV irradiation limit pure water flux decline and enable high fluxes.

  • UV cleaning of fouled composite membrane enables total recovery of performances.

Abstract

In order to obtain low-fouling membranes, TiO2 nanoparticles were entrapped in PVDF membranes prepared by the NIPS wet-process. Typical asymmetric membrane structure was obtained. Membrane structure, hydrophilic properties and permeability were improved in comparison with PVDF neat membrane when increasing TiO2 concentration up to an optimum concentration of 25%wt. Maximum permeate flux of 150 L/h/m2 was successfully obtained. For TiO2 content beyond 25%wt, TiO2 particles agglomeration prevents the improvement of hydrophilic properties and permeability. Under UV irradiation, phenomena of super-hydrophilicity due to presence of TiO2 in the composite membrane permits to suppress pure water permeate flux decline and reach higher fluxes. Fouled composite membranes after BSA filtration were successfully cleaned using water and UV irradiation. Permeate flux was totally recovered after this cleaning.

Introduction

If membrane technologies have emerged as advanced separation processes in water treatment over the last decades as they can be operated with minimal chemical, low energy, easy automation and optimal quality of treated water, membrane fouling remains today as the main limitation of the process. In particular, the very little size of organic pollutants causes rapid and severe internal and surface fouling which results in a strong decrease of membrane permeate flux and separation performances. PVDF is a common ultrafiltration and microfiltration membrane material because of its good mechanical properties, thermal stability and chemical resistance but its hydrophobicity induces a high tendency of these membranes to fouling.

Recent studies have investigated the possibility of coupling membrane filtration and a photocatalyst, because of interesting property of these membranes to mitigate membrane fouling (Cao et al., 2006, Mozia, 2010). Among photocatalysts, anatase-type titanium dioxide (TiO2) presents several advantages: an important photocatalysis activity under UV irradiation, a high stability, a low environmental impact, a low cost and an important availability (Mills A. and Lee 2002). Two main approaches associate membranes and catalytic TiO2 nanoparticles to form composite membranes are possible: blending nanoparticles in the membrane matrix or coating the nanoparticles on the surface of the membrane (Mozia 2010). Nonetheless, when using this second configuration, a release of catalyst nanoparticles could be observed (Bian et al., 2011, Alaoui et al., 2009, Reijnders, 2009) due to the difficulty to immobilize them on membranes without using binding mediums to form covalent bonds between nanoparticles and membrane. Despite many attempts made to find appropriate organic binders, the residual release of nanoparticles from the membrane may still raise questions about the properties of the membrane during long filtration period. The first configuration where TiO2 nanoparticles are entrapped inside the membrane matrix presents thus practical advantages: (i) the particles release should be limited, (ii) all advantages of membrane process such as easy scale-up and modularity are maintained. PVDF is a good candidate for such coupling because of its high resistance to UV degradation and photocatalytic activity.

Entrapped TiO2 composite membranes can be prepared by the induced phase separation process (usually Non-solvent Induced Phase Separation wet-process, or NIPS wet-process) (Damodar et al., 2009), where nano-sized TiO2 particles are added to the polymer-solvent solution, which is cast on an appropriate plate and then immersed in a coagulation bath of non-solvent (usually water) to induce phase separation. Different studies from literature have shown that entrapped TiO2 / polymer membranes prepared by NIPS process could present higher permeabilities and self-cleaning capacities than neat polymer membranes.

As observed on Table 1, literature review shows that TiO2 can have a beneficial effect on membrane properties (hydrophilicity for example) and performances (permeability for example). However, some contradictory results can be observed and the exact influence of the operating parameters of preparation on composite membrane properties may still be unclear. Some authors found a great improvement of membrane surface hydrophilicy with TiO2 (Yang et al., 2006, Yu et al., 2009, Hamid et al., 2011, Yuliwati and Ismail, 2011) while others observed only a slight increase (Alaoui et al., 2009, Damodar et al., 2009, Bae and Tak, 2005, Oh et al., 2009). Few authors also pointed out a decrease of this hydrophilicity for TiO2 concentration higher than 6–10 wt% TiO2/PVDF (Damodar et al., 2009, Yu et al., 2009, Yuliwati and Ismail, 2011). Same trends and contradictory results are observed concerning membrane permeability. Although authors seem to agree about the existence of an optimum TiO2 concentration for the improvement of permeability, this value can change dramatically according to the studies: 3-4 wt% TiO2/PVDF for Song et al. and Wu et al. Song et al., (2012; Wu et al., 2008); 6–11 wt% TiO2/PVDF for Yang et al., Yu et al. and Yulliwati et al. Yang et al., (2006; Yu et al., 2009; Yuliwati and Ismail 2011); or 24–25 wt% TiO2/PVDF for Li et al. Li et al., (2009).

