1. Introduction
1.2 Outlines
In this study, the impact of sludge characteristics on membrane fouling in a submerged MBR was investigated. The cause for membrane fouling and the fouling
operated at SRT 10, 30 and 60 days to evaluate the effect of SRT on sludge characteristics and membrane fouling. Sludge properties at three SRTs, including MLSS, specific cake resistance, EPS, SMP and particle size distribution, were investigated to evaluate their impact on membrane fouling.
For fouling mitigation, a TiO2 composite membrane was prepared. A neutral sol containing 1% of TiO2 nanoparticles was prepared by chemical copreciptization-peptization. Membranes were dip-coated with the neutral sol to form the TiO2 composite membranes which were characterized by X-ray photoelectron spectroscopy (XPS). The fouling mitigations of the TiO2 composite membranes made from acidic TiO2 suspension and neutral TiO2 sol were compared. The fouling mitigation was explored on two kinds of membranes, cellulose acetate (CA) and mixed cellulose ester (MCE). Finally, ultrasonic washing was performed to evaluate the stability of TiO2 composite membranes. The schematic diagram of this study is illustrated in Figure 1.1.
Chapter 2 reviews literatures related to this study. It includes membrane fouling and mitigation of fouling in MBRs. For membrane fouling in MBRs, the effects of membrane characteristics, sludge characteristics, and environmental and operational conditions on membrane fouling are examined. For mitigation of membrane fouling in MBRs, modification of membrane characteristics is reviewed.
Chapter 3 describes the material and analytical methods used in this study.
Chapter 4 and 5 show the experimental results and discussion. In chapter 4, performance and fouling characteristics of MBR under different sludge characteristics (filamentous and floc-forming sludge) are presented along with the effect of SRT on sludge characteristics and membrane fouling.
Chapter 5 states the results concerning antifouling ability of TiO2 composite membranes for MBRs. The preparation and characterization of TiO2 composite membranes as well as the stability of TiO2 nanoparticles on composite membranes are included.
Finally, the conclusions and recommendation are provided in Chapter 6.
Figure 1.1. Schematic diagram of this study.
Chapter 2 Literature review
Membrane bioreactors (MBRs) have attracted a lot of attention in the last decade because of their advantages over conventional activated sludge processes.
However, membrane fouling resulting in increase of operation and maintenance costs has become one of the most significant factors hindering the widespread application of MBRs. In this study, references considering membrane fouling in MBRs and the strategies to mitigate membrane fouling in MBRs are reviewed.
2.1 Membrane fouling in MBRs
Membrane fouling in MBRs is due to the deposition of materials into/onto the membrane, which is attributed to the interactions of membranes and activated sludge.
Membrane fouling depends on nature of feed, sludge characteristics, operation conditions, etc. Although many investigations concerning fouling in MBRs have been performed, inconsistent results are usually observed in studies (as shown in Table 2.1) due to the complication of mixed liquor and influent, different modules and different operation conditions implemented in studies. Table 2.1 shows that most studies identified extracellular polymeric substances (EPS), especially the carbohydrates fraction, as the main foulant in MBRs (Cho & Fane, 2002; Kimura et al., 2005; Rosenberger et al., 2006)
. However, as shown in Table 2.1, some studies pointed out that other materials like biopolymer cluster, smaller particles or fatty acids were the major foulants on membranes (Bai & Leow, 2002; Wang & Li, 2008; Al-Halbouni et al., 2009)
.
According to Chang et al (2002), membrane, biomass, and operating conditions are three key factors influencing membrane fouling in MBRs. Recently, Liao et al (2004) also proposed four main parameters which have significant impact on membrane fouling in MBRs (as shown in Figure 2.1). Although some of these parameters in Figure 2.1 directly influence membrane fouling, others interact with other parameters and subsequently influence membrane fouling indirectly. Due to the complex interaction of parameters affecting membrane fouling, a comprehensive understanding and investigation of membrane fouling in MBRs should be provided to properly control fouling in MBRs. Some important factors affecting membrane fouling in MBRs are reviewed as follows.
Table 2.1. Membrane fouling in MBRs.
Influent Pore size (µm)
Membrane
configuration Major foulants Test
duration Other References
UASB effluent 0.22 Flat sheet,
Hydrophilised PVDF
wastewater 0.1-0.2 Hollow fiber Polysaccharides
fraction 1 year Pre-denitrification and post-denitrification in MBR
2 years Pre-denitrification in MBR Lyko et al.
