Membrane bio-reactor (MBR): Effect of operating parameters and nutrients removal
by
MD. ABU HASAN JOHIR
A thesis submitted in fulfillment for the degree of
Doctor of Philosophy
School of Civil and Environmental Engineering Faculty of Engineering & Information Technology
University of Technology Sydney (UTS) New South Wales, Australia
September 2015.
I certify that the work in this report has not previously been submitted for any degree
nor has it been submitted as part of requirements for a degree except as fully
acknowledge within the text.
I also certify that this report has been written by me and the help that I have received in
my research work and the preparation of the report itself has been acknowledged. In
addition, I certify that all information sources and literature used are indicated in the
thesis.
Signature of Candidature
Md. Abu Hasan Johir
Sydney, September 2015
I would like to express my sincerest gratitude to Professor Saravanamuthu
Vigneswaran, Professor of Faculty of Engineering and Information Technology (FEIT),
University of Technology, Sydney (UTS), Australia, for his continuous guidance,
valuable suggestions, spontaneous encouragement and his various co-operation and
efforts throughout the project work. The author should remain ever grateful for his
super co-operation by inspecting every phase of the work and for providing his valuable
time throughout the project work. The author would like to express his humble respect
to A/Prof. Dr. Jaya Kandasamy, School of Civil and Environmental Engineering
(FEIT), UTS, for his kind help and encouragement to complete this study. I would like
to thank Professor H. H. Ngo and Mr. Rami Haddad for their support while working in
the Environmental and Hydraulic laboratories.
I appreciate the great help of Dr. P. Loganathan, Dr. A. Sathasivan, Prof. A. Grasmick,
Prof R. BenAim and Dr. H. K. Shon, Dr Robert, Dr P. Hagare and would like to thank
them for their valuable advice, discussions and support during my work. I would like to
thank Dr. Rupak for his advices and guidance’s. Sincere thanks are given to Dr. Vinh,
Nadnita, Dr Sherub, Dr Leonard, Dr Muna, Dr. Thamer, Dr. Ben, Dr. Nur, Dr. Jeong,
Dr. Gayathri, Jasmin, Danious, Sukanyah, Woo, and staffs in the Research Office for
their generous help, friendship and companionship.
I wish to acknowledge UTS (APA) and NCED for their financial support during my
study. I wish to express my deepest appreciation, gratitude and thanks to my beloved
family members for their endless love, encouragement and spiritual support.
Md. Abu Hasan Johir
Sydney, 2015
Table of Contents
Certificate of authorship ................................................................................................ ii
Acknowledgement .......................................................................................................... iii
List of Figures ................................................................................................................. ix
List of Tables ................................................................................................................ xiii
List of Journal Publications ........................................................................................ xvi
Nomenclature............................................................................................................... xvii
Abstract ......................................................................................................................... xix
CHAPTER 1 ................................................................................................................. 1-1
INTRODUCTION ........................................................................................................ 1-1
...................................................................................................... 1-2
................................................................................. 1-6
.............................................. 1-6
CHAPTER 2 ................................................................................................................. 2-1
LITERATURE REVIEW ............................................................................................ 2-1
...................................................................................................... 2-2 ................................................................. 2-2
.............................................. 2-5
.............................................................................. 2-8 ... 2-10
..................................................... 2-11
........................... 2-12
............................................ 2-13
....................................................................................... 2-16 .................................................... 2-16
..................................................... 2-17
.................................................... 2-19
........................................................................... 2-24 ..................................................................................................... 2-24
................................................................................. 2-27 ....................................................... 2-28
...... 2-29
CHAPTER 3 ................................................................................................................. 3-1
EXPERIMENTAL MATERIALS AND METHODS ............................................... 3-1
...................................................................................................... 3-2
................................................................................ 3-2 ......................................................................................... 3-2
................................................................................ 3-2
.................................................................................. 3-3 .
........................................................................................................... 3-3 ....... 3-8
........................................................ 3-12 ................ 3-15
................................................ 3-17 .......................................................... 3-17
....................................................................................... 3-19
..................................................................... 3-19 ..................................................................... 3-20
.................................................... 3-20
.................................................................. 3-21 ................. 3-22
................................ 3-22 ............... 3-23
................................................................................ 3-23 ....................................................... 3-23
.................................... 3-24.................. 3-24
............................................................................ 3-24
................................................. 3-24 .......................................................... 3-25
................................................................................. 3-25
............................................................................................. 3-26
CHAPTER 4 ................................................................................................................. 4-1
INFLUENCE OF ORGANIC LOADING RATE, IMPOSED FLUX AND SALINITY ON THE PERFORMANCE OF MEMBRANE BIO-REACTOR (MBR) ............................................................................................................................ 4-1
...................................................................................................... 4-2
.............................................. 4-24.2.1.
...................... 4-24.2.2. ....................................................... 4-64.2.3.
.................................................................................... 4-104.2.4.
................................................................................................ 4-12
............................................. 4-164.3.1. ............................................... 4-164.3.2. .... 4-214.3.3. ................................ 4-254.3.4. ............................................. 4-304.3.5. ............................................................................. 4-33
................................. 4-344.4.1. .......................................... 4-344.4.2. ................................................................................... 4-394.4.3. ...................................................... 4-424.4.4. ................. 4-484.4.5. ........................................................................................ 4-51
........................................................................................................ 4-53
CHAPTER 5 ................................................................................................................. 5-1
INFLUENCE OF SUPPORT MEDIA IN SUSPENSION FOR MEMBRANE FOULING REDUCTION IN SUBMERGED MEMBRANE BIOREACTOR (SMBR) ......................................................................................................................... 5-1
...................................................................................................... 5-2
.................................... 5-25.2.1. .... 5-25.2.2. .......... 5-45.2.3. ................................................................ 5-65.2.4.
..................................................................... 5-95.2.5.
................................................................................... 5-12
..................................................................................................... 5-155.3.1. .................................. 5-155.3.2. ....................................... 5-165.3.3.
