Effects of surface crosslinking on sieving characteristics of chitosan/poly(acrylonitrile) composite nanofiltration membranes

https://doi.org/10.1016/S1383-5866(00)00188-XGet rights and content

Abstract

Effects of surface crosslinking of chitosan/poly(acrylonitrile) (PAN) composite nanofiltration membranes at different crosslinker (glutaraldehyde) concentrations and crosslinking times on their surface chemical composition and sieving properties such as pure water permeation, molecular weight cut-off and the rejection of mono/divalent salts and mono/oligosaccharides were investigated. Fourier transform infrared-attenuated total reflectance spectroscopy (FTIR-ATRS) and X-ray photoelectron spectroscopy (XPS) studies revealed the crosslinking of chitosan with glutaraldehyde as well as variations in chemical composition with glutaraldehyde concentration and crosslinking time. Pure water permeation/swelling in water decreased and rejection of salts and sugars increased with increasing glutaraldehyde concentration, indicating pore contraction and increase in hydrophobicity as well as pore tortuosity due to crosslinking. Molecular weight cut-offs of surface crosslinked membranes were in the range of 550–700 Da, a characteristic of nanofiltration membranes, whereas uncrosslinked membrane had cut-off of >1500 Da. The crosslinked membranes were found to be stable over 10-h operation for pure water permeation and the stability increased with increasing glutaraldehyde concentration.

Introduction

Nanofiltration (NF) is relatively a new membrane process and is increasingly being used in a range of applications that involve removal of salts from small molecular weight organics or separation of small molecular weight solutes from macromolecules. Most of the NF membranes developed so far are composite in nature, with a selective layer on top of the microporous substrate. Various membrane materials have been used for a selective layer on polymeric or inorganic microporous substrates. The choice of materials depends on both chemical and physical compatibility of selective layer with the substrate, which in turn determines the stability and performance of the resulting composite membranes. Poly(acrylonitrile) (PAN), one of the versatile polymeric materials, in addition to its well established use for asymmetric ultrafiltration (UF) membranes has been used as a microporous substrate for making composite membranes [1], [2], [3]. Chitosan, a natural hydrophilic biopolymer, has been used as a membrane material for reverse osmosis [4], pervaporation (PV) [1], [5], and gas separation membranes [6], and recently we have demonstrated its use in composite UF membranes [7].

Since PAN is more hydrophilic compared to the commonly used polysulfone (PSF) substrate, the composite NF membranes of PAN with hydrophilic chitosan may be relatively more stable than chitosan/PSF composite membranes, during long-term operation. Although work on chitosan/PAN composite PV [1], [2], [8] and UF [7] membranes has been reported, to the best of our knowledge there is no literature on chitosan/PAN composite NF membranes. The hydrophilic chitosan/PAN composite NF membranes would be of interest in applications such as demineralization of whey, where hydrophilicity of the membranes would help to minimize fouling caused by protein adsorption. Most of the commercial NF membranes are charged, which is beneficial for removal of divalent salts by Donnan exclusion. However, charged membrane surface may cause significant fouling due to electrostatic interactions, particularly in processing protein containing feeds. Krajewska and Olech [9] have reported that crosslinking of chitosan gel membranes results in reduction in pore size and water absorption with increasing crosslinker (glutaraldehyde) concentration. Therefore it is of interest to study the formation of composite NF membranes with a hydrophilic and neutral (at pH ≥7.0) selective layer of chitosan on PAN substrate and subsequent surface crosslinking.

The present study reports the effect of surface crosslinking at different crosslinker (glutaraldehyde) concentrations and crosslinking times on surface chemical composition and sieving characteristics such as pure water permeation, molecular weight cut-off and rejection of model solutes, of chitosan/PAN composite NF membranes.

Section snippets

Materials

PAN (MW≈150 kDa) and chitosan (MW≈116 kDa and degree of deacetylation (d.d.) ≈86%) were obtained from Polysciences, USA. Polyethylene glycols of different molecular weights and glutaraldehyde were obtained from Fluka, USA, and Aldrich, USA, respectively while all other solutes and reagents were obtained from Anachemia, Canada and used as received. Reverse osmosis treated water with a conductivity of ≈5×10−4 S m−1 was used for membrane preparation, swelling and permeation studies.

Preparation of chitosan/PAN composite NF membranes (PANCHINF)

The details of

Results and discussion

The average pore size of the composite membranes calculated using the PEG rejection data (Fig. 5) and Ferry's modified equation [12] was found to be of the order of ≈10−9 m. Therefore, after crosslinking the exterior surface, the diffusivity of glutaraldehyde molecules inside the top crosslinked chitosan layer may be reduced and hence we believe that crosslinking might have taken place only on the exterior surface of chitosan layer.

The reaction scheme for the crosslinking of chitosan with

Conclusions

This study concluded that crosslinking of top surface of chitosan/PAN composite NF membrane with glutaraldehyde could improve the membrane stability and lower molecular weight cut-offs.

The increase in C/N ratio on the membrane surface with increasing glutaraldehyde concentration as well as CT, observed by XPS spectroscopy, indicated an increase in crosslinking. FTIR-ATR spectra showed peaks characteristic of chitosan in the control membrane, whereas in case of crosslinked membranes, a distinct

Acknowledgements

The authors would like to thank G. Pleizier for SEM and XPS analysis, S. Argue for atomic absorption spectroscopic analysis and M.M. Dal-Cin for operation of automated casting machine for base membrane preparation.

References (16)

  • X.P Wang et al.

    J. Membr. Sci.

    (1996)
  • X Feng et al.

    J. Membr. Sci.

    (1996)
  • Y Mizushima

    J. Non-Cryst. Solids

    (1992)
  • B Krajewska et al.

    Polymer Gels Networks

    (1996)
  • G Capannelli et al.

    Ultrafiltration membranes: characterization methods

    J. Membr. Sci.

    (1983)
  • T Uragami et al.

    J. Membr. Sci.

    (1994)
  • S.Y Nam et al.

    J. Membr. Sci.

    (1997)
  • J Schaep et al.

    Sep. Purif. Tech.

    (1998)
There are more references available in the full text version of this article.

Cited by (88)

  • Bio-sourced and biodegradable materials for membrane fabrication

    2023, Green Membrane Technologies towards Environmental Sustainability
  • Nanofiltration membrane technologies

    2022, Advancement in Polymer-Based Membranes for Water Remediation
  • Carboxymethyl cellulose/polyethersulfone thin-film composite membranes for low-pressure desalination

    2021, Separation and Purification Technology
    Citation Excerpt :

    Indeed, by increasing the degree of crosslinking, more hydroxyl groups of the CMC react with aldehyde groups of GA, resulting in a decreasing number of free hydroxyl groups. In addition, more hydrophobic GA moieties were inserted into the selective layer of the NF membrane by enhancing the crosslinking degree, leading to a decrease in hydrophilicity of the membranes [47,48]. The stable permeate flux of the membranes after 75 min was observed, as shown in Fig. 5 ((a)-(f)).

View all citing articles on Scopus

Issued as NRCC No: 42009

View full text