Myongji University - The Graduate School Department of Environmental Engineering and Biotechnology

Removal of Natural Organic Matters (NOMs) Using Functional Magnetic-impregnated Ion Exchange Resin (FMIEX) and Its Effects on Membrane Fouling A thesis in Environmental Engineering by

Chu Xuan Quang Thesis advisor: Kisay

Lee, Ph.D.

June 2004

This thesis is submitted in partial fulfillment of the requirements for the degree of Master of Engineering © Chu Xuan Quang 2004. All rights reserved. No part of this publication may be reproduced without written permission of the copyright owner.

공학석사 학위논문 서기 2003 학년도

Removal of Natural Organic Matters (NOMs) Using Functional Magnetic-impregnated Ion Exchange Resin (FMIEX) and Its Effects on Membrane Fouling

지도교수

이 기 세

명지대학교 대학원 환경생명공학과 Chu Xuan Quang

Removal of Natural Organic Matters (NOMs) Using Functional Magnetic-impregnated Ion Exchange Resin (FMIEX) and Its Effects on Membrane Fouling

이 논문을 공학석사학위 논문으로 제출함

2004 년 06 월

23 일

명지대학교 대학원 환경생명공학과 Chu Xuan Quang

Chu Xuan Quang 의 공학석사학위 논문을 인준함

주심위원

장 덕 진

인

부심위원

이 용 훈

인

부심위원

이 기 세

인

2004 년 06 월 30 일

명지대학교 대학원

ACKNOWLEDGMENTS

First and foremost, I would like to express my sincere gratitude to my academic advisors, Professor Lee Yonghun and Professor Lee Kisay, for all their advice, support and encouragement during the course of my studies. They have been the great advisors all through my academic and nonacademic works. I am also grateful to chair of committee, Professor Jahng Deok-jin, for his valuable advices as well as his time to review my manuscript thesis. Besides, I would like to thank the MyongJi University for providing scholarship. And my appreciation also goes to the Faculty of Chemistry, Vietnam National University, Hanoi and especially Professor Nguyen Van Noi. Without their oversea-studying program, I would not be here today. Finally, I extremely owe my family and especially my dad for their understanding, support and encouragements. Also, I express my thankfulness to my little sister in France for her assistance to find out some referential documents.

i

TABLE OF CONTENS ACKNOWLEDGMENTS ........................................................................................ i TABLE OF CONTENS ............................................................................................ ii LIST OF FIGURES .................................................................................................. v LIST OF TABLES .................................................................................................... vii ABBREVIATIONS ................................................................................................... iix

Chapter 1. INTRODUCTION............................................................... 1 1.1. Overview ................................................................................................. 1 1.2. Objectives and Research scope ............................................................... 3

Chapter 2. LITERATURE REVIEW................................................... 5 2.1. Ultrafitration and Microfiltration.......................................................... 5 2.1.1. Definition of UF/MF processes....................................................... 5 2.1.2. Basic principles of UF/MF processes ............................................. 9 2.2. Characterization of NOMs...................................................................... 13 2.3. NOM fouling on UF/MF membranes..................................................... 16 2.4. Control of NOMs in water treatment..................................................... 23 2.4.1. Coagulation ...................................................................................... 23 ii

2.4.2. Adsorption ........................................................................................ 26 2.4.3. Magnetic Ion Exchange Resin ........................................................ 29

Chapter 3. MATERIALS AND METHODS ....................................... 34 3.1. Materials and Methods............................................................................ 34 3.1.1. Functional magnetic-impregnated ion exchange resin (FMIEX) ................................................................................. 34 3.1.2. NOM and inorganic solutions ......................................................... 37 3.1.3. Membranes....................................................................................... 38 3.2. Experimental Methods ............................................................................ 41 3.2.1. Jar testing experiment ..................................................................... 41 3.2.2. Dead-end filtration........................................................................... 43 3.2.3. Analytical methods........................................................................... 48

Chapter 4. RESULTS AND DISCUSSION ......................................... 52 4.1. Preliminary investigation on FMIEX .................................................... 52 4.1.1. Removal of humic acids (HA) using FMIEX .................................52 4.1.2. Removal of inorganic anions using FMIEX .................................. 55

iii

4.1.3. Evaluation on FMIEX resin regeneration...................................... 61 4.1.4. Preparation of FMIEX resin from the uninvolved materials ........ 65 4.2. Development of FMIEX process for HA removal................................. 69 4.2.1. Effects of contact time on HA removal ........................................... 69 4.2.2. Effect of pH on HA removal............................................................ 70 4.2.3. Determination of appropriate FMIEX dosage................................ 75 4.3. Effects of FMIEX pretreatment on membrane filtration performance ............................................................................................ 78 4.3.1. Analysis of flux decline test results ................................................. 78 4.3.2. Analysis of resistance components.................................................. 85

Chapter 5. CONCLUSIONS ................................................................. 88 REFERENCES.......................................................................................................... 90

iv

LIST OF FIGURES

Figure 2-1.

Overview of UF/MF membrane treatment processes and solute, particle dimensions……..…………………………..8

Figure 2-2.

Mechanism of DOC exchange by FMIEX…….……………….31

Figure 3-1.

The structure of FMIEX resin…………………………….……35

Figure 3-2.

Structure of polyethersulfone membrane……….…………….. 40

Figure 3-3.

The relationship of UV254 and DOC with NOM concentration…………………………………………………..42

Figure 3-4.

Schematic diagram of the deadend filtration system for flux decline tests……………………………….…………..46

Figure 3-5.

The ultrafiltration protocols with PES membrane…………..…47

Figure 4-1.

UV254 removal percent as a function of contact time for different FMIEX resin dosages……………….………….. 54

Figure 4-2.

Nitrate removal as a function of contact time for different FMIEX dosages………………………….….….. 59

Figure 4-3.

Phosphate removal as a function of contact time for different FMIEX dosages…………….………………...….60

Figure 4-4.

Inorganic anion removals versus regenerated resin reuse.…….62

Figure 4-5.

Effect of different protocols of AHA loaded FMIEX resin regeneration………………………………….………….62 v

Figure 4-6.

Water quality after 60 min. contact with different FMIEX dosages………………………..……………...…….…66

Figure 4-7.

Preparation variables for FMIEX resin………...………….…...67

Figure 4-8.

UV254 removal percent as a function of contact time……..…....71

Figure 4-9.

Influence of pH on HA removal………………………………..72

Figure 4-10.

Impact of FMIEX resin dosage on UV254 and DOC removals....77

Figure 4-11.

Flux decline test results on FMIEX treated water with different pH………………………………………...……..80

Figure 4-12.

Flux decline test results on FMIEX treated water with different FMIEX dosages…………………………………81

Figure 4-13.

Resistance components of UF membrane after filtration of FMIEX treated water with different pH……………………..86

Figure 4-14.

Resistance components of UF membrane after filtration of FMIEX treated water with different FMIEX dosages………87

vi

LIST OF TABLES

Table 2-1.

Characteristics of membrane processes…………………………7

Table 3-1.

The characteristics of FMIEX resin……………………………36

Table 3-2.

Properties of the membrane used……………………...……….40

Table 4-1.

The capacity of FMIEX resin towards several inorganic ions…58

Table 4-2.

The water quality summary for treated water under different pH value………………………………………………83

Table 4-3.

The water quality summary for various treatments with different FMIEX dosages ……………………………...…84

vii

ABBREVIATIONS

AHA AMW DOC FA FMIEX GAC GPC HA HPIA HPOA KHP MF MIEX MW MWCO NF NOM PAC PACl PES SEC SRHA SUVA TMP UF UV

Aldrich humic acid Apparent molecular weight Dissolved organic carbon Fulvic acid Functional magnetic-impregnated ion exchange resin Granular activated carbon Gel-permeation chromatography Humic acid Hydrophilic acid Hydrophobic acid Potassium hydrogen phthalate Microfiltration Magnetic ion exchange resin Molecular weight Molecular weight cut-off Nanofiltration Natural organic matter Powdered activated carbon Polyaluminium chloride Polyethersulfone Size exclusion chromatography Suwannee river humic acid Specific ultraviolet absorbance Transmembrane pressure Ultrafiltration Ultraviolet

viii

Chapter 1.

INTRODUCTION

1.1. Overview Natural organic matters (NOMs) are present in all types of surface and underground waters. NOMs cause a number of identified and non-identified problems for the drinking water treatment. NOMs have been known not only as an unaesthetic cause of color in water but also as an indirect precursor of disinfection by-products (DBP) [1,2]. Nowadays, the low-pressure membrane filtration has been in progress as an alternative to conventional treatments for drinking water production. Although the use of ultrafiltration/microfiltration (UF/MF) technologies have been attracted attention because of its ability to remove particles, turbidity, and microorganisms, but UF/MF membranes are in most part incapable of removing natural organic compounds, especially dissolved organic carbon (DOC) fraction and small colloid from the raw water [3-5]. In addition, many studies have demonstrated that NOMs are major foulants during ultrafiltration and microfiltration of drinking water treatment [6-8]. Therefore, removing NOMs imposes significant importance not -1-

only in fouling control of membrane filtration, but also in meeting the stringent DBP regulations proposed. The addition of a pretreatment step such as coagulation and adsorption processes prior to the UF/MF unit provides potential to solve problems mentioned above [9-12]. However, these processes are complicated, and performance of treated water is varied by both of pretreatment conditions and raw water characteristics. In order to improve efficiency of each treatment process, the operating conditions should be optimized for each NOM water source. Generally, hydrophobicity and molecular weight are the distinct difference among different NOM waters [13]. Degree of aromaticity of NOMs also has great effect on the treatment process, especially on membrane fouling control [14,15]. Therefore, the humic fraction of NOMs, which is more hydrophobic and larger in its size, needs to concentrate because of its adsorptive capacity on the membrane. With respect to control the contaminated anions in drinking water treatments, especially NOMs, the magnetic ion exchange (MIEX) and the functional magnetic-impregnated ion exchange resin (FMIEX) were provided as -2-

the innovative materials [16,17]. These resins are a strong base anion exchange combined with a magnetized component. The base material of the resin is polyacrylic or polystyrene with cross-linkage by divinyl benzene. The resins have small particle size of around 200 µm; therefore, they provide a high surface area which allows rapid adsorption kinetics. Besides, these particles are able to settle rapidly by magnet. In order to use this innovative material successfully for removing NOMs in conjunction with ultrafiltration or microfiltration, the ability of FMIEX resin and its effect on membrane fouling should be assessed and evaluated.

1.2. Objectives and Research scope An innovative functional magnetic-impregnated ion exchange resin (FMIEX) was produced in the 2003 year. This FMIEX resin has never been explored for removal of natural organic matters (NOMs), both cases of individual NOM treatment and combination of NOM removal process with membrane filtration. Therefore, the specific objectives of this study were (i) to investigate -3-

NOM removal effectiveness of the FMIEX resin and (ii) to evaluate the impact of the FMIEX pretreated water on membrane fouling. These would be achieved by the followings scope: To identify the humic acid removal capacity of the FMIEX resin based on the reduction of the aromatic level. In addition, the ability of the FMIEX resin for removal of a number of the inorganic anions contaminated water, and the potential of this resin to reuse would be also observed. To develop the FMIEX process for humic acid removal. This study would focus on the finding out the appropriate treatment conditions. The effects of contact time, pH as well as FMIEX resin dosage would be investigated. To estimate the performance of the ultrafiltration with untreated and FMIEX pretreated waters as well as to establish the impact of FMIEX pretreatment on membrane fouling. It would also prove the effects if residual particles in FMIEX treated water had any adverse impact on membrane performance.