......................................................................... 5-215.3.4. ........... 5-24
........................................................................................................ 5-27
CHAPTER 6 ................................................................................................................. 6-1
REMOVAL AND RECOVERY OF NUTRIENTS BY ION EXCHANGE FROM HIGH RATE MEMBRANE BIO-REACTOR (MBR) EFFLUENT ...................... 6-1
...................................................................................................... 6-2
................................ 6-4 ........................................................... 6-4
........... 6-5 ................................................... 6-6
......................................................... 6-9 ............................................................... 6-12 ................................................................ 6-13
................ 6-14
......................................................................................................... 6-16
........... 6-19 ..................... 6-19
............................ 6-22
.................................................................................... 6-24
...................................................................................................... 6-25
....................................................... 6-26
.......... 6-27
........................................................................................................ 6-28
CHAPTER 7 ................................................................................................................. 7-1
CONCLUSION AND RECOMMENDATIONS ....................................................... 7-1
............................................ 7-2
........................................................................................ 7-2
.......................................................................................................... 7-4
..................................................................................................... 7-4 .......... 7-5
......................................................... 7-6
References .................................................................................................................... R-1
List of Figures
Figure 2.1. MBR configurations (a) side-stream and (b) submerged............................ 2-3
Figure 2.2. Nitrogen transformations in biological treatment process (Source:
Sombatsompop, 2007) ............................................................................................ 2-7
Figure 2.3. Fouling mechanisms for MBR operated at constant flux is presented in
(Source: Le-Clech et al., 2006) ............................................................................. 2-17
Figure 2.4. Schematic representation of different fouling rates during long-term
operation of full-scale MBRs (Drews, 2010)........................................................ 2-18
Figure 2.5. Schematic illustration of the formation and removal of removable and
irremovable fouling in MBRs (adapted from Meng et al., 2009). ........................ 2-18
Figure 2.6. Factors influencing membrane fouling in the MBR process (Adapted from
Chang et al., 2002) ................................................................................................ 2-20
Figure 2.7. Inter-relationships between different operating factors and permeability loss
in lab scale MBRs (Source: Drews, 2010). ........................................................... 2-22
Figure 3.1. Laboratory scale membrane bioreactor (membrane area = 0.2 m2, pore size
= 0.14 m, volume of reactor = 10 L) .................................................................... 3-4
Figure 3.2. Experimental set up of membrane bioreactor (MBR) ................................ 3-8
Figure 3.3. Profile of gradual loading of salt in MBR ................................................ 3-12
Figure 3.4. Experimental set up (membrane bioreactor (MBR) followed by purolite
(A500P and A520E) ion-exchange column) ......................................................... 3-13
Figure 3.5. Extraction procedure of SMP and EPS from mixed liquor samples......... 3-22
Figure 4.1. Effect of OLR on the conversion of NH4-N into NO3-N (HRT = 8 h, SRT =
40 days, volume of the reactor = 4 L) ..................................................................... 4-6
Figure 4.2. Effect of OLR on membrane fouling (HRT = 8 h, SRT = 40 days, volume of
the reactor = 4 L) .................................................................................................... 4-8
Figure 4.3. Correlation between OLR with membrane fouling and hydrophobic and
hydrophilic fraction of organic (HRT = 8 h, SRT = 40 days, volume of the reactor =
4 L) .......................................................................................................................... 4-9
Figure 4.4. LC-OCD chromatogram of MBR effluent, SMP, EPS and foulant (OLR =
1.0 kgCOD/m3.d, HRT = 8 h, SRT = 40 days, volume of the reactor = 4 L) ....... 4-14
Figure 4.5. Temporal variation of membrane resistance at different imposed flux and
aeration rates (membrane area = 0.2 m2; membrane pore size = 0.14 μm; volume of
reactor = 10 L; SRT = 20 days; OLR = 1.5 kg COD/m3.day). ............................. 4-24
Figure 4.6. Correlation between filtered volume before getting rapid TMP rise with
imposed flux and aeration rates (membrane area = 0.2 m2; membrane pore size =
0.14 μm; volume of reactor = 10 L; SRT = 20 days; OLR = 1.5 kg COD/m3.day) ....
.............................................................................................................................. 4-25
Figure 4.7. MWD of SMP and EPS in MLSS at different fluxes (membrane area = 0.2
m2; membrane pore size = 0.14 μm; aeration rate = 1.2 m3/m2.membrane area.h; volume
of reactor = 10 L; SRT = 20 days; OLR = 1.5 kg COD/m3.day) .......................... 4-29
Figure 4.8. MWD of organic matter of filtrate and foulant (backwash water) (Flux = 20
L/m2.h; aeration rate = 1.2 m3/m2.membrane area.h; membrane area = 0.2 m2; membrane
pore size = 0.14 μm; volume of reactor = 10 L; SRT = 20 days; OLR = 1.5 kg
COD/m3.day) ........................................................................................................ 4-32
Figure 4.9. Profile of specific removal (average) of organic (DOC) and NH4-N with
different salt concentrations. ................................................................................. 4-38
Figure 4.10. Profile of MLSS, MLVSS, and SOUR (average) with different salt
concentrations. ...................................................................................................... 4-38
Figure 4.11. Dissolved organic nitrogen (DON) concentration present in biopolymer at
different salt concentration in MBR mixed liquor. ............................................... 4-48
Figure 4.12. TMP development with time in MBR at different salt concentration (Flux
2.5 L/m2.h; 1 m3/m2membrane area.h) .......................................................................... 4-50
Figure 4.13. Cluster analysis of DOC concentration at different load of salt (a) MBR
