Preparation, characterization and scaling propensity study of a dopamine incorporated RO/FO TFC membrane for pesticide removal

Highlights

• A new high-flux dopamine incorporated RO/FO TFC membrane was synthesized.

• The optimal amount of MPD and DA in IP was found to be 2.0 and 0.1 wt%, respectively.

• Modified TFC membrane evinced five-fold water flux in FO than the control membrane.

• The optimal membrane evinced high pesticides rejection (>91%) in both RO and FO.

• FO is per se more resistant against scaling than RO, regardless of membrane material.

Abstract

The poor permeability of prevalent thin film composite (TFC) membranes impedes their application in forward osmosis (FO). In this study, a high-flux TFC membrane was prepared by incorporating dopamine (DA) into an aqueous solution with various concentrations of m-phenylendiamine (MPD) in interfacial polymerization with trimesoyl chloride (TMC) on a fabricated polysulfone (PSF) substrate. SEM, AFM, XPS, ATR-FTIR, and water contact angle measurement (WCA) and Zeta potential were exploited to characterize synthesized membranes. The optimized TFC membrane (TFC-2; MPD: 2 wt%, DA: 0.1 wt%) attained a nearly five-fold water flux improvement (33.3 LMH (L.m⁻².h⁻¹) versus 7.1 LMH for control membrane), an acceptable reverse salt flux of 4.1 g/m²h and a reduced structural parameter (125 μm) in FO with 1 M NaCl draw solution. Membranes were then employed for pesticide removal from water in both reverse osmosis (RO) and FO resulting in high rejection values of >92% in RO and >91% in FO by the optimal membrane for all studied pesticides. Furthermore, the optimized DA-incorporated sample exhibited a better performance compared to the control membrane in terms of anti-scaling behavior and the scaling propensity of RO and FO processes was studied through flux decline measurement in supersaturated solutions with different concentrations of gypsum as model scalant. This study suggests that the FO process, regardless of the membrane material, is per se more resistant against scaling with a threshold gypsum concentration of 25–30 mM compared to RO starting to scale within 20–25 mM of gypsum solution within 24 h.

Introduction

Water pollution has become a continuously increasing worldwide problem causing adverse effects on human health [1]. Micropollutants cause a high level of contamination of groundwater resources because of the widespread use of pesticides to enhance the production of crops [2]. Even though the use of pesticides has dramatically been restricted over the last decades, there are still substantial recalcitrant pesticide residues or pesticide transformation products (PTP) leaching to groundwater aquifers [3,4].

Among various water treatment processes, membrane separation through reverse osmosis (RO) and nanofiltration (NF), has previously been found to be capable of rejecting most pesticides depending on both membrane and solute characteristics [[5], [6], [7]]. However, these membrane processes suffer from drawbacks such as propensity to membrane fouling and scaling, which can diminish the lifetime of the membrane and impose additional costs to the treatment process through the utilization of chemical cleaning or the addition of scale inhibitors [8]. Forward osmosis (FO), however, has gained considerable heed owing to its advantages over pressure-driven membrane processes (PDMP), NF and RO. Exploiting the osmotic pressure gradient, FO requires no hydraulic pressure, potentially leading to lower energy requirements and less fouling/scaling propensity [9]. Moreover, it can be used for the removal of a broad range of organic contaminants, including pesticides [[10], [11], [12]]. Nevertheless, the implementation of FO is hindered by the lack of specifically-tuned membranes for FO, as cellulose triacetate (CTA) membranes cannot withstand a wide range of pH and conventional thin film composite (TFC) membranes have a low permeability when used in FO [13]. In particular, prevalent commercial TFC RO membranes comprising a thin selective film covered on a porous support layer, generate low osmotic pressure due to an extremely tight selective layer and high internal concentration polarization primarily caused by a fairly high structural parameter, S, which quantifies the mass transport length scale across the support layer [14,15].

A tremendous amount of research has been carried out to develop TFC membranes compatible with the FO process by focusing on either layer of the TFC membranes. Most efforts were directed to the support layer of the TFC membrane, in order to minimize the membrane structural parameter (S) through lowering thickness, increasing porosity, or reducing the tortuosity of the support layer. Alternatively, the selective polyamide layer has also undergone different modification strategies such as the incorporation of nanoparticles such as zeolite [16], graphene oxide (GO) [17], titanate nanotubes [18], multi-walled carbon nanotube (MWCNT) [19], and halloysite nanoparticles [20]; such membranes are categorized as thin film nanocomposite (TFN) membranes. However, these methods could barely be transferred from lab-scale to commercial membrane production because of the high possibility of severe aggregation of the nanoparticles in the organic solvent during the interfacial polymerization, as well as weak compatibility of the nanoparticles with the formed polyamide matrix. The non-uniform dispersion of nanoparticles forms non-selective voids at the interface of the nanoparticles and polymer, which considerably compromises the rejection [21,22]. Therefore, the uncertainties associated with the formation of a defect-free and integrally-skinned active layer due to the presence of nanoparticles restrict further development of TFN membranes.