-4-

Chapter 2.

LITERATURE REVIEW

2.1. Ultrafitration and Microfiltration 2.1.1. Definition of UF/MF processes Membranes can be characterized according to the several major factors such as particle or molecular size, mechanism of separation, and a primary factor affecting the separation process - driving force. Table 2-1 shows the main membrane processes and their characteristics. Among these processes, UF and MF processes are most often associated with the term “membrane filtration” [18]. UF and MF fall into the separation process group with sieving mechanism based on the size of the membrane pores relative to that of particulates. In the term of driving force, they are defined as the two low-pressure-driven membrane processes. Besides, membranes are also usually categorized by the size of the separated components [19]. In general, UF and MF membranes remove particulate contaminants via a size exclusion mechanism and these membranes are characterized according to pore size with respect to microbial and particulate removal capabilities. In its -5-

definition, MF retains only micron size particles, but UF able to retain both of macromolecules and particles that have size lager than its pore size. MF membranes are generally considered to have a pore size range of 0.1 to about 5 µm; and for UF membranes, typical pore sizes lie between 0.001 and 0.02 µm. However, some UF membranes have the ability to retain larger organic macromolecules, and thus the concept of pore size in microns becomes inappropriate. Therefore, it is convenient to refer to the “molecular weight cut off” (MWCO) which expressed in Daltons - a unit of mass rather than the particular pore size to define UF membranes when discussed in reference to these types of compounds. The concept of the MWCO is a measure of the removal characteristic of a UF membrane in terms of atomic weight (or mass), that is, the membranes with a specified MWCO are presumed to act as a barrier to compounds or molecules with a molecular weight exceeding the MWCO. With this conceptual term, UF covers particles and molecules that range from about 1000 in molecular weight to about 500,000 Daltons [18,19]. An overview of the correlation

between

UF/MF

membrane

filtration

solute/particle dimensions are illustrating in Figure 2-1. -6-

processes

and

several

Table 2-1.

Characteristics of membrane processes [18,19]

Process

Driving force

Mechanism of separation

Retentate

Permeate

Microfiltration

Pressure

Sieve

Suspended particles, water

Dissolved solutes, water

Ultrafiltration

Pressure

Sieve

Nanofiltration

Reverse osmosis Electrodialysis Pervaporation

Lager molecules, water Pressure Sieve, Small Diffusion molecules, divalent salts, dissociated acids, water Pressure Diffusion All solutes, water Voltage/current Ion exchange Nonionic solutes, water Pressure Evaporation Nonvolatile molecules, water

-7-

Small molecules, water Monovalent ions, undissociated acids, water Water Ionized solutes, water Volatile small molecules, water

10-4

10-2

10-3 Salts

10-1

1.0

101

102 Size [µm] Sand

HA Viruses

FA

Bacteria

HPIA

Colloids Ultrafiltration

-8-

Microfiltration

Size of water molecules

101 102 HA Humic Acid Figure 2-1.

0.45 micron boundary 103

104 FA

105

106

Molecular Weight [Da]

107 108

Fulvic Acid

HPIA

Hydrophilic Acid

Overview of UF/MF membrane treatment processes and solute, particle dimensions [2,19,20]

2.1.2. Basic principles of UF/MF processes Membrane filtration system throughput or productivity is typically characterized by the system flux, which is defined as the filtrate flow per unit of membrane filtration area. Therefore, numerous flux models described the effects of system operating parameters and physical properties on membrane performance. They not only discussed the entire pressure-flux behavior observed during typical UF or MF, such as pressure-controlled at low pressures and pressure-independent at high pressures, but also proposed a approach base on Dacry’s law to use the “resistance-in-series” concept that is common in heat transfer. The latter one is useful for both of modeling purposes and evaluating the effectiveness of the cleaning procedures [19]. For an ideal membrane and feed solution, the relationship among the permeate flux, the driving force, and the resistances is described in Equation 2.1 [2,19,21]. According to this model, the driving force for the transport of water across a porous membrane, transmembrane pressure (a pressure gradient across the membrane), is defined by the pressure on the feed side of the membrane minus -9-

the filtrate pressure (backpressure).

J≡

1 dVp ∆P = (2.1) A m dt µR m

where J = permeate flux (LMH) Am = membrane filtration area (m) Vp = total volume of permeate (L) t = filtration time (hr) ∆P = transmembrane pressure (psi) µ = viscosity of permeate (centipoises) Rm = intrinsic membrane resistance (m-1)

In fact, the resistances not consist only intrinsic membrane resistance components, but this is accounted for by adding the resistance attributable to the accumulated foulant layer at the membrane surface. The total resistance is represented by the sum of these two components, as shown in Equation 2.2:

-10-

R t = R m + R f (2.2)

where: Rt = total membrane resistance (m-1) Rm = intrinsic membrane resistance (m-1) Rf = resistance of the foulant layer (m-1)

However, while the intrinsic resistance of the membrane should remain constant for all practical purposes, the fouling resistance may be reduced by a result of backwashing and chemical cleaning. Thus, the resistance to flow acting in opposition to the driving force and inhibiting the transport of water across the membrane can also be quantified. The flux equation becomes:

J=

∆P (2.3) µ( R m + R hr + R cr + R ir )

where J = permeate flux (LMH) ∆P = transmembrane pressure (psi) µ = viscosity of permeate (cp)

-11-

Rm = intrinsic membrane resistance (m-1) Rhr = the resistance which can be removed by hydraulic solution (m-1) Rcr = the resistance which can be removed by chemical cleaning (m-1) Rir = the resistance due to irreversible fouling (m-1)

Note that the viscosity of water increases with decreasing temperature, so the values for water viscosity are required to correct [18]. It can be found in the literature or approximated using the empirical relationship expressed in Equation 2.4:

µ T = 1.784 − 0.0575T + 0.0011T 2 − 10 −5 T 3 (2.4)

where: µT = viscosity of water at temperature T (cp) T = water temperature (oC)

In order to quantitatively estimate the relative degree of purification in UF/MF processes, the concept “rejection”, which shows how well a membrane -12-

retains or allows passage of solutes, can also used. In this case, it was consumption that the probability of a component of the feed solution passing through the membrane is highest. The rejection percentage at any point in the process is computed as shown in Equation 2.5:

 Cf − Cp %R = 100  Cf

  (2.5) 

where Cf is the solute concentration in the feed and Cp is the solute concentration in the permeate [2, 4, 8].

2.2. Characterization of NOMs Natural organic matters (NOMs) have known as the complex mixture of compounds that occurs ubiquitously in all waters. The mixture is the product of the complex biogeochemistry of carbon, which formed due to the breakdown of animal and plant materials in the environment. According several reports, biopolymers and metabolic products of organisms are transformed biotically and

-13-

abiotically to produce over time macromolecular, polymorphic condensation products having both aliphatic and aromatic nature, acidic functionality and the ability to complex and partition a wide range of chemical species [20,22]. The background about structure of NOMs is undefined because the composition of NOMs is strongly dependent on the environmental source, and characterization of the compounds present in the mixture is extremely diverse. However, if the complexity of the NOMs is kept in mind, the broad understanding of the character is possible [20]. In general, NOMs involving mainly of carbon, oxygen and hydrogen; and nitrogen or sulfur can also be present. Numerous of investigation reported that a range of compounds, from small hydrophilic acids, proteins and amino acids (nonhumic components) to larger humic and fulvic acids (humic substances), are constituents of most NOMs [1,20]. Hence, NOMs consist both of aliphatic and aromatic, highly charged and uncharged compounds. Most characterization studies indicate that NOMs in natural water on average have a significant charge due to carboxylic acid groups, and some aromatic/hydrophobic character. The -14-

whole water NOMs have a high carboxyl, aromatic and aliphatic carbon content, whereas the higher molecular weight fraction of the same NOMs display a relatively low carboxyl, aromatic and aliphatic character, and high O-alkyl character, indicative of carbohydrates. This is a good illustration of the diversity of the compounds present in NOMs, and the difficulty of describing the mixture adequately. On the other hand, NOMs can also divided into particulate and dissolved organic components according to the 0.45 µm boundary in diameter. Particulate organic matter generally consists of bacteria and planktonic organisms such as zooplankton, phytoplankton, while dissolve organic matters cover humic substances, viruses, carbohydrates, and hydrocarbons. Dissolved organic carbon, which predominantly comes from polymeric organic acids called humic substances, gives the yellow or yellowish-brown color of nature surface water. There have polyelectrolytes of carboxylic, hydroxyl, and phenolic functional groups (1000 to 2000 molecular weight), dissolved molecules of fulvic acids (approximately 2 nm in diameter and at least 60 nm apart). There are also some -15-

colloidal organic matters which are the humic acid fraction. These fractions are can range in molecular weight from 2000 to 100,000 Daltons and contain fewer carboxylic and hydroxyl functional groups than the fulvic acid fraction [20].

2.3. NOM fouling on UF/MF membranes Low-pressure

membrane

systems,

such

as

microfiltration

and

ultrafiltration, can be successfully used in water treatment. However, one biggest obstacle in most applications of membrane technologies including UF/MF is fouling which impede the flux of solution through the membranes, and limit the contaminant rejection, durability and chemical resistance. Natural organic matters (NOMs) have been known as the major foulants during UF/MF processes. Generally, NOM concentration appears as one of factors affecting membrane performance because increasing NOM concentration cause increasing NOM accumulation on membrane surface; and thus increased solution flux decline. Waite and co-workers reported that increasing fulvic acid concentration in the present of hematite resulted in increasing cake resistance except for the case -16-

when zeta potential was close to zero and rapid aggregations occurred [21]. However, Amy and Cho found that flux decline was not significantly affected membrane properties or by NOM characteristics under comparable hydrodynamic conditions, except for high-humic sources. They suggested that the negative charge density of hydrophobic (fulvic and humic) acids promoted NOM rejections [20]. Therefore, the loss of flux through the membrane not only caused by NOM concentration, but also strongly depends upon the complex interactions between the membrane and components in the raw water. There have been many findings in terms of the impacts the membrane properties (hydrophobicity, charge, and morphology) and the organic matter properties (hydrophobicity, molecular weight, and charge density) on UF/MF filtration performance. Schäfer et al. studied effects of fouling on rejection with different membrane processes, membrane pore sizes, and NOM types in the present of calcium cation [2]. Their results have showed that pore plugging and cake information caused fouling in MF, while internal pore adsorption of calciumorganic flocs was effective in UF. Based on these observations, they proposed that -17-