effluent and (b) mixed liquor (S.C = salt concentration) ...................................... 4-52
Figure 5.1. Effect of filtration flux and aeration of membrane resistance (membrane
area = 0.2 m2; reactor size = 10 L; SRT=20 days) .................................................. 5-3
Figure 5.2. Effect of suspended media on membrane resistance (membrane area = 0.2
m2; reactor size = 10 L; SRT = 20 days) ................................................................ 5-5
Figure 5.3. MWD distribution of organic matter in the MBR A) effluent; B) SMP; C)
EPS and D) foulant with and without suspended medium (flow rate = 25 L/m2.h;
aeration rate = 1.0 m3/m2membrane area. h; suspended medium (GAC) @ 2g/L of
volume of reactor) ................................................................................................. 5-11
Figure 5.4. EEM distribution of A) effluent; B) SMP; C) EPS and D) foulant (flow rate
= 25 L/m2.h; membrane area = 0.2 m2; aeration rate = 1.0 m3/m2membrane area.h;
reactor size = 10 L; SRT = 20 days) ..................................................................... 5-13
Figure 5.5. EEM distribution of MBR A) effluent; B) SMP; C) EPS and D) foulant
(flow rate = 25 L/m2.h; aeration rate = aeration rate = 1.0 m3/m2membrane area.h; with
suspended media GAC @ 2g/L of volume of reactor). ........................................ 5-14
Figure 5.6. Transmembrane pressure (TMP) development profile with time with and
with the addition of different particle sizes of GAC (A = without GAC; B = with
GAC particle size of 300 - 600 μm; C = with GAC particle size of 150 - 300 μm; D
= with GAC particle size of 600 - 1200 μm). ....................................................... 5-24
Figure 6.1. XRD pattern of Zr hydroxide...................................................................... 6-4
Figure 6.2. FTIR spectrum of Zr hydroxide .................................................................. 6-5
Figure 6.3. Effect of contact time and adsorbent dose (doses are shown as legends
within the figure) on the removal of phosphate by Zr hydroxide (initial phosphate
concentration 10 mg-P/L) ....................................................................................... 6-6
Figure 6.4. Effect of pH on phosphate adsorption by Zr hydroxide (Zr hydroxide dose
0.1 g/L).................................................................................................................... 6-8
Figure 6.5. Equilibrium phosphate adsorption isotherms as influenced by (a)
temperature, (b) pH, and (c) co-existing anions and Langmuir adsorption model
fitting. ...................................................................................................................... 6-9
Figure 6.6. Equilibrium phosphate adsorption isotherms as influenced by (a)
temperature, (b) pH, and (c) co-existing anions and Langmuir adsorption model
fitting. .................................................................................................................... 6-11
Figure 6.7. Effect of time on P concentration in the effluent in MFAH system with
addition of Zr hydroxide (1 g/L) for different initial P concentrations ................ 6-15
Figure 6.8. Effect of time on P removal efficiency in MFAH system with with addition
of different doses of Zr hydroxide (Inlet concentration 10 mg-P/L). ................... 6-16
Figure 6.9. Effect of repeated additions of Zr hydroxide (5 g/L) to MFAH system on
the phosphate removal at (a) 5 L/m2.h filtration flux and two inlet P concentrations
and (b) 10 mg/L inlet P concentration and two filtration fluxes ........................... 6-18
Figure 6.10. PO43- removal by HFO from SMBR effluent (0, 1, 5 and 10% of HFO by
mass with anthracite coal as inert material was used as filter medium; influent PO43-
oncentration to the post treatment HFO adsorption column was 2.2 mg/L) ......... 6-21
Figure 6.11. Comparison between purolite A520E and A500P (bed height = 6 cm;
velocity = 2.5 m/h, the concentration of PO4-P and NO3-N of the MBR effluent was
3.1 and 11 mg/L respectively) .............................................................................. 6-23
Figure 6.12. Effect on nutrient removal of two types of purolite ion-exchange resin
columns in series (velocity = 2.5 h, the concentration of PO4-P and NO3-N of MBR
effluent was 4.18 and 9 mg/L respectively) .......................................................... 6-24
List of Tables
Table 2.1. Comparison between MBR configurations (Source: Sombatsompop, 2007) ...
................................................................................................................................ 2-4
Table 2.2.Characteristics of different types of membranes (Adapted from Fane, 2002) ...
................................................................................................................................ 2-5
Table 2.3. Summary of operation conditions of aerobic membrane bioreactor for
different wastewaters (Adapted from Sombatsompop, 2007) ................................ 2-9
Table 2.4. Typical ranges of the different fouling rates occurring at full-scale MBR
(Source: Drews, 2010) .......................................................................................... 2-19
Table 2.5. Relationship between various fouling factors and membrane fouling
(Adapted from Meng et al., 2009) ........................................................................ 2-21
Table 2.6. Comparison of relevant conditions and fouling results (HF, hollow fibre; FS,
flat sheet) (Source: Drews, 2006). ........................................................................ 2-23
Table 2.7. Aeration conditions for different full-scale MBRs (Source: EUROMBRA,
2006) ..................................................................................................................... 2-26
Table 3.1. Membrane characteristics used in this study ................................................ 3-5
Table 3.2. Operating conditions of laboratory scale SMBR with and without suspended
media ....................................................................................................................... 3-6
Table 3.3. Composition of synthetic wastewater........................................................... 3-7
Table 3.4. Membrane characteristics used in this work ................................................ 3-9
Table 3.5. Laboratory scale hollow fibre MBR operated at different OLRs and salinity
................................................................................................................................ 3-9
Table 3.6. Composition of synthetic wastewater......................................................... 3-10
Table 3.7. Typical chemical and physical characteristic of A-500P and A520E ........ 3-14
Table 3.8. Concentrations of the constituents of synthetic wastewater ....................... 3-16
Table 4. 1. Effect of OLR on the removal of DOC and nutrients (nitrogen and
phosphorous) (HRT = 8 h; SRT = 40 days). ........................................................... 4-5
Table 4.2. Fractionation of OM by LC-OCD of bio-reactor effluent, SMP and EPS
operated without membrane (OLR = 1.0 kgCOD/m3.d) ....................................... 4-15