Compared to neat polymer (PVDF, PES) membrane under UV light, a photocatalytic property due to the presence of TiO2 was put in evidence despite this photocatalytic property was not enhanced significantly by an increase of TiO2 content from 0 to 20 wt% TiO2/PVDF. Ngang et al. observed a great improvement of self-cleaning capacity of the membrane during filtration of dye solution (Ngang et al., 2012). Song et al. did not observe any clear effect of TiO2 on pure water flux (Song et al., 2012) but a self-cleaning ability was offered by UV irradiation prior filtration experiments (Vatanpour et al., 2012) and enhanced by a continuous UV irradiation (Song et al., 2012). Finally, Yang et al. put in evidence that mechanical strength of PSF membrane was enhanced through addition of TiO2 (Yang et al., 2007).

Comparison between these studies is not easy as a large range of mass of polymer (from 10 to 27% of the total mass of casting solution) and type of porogen were used leading to great differences in viscosity of the casting solution and in TiO2/polymer ratios. In this context, to better understand the real effect of nanoparticles addition to polymer membranes, a complete study with investigation of a wide range of TiO2 concentrations is necessary. Moreover, the effect of UV irradiation was always tested during foulants filtration and there is no result in literature about UV irradiation during pure water filtration despite this last point is of primary importance to well understand basic phenomena responsible for the increase of composite membrane performances and hydrophilicity. Cleaning of TiO2 membranes after filtration has been studied only a few times and for very different pollutants. Rahimpour et al. show an improvement of permeate flux recovery ratio for TiO2 composite membranes after simple water cleaning following a BSA filtration (Rahimpour et al., 2011). Rahimpour et al. observed same results for milk filtration and a further limited improvement after static UV irradiation (Rahimpour et al., 2008). Same tendency are also observed by Samblante 2013 but for colorant filtration (Semblante et al., 2013).

The aim of this article is to bring new insights in composite PVDF/TiO2 membranes preparation by NIPS process. The originality of this article lies in three main points. First, the effect of TiO2 concentration was tested over a wide range of concentration in order to deeply understand the influence of this parameter. Second, behavior of water flux along time under UV irradiation was investigated, for the first time to our knowledge, for neat polymer and composite membranes. Finally, a study of water flux and self-cleaning properties under UV irradiation was performed.

The general objective is the development of composite membranes to be used in a UV-TiO2 photocatalytic system with an objective to limit fouling and thus reduce the frequency of chemical cleaning.

Section snippets

Material

Polymer polyvinylidene fluoride (PVDF, average molecular weight=275,000 g mol−1), solvent N-N-Dimethylacetamide (DMAc, assay >99.5%) and Bovine Serum Albumine (BSA) (assay >98%) were purchased from Sigma Aldrich. Additive polyethylene glycol 200D (PEG200, average molecular weight=200 g mol−1) was purchased from Merck. Nanoparticles were Aeroxide ® TiO2 P25 nano-powder (about 85% anatase-15% rutile, size of c.a. 21 nm, assay ≥99.5%) purchased from Sigma Aldrich.

Membrane preparation protocol

Membranes were prepared by the

Membrane structure

Membranes were first prepared with different TiO2 concentrations (0 to 30 wt. % TiO2/PVDF). Membranes cross-section and top surface were observed by SEM. Fig. 1 present these results for several TiO2 concentrations: 4 wt. % TiO2/PVDF (membrane T4P5, Fig. 1b), 15 wt. % TiO2/PVDF (membrane T15P5, Fig. 1c) and 30 wt. % TiO2/PVDF (membrane T30P5, Fig. 1d). The observation of a neat polymer membrane is also presented (membrane T0P5, Fig. 1a). Top surface is membrane side in contact with water interface

Conclusion

TiO2 entrapped PVDF membranes were successfully prepared by phase inversion method. Influence of TiO2 content over a wide range of concentration was investigated through membrane structural characterization, surface hydrophilicity determination and water flux measurement.

Opposite results were obtained according to the TiO2 content added to the PVDF membrane. For low concentration, no improvement of membrane properties was observed. For intermediate TiO2 content, permeability and flux

References (27)

Cited by (0)

View full text