Membrane bioreactor (MBR)
Membrane filtration, widely used in chemical and biotechnology processes, is already established as a valuable means of filtering and cleaning wastewater. Membrane filtration denotes the separation process in which a membrane acts as a barrier between two phases (Fig. 1.1). The membrane separates on the basis of particle size. It retains constituents bigger than the pore size. According to the pore size of the membrane, the filtration process can be classified as microfiltration (MF), ultrafiltration (UF), nano-filtration (NF) or reverse osmosis (RO). Figure 1.2 introduces the overview of existing membrane types for water treatment and their corresponding rejection behavior of common pollutants. As shown in Fig. 1.2 the microfiltration membrane can reject the bacteria, while the ultrafiltration membrane may also retain the viruses. In addition to the bacteria and viruses, the nano-filtration membrane can also reject the multivalent ions. Reverse osmosis membrane even reject both of the monovalent ions and multivalent ions.
The MBR technology mostly employs microfiltration (MF) and ultrafiltration (UF). The membrane can be configured in membrane modules in different ways, such as hollow fiber membranes (outside-in), tubular membranes (inside-out) and flat-sheets (plate and frame membranes). The membrane configurations were commonly used hollow-fiber and plate and frame modules. The driving force for permeation is a transmembrane pressure (TMP) in most water treatment membrane filtration applications. The concentration polarization can decrease the membrane flux (increase resistance) for instance in ultrafiltration as it can lead to gel-layer formation causing additional resistance to filtration. In NF or RO concentration polarization can lead to decreased salt rejections or scaling at the membrane surface. In addition, concentration polarization causes increased osmotic pressures. The rate of permeate flux (J) may be described by Darcy’s law as follows:
J = TMP = TMP (1.1)
J: permeation flux [L.m-2.h-1] or [m.s-1]
TMP: transmembrane pressure [Pa], or [bar]
µ: dynamic permeate viscosity at 20oC [Pa.s]
Rt: total filtration resistance [m-1]
Rm: membrane resistance [m-1]
Rf: fouling resistance [m-1]
The permeability of the membrane (at T = 20oC) can be calculated as the ratio between the flux and the TMP, see eq. 1.2:
Permeability, LP = J [L.m-2.h-1.bar-1] (1.2)
The rejection R (%) of a membrane for a specific compound can be calculated by:
R = 1 × 100% (1.3)
Cper: permeate concentration (mol.L-1)
Cfeed: feed concentration (mol.L-1)
All membranes retain the solids that are present in the reactor, so all MBR permeates are solid-free. In order to reuse easily the MBR permeate, tertiary treatment is still necessary to remove recalcitrant compounds like viruses, endotoxins, pesticides, micropollutants or heavy metals.
Membrane fouling in MBRs
Membrane fouling is a process whereby the solution or the particle is deposited on the membrane surface or in the membrane pores such as in the membrane bioreactor, so that the performance of the membrane is degraded. The increase rate of TMP or the decline in permeate flux is an important factor to evaluate the system performance in submerged MBR because it is directly related to the rate of membrane fouling.
Fouling mechanisms in the MBR
Normally, the different fouling mechanisms can be described: (i) pore blocking, the particles enter the pore and get stuck in its opening, reducing the number of pore channels available for permeation; (ii) pore narrowing by adsorption, the substances or particles enter the pore and are adsorbed to the pore wall, thus narrowing the pore channel, reducing the permeate flow; (iii) cake layer formation, the particles and macromolecules accumulate at the membrane surface, forming a more or less permeable layer; (iv) the biofilm formed on the membrane surface in MBR consists of both microorganisms growing and microbial flocs deposited on the membrane, the formation of a biofilm on the membrane surface was mainly responsible for the loss of filterability, therefore, it became necessary to further investigate the properties of biofilms which are closely associated with membrane filterability. In both cases the fouling mechanism will lead to an increase in total filtration.
The various factors affecting the membrane fouling in the MBR could include three categories: membrane characteristics, operating conditions and biomass characteristics (see Table 1.1).