Table 4.3. Removal of organic matter and nutrients by MBR operated at different
imposed fluxes (membrane area = 0.2 m2; membrane pore size= 0.14 μm; volume
of reactor = 10 L; SRT = 20 days; OLR = 1.5 kg COD/m3.day; aeration rate = 1.2
m3/m2.membrane area.h) ............................................................................................... 4-20
Table 4.4. Removal of DOC and NH4-N by MBR with and without salt (gradual
addition) concentration (HRT = 8 h) .................................................................... 4-37
Table 4.5. Characterization of organic matter in MBR effluent operated at different
gradual loading of salt........................................................................................... 4-46
Table 4.6. Characterization of organic matter in MBR mixed liquor operated at
different gradual loading of salt ............................................................................ 4-47
Table 5.1. Rsc; Rpb, Rm and Rt at different operating flux and aeration rate (membrane
area = 0.2 m2; reactor size = 10 L; SRT = 20 days) ................................................ 5-4
Table 5.2. Rsc; Rpb, Rm and Rt at operating flux of 25 L/m2.h with and without GAC in
suspension at different dose and aeration rates: ...................................................... 5-6
Table 5.3. Effect of operating flux on organic matter removal and on nitrification at an
aeration rate of 1.5 m3/m2.membrane area. h (membrane area = 0.2 m2; reactor size = 10
L; SRT = 20 days): ................................................................................................. 5-8
Table 5.4. Effect of suspended media on nutrients removal operated at a aeration rate of
1.5 m3/m2.membrane area.h (membrane area = 0.2 m2; reactor size = 10 L; HRT = 2 h;
SRT = 20 days): ...................................................................................................... 5-9
Table 5.5. Biomass concentration and sludge properties with and without the addition
of GAC in SMBR ................................................................................................. 5-16
Table 5.6. Removal of organic, ammonia and phosphate with and without the addition
of GAC in SMBR (all the concentrations are in mg/L) ........................................ 5-20
Table 5.7. Organic characteristics of SMBR effluent with and without the addition of
GAC in SMBR (all the units are in mg/L) ............................................................ 5-23
Table 5.8. Rt, Rc and Rp with and without the addition of different particle of GAC
(membrane resistance, Rm = 0.59 x 1012 m-1) ....................................................... 5-26
Table 6.1. Langmuir model parameters for phosphate adsorption at different
temperatures, pHs, and in the presence of nitrate and sulphate ............................ 6-12
Table 6.2. Pseudo first- and second-order adsorption rate constants and calculated and
experimental qe values for different Zr hydroxide doses (initial phosphate
concentration 10 mg P/ L) .................................................................................... 6-13
Table 6.3. The thermodynamic parameters for the adsorption of phosphate on Zr
hydroxide .............................................................................................................. 6-14
Table 6. 4. Estimation of retention of nutrients in the purolite ion-exchange column .......
.............................................................................................................................. 6-26
Johir, M. A. H., Aryal, R., Vigneswaran, S., Kandasamy, J., & Grasmick, A.
(2011). Influence of supporting media in suspension on membrane fouling
reduction in submerged membrane bioreactor (SMBR). Journal of Membrane
Science, 374(1), 121-128.
Johir, M. A. H., George, J., Vigneswaran, S., Kandasamy, J., & Grasmick, A.
(2011). Removal and recovery of nutrients by ion exchange from high rate
membrane bio-reactor (MBR) effluent. Desalination, 275(1), 197-202.
Johir, M. A. H., Vigneswaran, S., Sathasivan, A., Kandasamy, J., & Chang, C. Y.
(2012). Effect of organic loading rate on organic matter and foulant characteristics
in membrane bio-reactor. Bioresource technology, 113, 154-160.
Johir, M. A., George, J., Vigneswaran, S., Kandasamy, J., Sathasivan, A., &
Grasmick, A. (2012). Effect of imposed flux on fouling behavior in high rate
membrane bioreactor. Bioresource technology, 122, 42-49.
Johir, M. A., Shanmuganathan, S., Vigneswaran, S., & Kandasamy, J. (2013).
Performance of submerged membrane bioreactor (SMBR) with and without the
addition of the different particle sizes of GAC as suspended medium. Bioresource
technology, 141, 13-18.
Johir, M. A. H., Vigneswaran, S., Kandasamy, J., BenAim, R., & Grasmick, A.
(2013). Effect of salt concentration on membrane bioreactor (MBR) performances:
detailed organic characterization. Desalination, 322, 13-20.
A = The membrane surface area (m2)
ASTM = American Standard Testing and Methods
BOD = Biochemical Oxygen Demand
BTSE = Biologically treated sewage effluent
BOM = Biodegradable Organic Matter
COD = Chemical Oxygen Demand
Da = Dalton
DOC = Dissolved Organic Carbon
DOM = Dissolved Organic Matter
kDa = Kilo Dalton
EfOM = Effluent Organic Matter
GAC = Granular Activated Carbon
EPS = Extracellular Polymeric Substances
HPSEC = High Pressure Size Exclusion Chromatography
kPa = Kilo Pascal
m.bar = Millibar
MWD = Molecular Weight Distribution
MF = Microfiltration
UF = Ultra filtration
NF = Nanofiltration
NOM = Natural Organic Matter
NTU = Nephelometric Turbidity Unit
PAC = Powdered Activated Carbon
Rm = Membrane resistance
RO = Reverse Osmosis
SEC = Size Exclusion Chromatography
SS = Suspended Solids
t = Time
TDS = Total Dissolved Solid
TMP = Trans-membrane Pressure
V = Total permeate volume (l)
P = Applied trans-membrane pressure (Pa)
= Water viscosity at 200C (N s/m2)
= The specific resistance of the cake deposited
= Polydispersity 0C = Degree Celsius
Membrane bio-reactor is an efficient, cost effective and reliable treatment process to
produce high quality water from wastewater. In this study, a number of submerged
membrane bio-reactors (SMBRs) experiments were conducted at different organic
loading rates (OLRs) and fluxes (ranging from 2.5 - 40 L/m2.h and corresponding
hydraulic retention time of 10 - 1.5 h) to investigate their influence on organic and
nutrient removal and on membrane fouling. A second set of experiment was also carried
out with gradual increase of salt concentration in continuous MBR to assess its
performances in this particular scenario (which may occur in coastal areas and in certain
industries). The operation of MBRs at low HRT resulted in sudden rise of trans
membrane pressure (TMP). The sudden development of TMP was minimized by
introducing granular activated carbon (GAC) in MBR as suspended medium. The
incorporation of GAC reduced TMP or total membrane resistance by 58% and also
helped to remove an additional amount of dissolved organic matter. Further, a set of ion
exchange adsorption study was conducted for the removal and recovery of the nutrients
from the effluent of high rate MBR. The major findings are summarizes below.