fouling reduced the internal pore diameter of membranes and subsequently increased rejection. In similar trend of study, but focused on NOM hydrophobicity, Aoustin et al. concluded that the larger and more UV-absorbing fraction of humic acid was shown to be responsible for irreversible pore adsorption and plugging, that are responsible for flux decline [23]. They also reported that the fulvic acid and hydrophilic fraction showed a smaller and mostly reversible flux decline; but fulvic acid seems to be a bigger and to interact more with calcium, thus causing more pore blocking. Besides, Taniguchi and co-workers showed a different order in relationship of fouling and membrane pore size. According to their study, the rate of fouling by cake formation was independent of the membrane pore size [24]. In their investigation, cake formation was dominant mode of fouling by the unfiltered feed, which contained aggregates. The results was identified by a constant rate of increase in membrane resistance with permeate throughput and was independent of pore size over a 10-1000 kDa MWCO range. The authors also reported that prefiltration (to remove aggregates) and dilution (to reduce aggregate -18-

concentration) reduced the rate of increase in membrane resistance for the low MWCO membranes but did not change fouling mode. In contrast, such pretreatment prevented cake formation on the larger MWCO membranes and shifted the mode of fouling to pore blockage. With respect to effects of molecular weight and NOM fractions, Lin and co-workers fractionated humic substances according to molecular weight (MW) with gel-permeation chromatography (GPC) and hydrophobicity with resins. They found that the highest MW components for both hydrophobic and hydrophilic fractions caused the greatest flux decline [3,15]. The results of MW fractions of both hydrophobic and hydrophilic fractions further indicate that those molecules with the largest MW (6.5-22.6 kDa) exhibit the worst flux decline. The hydrophilic fraction exhibits the worst flux decline despite little solute rejection. It was also reported by Fan et al. about mechanism for membrane fouling with involved a combination of adsorption of small molecules on the membrane pore wall and pore blockage by colloidal organics. Natural surface water and membrane ranging from 0.45 µm to 10 kDa were used in their studies. The results -19-

have demonstrated that the high molecular weight fraction of NOM (>30 kDa) was responsible for the major flux decline as committed tests to hydrophobic MF membranes. They also note that the neutral hydrophilic fraction had a greater concentration of colloidal DOM (defined as >30 kDa in their study) and postulated that the accumulation of colloidal DOM within the membrane pore structure may have been a significant factor in flux decline [7]. Base upon observations, they have been purposed several comments regarding the effects of NOMs and membrane hydrophobicity on membrane fouling. The fouling rate for the hydrophobic membrane was considerably greater than for hydrophilic membrane. The fouling potential for the fractions was hydrophilic neutral > hydrophobic acids > transphilic acids > hydrophilic charged. The low-aromatic hydrophilic neutral compounds were the main determinant of the rate and extent of flux decline. Specially, the higher aromaticity of the NOMs caused the greater flux decline [7]. This result was the same which was provided by Carroll et al., who found that the greatest degree of fouling was by the neutral hydrophilic fraction of DOM and characterized each fraction by size exclusion chromatography (SEC). They -20-

attributed the neutral hydrophilic fraction to smallest MW distribution [1]. Similarly, Yuan and Zydney concluded that the water fractioned through the smaller pore membrane caused less fouling of a subsequent membrane. They reported that large humic acid aggregates/particles have only relatively small effect on fouling during ultrafiltration, in contrast to the very large effect seen with MF membrane [6,25]. The large flux decline observed during constant pressure MF was caused by the formation of humic acid deposit located on the upper surface of the membrane. Prefiltration of the humic acid solution dramatically reduced the rate of fouling through the removal of large humic acid aggregates [25]. A results which consistent with above findings were also recognized by Howe and Clark, who observed most significant membrane fouling occurs in the presence of small colloidal matter, ranging from about 3 to 20 nm in diameter. They proposed that the majority of dissolved organic matter (DOM), by itself does not cause membrane fouling; the actual foulant is a relatively small fraction of bulk DOM [8]. -21-

Crozes et al. concluded that a polar molecular was more adsorbable than a non-polar molecule and thus, adsorption phenomena must be governed by membrane hydrophilicity and organic compound polarity. It appeared that adsorption of low MW molecules, much smaller than the membrane pore size, could lead to significant membrane fouling, irreversible by classical hydraulic backwash. Adsorption phenomena are driven by the nature of interaction between the organic compounds and a given membrane [14]. Other researches on humic acid fouling of microfiltration membranes has suggested that the convection transport and deposition of larger aggregates on the membrane surface dominates the fouling by humic components [26]. The tighter packing of charged humic acids within the fouling layer caused by the increased electrostatic shielding was recognized as reasons of fouling at high salt concentration. The effects of solution pH were more complex and likely reflect alterations in humic acid charge, hydrophobicity, and extend of aggregation.

-22-

2.4. Control of NOMs in water treatment 2.4.1. Coagulation Conventional treatment to produce drinking water normally involves coagulation, flocculation, sedimentation, and final filtration. Coagulation has been known as one of the most important processes. It is useful for destabilize suspended particles, and it also effective for removing dissolved organic matter. Coagulation is achieved by the use of conventional coagulants such as alum, ferric compound as well as newer coagulants polyaluminum chloride, polyaluminum sulfate, etc. Many studies have been conducted to evaluate the performance and effects of factors of coagulation process, especially aluminum and iron salts. Guigui et al. observed the UV removal was higher than DOC removal in the same conditions of coagulation [10]. According to their results, a pH of around 5 or 6 led to the highest DOC removal whereas a pH around 6 or 7 induced the highest UV removal. They also reported that good coagulation conditions (coagulant, pH and dosage) used in a conventional coagulation/settling process leads to good -23-

performance in terms of water quality. They found that the level of NOM removal by ferric coagulant was depended on the pH and was slightly higher for a pH around 7.0. In addition, impact of the pH on the coagulation efficiency was greater for the highest dosage of coagulant. In contrast, Bérubé et al. found that there was no benefit to using a greater concentration of polyaluminum chloride (PACl) and aluminum sulphate (alum) coagulants to achieve relatively high removal efficiencies for organic matter [27]. The reason was the characteristics of the reservoir water used in their study that was naturally soft, low in pH, alkalinity, and color, low turbidity as well as low concentration of organic material. With respect to affecting factors during coagulation using alum, Kuo and Amy has demonstrated that treatment of alum resulted in significant reductions in nonpurgeable organic carbon and UV absorbance. However, alum dosage was observed as a dominant factor affecting the removal of dissolved organic matter. They suggested that the greater reductions in nonpurgeable organic carbon could be gained with high alum dosage (25-50 mg/L). They also found that the most effective organic matter removal was occurred at the lower pH condition (pH=5.5) -24-

in the presence of a moderate amount of turbidity. Similarly, Harsha et al. provided results from a more extensive investigation on the coagulation behavior of eight natural water samples containing NOMs. The aluminum and iron salts have been applied in their experiments; and the coagulants dosages of 0.5 mgmetal/L and pH range from 5 to 6 were selected as optimum values. Besides, their data showed that larger NOM molecules required fewer amounts of coagulants to removed color unit in comparison with smaller NOM molecules [28]. Although optimum coagulation is significantly reduces the potential membrane fouling caused by particles and dissolved organic matter, but formed aggregates may also affect membrane filtration performance. Judd and Hillis recognized that pre-coagulant dosing upstream of microfiltration can improve the performance of the latter, with specific reference to the lowering of hydraulic resistance, provided floc growth has proceeded beyond a critical stage [11]. Their observation showed that low coagulant dosage tend to cause incomplete aggregation of colloidal particles and precipitated humic materials such that internal fouling of the membrane. Because of that, they emphasized that floc need -25-

to reach a certain critical floc size prior to microfiltration, otherwise membranes can be irreversibly clogged by the flocculant solids. Lee et al. also observed that the effect of coagulation condition was raising rate of transmembrane pressure at constant flux for MF membrane [29]. Besides, they have shown the smaller specific resistance with charge-neutralization than with sweep floc, which resulted much better on membrane permeability with charge neutralization. Carroll et al. investigated the impact of coagulation pretreatment on the microfiltration of a low turbidity and high-NOM surface water. They reported that 46% of DOC and 69% of UV254 were removed with alum at the optimum dosage. They also concluded that the rate of fouling caused by the main components of the residual NOMs could be controlled by the neutral hydrophilic substances [1].

2.4.2. Adsorption Adsorption materials such as kaolin, metal oxide, activated carbon, etc have potential to achieve removing NOMs. Zhang et al. have been demonstrated that the use of adsorbent particles in conjunction with UF or other membrane -26-

process can potentially provide benefits in terms of both NOM removal and fouling reduction [30]. Jekel provided results with granular ferric hydroxide as a new adsorption material, which can remove UV-active compounds preferentially [31]. However, among the above materials, activated carbon is the most widely applied adsorption material used in drinking water treatment and is has been proposed by US Environmental Protection Agency [32,33]. Carroll et al. have been shown that 68% DOC in water was removed with an excess of powdered activated carbon (100 mg/L for 68h) [17]. There are also detailed results about NOM adsorbed on an activated carbon, which have been reported by Speth [12]. According to the author, aromatic compounds are generally more strongly adsorbed than nonaromatic compounds. Therefore, terrestrial-derived organic compounds which tend to have greater aromatic character are expected to adsorb to a greater extent than aquatic-derived ones. Also, the greater adsorption was observed with the greater the hydrophobicity of the NOMs. Besides, smaller NOM molecules adsorb more strongly than larger ones because pore blockage of large molecules can limit the accessibility to adsorption sites [34]. Related to the adsorption mechanisms, -27-

Newcombe et al. suggested that at high surface concentrations an increase in ionic strength increased the adsorption onto activated carbons at all pH [35]. In the terms of adsorption treatments, powdered activated carbon (PAC), granular activated carbon (GAC), and iron oxides adsorption were known to remove NOMs but the added particles themselves could cause membrane fouling. Tsujimoto et al. recognized that the GAC pretreatment is effective to prevent irreversible fouling [5]. However, according to the results from Carroll et al., the rate of fouling by the PACtreated water was similar to the raw water though the PAC particles themselves did not contribute a resistance to filtration [1]. A similar result also reported by Bérubé et al., who reported that pre-adsorption with PAC did not significantly improve the ability of an UF membrane to removal organic matter contained in the reservoir water in their investigation [27]. Zhang et al. also recommend the effect of adsorbed particles depends strongly on the structure of the cake layer and its interaction with both NOMs and the membrane surface [30]. In addition, Yiantsios and co-worker suggested that the efficiency of the backwash process in removing the PAC deposits on the membrane was very limited especially when humic acids are absent [34]. -28-

2.4.3. Magnetic Ion Exchange Resin Recently decades, various advanced NOM treatment technologies have been proposed besides coagulation and adsorption processes. Among them, the magnetic polymer micro spheres, which have been successfully applied to biotechnology and medical science since 1970s, were developed to use in environmental technology to purify contaminated water [17]. Since the mid-1980s, MIEX (Magnetic Ion EXchange) has been designed as a novel material by ORICA (Australia) to improve DOC removal ability from raw water [36,37]. Their MIEX resins were acrylic, strong base anion exchange with achieved magnetic property by incorporation of permanently magnetized iron particles within the resin [16,38]. The MIEX process differs significantly from conventional ion exchange processes. Conventional ion exchange processes are occurred in fixed bed column, while the MIEX resin is suitable for use with stirred contactor which like a mixer in conventional water treatment plant; and this process does not require pre-treatment for solids removal [39,40]. In conventional ion exchange columns, the column has to be taken off-line and the -29-

resin regenerated, but in contrast, the loaded MIEX resin is continuously maintained. Also, the very small particle size of around 180 µm provides a high surface area allowing rapid adsorption kinetic. In addition, these magnetic particles agglomerate into rapidly settling resin floc, therefore it can separate completely [32,41]. The removal mechanic of DOC was demonstrated by the exchanging of negative charge DOC with a chloride ion on active site on the resin surface [39,42,43]. MIEX resin can be regenerated in a brine solution where adsorbed organics are substituted for chloride ions. The overview on the mechanisms of the MIEX resin loading and the regeneration process is illustrated in Figure 2-2.