Classification of MBR fouling
Fouling of membranes is characterized in general as a reduction of permeability or as an increase of transmembrane pressure (TMP) during the MBR operation. Different fouling mechanisms can occur during the filtration of the membrane, including three phenomena: gel or cake layer formation (particles and macromolecules accumulate at the membrane surface, forming a more or less permeable layer); pore blocking (particles enter the pore and get stuck in its opening, reducing the number of pore channels available for permeation); and adsorption (substances and/or particles enter the pores and are adsorbed to the pore wall thus narrowing the pore channels, reducing the permeate flow). Fouling can be classified into three major categories: inorganic fouling which refers to the deposit of inorganic material like salts, clay and metal oxides; organic fouling which includes all kind of deposit of organic material like grease, oil, surfactants, proteins, polysaccharides, humic substances and other organic biopolymers and biofouling which designates the formation of biofilms by compounds and microorganisms attached and growing at the membrane surface.
Biofouling is defined as undesirable accumulation of microorganisms, which may occur by deposition, growth and metabolism of bacteria cells or flocs on the membrane . In addition, the microbial community structure and the characteristics of individual microbial groups could play an important role in biofouling . Biofouling mechanisms include (i) the adsorption of soluble and suspended extracellular polymers on membrane surfaces and in membrane pores; (ii) the clogging of membrane pore structure by fine colloidal particles and cell debris; (iii) the adhesion and deposition of sludge cake on membrane surfaces . The biofouling layer formed on the membrane in an MBR mostly consists of microbial flocs deposited from the bulk phase. These microbial flocs are very heterogeneous, irregular structures consisting of microorganisms, organic and inorganic adsorbed particles, extra polymeric substances (EPS) and organic fibers [29,59,73]. Based on a few studies, biofouling can have several effects on the membrane systems. For examples, Yun et al. reported in the aerobic MBR, the accumulation of biomass with the growth of biofilm was the main factor for the loss of permeability, so, the TMP rises up continuously with the growth of biofilm on the membrane surface. Miura et al.  demonstrated that the biofilm development on hollow-fiber MF membrane surfaces caused severe irreversible fouling during a long-term operation of pilot-scale MBRs treating municipal wastewater. As the bio-cake accumulates on the surface of the membrane, endogenous decay or cell-lysis at the bottom layer would be expected to occur due to poor oxygen and substrate transfer. It gave rise to the excretion of EPS (polysaccharides and proteins) which could reduce the permeability of the MBR membrane .
Organic fouling by extracellular polymeric substances (EPSs) and soluble microbial products (SMPs)
Many researchers indicates that soluble microbial products (SMP) and extracellular polymeric substances (EPSs) excreted by microorganisms, produced from cell lysis and adsorbed organic matter from wastewater also play important roles in the formation of biological foulants and cake layer on the membrane surface [30,76,98]. The key components of EPS have been believed to include carbohydrates, proteins and humic-acid [22,98]. Extracellular polymeric substances (EPSs) are key biological substances, which can block membrane pores, adhere to membrane surface, affect cake structure and induce osmotic effect, and therefore affect significantly membrane fouling in the MBR . According to Gao et al. , the protein concentrations of extracellular polymeric substances in conjunction with the carbohydrates of soluble microbial products were the main factors that accelerate the membrane fouling. Furthermore, Campo et al.  demonstrate that the fouling rate of the MBR membrane is closely related to the protein contents of the mixed liquor expressed by ratio proteins/polysaccharides. Due to their hydrophobic nature, proteins tend to adhere to the surface of the flocs, effectively constituting EPSbound, and contributing to the irreversible cake deposition. Additionally, they also reported that an increase of EPSs in the MBR caused the de-flocculation of the sludge. Consequently, the specific cake layer resistance (α) is influenced by de-flocculation. In particular, the superficial cake deposition increased when the floc size decreased. Earlier study showed that EPS are recognized as the most direct and significant factor affecting biofouling in MBRs [105,112]. They indicated that the adsorption tests and atomic force microscopy observation confirmed that the EPS properties played an important role in membrane adsorption, eventually causing the different fouling behavior.
