Link between interfacial interaction and membrane fouling during organic solvent ultrafiltration of colloidal foulants

https://doi.org/10.1016/j.memsci.2020.118369Get rights and content

Highlights

  • Flux decline trends agreed with DLVO-based foulant-membrane interaction energy.

  • Some cake resistance agreed with DLVO-based foulant-foulant interaction energy.

  • Cake compressibility disagreed with foulant-foulant DLVO, XDLVO and zeta potentials.

  • XDLVO known to predict fouling well for aqueous, but did poorly for organic solvent.

  • DLVO provided better membrane fouling predictability than XDLVO or zeta potential.

Abstract

Membrane fouling is inevitable for liquid membrane-filtration processes. While the relationship between surface free energy and membrane fouling is now well-established for aqueous feeds, such a link remained missing for the emerging organic solvent applications. In this study, three organic solvents (namely, methanol, ethyl acetate and acetonitrile) and three colloidal foulants (namely, silicon dioxide (SiO2), titanium dioxide (TiO2) and aluminum (Al)) were investigated using the same polyacrylonitrile (PAN) ultrafiltration membrane at five transmembrane pressures (100–500 kPa). The zeta potentials were measured, and the Gibbs free energy of interfacial interactions for both foulant – membrane and foulant – foulant were quantified via the DLVO and XDLVO models to correlate with the fouling trends. Surprisingly, although the XDLVO-based interaction energy is generally acknowledged to correlate well with membrane fouling trends for aqueous feeds, the results here indicate that qualitative agreement with flux decline, cake resistance and cake compressibility was poor. Zeta potentials also exhibited poor predictive capability. Only the DLVO-based interaction energy values agreed with the flux decline trends, but nonetheless failed for cake compressibility trends. This study underscore the different fouling behaviors between aqueous versus organic solvent environments, particularly in the limited capacity of the XDLVO model and zeta potential values to provide insights on the fouling mechanisms.

Introduction

Fouling is an inevitable and challenging issue in membrane filtration processes that is detrimental due to the associated permeate flux decline and greater energy needed to maintain the same flux [1]. As one of the key membrane fouling mechanisms, the cake layer formed on the membrane surface has been extensively characterized and elucidated experimentally for aqueous feeds [2,3], and also well-described by mathematical models [4,5]. The resistance to permeation caused by a cake layer not only relies on the thickness of the layer, but also on its compressibility (i.e., density) [6]. Unsurprisingly, many studies on various aspects of cake compressibility are available, though the focus has been on aqueous systems in the field of wastewater treatment [7,8]. As organic solvent membrane-filtration gains momentum [9,10], an analogous effort is needed, which formed the motivation for the current study.

Understanding on the cake layer formed during membrane-based filtration of aqueous feeds has advanced progressively. As is well-acknowledged, the specific resistance of an incompressible cake layer is constant, but increases as the cake structure is compressed [11,12]. Changes in cake compressibility, which is a measure of the change in the cake structure (e.g., compaction, rearrangement) in response to an increase of pressure, have been reported to be independent of the membrane surface properties, but related to the aqueous environment (e.g., pH, ionic strength) and foulant properties (e.g., size, shape, rigidity, tendency to flocculate) [8,[13], [14], [15], [16]]. Compared to deformable particles (e.g., suspended solids in activated sludge), the cake layer formed by rigid colloidal particles (e.g., mineral oxides) is less compressible [13]. Bourcier et al. [16] found that the cake layer formed by small particles with wider size ranges was more compressible, and developed a model that incorporated factors like cake porosity, average particle size and size range to predict cake compressibility. Along the same vein, larger flocs resulting from foulant coagulation have been reported to form looser and more compressible cake layers [17]. Regarding the aqueous environment, the compressibility of the cake layer formed by colloidal particles has been reported to decrease with the increase of ionic strength or decrease of pH [13], which is attributed to the change in interfacial interactions. Unfortunately, an analogous understanding of the characteristics of the cake formed in organic solvents remains poor.