-30-

Figure 2-2.

Mechanism of DOC exchange by FMIEX

-31-

Pelekani and his co-worker suggested that direct filtration of MIEX treated river water with an average turbidity up to 55 NTU is feasible [44]. Cook et al. have been performed a laboratory study to comparison efficiencies of conventional alum treatment and MIEX treatment. Their results showed that MIEX removed a range of compounds of all molecular weights, while alum coagulation only removed UV absorbing compounds greater than 2000 apparent molecular weights [45]. Singer et al. also found that MIEX pretreatment in conjunction with alum treatment provided the lower potentials of the residual TOC concentration and UV absorbance as compared to alum coagulation alone [46]. Specially, the process was very effective for waters with low TOC concentration, low specific UV absorbance values, and high alkalinities - waters where coagulation of TOC is usually not effective. Morran et al. have been used MIEX resin to removal DOC in water resource containing 11 mg/L DOC concentration. They recognized that the removal rates were rapidly achieved within 10 minutes contact time. Besides, the reduction in DOC levels also resulted in decrease in the alum dosage of up to 75% -32-

and permitted the application of direct filtration with a maximum alum dosage of 20 mg/L [16]. Brouke et al. reported that MIEX could be removed up to 80% of raw water NOM prior to MF or UF. Also, MIEX was also highly effective at removing low and medium molecular weight NOM. Because of that, they proposed MIEX process as an alternative pretreatment to apply before MF, UF or NF [47]. Hammann et al. have been carrying out a trial facility that MIEX pretreatment were installed [48]. The results showed that an average of 71% raw water DOC (from 11.8 to 3.4 mg/L) and 95% true color (from 27 to 1.3 Pt-Co units) were removed by MIEX treatment.

-33-

Chapter 3.

MATERIALS AND METHODS

3.1. Materials and Methods 3.1.1. Functional magnetic-impregnated ion exchange resin (FMIEX) The NOM removal process in this study was based on the novel FMIEX resin developed by Lee et al. [17]. This resin composed of strong base anion exchange shell and magnetic component. It has macroporous structure made from a moderately cross-linked styrene skeleton, and contains a high concentration of strong base, quaternary ammonia functional groups. Because of that, the FMIEX resin can be exchange with the weak organic acid ions. The FMIEX resin also has a magnetic component incorporated into its polymeric structure. This is very special feature which enables the settling of the resin during particles/water separated strategy to be better. Besides, this resin has a lot more external bead surface because of the small particles with average diameter about 219 µm. The segment unit model and the characters of this FMIEX are showed in Figure 3-1 and Table 3-1, respectively [17].

-34-

CH2

CH

CH2

CH

x

CH2 Cl

CH2

CH

y

CH2 N

H5C2

H C2H5

Cl C2H5

Figure 3-1.

CH2 10-x-y

The structure of FMIEX resin [17]

-35-

CH n

10

CH

CH2

Table 3-1.

The characteristics of FMIEX resin [17]

Matrix

Polystyrene

Functionality Counter ion Average particle size (µm) Specific surface area (m2/g) Average Exchange capacity (meq/g of resin) Average Magnetite fraction (wt%) Cross linkage (%) Available pH

-36-

Quaternary ammonium Chloride ion 219 0.0285 0.21 4.4 30 1-14

3.1.2. NOM and inorganic solutions Synthesis raw waters were created to supply constant quality of water sample by simulating real water in aspect of natural organic materials (NOMs). To investigate the efficiency of FMIEX resin for the purification of water containing NOMs, sodium salts of Aldrich humic acid (AHA) was used. The impacts of FMIEX resin on inorganic contaminated waters were also performed with nitrate and phosphate solutions. For the experiments to develop FMIEX process for NOM removal, Suwannee river humic acid standard (SRHA) was used. AHA was prepared with the commercial humic acid purchased from Aldrich Chemicals (Sigma-Aldrich, Milwaukee) and SRHA was prepared with the natural organic materials purchased from International Humic Substances Society (IHSS). All the humic acids were prepared by dissolving the organic materials in the MilliQ water. The background of the stock solutions (100 mg/L and 25 mg/L for AHA and SRHA, respectively) was set with 10 mL of 0.01N NaHCO3 solution per liter as a natural buffer system. The pH of the stock solution was brought to around 7 using 0.02N NaOH or 0.05N HCl solution. The stock -37-

solutions were also filtered through 0.45 µm membrane filter (Durapore, Millipore, Ireland). The filtered solution was kept in the brown bottles and under 4oC in the refrigerator. The humic acid stock solutions were used within one month. The 1000 mg/L nitrate and 1000 mg/L phosphate stock solutions were prepared with dried NaNO3 (Showa Chemical Co., Japan) and anhydrous KH2PO4 (Showa Chemical Co., Japan), respectively. The raw water was prepared every morning before the experiment started by dilution above stock solutions with Milli-Q water, and the solution was mixed rapidly for an hour with magnetic stirrer. The components and characteristics of the raw water can be adjusted according to the experimental requirements.

3.1.3. Membranes Membrane properties A polyethersulfone (PES) membrane with 100 kDa molecular weight cut off (MWCO) was used in this study, and its properties are summarized in Table 3-2. The membrane purchased as a cycle (63.5 mm in diameter, which provides about -38-

4.1 cm2 of filtration area). The structure of polyethersulfone with diphenylene sulfone repeating units is showed in Figure 3-2. The PES membrane has wide pH tolerance (from pH 1 to 13), fairly good chlorine resistance (up to 200 ppm chlorine for cleaning and 50 ppm chlorine for short-term storage of the membrane), and good chemical resistance to aliphatic hydrocarbons, fully halogenated hydrocarbons, alcohols, and acids. Disadvantages of PES are less resistance to aromatic hydrocarbons, ketones, ethers, and esters; and hydrophobicity which leads to an apparent tendency of membrane fouling by stronger interactions with a various solutes [19]. Also, this membrane which has NMWL=100,000 has low pressure limit (the maximum recommended pressure is 10 psi according to suggestion of manufacturer).

Preparation of virgin membrane for deadend filtration The virgin membrane sheet was soak in 1,000 mL of Milli-Q water for 24hrs and the water was changed three times every hour. The rinsed membrane was kept in the Milli-Q water in the refrigerator at 4oC.

-39-

Table 3-2.

Properties of the membrane used in this study

Membrane ID

PBHK BiomaxTM, Millipore UF 100 k MWCO Polyethersulfone Hydrophobic 50oC 10 psi (0.7 atm)

Manufacturer Membrane type Nominal pore size Base material Hydrophilicity Temperature limit Maximum recommended pressure

O S

O

O Figure 3-2.

n

Structure of polyethersulfone mebrane

-40-

3.2. Experimental Methods 3.2.1. Jar testing experiment A useful laboratory experiment for the evaluation of treatment process of raw water is the jar test. This test provides information on the effects of the condition of the process, such as chemical dosage, required contact time, mixing rate, and the water quality parameters on the process. The jar test is simulating a water treatment process and often used before the design of treatment facilities. In this study, jar testing was employed to evaluate the effects of parameters as well as efficiency of FMIEX resin process base upon investigation of NOM removal process. A jar tester (C-JT, Chang shin Science Co.) with 6 variable-speed-paddles overhead stirrer was used. Jar testes on FMIEX resin process were performed as follows. Jar tests were carried out in 500 mL (sometimes 1000 mL) beakers. One container acts as a control while the operating conditions can be varied among the remaining the other containers. The solution was mixed rapidly at 180 rpm to keep the resin in suspension. Mixing was then stopped and the resin was allowed to settle for 5-10 minutes by a magnetic force. -41-

1.0

12 10

0.6 8 0.4

6

DOC (mg C/L)

0.8 UV254 (cm-1)

14

UV254 DOC

4 0.2 2

(a) 0.0

0 0

10

20

30

40

NOM concentration (mg AHA/L)

0.7 14

UV254 DOC

12

UV254 (cm-1)

0.5

10

0.4

8

0.3

6

0.2

DOC (mg C/L)

0.6

4

0.1

2

(b)

0.0

0 0

5

10

15

20

25

NOM concentration (mg SRHA/L)

Figure 3-3.

The relationship of UV254 and DOC with NOM concentration (a) Aldrich humic acid; and (b) Suwannee river humic acid

-42-

Supernatant were sampled for “treated water” quality tests such as pH, temperature, turbidity, color, DOC and UV254 measurements. Samples for determination of color, DOC and UV254 qualities were filtered through 0.45 µm membrane to remove the suspended materials. Supernatant were also collected to perform in the protocol for UF filtration with FMIEX pretreated water. These solutions were not prefiltered with other membranes filters. During the preliminary investigations, the effectiveness of treatment processes in term of NOM removal was estimated only by the UV254 improvements. However, both UV254 and DOC parameters were focused in experiments for development of FMIEX process. The relationship between these parameters (UV254 and DOC) and NOM concentration are showed in Figure 3-3.

3.2.2. Dead-end filtration A stirred deadend filtration unit (Amicon, model 8200) was used for the membrane filtration system. The internal diameter of the filtration unit was 63.5 mm and the volume was 200 mL with a stirring bar. All the experiments were carried -43-

out in stirred mode. The filtration unit was filled with the feed solution before each test. The driving force was provided by pressure of nitrogen gas. The pressure was controlled using the pressure regulator with control gauges attached to the gas line. After fluctuated pressure was stabilized, the pressure valve was opened to feed the provided pressure. The permeate flux of the system was measured automatically with the digital balance (AND, A&D Co., Ltd) connected to the computer (Figure 3-4). The data collected during the experiments were processed with computer programs. The experiments were carried out at room temperature. A new membrane was used for each test. All flux decline tests were performed with similar permeate flux. Before the ultrafiltration test was carried out with the raw waters, the pure water flux was measured for each membrane with different provided pressures. This order have given the right pressure and assisted to compute the intrinsic membrane resistance value. Besides, the new membrane surface could be compacted (i.e., compaction between membrane surface and supporting layer) after this filtration step; therefore, flux was also -44-

stabilized. Flux decline tests were then performed with untreated or FMIEX treated waters. Each water was filtrated through the membrane to obtain around 120 mL of permeate solution. The efficiency of hydraulic and chemical cleaning was carried out. The ultrafiltration protocols are illustrated in Figure 3-5.