It is known that MBR fouling is related to the group of soluble microbial products (SMP), this group contains soluble and colloidal biopolymers, mostly polysaccharides or carbohydrates and proteins, the major organic constituents of sludge supernatant [65,76,97,102]. Generally, polysaccharides and proteins are assumed to be the major fractions [27,114] that may have a significant influence on the physicochemical properties of the fouling layer. The formation of SMPs is also dependent on several factors, e.g., the type of food source, diversity of microorganisms and adaptation time of MLSS to certain SRT [43, 56]. SMP could be divided into two categories: biomass associated products (BAP), which were formed by biomass decay and utilization associated products (UAP), which were produced directly substrate metabolism . More recent studies indicated that SMP, especially polysaccharides, adsorbed on the membrane surface, blocked membrane pores and formed a cake layer, resulted in a hydraulic resistance to permeate flux [79,83,95]. Besides, Tian et al  demonstrated that the protein, polysaccharides of BAP (biomass-associated products) concentrations increased throughout the experiment. The content of polysaccharides released by biomass was greater than that of proteins with the operation time increased. The polysaccharide concentrations were also effectively retained by the membrane, ranging from 26.6% to 46.8%. Therefore, the presences of polysaccharides and proteins in the supernatant would make serious contributions to membrane fouling. Recently, Dalmau et al.  demonstrate that the SMP concentration had an influence on the transmembrane pressure (TMP) in the MBR. The direct impact on TMP after an increase of SMP concentrations was observed in the MBR supernatant. Previous researches highlighted reported that the proteins concentration in SMP significant influenced on membrane fouling in terms of the specific resistances increase, when the value of proteins concentration in the supernatant changes from 30 to 100 mg.L-1, the value of specific resistance increased by a factor of 10 . Meanwhile, Chu and Li  and Yigit et al  showed that polysaccharide-like substances contributed to fouling more significantly than protein-like substances. Similarly result was found by Dvořák et al , they reported that carbohydrates were the majority component of SMP retained by the MBR membrane (about 60%). Therefore, it can be assumed that the concentration of carbohydrates in the MBR supernatant play an important role in membrane fouling.
On the other hand, the molecular weight (MW) distribution of SMP were measured in membrane bioreactor using membranes with nominal molecular weight cut-offs of 1, 10, 100 kDa (see Table 1.2) to investigate the impact of each of the fractions, i.e. >100, 10-100, 1-10, <1 kDa on fouling in MBRs . Arabi and Nakhla  showed that no correlation between fractions of >100 kDa and 10-10 kDa and fouling in MBRs was found. However, a strong correlation was observed between fouling and SMP concentration in the range of 10-100 kDa with higher fouling rate associated with higher concentration of 10-100 kDa fractions. The high molecular weight compounds play an important role in creating high resistance of the membrane, leading to a reduction of permeate flux. The fluorescence spectra of membrane foulants also exhibited two peaks of protein-like substances, confirming that proteins played an important role in membrane fouling .
Pan et al  reported that membrane fouling is related to the reduction of SMP larger than 30 kDa in the effluent, which is due to the retention of large SMP by the formed cake layer. Hydrophilic fraction is the dominant species in SMP, which accumulates in the mixed liquor, so hydrophilic carbohydrates are most likely the major foulants in the MBR. In addition, Li et al  reported that the 100-1000 kDa SMPs could be completely retained by MF/UF membrane or MBR membrane during sludge filtration, leading to formation of biofilm in bio-cake, which contributes to higher sludge fouling propensity and steady increase of MBR fouling. The 10-100 kDa protein-like SMPs could be partially retained by MF/UF membrane or MBR membrane during sludge filtration, leading to the reduction of the empty space of bio-cake and result in change of bio-cake structure, that could lead to higher sludge fouling propensity and higher fouling rate and TMP jump in MBR.
Furthermore, the composition and concentration of SMP is significantly affected by operating conditions. For example, the membrane fouling rate increased 20-fold (from 0.18 to 3.65 LMH.bar-1.day-1 as the organic loading rates (F/M) increased from 0.34 to 1.41 gCOD.gVSS-1.day-1 . When the sludge retention time increased from 10 to 32-37 days, the protein and polysaccharide concentrations decreased in the supernatant of the MBR . Stricot et al  reported that high shear stresses (72 Pa) induce a very high release of soluble microbial products, mainly as protein-like substances. This release could be induced by the breakage of internal high strength bonds (fragmentation and erosion). The concentrations of protein rose strongly from the second day (the sludge had passed through the pump 2000 times) and stabilized at around 600 mg L-1.