The increase of OLR, flux and salt concentration resulted in lower removal of organic
and nutrients and also caused higher membrane fouling (i.e. increased transmembrane
pressure (TMP) development). The removal efficiency of DOC decreased from 93 –
98 % to 45 - 60 % when the OLR increased from between 0.5 – 1.0 to 2.75 – 3.0 kg
COD/m3d. Similarly the removal of ammonia decreased from 83–88% to less than 67%
when the OLR was increased to 2.0 – 3.0 kg COD/m3d. The increase of flux (i.e.
reducing of HRT) also resulted in 30 - 40 % lower removal of organics and nutrients.
The removal of organic and nutrient decreased when the salt concentration was
increased from 0 to 35 g/L. Based on the operating conditions of this study, the
suspended media had less effect on nitrification but had an influence on organic
removal. However, changing the operating parameters (such as increase of SRT) may
improve nitrification rate.
The increase of OLR and salt concentration resulted in higher membrane fouling.
Similarly flux and aeration rate also played a major role in membrane fouling reduction.
However, the effect of flux on the reduction of membrane fouling was much higher than
that caused by aeration rate. A lower flux of 20 L/m2 h produced 75 times more water
than a higher flux of 40 L/m2h with an aeration rate of 0.6 m3/m2membrane area.h. The
reduction of aeration rate from 1.5 to 1.0 m3/m2membrane area.h caused a sudden rise of
TMP. The sudden rise of TMP can be minimized by incorporating the medium in
suspension in the reactor (to induce surface scouring of the membrane). The
incorporation of suspended medium prevented a sudden rise of TMP (total membrane
resistance reduced by ~ 58%) by creating an extra shearing effect onto the membrane
surface produced by suspended media. It reduced the deposition of particles on the
membrane surface by scouring. The addition of GAC also adsorbed some organic
matter prior to its entry to the membrane. Nevertheless it is also important to apply a
sufficient aeration rate (in our case 1 m3/m2membrane area h) to maintain a good functioning
of suspended media in MBR. The aeration helped in scouring and provision of oxygen
to microorganisms and maintained the media in suspension. Additionally, the amount
and sizes of the suspended medium played major role in fouling reduction. In this study,
we found the concentration of suspended media of 2 g/L and GAC size of 300-600 μm
was effective in reducing membrane fouling. Therefore a suitable amount and size of
suspended medium needed depends on the flux and aeration (or air scour) rate used.
The characteristics of organic matter of SMBRs effluent showed that a range of organic
matter (such as amino acids, biopolymers, humics and fulvic acids type substances) was
removed by the GAC both by scouring and adsorption mechanisms. A detailed organic
matter characterization of membrane foulant, soluble microbial product and
extracellular polymeric substances showed that bio-polymer together with humic acid
and lower molecular neutral and acids were responsible for membrane fouling along
with the deposition of floc particle onto the membrane surface.
MBR usually removes both organic matter and nitrogen from water. However, the
removal of nitrogen and phosphorus using a high rate MBR system is not sufficient. It is
equally practical to remove nitrogen and phosphorus by physico-chemical processes as
post-treatment such as ion exchange/ adsorption. In this study, different ion exchange
materials such as purolite (A520E and A500P), hydrated ferric oxide (HFO) and
zirconium (IV) hydroxides were used to remove nitrogen and phosphorus from MBR
effluent. They all showed ~ 90% removal of nutrients. The nutrients captured on the ion
exchanger were later recovered when the ion-exchange was regenerated.
Efficient, cost effective and reliable treatment processes are needed to produce high
quality water from wastewater that can be reused without detrimental effects. One of the
most promising technologies in wastewater treatment is the membrane bioreactor
(MBR). MBR is a combination of an activated sludge process and membrane separation
process. A small footprint, complete solid liquid separation, superior removal of
organics and production of high quality of water are key advantages of the MBR
process.
The performance of MBR depends on different operating parameters such as filtration
flux, hydraulic retention time (HRT), solid retention time (SRT), organic loading rate
(OLR), etc. Many studies have been conducted to observe the effect of SRT on MBR
performance. For example Grelier et al. (2006) reported that the increase of SRT from 8
to 40 d elevated membrane performance while Al-Halbouni et al. (2008) reported that
higher concentration of floc bound exo-polymeric substances (EPS) at lower SRT of 23
d than higher SRT of 40 d. Further, Laera et al. (2009) recommended SRT of >40 d for
reliable operation of MBR when treating municipal wastewater. Al-Halbouni et al.
(2008) confirmed that SRT’s negative impact on MBR performance was caused by high
concentrations of soluble microbial product (SMP) or bound EPS. A details explanation
on the effect of STR along with soluble microbial product (SMP) or bound EPS is
presented in chapter 2, section 2.3.1. Along with SRT, HRT, filtration flux, mixed
liquor suspended solids (MLSS) and ORL also play important roles in the performance
of MBR.
Several studies have been conducted on the effects of HRT and ORL on membrane
fouling (Galapate et al., 1999; Ren et al., 2005; Birima et al., 2009; Kornboonraksa and
Lee, 2009; Khoshfetrat et al., 2011). Kornboonraksa and Lee (2009) reported that an
increase in chemical oxygen demand (COD) concentration from 1150 to 2050 mg/L led
to poorer treatment efficiency (COD removal fell from 96% to 92%). They also reported
that increase of MLSS resulted in higher sludge viscosity and reduced membrane
filterability. A detailed comparison of various operating conditions of MBR and their
influences are presented in Chapter 2 (Table 2.6). In addition, many types of wastewater
(especially industrial wastewater from the cheese industry) and water from coastal areas
(where seawater seeps through the ground) are very saline and a detailed study on MBR
under saline conditions is critical.
The MBR system has also been effectively applied to the treatment of saline water such
as seawater. For example, Artiga et al. (2008) studied the performance of hybrid MBR
for the treatment of saline wastewater from a fish canning factory (salt concentration
was up to 73-83 g/L) as while Dan et al. (2002) conducted their analysis using yeast
MBR and biological MBR at a high salt concentration of 32 g/L. However, the
treatment efficiency of saline water depends on salt concentration and the removal of
organic and nutrient is affected by high salt concentration (Lay et al., 2010).