(2007)
Table 2.1. Membrane fouling in MBRs (continued) Influent Pore size
(µm) Membrane configuration Major foulants Test duration Other References Municipal
wastewater 0.2 Hollow fiber,
PVDF (Zizheng, China)
Not only EPS but also other organic
substances
160 days Pre-denitrification in MBR
Wang et al.
(2008) Synthetic
wastewater 0.4 Hollow fiber,
(Mitsubishi Rayon) Biopolymer clusters 28 days 18-min-operation and 2-min-off
Wang & Li (2008) Municipal
wastewater 0.04 Hollow fiber,
PVDF (Zenon) Fatty acids 3 years - Al-Halbouni et al.
(2009)
Figure 2.1. Factors affecting membrane fouling in MBRs. (adapted from Liao et al., 2004)
Membrane Fouling
Sludge characteristics (particle size, extracellular polymeric
substance, MLSS, hydrophobicity and surface charge, floc
morphology)
Membrane characteristics (hydrophobicity,
surface charge, roughness, pore size) Fluid Mechanics
(Shear stress on membrane surface)
Membrane module design
(fiber packing density, space distance,
aerators)
Environmental/Operating Conditions (pH, ionic strength, growth rate, nutrient, hydraulic residence, Food/Microorganism, solids retention time, membrane flux, operating cycle length)
2.1.1 Membrane characteristics
Membrane characteristics, such as roughness, pore size, porosity, surface charge, and hydrophobicity have impact on membrane fouling.
2.1.1.1 Roughness thin-film composite reverse osmosis membranes (RO), and concluded that surface roughness increased membrane fouling by increasing the rate of colloid attachment onto the membrane surface. In the consequent study, similar results were discovered in RO and nanofiltration (NF) membranes that rough membranes had higher fouling propensity because particles preferentially accumulated in the “valleys” of rough membranes, resulting in “valley clogging” which caused more severe flux decline
(Vrijenhoek et al., 2001)
. He et al (2005) used a series of polyethersulfone (PES) UF membranes of various molecular weight cutoff (MWCO) and roughness to evaluate the effect of MWCO and membrane roughness on flux decline in an anaerobic MBR.
They found that smoother membranes had less permeate flux decline because foulants found fewer crevices to fill in and to buildup the fouling layer. Recently, Choi & Ng (2008) also concluded that in a submerged MBR phase-inversed polytetrafluorethylene (PTFE) membranes had higher total filtration resistance than track-etched polycarbonate (PCTE) and track-etched polyester (PETE) membranes due to its higher degree of roughness.
2.1.1.2 Hydrophobicity and hydrophilicity
For membrane, hydrophilicity affects its wettability and the pressure to drive the liquid through the membrane. It also influences the adhesion characteristics of contaminants to the membrane materials. In most cases, membranes with hydrophobic characteristics have been found more prone to membrane fouling because of the hydrophobic-hydrophobic interaction between solutes, microbial cells and membrane materials (Madaeni et al., 1999; Choi et al., 2002; Yu et al., 2005)
. However, Maximous et al (2009) observed that hydrophilic membranes did not benefit membrane flux although hydrophilic membranes did show better reversibility in cake resistance. This result is consistent with the finding observed by Parsmore et al (2002) that young biofilms attached on the hydrophilic membranes are more facile to be removed than hydrophobic
membranes. Nevertheless, it is difficult to show the correlation between membrane hydrophilicity and fouling, because membrane hydrophilicity usually accompanied by the changing of other membrane characteristics such as roughness. For example, Zhang et al (2008), reported that PES membranes fouled more easily than polyacrylonitrile (PAN) and polyvinylidene fluoride (PVDF) membranes due to higher degree of roughness and hydrophilicity.
2.1.1.3 Pore structure and size
Pore structure of membranes is also one of the important membrane characteristics affecting membrane fouling. Ho & Zydney (1999) filtered protein with various membranes such as track-etched, isotropic, and asymmetric microfiltration (MF) membranes. The result showed that membranes with interconnected pores fouled more slowly since the fluid could flow around the blocked pores. Hwang & Lin (2002) have shown that membranes with cylindrical pore structures had better filtration than those with sponge pore structures for filtration of suspension containing polymethyl methacrylate (PMMA) spherical particles. Fang & Shi (2005) also demonstrated that PES membrane which has the sponge-like microstructure with large pore openings had the highest pore resistance for filtration of activated sludge due to its large pore openings (18-20 µm). As a result, MBR system should use cake-resistance-dominant membranes, such as track-etched polycarbonate (PC), PVDF and mixed cellulose esters (MCE) rather than pore-resistance-dominant membranes like PES.