Conclusion, the group of soluble microbial products (SMP) could be the major fractions which contribute to the membrane fouling. The accumulated SMP not only inhibited the metabolic activity of microorganisms, but also greatly decreased the membrane permeability as a result of organic fouling.
c) Inorganic fouling
In general, the membrane fouling in the MBR was the inorganic fouling can form through two ways (see Figure 1.4): chemical precipitation and biological precipitation . Chemical precipitation occurs when the concentration of chemical species exceeds the saturation concentrations due to concentration polarization. Biological precipitation is another contribution to inorganic fouling. The biopolymers contain ionisable groups such as COO-, CO32-, SO42-, PO43-, OH-. According to Guo et al  and Iorhemen et al , inorganic compound can cause fouling when the precipitation occurs on the membrane due to hydrolysis and oxidation during filtration. Examples such substances include cations and anions such as Ca2+, Mg2+, Fe3+, Al3+, SO42-, PO43-, CO32-, OH-, etc. Thus, high presentation of salt concentration in the feed water induces greater inorganic scaling propensity and more serious colloidal fouling. Early research has shown that the inorganic precipitation coupled the organic foulants enhanced the formation of gel layer and thus caused membrane fouling in the MBR .
Recent studies have shown that an increase in salt concentration (from 0 to 20 g.L-1 NaCl) did not significantly change the dissolved organic carbon (DOC) removal efficiencies in the MBR (>95%) [40-41]. Besides, Luo et al [71-72] reported that the removal of total organic carbon (TOC) had decreased from 98% to 80% when the salinity was higher than 10 g.L-1 NaCl. This phenomenon could due to the inhibitory effect of high salinity on the biomass, resulting in cell plasmolysis and /or loss of metabolic activity. Nevertheless, the removal efficiency recovered to approximately 99% when salt concentration was maintained at 10 g.L-1 NaCl for two weeks. Small variations in TOC removal were also obtained (around 90 to 98%) during the increase of salinity loading from 10 to 16.5 g.L-1. In contrast, Johir et al  reported that DOC removal efficiency decreased from 72% to 35% when salt concentration increased from 1 g NaCl.L-1 to 10 g NaCl.L-1. The lower removal of dissolved organic carbon with high salt concentration could be due to the adverse effect of salt on microbial activity.
According to Di Bella et al , after the addition of 10g NaCl.L-1 in the inlet wastewater, it was observed a slight decrease of chemical oxygen demand (COD) removal efficiency (from 97% to 87%). The suggestion is that MBR can respond well to salt increase in term of organic matter removal in the MBR, as confirmed by previous study . Similarly, Arabi and Nakhla  and also found that no impact of salinity was observed on the COD removal rate in the MBR. Besides, Yogalakshmi and Joseph  resulted under sodium chloride shock of 5- 30 g.L-1, the COD removal efficiency dropped and was in the range of 83-87%.
On the other hand, salinity exhibited a significant negative impact on the nitrification process, ammonia removal efficiency decreased from 95 to 46%  and 87 to 46% . Meanwhile, Hong et al  reported NH4+ removal efficiency decreased no significantly from 97 to 91% when the salinity increased from 0 to 35 g.L-1; its removal rate was about 99% after one week of addition of 4 g.L-1 NaCl . It is of interest to note that the removal efficiency of total nitrogen decreased as the NaCl loading increased from 2 to 4 g.L-1 and then its removal efficiency recovered to the normal condition (approximately 97%) despite the high NaCl loading of up to 12 g.L-1, indicating that the microbial population has adapted to the hyper-saline condition .
In addition, the high salinity greatly affects the physical and biochemical properties of activated sludge, increasing proteins concentrations as well as decreasing membrane permeability. Sun et al  indicated that when the salt concentration increased from 5 to 15 g.L-1, a larger impact was observed where protein concentration increased by 41% and carbohydrate concentration increased by 106%. This confirm that significant increase of salinity affected seriously the microorganisms, consequently more SMPs are released due to the stress and biomass detachment occurs, resulting in higher membrane fouling rate. Similarly, Tadkaew et al  reported that at NaCl concentration of 4 g.L-1, the highest concentration of protein for SMP was observed. It then decreased gradually to a normal level as the salinity in the reactor to 12 g.L-1, indicating that the microbial population has adapted to the hyper-saline condition; whereas, the concentration of carbohydrate fluctuated slightly throughout their study. They also reported that a gentle increase in the transmembrane pressure (TMP) was observed as the salinity increased to 4 g.L-1. This was followed by severe membrane fouling as the NaCl concentration increased to 8 and 12 g.L-1. Previous study by Arabi and Nakhla  showed higher membrane permeability and lower fouling rate at the highest magnesium concentration of 96 mg.L-1 (Mg : Ca ratio of 5 : 1), as compared to the other reactors at Mg:Ca of 1:5 and 1:1. This was due to magnesium bridging of negatively charged biopolymers thus enhancing bio-flocculation, and decreasing membrane fouling. Furthermore, an increased influent Mg concentration decreased the concentration of the fraction > 100 kDa and < 10 kDa for SMPs. Moreover, with increasing sodium chloride shock loads, an increase in soluble EPS was observed in the MBR. During higher sodium chloride shock loads (30-60 g.L-1), the soluble EPS and its constituents (protein and carbohydrate) were almost doubled . Recently, Zhang et al  demonstrate that the presence of 10 g.L-1 NaCl, which was added into the MR feed, caused an obviously increase of polysaccharides in the supernatant (from 5.6 to 13.2 mg.L-1 on average), whereas the proteins content had a slightly increase, suggesting the salt stress resulted in more polysaccharides than proteins in the supernatant, which was consistent with the previously report’s Di Bella .