Nonetheless the removal of organic matter and nutrients may be improved by adding
salt tolerant culture such as halobacter halobium to the biomass or by providing a longer
acclimation period to the biomass (Lay et al., 2010).
Rapid worldwide commercialisation of membrane technology is limited by the major
problem of membrane fouling. Fouling is the deposition of soluble and particulate
matter onto and into the membrane surface due to physicochemical and biological
interactions between the membrane and the sludge. This compromises the efficiency of
membrane filtration and reduces the permeate quality requiring more frequent
membrane cleaning and membrane replacement leading to higher system operation
costs (Jefferson et al., 2004). Membrane fouling can be categorised as reversible,
irreversible and irrecoverable fouling (Drews, 2010). Different strategies have been
applied to minimise membrane fouling. Of these the use of aeration across the
membrane surface is the most common method for minimising membrane fouling. This
helps to scour deposited particles on the membrane surface. The use of aeration can
control reversible fouling but it cannot control irreversible membrane fouling (Drews,
2010). Furthermore, the use of higher aeration forms a large part of the operating costs
of the MBR (Cui et al., 2003; Judd, 2007).
Besides the use of more aeration, cost effective reliable technologies are required to
reduce membrane fouling. Membrane fouling could be minimised by utilising a
medium in suspension in the MBR (such as activated carbon) which could help to
adsorb organic matter before it reaches the membrane surface. The medium can also
provide higher shearing stress on the membrane surface and this will prevent the
deposition of sludge particles on the membrane surface. Several studies have been
conducted on the addition of medium in suspension to reduced membrane fouling (Li et
al., 2005; Akram and Stuckey, 2008; Siembida et al., 2010; Pradhan et al., 2012).
Siembida et al. (2010) used granular material (polypropylene) in their study, whereas Li
et al. (2005) and Akram and Stuckey (2008) used powdered activated carbon (PAC).
PAC due to its small size could foul the membrane by producing a more compact cake
layer on the membrane surface along with sludge particles. Thus, employing a larger
medium could be useful. On the other hand the study conducted by Pradhan et al.
(2012) used GAC as the suspended medium. However, they used inorganic kaolin clay
suspension as feed water and it needs to be tested with wastewater. The emphasis
should be not only on scouring but also the prior removal of organics on GAC.
The MBR system usually removes both organic matter and nitrogen from water.
However, the removal efficiency of nitrogen and phosphorus by the MBR system
depends on operating parameters. It is therefore important to remove nitrogen and
phosphorus through physicochemical processes such as ion exchange, reverse osmosis,
electro-dialysis and catalytic reduction (Nur et al., 2012). Of these methods the
adsorption and ion exchange processes for the removal of nitrate and phosphate are
more consistent due to their simple and economical operation (Nur et al., 2012). Many
studies have been done on the removal and recovery of phosphate and nitrate using ion
exchange resin (Bae et al., 2002; Samatya et al. 2006; Jung et al., 2006; Nur et al., 2012;
2014). Such studies have used different ion-exchange resin such as Purolite A500PS,
Purolite A100, Purolite A520E, Purolite A300, macroreticulated Amberlite IRA900,
Dowex SBRP, aluminium oxide, iron oxide, zirconium oxide, hydrotalcite and layered
double hydroxides, hybrid anion exchanger (HAIX) and hydrated ferric oxide (HFO)
nanoparticles (Chen et al., 2002; Blaney et al., 2007; Lee et al., 2007; Terry, 2009; Nur
et al., 2012). Thus, a strategy employing MBR at a high rate especially to remove
organics and ion exchange as post-treatment to remove and recover nutrients should be
explored.
This thesis examines the effect of organic loading rate (OLR) in an MBR in terms of
effluent quality and membrane fouling in order to optimize the OLR for a given
wastewater. It also examines the effect of shock and gradual loading of salt on sludge
properties and on the membrane fouling (to observe the capability in MBR in handling
salty industrial wastewater or wastewater in coastal area). Another area of interest is an
investigation into the influence of imposed flux (hydraulic retention time; HRT) on
sludge properties and on membrane fouling. In terms of evaluation this study looks at
the effectiveness of the large suspended medium (such as GAC) as adsorbent and
scouring medium with reference to organic removal and membrane fouling reduction.
Different ion exchangers in capturing optimum amount of nutrients are explored.
Finally, the thesis investigates a MBR-ion-exchange hybrid system for its removal of
organic matter by optimising the MBR at the lowest feasible HRT to remove and
recover nitrogen and phosphorus.
This thesis has been divided into 7 chapters as follows:
• Chapter 1: This is an introductory chapter containing background, motivation
and objectives of this study.
• Chapter 2: Detailed literature review on the status of objectives proposed in
this study and their limitations. This chapter discusses the fundamentals of the
MBR treatment system, factors affecting MBR process, effect of salinity on
MBR performance, mitigation of membrane fouling and physicochemical
processes available for the removal and recovery of nutrients from
wastewater.
• Chapter 3: This chapter presents the experimental methods used in this study
such as experimental set-up of MBR systems and different analytical
approaches.
• Chapter 4: This chapter presents and discusses the results of MBRs operating
at different operating parameters such as different imposed fluxes, different
organic loading rate and most importantly the effect of salt concentration on
the MBR performance.
• Chapter 5: This chapter discusses the results of experimental investigation of
MBR systems operating with and without the addition of granular activated
carbon (GAC) as suspended medium. This also includes the effect of different
particle sizes of GAC on the membrane fouling reduction.
• Chapter 6: This chapter analyses the experimental findings for the removal
and recovery of nutrients, mainly nitrate and phosphate from a high rate MBR
system. The phosphate and nitrate removal by different ion exchangers are
also discussed in this chapter.
• Chapter 7: Summary of key findings of this research is presented here
together with recommendations for future analyses on this topic.