-45-

Pressure gauge

Data collection

Nitrogen gas Membrane

Magnetic stirrer

Digital Balance Figure 3-4.

Schematic diagram of the deadend filtration system for flux decline tests

-46-

Measuring J0 (Pure water flux) J0 = NOM flux decline test (filtration of untreated or FMIEX treated water using PES100 kDa membrane)

∆P µRm

J end =

∆P µ (R m + R t )

J 0,b =

∆P µ (R m + R b )

J 0 ,c =

∆P µ (R m + R i )

Backwashing

Measuring J0,b (Pure water flux)

Chemical cleaning (using 0.1N NaOH)

Measuring J0,c (Pure water flux)

Figure 3-5.

The ultrafiltration protocols with PES membrane

-47-

3.2.3. Analytical methods pH The pH was measured with the HACH pH meter, model SenIon-4 or SensIon-156. Every morning before the experiments started, the pH meter was calibrated with two standard buffer solutions with pH of 4.01 and 7.00 (Singlet, HACH). After using the pH meter, the electrode was stored in storage solution provided by the manufacturer. Turbidity A HACH turbidimeter (model 2100AN) with HACH filter module (suitable for USEPA method 180.1) was used to give a direct reading of the turbidity of a sample in nephelometric turbidity units (NTU). The turbidity calibration was performed with Stablcal formazin solution. Color Color was determined spectrophotometrically by comparing the absorbance of the sample at 455 nm with the platinum-cobalt standard with HACH turbidimeter (model 2100AN) giving a direct reading of the color. All samples -48-

were filtered through a 0.45 µm membrane (Durapore membrane, Millipore). UV absorbance Ultraviolet

absorbance

was

measured

on

a

UV-VIS

reading

spectrophotometer (model UV2401-PC, Shimadzu) at 254 nm wavelength with 1-cm quartz cell. The MilliQ water (Millipore) was used as the blank. Sample solutions were prefiltered through 0.45 µm membrane. pH was not adjusted and measurement was carried out at room temperature. Ultra violet absorbance at 254 nm (UV254) was focused because it has been known to be directly related to the aromatic fraction of NOM [49,50]. DOC Shimadzu TOC-5000A Total organic carbon analyzer employing a hightemperature combustion method was used to measure the DOC. Samples were prefiltered through 0.45 µm membrane. Samples for DOC analysis were not acidified. The sample solution was stored in the refrigerator at 4°C until measurement. Organic carbon samples were quantified by comparison with Potassium Hydrogen Phthalate (Showa Chemicals Co., Japan) standard solution -49-

and reported as carbon in mg/L. The DOC content was calculated from the difference of the total carbon minus inorganic carbon and was read directly from the analysis instrument. SUVA SUVA is simple parameter to characterizing the nature of organics by generating a value dependent on the level of aromaticity per carbon unit of the NOM. In order to calculate SUVA, the UVabs of sample in cm-1 was divided by the DOC of the sample, multiplied by 100cm/M and reported in units of L/mg-M.

SUVA (L/mg-M) = UV254(cm-1) / DOC (mg/L) * 100 cm/M

UV254 calculation: UV254 = A /d

where: UV254 = the calculated UV absorbance at 254 nm wavelength of the sample in absorbance units (cm-1); A=the measured UV absorbance at 254 nm of the sample that is filtered through a 0.45-µm filter media; and d = the quartz cell path length in cm.

-50-

Inorganic anions Inorganic contents were measured by use of a Dionex ion chromatography (model DX-80) employing an anionic column (model IonPACAS14A) with flow rate 0.5 ml/min. Mobile phase used was mixture solution of 8 mM/L Na2CO3 and 1 mM/L NaHCO3. Nitrate and phosphate concentrations were quantified using NaNO3 and KH2PO4 standard solutions, respectively.

-51-

Chapter 4.

RESULTS AND DISCUSSION

4.1. Preliminary investigation on FMIEX 4.1.1. Removal of humic acids (HA) using FMIEX Preliminary experiments were first conducted for raw water with Aldrich humic acid (AHA) concentration of 10 mg/L to investigate the impact of FMIEX resin on NOM removal. The range of FMIEX dosages of 4 g/L, 8 g/L and 12 g/L were selected. Raw waters, which were brought to pH 7.0, were placed in 1-liter beakers, and dosed with desired amounts of FMIEX resin, and mixed at 180 rpm. While the waters were being mixed, samples for water quality measurements were taken from each beaker at mixing times of 5, 10, 20, 30, 60, 120 and 180 minutes. The impact of FMIEX dosage and contact time on the removal of UV254 and pH of treated water are showed in Figure 4-1. This figure illustrates that UV254 removal ability increased with increasing FMIEX dosage and contact time. The increasing resin dosage provided a lager particle density; therefore, the rate of ion exchange reaction was improved. However, there are quite different levels of the impact increases. The UV254 removal efficiency was twice rising when resin -52-

dosage was increased from 4 g/L to 8 g/L, but a less impact was recognized when the dosage was increased from 8 g/L to 12 g/L. For the effects of contact time, the results show that most of the removals occurring within the first 60 minutes of contact time. The figure also shows that the resin seems to be fouled by AHA after 180 minutes towards the case of 4 g/L FMIEX dosage, while a larger amount of the resin appears to be able to react. Hence, the illustrated results in Figure 4-1 clearly demonstrated the capability of new FMIEX resin to remove humic acid. The AHA removal capacity of this FMIEX resin was approaching 0.84 mg AHA/g resin. Removal achievements are depending upon the dosage of resin as well as mixing time. The slight drops in the pH of the water as a result of the FMIEX treatment were also observed. pH of treated water with in case of a resin dosage of 12 g/L was lowest among three cases of different used FMIEX dosages.

-53-

dosage=4g/L (UV254 removal percent) dosage=12g/L (UV254 removal percent)

100

14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

dosage=4g/L (pH of finished water) dosage=8g/L (pH of finished water) dosage=12g/L (pH of finished water)

UV254 Removal (%)

80

60

40

20

0 0

20

40

60

80

100

120

140

160

180

pH

dosage=8g/L (UV254 removal percent)

200

Contact Time (min.) Figure 4-1.

UV254 removal percent as a function of contact time for different FMIEX resin dosages (AHA concentration=10mg/L; pH=7; FMIEX resin: virgin particles; mixing rate: 180rpm)

-54-

4.1.2. Removal of inorganic anions using FMIEX As a previous study, the FMIEX resin has been made to purify contaminated upon the ion exchange mechanism [17]. Therefore, the resin would be also effect on the inorganic anions removal. Potential for application of FMIEX resin for removal of nitrate and phosphate, which can create water quality or treatment issues when they are present in raw water, has been observed in this study. Removals of these inorganic anions were performed in separated experiments. Raw waters with nitrate, or phosphate concentration approximately 15 mg/L as NO3--N, or 5 mg/L as PO43--P was used, respectively. The operation condition of jar testing was same in the experiments for AHA removal. Water samples were taken to measure nitrate or phosphate using ion chromatography analyzer. Figures 4-2 and 4-3 show the impact of contact time and FMIEX dosages towards nitrate and phosphate removals, respectively. Both figures show that the required contact time to reduce most nitrate and phosphate were approximately 30 minutes, except the case of 12 g/L FMIEX resin was used. Observably, FMIEX -55-

resin was more effective on nitrate removal than phosphate removal. Expectedly, the contact time required was reduced by increasing the resin dosage, but not directly in proportion. The increasing FMIEX resin dosage was made phosphate removal capacity to increase in a regular rising, while the impact of nitrate removal relating to the increasing FMIEX resin dosage was not even. For nitrate removal, it seems that the higher density of the resin which provided much surface activity site, and the small molecular size of nitrate anion have been responded to higher kinetic activity. Therefore, nitrate anions were exchanged rapidly within first five minutes. Of course, the increasing FMIEX dosage also became available for further phosphate exchange, but it were slightly obstructed by the bulky phosphate ion which led the reaction occurred slowly. The removal capacity of nitrate and phosphate were calculated and summarized in Table 4-1. The ion exchange capacity of nitrate was 0.094 meqNO3--N/g resin, whereas that of phosphate was only 0.007 meqPO43--P/g resin (approximately ten times higher). The difference among these values might have been caused by the differently sized ions. This result is similar with that reported -56-

by previous study, which have been found that the ion exchange capacity of nitrite removal was smaller than the ion exchange capacity given by acid-base titration due to size of target ion as well as the pH [17].

-57-

Table 4-1. Inorganic ion

The capacity of FMIEX resin towards several inorganic ions Capacity

Data source

Hydroxyl

0.21 meq mol OH-/g resin

Lee et al. [17]

Nitrite

0.18 meq mol NO2-/g resin

Lee et al. [17]

Nitrate

0.094 meqNO3--N/g resin

this study

Phosphate

0.007 meqPO43--P/g resin

this study

-58-

100

14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

NO3- Removal (%)

80

60

40

20

0 0

20

40

60

80

100

120

140

160

180

pH

Nitrate removal (MAEX dosage = 4g/L) Nitrate removal (MAEX dosage = 8g/L) Nitrate removal (MAEX dosage = 12g/L) pH of finished water (MAEX dosage = 4g/L) pH of finished water (MAEX dosage = 8g/L) pH of finished water (MAEX dosage = 12g/L)

200

Contact Time (min.) Figure 4-2.

Nitrate removal as a function of contact time for different FMIEX dosages (nitrate concentration: 15mg/L as NO3--N; pH=6.7; stirring rate: 180rpm)

-59-

100

PO43- Removal (%)

80 60 40 20 0 0

20

40

60

80

14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

pH

Phosphate removal (MAEX dosage = 4g/L) Phosphate removal (MAEX dosage = 8g/L) Phosphate removal (MAEX dosage = 12g/L) pH of finished water (MAEX dosage = 4g/L) pH of finished water (MAEX dosage = 8g/L) pH of finished water (MAEX dosage = 12g/L)

100 120 140 160 180 200

Contact Time (min.) Figure 4-3.

Phosphate removal as a function of contact time for different FMIEX dosages (phosphate concentration: 5mg/L as PO43--P; pH=7; stirring rate: 180rpm)

-60-

4.1.3. Evaluation on FMIEX resin regeneration Generally, the resin loaded with anions can be regenerated with a concentrated sodium chloride (NaCl) solution. The two main parameters that control efficiency of the regeneration are contact time and regenerant dosage. In this study, up take FMIEX resin with nitrate or phosphate was regenerated by 10% w/w NaCl solution with different contact times. This concentration of regenerant solution was referenced from several previous studies, which were recommended to use the solution of 8 to 12% w/w brine [32,39]. Figure 4-4 clearly demonstrates the effectiveness of regeneration for reusing the resin about 15 times with approximately more than 80% recovery capacity. The inorganic ion exchange reaction is more effectively reversed with the resin preference changed from nitrate or phosphate to chloride ions under the high regenerant ionic strength and chloride concentration. With the regenerant was 100 g/L NaCl dose at 1 g NaCl per 1 g resin, the regeneration time required only 5 minutes. This caused by the inorganic ions had just exchange with chloride counter ion during loading strategy. In addition, short time regeneration required -61-

Regeneration using 10%NaCl regenerant for 10min. for 5min. for 15min.