Choo & Lee (1996) proposed that MF membranes with 0.1 µm had minimal fouling tendency compared to membranes with 0.02, 0.5, and 1 µm. However, researchers have found that membrane pore size have no significant impact on critical flux (Madaeni et al., 1999; Le-Clech et al., 2003b)
. Critical flux was affected only when membranes with small pore size and in low concentration of mixed liquor suspended solids. As a result, the effect of membrane pore size on membrane fouling may depend on both particle size of sludge and membrane.
2.1.1.4 Surface charge
Surface charge of the membrane is critical to membrane fouling because the interaction between organic compounds and membranes depends on membrane surface charge. Surface charge of the membrane can attract or repel charged species in water. It is generally accepted that negatively charged membrane is preferable in
greatly affect the adsorption of effluent organic matters (Jarusutthirak & Amy, 2001)
. Pasmore et al (2002) reported that biofilm attached on the membrane surface was easier to be removed when membrane surfaces bear neutral or slight negative charges.
2.1.2 Sludge characteristics
Sludge characteristics of MBRs directly relate to the feed characteristics and the environmental/operational conditions as illustrated in Figure 2.1. Different feed characteristics and the environmental/operating conditions result in different sludge characteristics. Membrane fouling is due to the interaction between sludge and membranes, and, therefore, sludge characteristics play a vital role in membrane fouling in MBRs.
2.1.2.1 Particle size
Sludge was generally divided into three components, suspended solids, colloids and solutes, by size to investigate their contributions on membrane fouling in MBRs.
Table 2.2 compares the results concerning the contribution of sludge components on membrane fouling in various studies. Inconsistent results were observed in these published studies. Wisniewski & Grasmick (1998) showed that solutes were the main causes for membrane fouling. Others showed that suspended solids were the main contributor for membrane fouling (Defrance et al., 2000; Lee et al., 2003; Bae & Tak, 2005a)
. Bouhabila et al (2001), on the other hand, found that colloids were the major component for membrane fouling. These contradictory findings may be caused by the differentce in operational conditions, sludge properties, methods of sludge separation and others.
In order to mitigate membrane fouling in MBRs, shear stresses along membranes are created by aeration in submerged MBRs or tangential flows in side-stream MBRs.
Therefore, activated sludge usually contains larger flocs, while MBR sludge contains primarily small flocs due to higher shear stress. High shear stress could break and sludge flocs and reduce their size, resulting in serious fouling (Wisniewski & Grasmick, 1998;
Cicek et al., 1999)
. Chang & Kim (2005) compared the filtration performance of a submerged MBR and a tertiary treatment plant with membrane unit. They found that the tertiary treatment plant with membrane unit had worse filtration performance due to the small particle size in the secondary effluent. Bae & Tak (2005a) further fractionated the MBR sludge into solutes, colloids and suspended solids, and concluded that the fouling contribution of each sludge fraction was strongly related to particle size because both permeation drag and back transport velocity are particle size-related functions. Therefore, it is important to control the operational conditions to avoid breaking sludge flocs and simultaneously provide suitable shear stress for membranes filtration.
Table 2.2. Relative Contribution of sludge components on membrane fouling (%)
Fraction Wisniewski &
Grasmick (1998)
Defrance et al.
(2000)
Bouhabila et al.
(2001) Lee et al. (2003) Bae & Tak (2005a)
Suspended solids 24 65 24 63-72 72-83
Colloids 24 30 50 4-14
Solutes 52 5 26
28-37a
13-14
a supernatant
2.1.2.2 Extracellular polymeric substances
As mentioned in 2.1, EPS is the most commonly discussed foulant in studies.
