Mechanism analysis of powdered activated carbon controlling microfiltration membrane fouling in surface water treatment

Highlights

• Blocking models and interaction energy theory were employed to evaluate PAC alleviating membrane fouling.

• Cake formation was the major mechanism of membrane fouling.

• PAC pre-treated water had a relatively low fouling rate.

• PAC enhanced the removal of NEU and SHA (based on fractionation), and BB&LMW-A (based on HPSEC).

• PAC could control membrane fouling through decreasing AB interaction energy.

Abstract

The mechanism of powdered activated carbon (PAC) mitigating microfiltration (MF) membrane fouling for surface water treatment was investigated via blocking models, resin adsorption fractionation, high performance size exclusion chromatography (HPSEC) with peak-fitting, and interaction free energy theory. The results identified cake formation as the major mechanism of membrane fouling. Also, PAC pre-treated water showed a relatively lower fouling rate than untreated raw water and PAC had a positive effect on improving membrane flux on account of the enhanced removal of neutral hydrophilic compounds (NEU) and strongly hydrophobic acids (SHA) (based on fractionation), and building blocks and low molecular weight acids (BB&LMW-A) (based on HPSEC). Analysis of the interfacial energy between membrane and foulants via the extended Derjaguin–Landau–Verwey–Overbeek (XDLVO) theory found that AB interaction energy played a dominating role in the components of interaction free energies during MF process. PAC, by decreasing AB interaction energy, could thus control membrane fouling.

Introduction

Numerous studies have focused on membrane fouling during the past decades, and they have achieved substantial progress towards a better understanding of the fouling behavior [1]. Natural organic matter (NOM), known as the major foulant to microfiltration (MF) membrane [2], was often classified based on molecular weight (MW) and hydrophobicity in relevant studies [3], [4], [5], [6], [7]. Improvements in analytical techniques make it possible to characterize NOM more efficiently. Resin adsorption fractionation and high performance size exclusion chromatography (HPSEC) are the most optimistic measurement methods available for the determination of organic fractions in surface water. With DAX-8/XAD-4/IRA-958 adsorbent resins, NOM could be fractionated into strongly hydrophobic acids (SHA), weakly hydrophobic acids (WHA), charged hydrophilic acids (CHA) and neutral hydrophilic acids (NEU) fractions [8]. MW distribution of NOM such as low MW neutrals, low MW acids, building blocks(BB), humic substances(HS), biopolymers (BP) could also be identified and isolated by HPSEC in conjunction with a relatively new technique named peak-fitting [9–11]. Based on the advanced analytical techniques, hydrophobic fractions (e.g. humic acid) [12], hydrophilic fractions [13], hydrophobic and hydrophilic compounds [14], SHA and NEU fractions [15], low MW compounds [16], macromolecules BP fractions [17] were all once believed to be responsible for membrane fouling. Notwithstanding the above efforts, there were obvious contradictory results on the fouling tendency of different fractions, a clear call for an in-depth and thorough elaboration about the characteristics of natural water that causes membrane fouling.

Pre-treatment of the raw water is known to be conductive to mitigating low-pressure membrane fouling [18], [19]. The powdered activated carbon (PAC) pre-adsorption was well known as a common procedure before membrane filtration process to remove NOM. Studies testified that PAC could enhance membrane performance [20], [21], [22] as it controlled membrane fouling by adsorbing a proportion of NOM [23], [24], [25], [26], and thus lowered the fouling load. Yet, the mechanism of PAC addition prior to membrane filtration on fouling mitigation is not well elucidated.

Membrane fouling is generally believed to be a decrease in permeate flux caused by foulants’ depositing on the membrane surface or blocking the membrane pores [27]. The mechanisms of blocking during fouling have been discussed in various studies according to Hermia’s four types of blocking mechanisms, i.e. complete pore blocking, intermediate pore blocking, standard pore blocking and cake layer formation [28]. Hwang and Lin [29] reported that various membrane presented different blocking types e.g., standard blocking, intermediate blocking and complete blocking were responsible for the MF-Millipore, Durapore, Isopore membrane, respectively. Lin et al. [30] used bovine serum albumin (BSA) to determine the membrane fouling behavior and found that standard blocking was the main fouling mechanism. Li et al. [31] investigated the effects of algae organic matter on the fouling of a poly(ether sulfone) membrane and the results suggested that cake formation was the major mechanism. Blocking mechanism is meaningful because of the guidance it provides to understanding the fouling behavior, and choosing the proper membrane with suitable cleaning methods. However, few studies have so far been reported about the evaluation of PAC on membrane fouling by blocking model.

In a microscopic aspect, the interaction force between foulants and membrane was the critical factor in membrane fouling [32], [33]. Recently, the extended Derjaguin – Landau – Verwey – Overbeek (XDLVO) theory has proven to be a powerful and rigorous approach to quantitatively unravel the membrane-foulants interaction [34], [35]. Wang et al. [35] found that XDLVO-model was sufficient to assess short-range membrane-foulant interactions and to predict soluble microbial products fouling. Huang et al. [36] stated that XDLVO theory could be used for algae organic matter membrane fouling analysis. Chen et al. [34] adopted the interfacial energy of membranes and foulants to explore the interaction energy profiles for different membranes, providing considerable insights into the role of interaction behavior between membrane and foulants. Lin et al. [37] used XDLVO theory to predict initial membrane fouling behavior and explore the mechanism of pre-chlorination for the mitigation of membrane fouling. To the best of our knowledge, the application of XDLVO theory in the study of membrane fouling mechanisms coupled with PAC pretreatment has not been reported yet.

The objectives of the study were (1) to observe the improvement of membrane flux by PAC addition and to explore the MF fouling behavior by blocking model; (2) to characterize the foulants by resin adsorption fractionation and HPSEC with peak-fitting and to identify the major foulants removed by PAC; (3) to unravel the interaction energy between membrane and foulants with XDLVO theory. The main novelty of this work was that blocking models and XDLVO theory were used to further comprehend the behavior of PAC mitigating membrane fouling during MF of surface water. The results were expected to provide useful information for understanding the mechanism of PAC controlling membrane fouling during surface water treatment.