100

% Removal

80

60

40

20 Nitrate Removal percent Phosphate Removal percent 0 virgin 1

2

3

4

5

6

7

8

9 10 11 12 13 14 15

Number of Regenerated Resin Reuse

Figure 4-4.

Inorganic anion removals versus regenerated resin reuse

Recovery of UV254 removal ability (%)

100

80

60

40

20

0 without pretreatment

with pretreatment

Regeneration

Figure 4-5. Effect of different protocols of AHA loaded FMIEX resin regeneration -62-

was gained by the kinetic activity which provided by suspended particles. By this FMIEX mechanism, the larger surface area of particles could be contacted with regenerant solution in the same time. In case of AHA loaded FMIEX resin, the reversed reaction was less effectively reversed with the resin preference changed from AHA to chloride ions. The results showed that approximately 57% removal capacity of the resin was recovery by only regeneration (Figure 4-5). The AHA removal ability of reuse resin still decreased with increasing of number of reuses. Also, the recovery capacities of resin towards DOC were not improved much with the increasing regeneration time. This behavior of the FMIEX resin was affected due to the characteristics of AHA. The information on AHA which have been characterized using gel-permeation chromatography (GPC) indicated that its apparent molecular weight is primarily to be 20,000 Daltons or less; and also ranges over 67,000 Daltons in size [51]. Thus, it demonstrates that the FMEX resin was formed stronger ionic bonds with larger AHA. This results is consistent with the previous data of Slunjski et al., who reported that large NOM (MW > 10,000 Daltons) was -63-

hard to extract during resin regeneration and had tend to foul resin [52]. In order to solve the problem that like in this case, Bourke et al. suggested that small additions of caustic soda or acid to the brine should be required [41]. Herein, one pretreatment strategy was applied prior to regeneration. This pretreatment procedure consists of three-times rinsing step, alternating from 0.1N NaOH to 0.1N HCl, and the resin was consecutively washed with DI water, interposingly. The AHA carryover resin was suspended in each agent solution with stir for 30 minutes. AHA parts were soluble in sodium hydroxide solution, and then, the resins were acidifying by hydrochloride. After pretreatment, FMIEX resin was then regenerated by 10% w/w sodium chloride solution which adjusted to pH=3. The acidic pH provides the FMIEX resin in form of the positive charge which easily accept an negative counter ion (chloride). The obtained results showed that approximately 78% Aldrich humic acid removal capacity was recovery. This value is adequate to the recovery value of the inorganic anion take up resin. Hence, pretreatment strategy has been eluting AHA completely; and the

-64-

regenerant ionic strength as well as chloride concentration by using of 10% w/w NaCl solution was high enough.

4.1.4. Preparation of FMIEX resin from the uninvolved materials The effectiveness of FMIEX resin for removing UV absorbance, which relates to aromaticity removal, was recognized. However, because base material of the resin is polystyrene and is cross-linked by divinylbenzen, which involves C = C bonds as well as aromatic rings, so a few of organic material may be released into the aqueous solution caused by FMIEX resin itself. An experiment, which conducted the same jar test conditions of HA water treatment mentioned above, was performed. Different dosages of 4 g/L, 8 g/L, 12 g/L, 16 g/L and 20 g/L virgin FMIEX resin were suspended in ultrapure (Milli-Q) water at 180 rpm for 60 minutes. Water samples were taken after resin settling within 5 to 10 minutes. The quality of the finished water was then determined. The pH values and turbidity were showed in Figure 4-6. The UV254 absorbances and color units were also compared with those of waters containing 10mg/L Aldrich HA (Figure 4-6). -65-

14

50 pH value Turbidity

8

30

6

20

4 10

2 0 0.25

UV254 (cm-1)

0.20 0.15

0 100 UVabs @254nm Color unit UV254 value of water containing 10mg/L Aldrich HA

80 60

Color unit of water containing 10mg/L Aldrich HA

0.10

40

0.05

20

0.00

0 0

5

10

15

20

Color (Co-Pt unit)

pH

10

40 Turbidity (NTU)

12

25

FMIEX resin dosage (g/L) Figure 4-6.

Water quality after 60 min. contact with different FMIEX dosages

-66-

100 UV254 removal

% Removal

80

Color removal

60

40

20

0

Turbidity (NTU)

20

15

10

5

0 Virgin particles

Prepared particles

FMIEX resin used

Figure 4-7.

Preparation variables for FMIEX resin (raw water: SRHA with concentration=5mg/L; pH=6.3; FMIEX dosage: 12g/L; contact time: 30 min; mixing rate: 180 rpm) -67-

There in order that the UV254 absorbances and color units of finished water were equivalent those of tap water. Because of that, this effect can be accepted. However, the results show that the increasing FMIEX resin dosage increases the turbidity. Hence, FMIEX process causes a slight increase in turbidity due to a small amount of resin carry over from the settler. This phenomenon also has reported by almost of previous studies [32,41,44]. From above results, virgin FMIEX resin was proposed to prepare before using. The prepared procedure employed the pretreatment and the regeneration processes that early mentioned in part 4.1.3. In order that the behavior of prepared resin would be characterized as well as to prepare for further experiments, the Suwannee river humic acid (SRHA) solution was treated using the prepared FMIEX resin. The concentration of SRHA was 5 mg/L and the pH value (approximately 6.3) of this water was not adjusted. The FMIEX resin was dosed at 12 g/L, and mixed at 180 rpm for 30 minutes. The results on FMIEX treated water quality illustrate that the removing UV254 and removing color capacity of the resin were not improved much, but the gained benefit was approximately -68-

70% turbidity reduction in comparison with turbidity caused by virgin particles (Figure 4-7).

4.2. Development of FMIEX process for HA removal 4.2.1. Effects of contact time on HA removal Development experiments were conducted for water containing Suwannee river humic acid (SRHA) to examine HA removal process. The prepared FMIEX resin was used in this series tests. To determine contact time required for raw water with SRHA concentration equal to 5 mg/L, FMIEX dosage of 8 g/L was added. The jar test conditions were kept as performed during preliminary experiments. Sorption kinetic curve of SRHA by FMIEX resin is showed in Figure 4-8. This figure illustrates that sorption was rapid, as demonstrated by removing UV254 percentage versus time. Although removing UV254 percentage still increased with increasing contact time, but most of the removal were recognized within the first 20 to 30 minutes. According to the previous study on the characterization of NOMs, size -69-

exclusion chromatography (SEC) provided a molecular weight distribution profile of SRHA which ranges from less than 1,000 Daltons to around 10,000 Daltons have been reported by Amy et al. [53]. Hence, results in this experiment correspond to the judgment of Slunjski et al. that smaller NOM (MW 1,00010,000 Daltons) are exchanged rapidly during resin loading [52].

4.2.2. Effect of pH on HA removal Experiments on the influence of pH during SRHA removal were conducted at pH 4, pH 6, pH 8, and pH 10 (pH adjusted with 0.05N HCl or 0.02N NaOH). The FMIEX resin dosed at dosage of 12 g/L, and allowed resin suspension at 180 rpm for 30 minutes. Figure 4-9 shows the results on UV254 absorbance removal and DOC removal at different operating pH. The efficiencies of removing UV254 absorbance as well as removing DOC were dramatically increased up to 80.41% and 65.07%, while at higher pH (e.g. pH 10) those values were only 35.14% and 16.73%, respectively.

-70-

UV254 removal pH value of finished water

UV254 Removal (%)

80

60

40

20

0 0

20

40

60

80

14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

pH

100

100 120 140 160 180 200 220

Contact Time (min.) Figure 4-8.

UV254 removal percent as a function of contact time (raw water: SRHA; concentration=5mg/L; pH=6.3; FMIEX resin: prepared particles; dosage = 8g/L; mixing rate: 180rpm)

-71-

100 UV254 removal DOC removal

% Removal

80

60

40

20

0 4

6

8

10

pH value Figure 4-9.

Influence of pH on HA removal (raw water: SRHA; concentration=5mg/L; contact time: 30 min.; FMIEX resin: prepared particles; dosage = 12g/L; mixing rate: 180rpm)

-72-

Generally, there have more negative charged HA and neutral species of anion exchange resin at high pH value. By this reason, Croué et al. have been discussed that increasing the pH from 4 to 10 could favor the anion exchange. However, they also recognized that adsorption was the predominant mechanism at acid pH [54]. As a rule, humic acids are weak organic acids, which do not completely dissociate, so the Henderson-Hasselbach equation (Equation 4.1) may be also useful to consider the behavior of HA removal at different pH.

[A − ] pH = pK a + log (4.1) [HA]

where: Ka - the equilibrium constant for the dissociation A- - conjugate base HA - humic acid

Evidently, when the pH value is higher than pKa value, the humic acids exist major in anionic form ([A-] >> [HA]) and they would be easy to replace with -73-

counter ions on the resin. It means that ion exchange is main mechanism. In opposition, when the pH of the solution is lower than pKa value, the humic acid exists undissociated form (neutral form) and is responsible for the attachment of the nonionic portion of a humic acid to the internal surface of resin. According to Fu et al., an organic acid could be removed by either ion exchange mechanism alone or adsorption mechanism alone, or by combination of both mechanisms [33]. Herein, the experimental results show a decreasing in activity of the FMIEX resin with an increasing pH value. Therefore, the behavior of resin with humic acids under the influence of pH can be proposed as follows: adsorption probably occurred in combination with anion exchange at low pH value, while only anion exchange was existed at higher pH value. At the acidic pH, the carboxyl groups of a humic acid molecule bind with resin amine functional groups (ion exchange) and the nonionic portion of the molecular attaches to the styrenic skeleton of resin or adsorb through hydrogen bonds between NOM and the nitrogen atom of the amine functional group (surface adsorption). This phenomenon may be caused by hydrophobic fraction of SRHA is majority; and -74-

the ratio between carboxyl and phenolic functional groups at different pH are quite difference.

4.2.3. Determination of appropriate FMIEX dosage The impacts of FMIEX dosage on the removal of UV254 absorbing substances and DOC under initial pH of SRHA are showed in Figure 4-10. Both of UV254 and DOC of the water decreased with increasing FMIEX dosage. Differ from the effects of FMIEX resin dosage which employing virgin particles towards AHA water, the relationship between the dosages of prepared particles and the removal percentage of UV254 from SRHA water appears to be a straight. The data were summarized in Table 4-3 also shows that color removal is generally slower than the removal of the UV-absorbing compounds. This may be caused by the colored species are of larger molecular size. In general, these results clearly illustrate the capability of the FMIEX resin to removal aromatic compound. Removals approaching 80% was achieved, and a higher removal may be gained depending upon the dosage of FMIEX resin used. In this tests, a dosage of 16 g/L -75-

FMIEX resin yielded good efficiency in terms of both UV254 and DOC removals. However, the results given by this test in comparison with that of preliminary test lead to suggest that characteristics of NOM types are significantly effect on the performance of FMIEX treatment.

-76-

100 UV254 removal DOC removal

% Removal

80

60

40

20

0 0

4

8

12

16

20

24

FMIEX resin dosage (g/L) Figure 4-10.