Membrane bio-reactor (MBR) is a most promising process in wastewater treatment
because it has: the ability to completely removing solids; superior removal of nutrient and
organic matter; high loading rate capabilities; low/zero sludge production; and it leaves
only a small footprint. This makes the MBR particularly suitable for water reuse. The
MBR system was first developed in the 1970s for treating sanitary wastewater, and
consisted of a suspended-growth biological reactor combined with a membrane unit
process into a single process. The growth rate of MBR treatment systems is high (almost
10.9% per annum) when compared to other advanced wastewater treatment processes and
more than any other membrane technologies (Drews, 2010). Ten years ago, the
operational cost of MBR was $0.90/m3 and this fell to $0.08/m3 in 2005 due to lower
membrane costs and also due to improved energy efficiency to below 0.4 kWh/m3
(Hermanowicz, 2011). The MBR system can function at higher mixed liquor suspended
solids (MLSS) concentration of around 12 - 18 g/L (Holler and Trösch, 2001). Recent
experimental studies tend to use lower MLSS concentrations in order to avoid sludge build
up in the membrane module (Jang et al., 2006; Kawasaki et al., 2011). Furthermore, the
MBR system has been effectively applied to the treatment of saline water such as seawater.
The membrane bio-reactor can be configured both in side-stream or submerged in the
bioreactor (see Figure 2.1). Their relative merits are summarized in Table 2.1. The
configurations of MBR are based on either planar or cylindrical geometry. There are five
principal membrane configurations currently employed in practice such as hollow fiber
(HF), spiral-wound, plate-and-frame/flat sheet (FS), plated filter cartridge and tubular
(Radjenovi et al., 2008).
A qualitative comparison between different membrane configurations is presented in Table
2.2 (Fane, 2002).
Figure 2.1. MBR configurations (a) side-stream and (b) submerged
Table 2.1. Comparison between MBR configurations (Source: Sombatsompop, 2007)
Submerged Side-stream
Advantages • Small footprint
• Feed-forward control of O2
demand
• Less frequent cleaning required
• Lower operating costs
• Low liquid pumping costs (28%
of total costs)
• Low energy consumption
• Combined COD, solids and
nutrient removal in a single unit
• Low/zero sludge production
• Rapid start up
• Sludge bulking not a problem
• Small footprint
• Complete solids removal from
effluent
• Effluent disinfection
• High loading rate capability
• Combined COD, solids and
nutrient removal in a single unit
• Low/zero sludge production
• Rapid start up
• Sludge bulking not a problem
Disadvantages • Susceptible to membrane
fouling
• High aeration cost
• Aeration limitations
• Membrane fouling
• Membrane costs
• High operating costs
• High pumping cost (60-80% of
total costs)
• High cleaning requirement
• Process complexity
Table 2.2.Characteristics of different types of membranes (Adapted from Fane, 2002)
CharacteristicFlat Plate
(FP)
Spiral-
Wound Tubular
Hollow Fibre
(HF)
Submerged
(FP or HF)
Packing
Density Moderate High Low High
Moderate (FP)
High (HF)
Energy
Low-
Moderate
(Laminar
flow)
Moderate
(Spacer
losses)
High
(Turbulent)
Low (Laminar
dead end)
Low
Dead-
end/Bubbling
Solids
handlingModerate Poor Good Moderate/Poor
Moderate/Good
(Bubbling)
Cleaning Moderate Can be
difficult
Good-
Physical
cleaning
possible
Back flushing
possible
Back flushing
possible (HF)
Replacement Sheet [or
cartridge) Element
Tubes or
element Element Element/bundle
In an aerobic MBR process, the removal of organic and nutrients is achieved through the
bio-degradation of these materials by micro-organisms. In a MBR, two types of bacteria
are found, heterotrophic and autotrophic, where the former is more predominant.
Heterotrophs obtain their energy from organic compounds and depending on the medium
oxygenation, oxygen (oxic condition) or nitrate (anoxic condition) is used as a terminal
electron acceptor. Unlike heterotrophs, autotrophic bacteria obtain their energy by
oxidizing inorganic compounds and are obligate aerobes, so only use oxygen as an electron
acceptor (Horan, 2003).
Because the oxidation of inorganic material does not yield as much energy as oxidation of
organic carbon sources, autotrophs have a much slower growth rate than heterotrophs.
Autotrophic bacteria are very sensitive and are inhibited by a wide range of toxic organic
compounds (Liu, 2007). Both communities live in bioreactors in equilibrium as the
autotrophs allow the heterotrophs to survive by decreasing the ammonia level.
On the other hand, in a MBR system the removal of ammonia occurs by converting
ammonia into nitrite and nitrate through the nitrification process. This phenomenon is
carried out in two steps. The first is the conversion of ammonia to nitrite by ammonia
oxidizing bacteria (AOB) which include Nitrosomonas (formerly Nitrosococcus mobilis
and Nitrosomonas) and Nitrosospira (formerly Nitrosospira, Nitrosovibrio, and
Nitrosolobus) bacteria. The second step is the nitrite conversion into nitrate achieved by
nitrite oxidizing bacteria (NOB) including Nitrobacter and Nitrospira bacteria (Halling-
Sørensen, and Jørgensen, 1993; Schramm et al., 1998). Although, MBR provides better
removal of organic matters, the major challenge in the membrane filtration systems is the
control of membrane fouling and its minimization during operation. There is a pressing
need to minimize the fouling potential and/or develop a simple method to measure and
predict the fouling potential of wastewater.
On this theme, although the MBR process can remove most of the organic matter and
nutrients from water but the removal efficiency may decrease when exposed to peak and
variable loads depending on the operating conditions. Other disadvantages are the need for
larger reactor volumes, higher operating costs, and waste sludge production when
phosphorus removal is achieved by chemical precipitation.
Phosphorus is present in wastewater in the form of phosphates such as orthophosphates,
condensed phosphates and organic phosphate fractions (Radjenovi et al., 2008). The
removal of phosphate can be achieved by precipitation and/or adsorption, or by luxury
uptake. Only a small amount of phosphorus is used for cell metabolism and growth (1–2%
of the total suspended solids (TSS) mass in the mixed liquor) (Radjenovi et al., 2008).