Table of contents :
CHAPTER I. BIBLIOGRAPHIC SYNTHESIS
1.1. Membrane bioreactor (MBR)
1.1.1. Membrane filtration
1.1.2. Membrane fouling in MBRs
1.1.3. Problematics of pharmaceutical micropollutants in the MBR
1.2. MBR – RO process
1.2.1. The combination of MBR-RO process
1.2.2. Reverse osmosis (RO)/nano-filtration (NF) process
1.3. Reverse osmosis (RO) concentrate
1.3.1. The characteristics of reverse osmosis concentrate
1.3.2. Reverse osmosis concentrate treatment
1.4. RO/NF concentrate recirculation to the MBR
CHAPTER II. MATERIAL AND METHODS
2.1. Membrane biorector set-up
2.2. The reverse osmosis (RO) pilot
2.3. Filterability test
2.4. Batch reactors
2.5. MBR-RO set-up
2.6. Analytical methods
2.6.1. Mixed liquor suspended solids (MLSS)
2.6.2. Particle size analysis
2.6.3. Dissolved organic carbon (DOC)
2.6.4. Total nitrogen and ammonia analysis
2.6.5. Chemical oxygen demand (COD)
2.6.7. Dosage of proteins
2.6.8. Dosage of polysaccharides
2.6.9. Cations and anions concentrations
2.6.11. Size-exclusion high performance liquid chromatography (HPLC-SEC) analysis
CHAPTER III. SHORT-TERM EFFECTS OF REVERSE OSMOSIS CONCENTRATE IN PEAK INJECTION ON THE ACTIVATED SLUDGE
3.2. Characteristics of the supernatant and RO concentrate
3.3. Results and discussion
3.3.1. Effect of RO concentrate on salts composition in the supernatant of batch reactor
3.3.2. Effect of RO concentrate on the biomass characteristics
3.3.3. Effect of RO concentrate on supernatant composition
3.3.4. Effect of RO concentrate on permeate flux decline
3.3.5. Effect of RO concentrate on fouling layer characteristics of sludge
CHAPTER IV. REVERSE OSMOSIS CONCENTRATE CONTINUOUS RECIRCULATION TO THE MBR-RO PROCESS FOR WASTEWATER TREATMENT
4.2. MBR-RO set-up
4.3. Characterization of reverse osmosis concentrate
4.4. Results and discussion
4.4.1. Effect of RO concentrate on the biomass characteristics
4.4.2. Effect of RO concentrate on the MBR performances
4.4.3. Effect of RO concentrate on salt composition of MBR supernatant
4.4.4. Effect of RO concentrate on the change in supernatant characteristics
4.4.5. Effect of RO concentrate on sludge and supernatant filterability test
4.4.6. Effect of RO concentrate on MBR fouling
4.4.7. Removal of selected pharmaceutical micropollutants in the MBR
CHAPTER V. EFFECT OF OPERATING PARAMETERS ON PROCESS EFFICIENCY
5.2. Lab scale MBR-RO set-up
5.3. Characteristics of the reverse osmosis concentrate and the MBR mixed influent after recirculation of concentrate
5.4. Comparison of the effects of RO concentrate recirculation on the biomass characteristics between the two sets of experiments
5.5. Comparison of MBR global performances between two experimental sets
5.6. Composition of supernatant in the MBR
5.7. Fouling propensity in the sludge and supernatant filterability test
5.8. Comparison of the effects of RO concentrate on the MBR fouling between two experimental sets
CONCLUSIONS AND PERSPECTIVES