Alternatively, in recent years, dopamine (DA) as a representative of mussel-inspired chemistry [23] has been employed for the fabrication of highly-efficient membranes. Dopamine converts to polydopamine (PDA) through self-polymerization in an alkaline environment in the presence of oxygen [24,25]. In fact, polydopamine (PDA) functions as an adaptable organic nanomaterial capable of homogeneous dispersion into the polymer matrix. Due to its superior properties such as being tightly adherent and highly hydrophilic because of the inclusion of catechol, quinone, imine and amine groups, It has been newly exploited for tailoring MF, UF [26], NF [27], and RO membranes [28]. FO membranes have also been subject to modifications using dopamine through different strategies. For instance, Arena et al. reduced the internal concentration polarization via the use of PDA in the support layer to improve its hydrophilicity resulting in an enhanced water flux [29]. Moreover, the surface coating of the TFC FO membrane using PDA has also been reported by Guo et al. [30]. They showed that the PDA coating considerably improved the water flux and antifouling performance of the FO membrane due to the enhancement of hydrophilicity and reduction in the surface roughness [30]. More interestingly, a few research groups used DA as a co-reactant to m-phenylenediamine (MPD) to be used in aqueous solution during interfacial polymerization (IP) [15,31]. Xu and his colleagues demonstrated that the incorporation of DA can result in an enhanced FO water flux but simultaneously increased reverse salt flux and PA film thickness in comparison to the control membrane [15]. Wang et al. also prepared DA incorporated FO membranes in a broader range of DA/MPD ratios [31]. Their efforts led to a minimized FO structural parameter (S) and consequently higher water permeability [31]. In the both cases, however, the MPD concentration was set at a constant level (2 wt%) and the DA content was changed. Moreover, the performance of the modified FO membranes was not investigated in removal of micropollutants or any other applications.

In the present work, dopamine was incorporated into the aqueous phase of the IP reaction and the concentration of MPD was manipulated to ascertain whether DA can substitute the MPD and minimize its content. Polysulfone (PSF) substrate was firstly fabricated via the phase inversion method. The co-deposition of MPD-DA mixture on the PSF substrate with different MPD/DA ratios was then carried out. Afterwards, a typical IP reaction including a cross-linking process at the interface of aqueous/organic phase followed by a heat treatment formed a thin polyamide (PA) layer. Furthermore, state-of-the-art characterization methods (SEM, AFM, WCA, Zeta potential, XPS, and FTIR) have been employed to gain a thorough insight into the synthesized material. More importantly, in continuation to our previous studies on the removal of pesticides of concern in Danish groundwater wells: 2–6 dichloro-benzamide (BAM), 2-methyl-4-chlorophenoxyaceticacid (MCPA), and 2-(4-chloro-2-methylphenoxy)propionic acid (MCPP) [4,5,32,33], the synthesized DA incorporated membranes were used in both RO and FO processes to evaluate it for drinking water treatment application. This is contrasting most studies where most efforts have been devoted to the synthesis and characterization of a modified FO membrane rather than examining it for real-life applications.

Additionally, since groundwater based drinking water treatment using membrane technology encounters scaling as a result of high ionic strength of groundwater matrix, the FO process might be a better choice with respect to the absence of hydraulic pressure. However, the hypothesis of FO being less prone to scaling compared to RO needs to be investigated. Most studies in this field have focused on organic fouling to make a comparison between FO and RO using model foulants such as alginate [[34], [35], [36], [37]]. In one of few studies on scaling propensity of RO and FO processes, Tow et al. reported lower scaling propensity of FO than RO using gypsum as a model scalant. However, they attributed it to the membrane material as they tested CTA FO membrane versus TFC RO membrane [38]. Therefore, in this study, in addition to investigation of the effect of DA incorporation on the anti-scaling behavior of TFC membranes, the authors, for the first time, have employed the same TFC membrane to compare the scaling propensity of RO and FO processes using an identical membrane module and experimental setup under the same operational conditions. This study, regardless of membrane material, can serve the membrane community with an important insight into the FO process in terms of scaling propensity.