The free energy of surface interactions has been used to interpret the membrane fouling behaviors in aqueous filtration systems, with attractive membrane - foulant and foulant - foulant interactions tending to exacerbate membrane fouling by facilitating foulant deposition onto the membrane and fouling layer [18,19]. The surface charge of suspended particulates is known to significantly impact the structure and resistance of the cake layer [20]. The Lifshitz–van der Waals (LW), Lewis acid-base (AB), electrostatic double layer interaction (EL) and Brownian forces are the well-accepted non-covalent interfacial interactions incorporated in the classical DLVO and extended DLVO (XDLVO) theories to quantify the surface free energy [21]. Different interfacial forces between foulants and membranes due to different surface properties have been observed and proven to affect the membrane fouling behaviors [18,19,22]. Generally, the surface interactions predicted by the XDLVO approach demonstrated higher qualitative agreement with membrane fouling behavior compared to that from the classical DLVO approach in aqueous media due to the dominance of the AB component [23,24]. Such understandings are largely limited to results obtained from aqueous feeds, so the question of whether the XDLVO model can be applied to predict membrane fouling during the filtration of organic solvents remained unaddressed.

The membrane fouling phenomena during organic solvent filtration have been reported in a few studies [[25], [26], [27], [28]], but none of these studies provided more in-depth mechanistic analysis on the fouling data. To bridge this gap, this study was targeted at understanding the membrane fouling phenomena during the filtration of different colloidal foulants in different organic solvents. Three colloidal foulants (namely, silicon dioxide (SiO2), titanium dioxide (TiO2) and aluminum (Al)) and three organic solvents (namely, methanol, ethyl acetate and acetonitrile) were investigated using the same polyacrylonitrile (PAN) ultrafiltration membrane. The evolution of flux decline and cake resistance, as well as cake compressibility, were characterized for each pair of colloidal foulant and organic solvent. The zeta potentials and Gibbs free energy of interfacial interactions for both foulant – membrane and foulant – foulant were quantified to attempt to correlate with the fouling trends.

Section snippets

The classical and extended DLVO theory

The total Gibbs free energy of surface interaction (ΔG) can be expressed as the sum of the Lifshitz–van der Waals (LW) and electrostatic double layer interaction (EL) interaction energies according to the classical DLVO theory (Equation (1)), while an additional Lewis Acid-Base (AB) interaction is considered in the extended DLVO (XDLVO) theory (Equation (2)). The Brownian contribution was assumed negligible in this study of particles sized on the order of hundreds of nanometers. Negative and

Chemicals and reagents

The acetonitrile (99.8%, anhydrous) and ethyl acetate (99.8%, anhydrous) used in this study were purchased from Sigma-Aldrich. The methanol (>99.9%, analytical reagent grade) was purchased from Fisher Scientific. Silica (SiO2) spheres (given by vendor as 400 nm in diameter; > 99% purity), titanium oxide (TiO2) spheres (given by vendor as 500 nm in diameter; 99.9% purity) and aluminum (Al) spheres (given by vendor as 500 nm in diameter; 99.9% purity) were purchased from Nanostructured and

Particle size distribution and zeta potential

The particle size distribution of each colloidal foulant is depicted in Fig. 1. The average diameters of SiO2, TiO2 and Al particles were determined to be 399.1 nm, 547.7 nm and 592.5 nm, respectively. The sizes of even the smallest foulants were much larger than the pore size of the selected ultrafiltration membrane, thus only external fouling was possible, with cake layer being the key membrane fouling mechanism.

The zeta potentials of the colloidal foulants (ζs) and the membrane surfaces (ζm)

Conclusion

In order to address the question on whether the understanding on membrane fouling for aqueous feeds can be directly applicable for organic solvent filtration, a series of ultrafiltration experiments involving three colloidal foulants and three organic solvents at five transmembrane pressures were carried out. The zeta potentials were measured, and the Gibbs free energy of interfacial interactions for both foulant – membrane and foulant – foulant were quantified via the DLVO and XDLVO models to

CRediT authorship contribution statement

Ziqiang Yin: Conceptualization, Methodology, Data curation, Writing - original draft. Rique Jie En Yeow: Investigation. Yunqiao Ma: Methodology, Validation. Jia Wei Chew: Supervision, Writing - review & editing, 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

We acknowledge funding from the Singapore GSK (GlaxoSmithKline) – EDB (Economic Development Board) Trust Fund and the Singapore Ministry of Education Tier 1 Fund (2019-T1-002-065).

References (43)

Cited by (33)

View all citing articles on Scopus
View full text