Extracellular polymeric substances is a complex mixture of macromolecular polyelectrolytes including polysaccharides, protein, nucleic acids, and humic compounds. It is generally subdivided into two categories: (1) bound or extractable EPS (sheaths, capsular polymers, condensed gel, loosely bound polymers, and attached organic material) and (2) soluble EPS (soluble macromolecules, colloids, and slimes) (Laspidou & Rittmann, 2002; Rosenberger & Kraume, 2002)
. According to Laspidou & Rittmann (2002) soluble EPS is the same as soluble microbial products (SMP) which is defined as the pool of organic compounds that are released into solution from substrate metabolism and biomass decay (Barker & Stuckey, 1999)
. However, Ramesh et al (2006) compared the physiochemical characteristics of soluble EPS and SMP and concluded that soluble EPS and SMP were not identical. Furthermore, since there is no standard extraction method of EPS, contradicting results were found in studies concerning fouling caused by EPS and SMP, which were listed in Table 2.3. Although different foulants were identified, recent studied have suggested that soluble fraction (soluble EPS and SMP), especially the carbohydrates, is the major foulants in MBRs (Table 2.3).
Not only the quantity of EPS or SMP, but also the composition of EPS or SMP is critical for membrane fouling in MBRs (Mukai et al., 2000)
. Carbohydrates and proteins are the most abundant components in EPS or SMP. Therefore, studies have focused on the effect of protein and carbohydrate ratio (protein/carbohydrate) on membrane fouling.
Preferential adsorption of solutes and sludge particles on membranes was observed when protein/carbohydrate ratios were high by Ji & Zhou (2006) and Choi et al (2009). They found that gradually increase the protein/carbohydrate ratio in bound EPS was likely to increase fouling propensity. Moreover, Kim & Nakhla (2009) reported that the higher protein/carbohydrate ratio in SMP was related with higher fouling rate. In summary, higher protein/carbohydrate ratio is associated with membrane fouling, whether they are SMP or bound EPS.
Table 2.3. EPS and SMP on membrane fouling.
Influent Membrane module Details Major foulants identified References
Synthetic wastewater Loop-type hollow fiber;
polyethylene; 0.1 µm - Bound EPS Nagaoka et al.
Synthetic wastewater U-shaped hollow fiber;
PS, 0.1 µm
Organic loading: 0.3-0.4 g
COD/g MLSS/d Soluble EPS Hernandez Rojas et al.
(2005) Municipal effluent,
municipal wastewater and Industrial wastewater
MF and UF Six pilot cases in Europe Polysaccharides, a part of bacterial EPS
Rosenberger et al.
(2005) SRT: 8, 20 and 80 days;
Table 2.3. EPS and SMP on membrane fouling (continued)
Influent Membrane module Details Major foulants
identified References Synthetic wastewater Hollow fiber; 0.1µm Constant pressure operation Extractable EPS Meng et al.
(2006c)
Synthetic wastewater Flat sheet; hydrophilic polypropylene; 0.2 µm
F/M: 0.13-0.21 kg COD/kg MLSS/d;
Organic loading:
Municipal wastewater Hollow fiber; 0.035 µm Constant flux operation SMP Trussell et al.
(2006) Municipal wastewater Hollow fiber;
PVDF; 0.04 µm
SRT: 12 days;HRT: 7 and 10 h; Organic
loading: 0.88 and 1.03 kg COD/m3/d Soluble EPS Geng & Hall (2007)
Table 2.3. EPS and SMP on membrane fouling (continued)
Influent Membrane module Details Major foulants identified References
Synthetic wastewater Flat sheet; polyethylene; 0.4 µm
HRT: 12 h;
Organic loading:
0.36 and 0.9 kg TOC/m3/d
SMP Jeong et al.
(2007)
Synthetic wastewater and
domestic wastewater Flat sheet; PAN and PES; UF
HRT: 12, 13 and 17 h;
SRT: 22-31 days;
F/M:
0.11-0.66 g COD/g VSS/d
SMP but only under certain conditions such as larger particle size and low sludge age
Drews et al.
(2008)
Synthetic wastewater Flat sheet; polypropylene; 0.1µm
Sequencing batch MBR;
HRT: 12 h;
SRT: 10, 20, 40 and 60 days Carbohydrate in SMP Dong & Jiang (2009)
2.1.2.3 Mixed liquor suspended solids
Concentration of mixed liquor suspended solids (MLSS) is considered to impact directly upon cake layer formation on the membrane surface. The deposited cake layer on the membrane surface can play an important role in flux decline or trans-membrane pressure (TMP) increase, although contradict results were observed in studies. Table 2.4 shows the effect of MLSS on membrane fouling. Nagaoka et al (1996a) found that higher MLSS in MBRs resulted in higher sludge viscosity which caused serious membrane fouling. The finding was in consistent with the result observed by Madaeni et al (1999). On the other hand, some studies reported that higher MLSS resulted in better filtration performance (Lee et al., 2001)
. Others have concluded that there was no significant impact of MLSS on membrane fouling or the impact depended on the concentration of MLSS (Defrance et al., 2000; Hong et al., 2002; Rosenberger &
Kraume, 2002; Le-Clech et al., 2003b; Fan et al., 2006)
. This may be explained by the complex sludge composition. For example, sludge taken from different MBRs may possess different sludge properties such as particle size distribution and EPS, which are important in membrane fouling. Moreover, some studies evaluated the impact of MLSS on membrane fouling by settling sludge or diluting with saline solution, which neglected the non-settable fraction of sludge and small particles in sludge. Hence, the relationship between MLSS and membrane fouling is difficult to establish without considering other important factors.
