Elsevier

Desalination

Volume 279, Issues 1–3, 15 September 2011, Pages 201-209
Desalination

Hot pressing of electrospun membrane composite and its influence on separation performance on thin film composite nanofiltration membrane

https://doi.org/10.1016/j.desal.2011.06.009Get rights and content

Abstract

The primary objective of this paper was to investigate the influence of hot pressing on electrospun nanofibrous membrane (ENM) properties and subsequently the separation of salt after interfacial polymerization was carried out on the surface of ENM. Polyacrylonitrile solution was electrospun on a Hollytex backing material and this composite layer was subsequently subjected to hot pressing at 87 °C at pressures of 0.14, 0.28 or 0.41 MPa. The bubble-point of the hot pressed ENM significantly reduced with increasing pressure. Overlapping fibers fused at 0.28 MPa and over fusing occurred at 0.41 MPa hence accounting for the drastic decrease in bubble-point. In addition, the thickness of the ENM layer decreased with applied pressure as it compressed the ENM layer. A decrease in bubble-point resulted in a drastic decrease in pure water flux. Interfacial polymerization was carried out on the surface of ENM. The interfacial polymerized control was not able to withstand higher pressures. However, with treatment of 0.14 MPa onwards, the composite membrane was resilient at higher pressure and there wasn't a drastic difference in the average separation of the interfacially polymerized ENM. When compared to NF90, ENM-1 and ENM-2 had fluxes more than 3 folds. However the rejection was compromised by 8–12%.

Research highlights

► The properties (pore size and flux) of ENM are affected by hot pressing. ► Interfacial polymerization was carried out on hot pressed and non-hot pressed ENMs. ► The treated membranes possessed high flux and improved smoothness. ► These membranes are expected to offer energy saving performance in liquid filtration.

Introduction

Nanofiltration (NF) membrane is a type of pressure-driven membrane with properties in between ultrafiltration (UF) and reverse osmosis (RO). The transmembrane pressure (pressure drop across the membrane) required is lower (up to 3 MPa) than in RO, hence the operating cost is reduced significantly. The breakthrough in NF took place with the invention of a thin film composite (TFC) structure which comprises of three essential layers (1) top ultra-thin selective barrier layer, (2) middle porous support and (3) bottom non-woven fabric. Layers (1) and (2) could be carefully altered to produce the optimal separation performance. Concomitantly layer (3) does not influence the separation characteristics but rather it offers handling strength [1]. Over the years, improvements in performance of TFC membranes for aqueous applications have taken place mainly in terms of selectivity (solute rejection) without any appreciable change in membrane productivity (flux). It has become imperative in developing membranes, which provide higher fluxes or productivities without severely affecting membrane selectivity. In particular, the demand for developing membranes with high water flux is enormous for applications to industrial wastewater treatment and ultra-pure water production [2].

The top ultra-thin selective barrier layer is commonly prepared by interfacial polymerization (polycondensation reaction of polyfunctional amine and acid chloride monomers at the interface of two immiscible solvents) technique. The performance (solute rejection and flux) of the barrier layer is generally improved by the addition of additives during polymerization, post treatments (for e.g. by-product removal) and pretreatment conditions. Most of the research has been devoted in optimizing the top barrier layer to achieve a desired combination of solvent flux and solute rejection. The middle microporous layer is typically prepared by phase-inversion process and possesses a dense surface skin thus adopting an asymmetric structure. It generally offers maximal mechanical strength and compression resistance, combined with a minimal resistance to permeate flow. The porous support also plays a pivotal role in the formation of the barrier layer and hence influencing the selectivity of the top barrier layer [3].

Singh et al. (2006) [4] studied the effect of pore size (70 nm and 150 nm, designated as type 1 and type 2 membrane, respectively) of the polysulfone support layer on the morphology and performance of thin film polyamide membrane. Their results show that larger pores of type 2 membrane favors effective formation of polyamide inside in the pores and thereby reduced the thickness of thin polyamide film, whereas in type 1 membrane surface defects and two fold enhancement in the thickness were observed. NaCl (2000 ppm) rejection efficiency of 95–96% and permeate flux of 0.14–0.16 L/(m2.h.psi) for the type 1 membrane was achieved; while rejection efficiency of 45–66% and permeate flux of 0.32 L/(m2.h.psi) for type 2 membrane was obtained.

Recently, it has been shown that if the middle layer was made of a nanofibrous network, produced by electrospinning technique (with a surface porosity close to 80%), thin film nanofibrous composite membrane (TFNC) has a permeate flux 38% higher than commercial NF270 membrane with similar rejection rate [5]. Some of the salient features of TFNC membrane are highlighted in a recent review article by Chu et al. [6].

The primary objective of this paper is to study the influence of heat treatment and pressure on the electrospun nanofibrous membrane (ENM, to be used as middle layer) and how this treatment influences the membrane property as well as separation performance after interfacial polymerization (as TFNC membrane). To the best of our knowledge, there are no extensive researches reported on the influence of the above factors on electrospun membranes for nanofiltration application. Hot pressed TFNC membranes showed improved rejection, better pressure tolerance and mechanical performance when compared to untreated membrane.

Section snippets

Chemicals/solvents

Polyacrylonitrile (PAN) average Mw 150,000 (Aldrich Product Number 181315), N,N-Dimethylformamide GR ACS (DMF) (Merck Ltd, Product code 1.03053), Piperazine (Sigmaaldrich Product Number P-45907), Bipiperidine dihydrochloride (Sigmaaldrich Product Number 180742), Triethylamine (Sigmaaldrich Product Number T0886), Sodium hydroxide (Sigmaaldrich Product Number S8045), Water (Milli Q), 1,3,5-Benzenetricarbonyl trichloride (trimesoyl chloride abbreviated hereafter TMC) (Sigmaaldrich Product Number

Surface morphology of ENMs

PAN solution was electrospun on backing material (BM). It is essential to have a BM as it provides mechanical strength to the entire membrane structure. The average fiber size of the BM was 32 μm which is several orders greater than the size of the electrospun fibers. In addition, the bubble-point was ~ 80 μm. The surface architecture and pore size-distribution of this BM is reflected in Fig. 1, Fig. 2, respectively. Since the electrospun fibers are extremely long [8], they deposit easily on the

Conclusions

TFNC membranes based on ENM supports subjected to hot pressing were fabricated. The hot pressing of the nanofibrous support layer and backing material had a significant influence on the flux of the membranes and their mechanical and structural integrity. This subsequently influenced the separation performance, pressure tolerance and handling ease of the developed TFNC membranes. Without hot pressing, ENM-control had a large pore size (5.6 μm) as compared to ENM-1 (2.3 μm), ENM-2 (1.1 μm) and ENM-3

Acknowledgments

This work is supported by the Environment and Water Industry (EWI) Development Council (Govt. of Singapore) through the funded project “Development of low pressure, high flux UF and NF membranes based on electrospun nanofibers for water treatment” and NUS Nanoscience and Nanotechnology Initiative (NUSNNI), National University of Singapore. The author (Satinderpal Kaur) is thankful for the fellowship for research work at University of Ottawa. The authors gratefully acknowledge the financial

References (9)

There are more references available in the full text version of this article.

Cited by (124)

View all citing articles on Scopus
View full text