Impact of FMIEX resin dosage on UV254 and DOC removals (SRHA; concentration=5mg/L; pH=6.32; FMIEX resin: prepared particles; contact time: 30 min.; mixing rate: 180rpm)

-77-

4.3. Effects of FMIEX pretreatment on membrane filtration performance 4.3.1. Analysis of flux decline test results Ultrafiltration of FMIEX pretreated waters (under different pH conditions: pH 4, pH 6, pH 8, and pH 10) was performed to investigate the effect of FMIEX process on the performance of the membrane filtration. Four different FMIEX treated waters (which were taken from jar test series in part 4.1.2) were prepared for the UF experiments. Figure 4-11 shows the flux obtained with PES 100 kDa membrane when filtering the four FMIEX pretreated waters mentioned above. The feed characteristics were similar in turbidity which caused by suspended FMIEX resin residual, but different in SRHA concentration as well as pH value. Here Jo was the initial FMIEX treated water flux. The evidence that flux decline has implication of HA concentration. The flux reduction increased with increasing aromatic content. However, pH also has significant impact. Although the FMIEX treated water at pH 4 had lowest UV254 absorbance, but the rate of flux decline was higher than that of the other ones. This result was able to demonstrate the -78-

importance of SRHA adsorption in the flux decline. The more neutral form of SRHA, which exists at acidic pH, was most responsible for fouling. The HA deposit on the upper surface of the membrane could be major reason. Similar results have been reported by Yuan and Zydney [6,26]. They have observed that the flux decline was most rapid at pH 3 in comparison with flux decline at pH 7 and 9 during filtration of 2 mg/L humic acid solutions through the both of UF (30 kDa) and MF (0.16 µm) polyethersulfone membranes. The declines in permeate flux during ultrafiltration of untreated water and the water after treatment with different FMIEX dosages are compared in Figure 4-12. In this test, pH of all FMIEX treated waters was adjusted to initial pH value of raw water. The direct filtration of untreated water led to rapid flux decline. The FMIEX pretreatment evidently led to significant reduction of flux decline compared to the untreated water. Also, the rate of flux decline decreased to be adequate to increase of FMIEX dosage. This is due to higher reduction of SRHA components through the FMIEX pretreatment.

-79-

1.00

0.95

J/Jo

0.90

0.85

TWP4: UV254=0.029cm-1; pH=3.88 TWP6: UV254=0.073cm-1; pH=4.82

0.80

TWP8: UV254=0.087cm-1; pH=6.78 TWP10: UV254=0.096cm-1; pH=6.98

0.75 0

20

40

60

80

100

120

140

Permeate volume (mL) Figure 4-11.

Flux decline test results on FMIEX treated water with different pH (filtration at permeate flux around 450 LMH with stir; TWP4, TWP6, TWP8, TWP10 are SRHA treated waters at pH=4, pH=6, pH=8, pH=10, respectively; FMIEX resin: prepared particles; FMIEX dosage= 12g/L)

-80-

1.00

0.95

J/Jo

0.90

0.85 Water without FMIEX pretreatment FMIEX pretreated water with dosage=12g/L FMIEX pretreated water with dosage=16g/L FMIEX pretreated water with dosage=20g/L

0.80

0.75 0

20

40

60

80

100

120

140

Permeate volume (mL) Figure 4-12.

Flux decline test results on FMIEX treated water with different FMIEX dosages (filtration at permeate flux around 450 LMH with stir; feeds are FMIEX pretreated waters; pH=6.3; FMIEX resin: prepared particles)

-81-

The water quality given by FMIEX treatment alone and FMIEX treatment combined with ultrafiltration are summarized in Tables 4-2 and 4-3. It was demonstrated that UV254, DOC, SUVA, and color removals were achieved. Besides, although FMIEX process caused a slight increase in turbidity, but it appears that FMIEX particles themselves did not contribute the membrane filtration performance.

-82-

Table 4-2. Procedure

The water quality summary for treated water under different pH value Raw water

FMIEX

FMIEX + UF

FMIEX

FMIEX + UF

FMIEX

FMIEX + UF

FMIEX

FMIEX + UF

pH of finished water

6.29

3.88

3.94

4.82

4.76

6.78

6.31

6.98

6.52

UV254 (cm-1)

0.148

0.029

0.008

0.073

0.037

0.087

0.043

0.096

0.021

80.41

94.59

50.68

75.00

41.22

70.95

35.14

85.81

0.896

0.462

1.53

1.212

2.05

1.639

2.136

1.206

65.07

81.99

40.35

52.75

20.08

36.10

16.73

52.98

UV254 (cm-1) % removed -83-

DOC (mg/L)

2.565

DOC % removed

pH = 4

pH = 6

pH = 8

pH = 10

SUVA (m-1mg-1L)

5.770

3.237

1.732

4.771

3.053

4.244

2.624

4.494

1.741

Color (Co-Pt unit)

61

18

3

36

14

54

21

42

13

70.49

95.08

40.98

77.05

20.08

66.67

16.73

79

6.86

0.145

6.31

0.229

5.94

0.162

6.34

0.159

Color % removed Turb. (NTU)

0.198

Table 4-3.

The water quality summary for various treatments with different FMIEX dosages

Procedure

UV254 (cm-1)

Raw water

0.147

UV254 (cm-1) % removed DOC (mg/L)

2.529

-84-

DOC % removed

FMIEX dosage=12g/L

FMIEX dosage=16g/L

FMIEX dosage=20g/L

FMIEX

FMIEX + UF

FMIEX

FMIEX + UF

FMIEX

FMIEX + UF

0.067

0.031

0.041

0.019

0.029

0.012

54.42

78.91

72.11

87.07

80.27

91.84

1.605

1.205

1.242

1.076

1.174

0.875

36.54

52.35

50.89

57.45

53.58

65.40

SUVA (m-1mg-1L)

5.813

4.174

2.57

3.301

1.77

2.470

1.37

Color (Co-Pt unit)

63

30

21

21

12

18

3

52.38

66.67

66.67

80.95

71.43

95.24

3.89

0.219

4.73

0.194

5.62

0.167

Color % removed Turbidity (NTU)

0.173

4.3.2. Analysis of resistance components The efficiency of hydraulic and chemical cleaning was performed. Fouled membrane by SRHA was rinsed and backwashing with MilliQ water (to obtain 120 mL permeate) at the same TMP which applied during flux decline test. The membrane was also chemical cleaning by immersed the membrane in 0.1N NaOH solution for 30 minutes. The resistance components were calculated and illustrated in Figures 4-13 and 4-14. Figure 4-13 shows that the irreversible fouling resistance tends to be decrease towards water samples have higher pH value. In case of treated water at pH=4, the irreversible fouling resistance value was minus due to the fluxes were quite pronounced at pH=4, with the MilliQ water flux at pH=6.3 measured after chemical cleaning. Figure 4-14 also clearly demonstrates that fouling resistance have reduced through the removal of humic acid by increasing dosage of FMIEX resin used. The irreversible fouling resistance makes up only approximately 5% of total resistance components. It appears that the fouling resistance by NOM was prevented. -85-

2.5e+5

intrinsic membrane resistance reversible fouling resistance irreversible fouling resistance

Resistance value (cm-1)

2.0e+5 1.5e+5 1.0e+5 5.0e+4 0.0

WOT1

TWP4

TWP6

TWP8

TWP10

Feeds Figure 4-13.

Resistance components of UF membrane after filtration of FMIEX treated water with different pH (TWP4, TWP6, TWP8, TWP10 are SRHA treated waters at pH=4, pH=6, pH=8, pH=10, respectively; FMIEX resin: prepared particles; FMIEX dosage= 12g/L; mixing rate: 180rpm; WOT1 is raw water without FMIEX pretreatment)

-86-

irrevesible fouling resistance chemically reversible resistance hydraulically reversible resistance intrinsic membrane resistance Percentage of resistance components (%)

100

80

60

40

20

0 WOT2

Figure 4-14.

TWD_12 TWD_16 Feeds

TWD_20

Resistance components of UF membrane after FMIEX treated water with different FMIEX dosages (TWD12, TWD16, TWD20 are SRHA treated waters with FMIEX dosage=12g/L, 16g/L, 20g/L, respectively; pH=6.3; FMIEX resin: prepared particles; mixing rate: 180rpm; WOT2 is SRHA raw water without FMIEX pretreatment)

-87-

Chapter 5.

CONCLUSIONS

The results of this research conducted for the investigation of NOM removal using the novel FMIEX resin and estimation of FMIEX pretreated water effecting membrane fouling have suggested. The specific contributions are as follows: 1. The innovative functional magnetic-impregnated ion exchange resin based on polystryrene, strong base anion exchange, and quaternary functional group showed a significant amount of humic acid removal. Removals approaching 80% of UV254 absorbance which related to aromaticity was achieved, and a higher removal can be gained depending upon the dosage of FMIEX resin used, operating pH as well as characteristics of raw water. 2. Humic acid could be removed by either alone ion exchange mechanism at base pH, or by combination of both ion exchange and surface adsorption mechanisms at acid pH. The FMIEX resin presents are also able to reversible exchange, which leads it to reuse for number of time. -88-

3. Potential for application of FMIEX for removal of a number on inorganic anions contaminated water such as nitrate, phosphate has been observed. The removal capacity of nitrate and phosphate were 0.094 meqNO3--N/g resin and 0.007 meqPO43--P/g resin, respectively. 4. The effects of FMIEX pretreated water on flux decline were investigated with the stirred deadend filtration unit, and the membrane used for this study was hydrophobic ultrafiltration (UF) membrane (polyethersulfone, MWCO=100kDa). The FMIEX pretreatment dramatically reduced the rate of memrbane fouling through the removal of humic acid. 5. Hydrophobic fraction of humic acid with high aromaticity level, which has been known to be the major foulants during membrane filtration, were removed by the FMIEX. The process using the new FMIEX material has been shown to be very effective for the NOM removal. The FMIEX pretreatment combined with membrane filtration technology can be a potential tool for advanced drinking water treatment.