Table 2.4. Relationship between MLSS and membrane fouling in MBRs.
Influent Membrane module MLSS
(mg/L) Effect of MLSS on fouling Details References
Synthetic wastewater Polyethylene, 0.1 µm,
loop-type hollow fiber -
Increasing MLSS resulted in decreasing filterability due to higher sludge viscosity
- Nagaoka et al.
(1996a)
- HVLP membrane (MIllipore),
0.45 µm 0 to 10,000 Critical flux was lower for higher concentration
Testing with batch a crossflow cell
Madaeni et al.
(1999)
Domestic water (75%) and industrial wastewater (25%)
Ceramic membrane, 0.1 µm 2,000 to 6,000 No significant effect
Adjusted settling time to obtain required SS
Defrance et al.
(2000)
Table 2.4. Relationship between MLSS and membrane fouling in MBRs (continued)
Influent Membrane module MLSS
(mg/L)
Effect of MLSS on
fouling Details References
Synthetic wastewater Polyethylene, 0.1 µm, u-shaped hollow fiber
24,000 No effect - Rosenberger & Kraume
(2002)
Synthetic wastewater PS membrane, 0.1 µm, hollow fiber
3,600 6,800 8,400
No effect was observed - Hong et al. (2002)
Table 2.4. Relationship between MLSS and membrane fouling in MBRs (continued)
Influent Membrane module
MLSS
(mg/L) Effect of MLSS on fouling Details References
-
Critical flux had no difference from 4 to 8 g/L, but critical flux significantly increased at 12 g/L
Changing SRT to obtain required MLSS
Specific cake resistance increased with MLSS decreased; Rc decreased with MLSS decreased
Adjusted settling time and diluted with saline solution to obtain required MLSS
Little impact was observed on critical
flux - Fan et al.
(2006)
2.1.2.4 Hydrophobicity and surface charge
Other sludge characteristics like hydrophobicity and surface charge of sludge also have significant impact on membrane fouling. Lee et al (2003) found that hydrophobicity and surface charge of flocs were closely related to composition and properties of EPS and were significantly related to composition and properties of EPS and could be used to estimate resistance caused by flocs. Arabi & Nakhla (2009) have reported that higher EPS concentration and relative hydrophobicity of flocs in nitrification and denitrification MBR resulted in increasing of cake resistance. Meng et al (2006b) concluded that higher hydrophobicity of flocs resulted from excess growth of filamentous bacteria would cause serious membrane fouling in MBRs. Ahn et al (2007) also found that severe fouling was due to the increased EPS and hydrophobicity of supernatant when using an anaerobic upflow bed filter combined with an aerobic MBR to treat high strength organic and nitrogen wastewater.
2.1.2.5 Floc morphology
By theory, all the biomass in MBRs can be retained by the membrane unit to maintain an excellent effluent quality regardless of the sludge seattleability. Hence, sludge bulking should not be a problem in MBRs. However, Chang & Lee (1998) reported that foaming sludge showed greater flux decline than the non-foaming sludge.
Later, Chang et al (1999) also reported that bulking sludge had higher fouling tendency than normal sludge and pinpoint sludge. Among these three sludge morphologies, normal sludge had the least fouling tendency. More recently, studies have focused on the effect of filamentous bacteria on membrane fouling in MBRs. They have demonstrated that bulking sludge caused by overgrowth of filamentous bacteria resulted in deterioration of MBR performance (Meng et al., 2006a, 2006b, 2007)
. Meng et al (2006b) reported that the excess growth of filamentous bacteria formed a non-porous cake layer on the membrane surface which interfered with the membrane filtration. Meng
. Meng et al (2006b) reported that the excess growth of filamentous bacteria formed a non-porous cake layer on the membrane surface which interfered with the membrane filtration. Meng