-89-

REFERENCES [1]

T. Carroll, S. King, S. R. Gray, B. A. Bolto, and N. A. Booker, The fouling of microfiltration membranes by NOM after coagulation treatment, Wat. Res., Vol. 34, No. 11 (2000), pp. 2861-2868

[2]

A.I. Schäfer, A.G. Fane, T.D. Waite, Fouling effects on rejection in the membrane filtration of natural waters, Desalination, 131 (2000), pp.215-224

[3]

Cheng-fang Lin, Tze-yao Lin, Oliver J. Hao, Effects of humic substance characteristics on UF performance, Wat. Res., Vol. 34, No. 4 (2000), pp. 1097-1106

[4]

Jaeweon Cho, Gary Amy, John Pellegrino, Membrane filtration of Natural Organic Matter: initial comparison of rejection and flux decline characteristics with ultrafiltration and nanofiltration membranes, Wat. Res., Vol.33, No.11 (1999), pp. 2517-2526

[5]

Mohamed Siddiqui, Gary Amy, Joseph Ryan, and Wilbert Odem, Membranes for the control of Natural Organic Matter from surface waters, Wat. Res., Vol.34, No.13 (2000), pp. 3355-3370

[6]

Wei Yuan, and Andrew L. Zydney, Humic acid fouling during ultrafiltration, Environ. Sci. Technol., 34 (2000), pp.5043-5050

[7]

Linhua Fan, John L. Harris, Felicity A. Roddick, and Nic A. Booker, Influence of the characteristics of Natural Organic Matter on the fouling of

-90-

microfiltration membranes, Wat. Res., Vol.35, No.18 (2001), pp.4455-4463 [8]

Kerry J. Howe, and Mark M. Clark, Fouling of microfiltration and ultrafiltration membranes by natural waters, Environ. Sci. Technol., 36 (2002), pp.3571-3576

[9]

Wataru Tsujimoto, Haruo Kimura, Tomohiro Iztl, Takashi Ilie, Membrane filtration and pre-treatment by GAC, Desalination, 119 (1998), pp.323-326

[10]

C. Guigui, J.C. Roucha, L. Durand-Bourlierb, V. Bonnelyeb, P. Aptel, Impact of coagulation conditions on the in-line coagulation/ultrafiltration process for drinking water production, Desalination, 147 (2002), pp.95-100

[11]

S. J. Judd and P. Hillis, Optimisation of combined coagulation and microfiltration for water treatment, Wat. Res., Vol. 35, No. 12 (2001), pp.2895-2904

[12]

Thomas F. Speth, Activated carbon and membrane processes for Disinfection By-Product (DBP) and microbial control, Controlling Disinfection By-Products and microbial contaminants in drinking water, EPA600R01110 (2001)

[13]

Gary Amy, Jaeweon Cho, Interaction between Natural Organic Matter (NOM) and membrane: rejection and fouling, Wat. Sci. Tech., Vol.40, No.9 (1999), pp.131-139

[14]

G. Crozes, C. Anselme and J. Mallevialle, Effect of adsorption of organic matter on fouling of ultrafiltration membranes, Journal of Membrane

-91-

Science, 84 (1993), pp.61-77 [15]

Cheng-fang Lin, Yuh-jay Huang, Oliver J. Hao, Ultrafiltration processes for removing humic substances: effect of molecular weight fractions and PAC treatment, Wat. Res., Vol.33, No.5 (1999), pp.1252-1264

[16]

Morran JY, Bursill DB, Drikas M, Nguyen HV, A new technique for the removal of Natural Organic Matter, AWWA Watertec Convention, Sydney, (1996)

[17]

Yonghoon Lee, Joonho Rho, Bumsuk Jung, Preparation of magnetic ionexchange resins by the suspension polymerization of styrene with magnetite, Journal of Applied Polymer Science, Vol. 89 (2003), pp.2058-2067

[18]

Membrane filtration guidance manual, EPA815-D-03-008, Proposal draft, EPA (2003)

[19]

Munir Cheryan, Ultrafiltration and microfiltration, Handbook, Technomic Publishing, (1998)

[20]

I.H. Suffet and Partrick Mac Carthy (edited), Aquatic humic substances Influence on fate and treatment of pollutants, American Chemical Society (1999), pp.93-106

[21]

T. David Waite, Andrea I. Schäfer Anthony, G. Fane, and Axel Heuer, Colloidal fouling of ultrafiltration membranes: Impact of aggregate structure and size, Journal of Colloid and Interface Science, 212 (1999), pp.264–274

-92-

[22]

Jerry A. Leenheer, Jean-Philip Pecroué, Characterizing of dissolved organic matter, Environmental Science and Technology, American Chemical Society (2003), pp.19A-26A

[23]

E. Aoustin, A.I. Schäfer, A.G. Fane, T.D. Waite, Ultrafiltration of natural organic matter, Separation and Purification Technology, 22-23 (2001), pp.63–78

[24]

Masahide Taniguchi, James E. Kilduff, and Georges Belfort, Modes of natural organic matter fouling during ultrafiltration, Environ. Sci. Technol., 37 (2003), pp.1676-1683

[25]

Wei Yuan, Andrew L. Zydney, Humic acid fouling during microfiltration, Journal of Membrane Science, 157 (1999), pp.1-12

[26]

Wei Yuan, Andrew L. Zydney, Effects of solution environment on humic acid fouling during microfiltration, Desalination, 122 (1999), pp.63-76

[27]

P.R. Bérubé, D.S. Mavinic, E.R. Hall, S.E. Kenway, and K. Roett, Evaluation of adsorption and coagulation as membrane pretreatment steps for the removal of organic material and disinfection-by-product precursors, J. Environ. Eng. Sci., 1 (2002), pp.465–476

[28]

Harsha Ratnaweera, Nicole Hiller, Ulrike Bunse, Comparison of the coagulation behavior of different Norwegian aquatic NOM sources, Environment International, Vol.25, No. 2-3 (1999), pp.347-355

[29]

Jeong-dae Lee, Sang-ho Lee, Min-ho Jo, Pyung-kyung Park, Chung-hak

-93-

Lee, Jong-won Kwak, Effect of coagulation conditions on membrane filtration characteristics in coagulation-microfiltration process for water treatment, Environ. Sci. Technol., 34 (2000), pp.3780-3788 [30]

Miaomiao Zhang, Chun Li, Mark M. Benjamin, and Yujung Chang, Fouling and natural organic matter removal in adsorbent/membrane systems for drinking water treatment, Environ. Sci. Technol., 37 (2003), pp.1663-1669

[31]

Martin Jekel, A new process for NOM removal with Granular Ferric Hydroxide (GFH) and oxidative regeneration, Utilization of NOM characteristics to improve process selection and performance, Workshop, Berlin, October 10-12, (2001)

[32]

Elizabeth Hamm, Michael Bourke, Application of magnetized anion exchange resin for removal of DOC at the Coldiron Watkins Memorial water treatment plant in Danville, KY, Presented at the AWWA Water Quality Technology Conference, 11-15 November, Nashville, TN, (2001)

[33]

I.H. Suffet and Partrick Mac Carthy (edited), Aquatic humic substances Influence on fate and treatment of pollutants, ACS (1999), pp.532-548

[34]

S.G. Yiantsios and A.J. Karabelas, An experimental study of humid acid and powdered activated carbon deposition on UF membranes and their removal by backwashing, Desalination, 140 (2001), pp.195-209

[35]

G. Newcombe, and M. Drikas, Adsorption of NOM onto activated carbon:

-94-

Electrostatic and non-electrostatic effects, Carbon, Vol. 35, No. 9 (1997), pp.1239-1250 [36]

Slunjski M, Nguyen HV, Ballard M, Eldridge R, Morran J, Drikas M, O'Leary B, Smith P, MIEX - Good research commercialised, Water, Vol.29, No.2 (2002), Melbourne, Australia, pp.42-51

[37]

Richard Collins, South Australia - Great leap forward in water treatment, Waste Management & Environment (WME), Vol.13, No.2 (2002), Sydney, Australia, pp.50

[38]

Morran JY, Bursill DP, Drikas M, Nguyen HV, A simple method to reduce disinfection by-product formation, 17th Federal Convention, Australian Water & Wastewater Association, Proceedings 16-21 March (1997), Melbourne Australia, pp.373-379

[39]

Cadee K, O'Leary B, Smith P, Slunjski M, Bourke M, Worlds first magnetic ion exchange water treatment plant to be installed in Western Australia, American Water Works Association Conference, Proceedings 11- 15 June (2000), Denver, USA

[40]

Semmens M, Burckhardt M, Schuler D, Davich P, Slunjski M, Bourke M, Nguyen H, An evaluation of Magnetic Ion Exchange (MIEX) for NOM removal, American Water Works Association Conference, Proceedings 1115 June (2000), Denver, USA

[41]

Michael Bourke, Marin Slunjski, Bernie O'Leary, Paul Smith, Scale-up of

-95-

the MIEX-DOC process for full scale water treatment plants, 18th Federal Convention, Australian Water & Wastewater Association, Proceedings 1114 April (1999), Adelaide Australia [42]

Nguyen HV, Slunjski M, Bourke MF & Drikas M, DOC removal by MIEX process - scaling-up and other development issues, 17th Federal Convention, Australian Water & Wastewater Association, Proceedings 1621 March (1997), Melbourne Australia, pp.373-379

[43]

Marin Slunjski, Michael Bourke, Bernie O'Leary, MIEX-DOC process for removal of humics in water treatment, Australian Chapter Seminar, 18 February (2000)

[44]

Pelekani C, Drikas M, Bursill D, Application of direct filtration to MIEX treated river Murray water, Proceedings of 19th Federal AWA Convention, April 1-4 (2001), Canberra, Australia

[45]

Chow C, Cook D, Drikas M, Laboratory study of conventional alum treatment versus MIEX treatment for removal of natural organic matter, Proceedings of 19th Federal AWA Convention, April 1-4 (2001), Canberra, Australia

[46]

Douglas Hammann, Michael Bourke, and Cory Topham, Evaluation of a magnetic ion exchange resin to meet DPB regs at the village of Palm Springs, Field report, Journal AWWA (2004), pp. 46-50

[47]

Bourke M, Harrison S, Long B, Lebeau T, MIEX resin pretreatment

-96-

followed by microfiltration as an alternative to nanofiltration for DBP precursor removal, Proceedings of AWWA on Membrane Technology Conference (2001) [48]

Slunjski M, Cadee K, O'Leary B, Tattersall J, MIEX resin water treatment process, Aquatech Amsterdam 2000, Proceedings 26-29 September (2000), Amsterdam, The Netherlands

[49]

Gregory V. Korshin, Chi-wang Li, Aark M. Benjamin, Monitoring the properties of natural organic matter through UV spectroscopy: A consistent theory, Wat. Res., Vol.31, No.7 (1997), pp.1787-1795

[50]

Jörg Dilling, Klaus Kaiser, Estimation of hydrophobic fraction of dissolved organic matter in water samples using UV photometry, Wat. Res., Vol.36 (2002), pp.5037-5044

[51]

Marin Slunjski, Michael Bourke, Hung Nguyen, Mat Ballard, Jim Morran, Don Bursill, MIEX-DOC process - A new ion exchange process, 18th Federal Convention, Australian Water & Wastewater Association, Proceedings 11-14 April (1999), Adelaide Australia.

[52]

Domenic Grasso, Yu-ping Chin, Water J. Weber, Structure and behaviour characteristics of a commercial humic acid and natural dissolved aquatic organic matter, Chemosphere, Vol.21, No.10-11 (1990), pp.1181-1197

[53]

Gary Amy, Nam Guk Her, and Chalor Jarusutthirak, Natural Organic Matter

(NOM)

fouling

of

pressure-driven

-97-

membranes:

Foulant

identification by size exclusion chromatography (SEC) and Fourier transform infrared (FTIR) spectroscopy, Utilization of NOM characteristics to improve process selection and performance, Workshop, Berlin, October 10-12 (2001) [54]

J. P. Croué, D. Violleau, C. Bodaire, and Legube, Removal of hydrophobic and hydrophilic constituents by anion exchange resin, Wat. Sci. Tech., Vol.40, No.9, pp.207-214, (1999)

[55]

Standard methods for the examination of water and wastewater, APHA, AWWA and WEF, 20th edition (1998)

-98-