Furthermore, phosphorus can also be removed through enhanced biological phosphorus
removal (EBPR) however the successful operation of EBPR depends on many process
operational factors, especially variation in wastewater quality (Zheng et al., 2014). Indeed
phosphorus can be removed by biological processes and although this process is an
environmentally friendly one, the mechanism is complex. One cannot remove the P below
a particular concentration and so the nitrogen and phosphorous can be removed by
physico-chemical treatment such as ion-exchange.
Figure 2.2. Nitrogen transformations in biological treatment process (Source:
Sombatsompop, 2007)
The treatment of wastewater by MBR mainly depends on its operating conditions (Le-
Clech et al., 2006) such as solid retention time (SRT), hydraulic retention time (HRT),
biomass concentration, and characteristics of biomass in-terms of soluble microbial
product (SMP) and extracellular microbial products (EPS), temperature, mode of
operation, and membrane morphology etc.
The operating conditions of aerobic membrane bioreactor for different wastewaters are
summarised in Table 2.3.
Table 2.3. Summary of operation conditions of aerobic membrane bioreactor for different wastewaters (Adapted from Sombatsompop, 2007)
Wastewater type Synthetic Municipal Synthetic Domestic DomesticOrganic
compoundsOil wastewater Tannery
Reactor Volume (L) 7 3900 2.5 4.5 66 30 - 2.25
Membrane area (m2) 0.1 13.9 0.03 4 0.24 0.04 0.087 0.27
HRT (h) 7.8 10.4-15.6 3.3 5 30 7.5 5-30 1
SRT (d) 20-60 - 10-30 5-40 - 60 5-30 10, 20, 550
MLSS (g/L) 2.4-5.5 18-20 17.2-27 - - 0.8 0.2 10-40
Initial COD (mg/L) 280 786 800-880 95-400 74-102 8200 36000-146000 1500-2200
COD loading (kg/m3.d) - 1.2-1.8 5.7 - 0.14-0.18 - 0.49-0.11 3-10
COD removal (%) >95 90-95 97 >90 >85 - >90 93
Flux (L/m2.d) 9 432-648 - - 28 300 - 3.5-6.7
TMP (kPa) - 18-26 - - - - 101 27
Many researchers studied the effect of SRT on membrane fouling. SRT is defined as the
total volume of reactor divided by the amount of sludge withdraws every day. For example
if the working volume of the reactor is 10 L and the amount of sludge discharge every day
is 0.5 L then SRT is 20 d (10 L/ 0.5 L.d). Some recommended a higher value of SRT for
successful operation of a MBR. For example Adham and Gagliardo (1998) suggested the
use of higher SRT for more than 30 days, while Cicek et al. (2001) stated that a MBR can
function at a lower SRT of less than 10 days. Other researchers found less fouling when
they increased SRT from 2 to 10 days (Trussell et al., 2006) and from 20 to 60 days
(Ahmed et al., 2007). Some researchers also recommended that a higher SRT gives a
better effluent quality. For example Grelier et al. (2006) observed good bio-degradation of
organic and nutrients with a higher SRT of 40 days. Ke and Junxin (2009) reported that a
higher sludge concentration at long SRT demostrated better organic removal efficiency,
and was favourable for growing nitrifiers. They also reported the highest fouling rate when
a MBR was operated at a SRT of 10 days and the lowest fouling rate occurred when there
was no sludge withdrawal (i.e. SRT = days). On the other hand Al-Halbouni et al. (2008)
observed the negative impact of operating low SRT on membrane performance. This
resulted in the production of high concentrations of SMP or bound EPS. Majority of the
studies recommend an SRT of 20-50 days as an optimum (Meng et al., 2009)
On the other hand, hydraulic retention time (HRT) also affects MBR performance in terms
of effluent quality and membrane fouling. HRT is defined as the volume of the reactor
divided by the filtration velocity. For example if the volume of the reactor is 10 L and
filtration flux is 5 L/h the HRT is 2 h (i.e. 10 L/ 5 L/h). In a fixed volume of reactor, the
increase of imposed flux decreases the hydraulic retention time (HRT). The HRT affects
not only the treatment efficiency of MBRs (Ren et al., 2004, 2005), but also the
characteristics of the biomass in MBRs (Yoon et al., 2004). Further, it has also been
reported that a MBR operated at a lower HRT resulted in poor effluent quality and a higher
membrane fouling than when operated at longer HRTs (Meng et al., 2007; Aryal et al.,
2009; Hong et al., 2011; Yuzir et al., 2011). For example, Meng et al. (2007) observed an
increase in particle size in the mixed liquor in the MBR with a decrease in HRT. The
particle size distributions (PSD) of mixed liquor had significant impacts on the
permeability of the fouling cake layer. Meng et al. (2007) also reported that a lower HRT
reduced COD removal, reduced biomass activity and dissolved oxygen concentration. Ren
et al. (2005) found that the removal of COD decreased with shorter HRT. Thus, a shorter
HRT while having an advantage of a smaller footprint, lower oxygen demand, greater
sludge production, which facilitates maximal recovery of nutrients in the sludge, and
greater reuse potential of carbon from grey water, has the disadvantage of increased
membrane fouling due to the larger amounts of EPS produced. Moreover, Meng et al.
(2007) reported that there was an increase in filamentous bacteria with increasing EPS
concentration.
In a similar manner to SRT and HRT, organic loading rate (OLR) also plays an important
role in the treatment of wastewater by a MBR. An increased OLR reduced the filterability
of the MBR. Kornboonraksa et al. (2009) reported that an increase of influent COD, BOD
and NH4-N from 1150 to 2050 mg/L, 683 to 1198 mg/L and 154 to 248 mg/L, respectively,
resulted in a decline in the removal efficiency of COD, BOD and NH4-N. Trussell et al.
(2006) reported an increased membrane fouling with higher OLR. They reported that
steady-state membrane fouling rates increased 20-fold over a four-fold increase in F/M.
Khoshfetrat et al. (2011) found a reduction of COD removal efficiency from 90% to 74%,
when OLR increased from 1 to 2.5 kgCOD/m3d. Shen et al. (2010) reported a higher
degradation of organic (glucose) of 98% at OLRs of 13 gCODL−1 d−1 than an OLR of 30
gCODL−1 d−1 which had a degradation of about 70%. It must be noted that in these cases
while the OLR was changed, the HRT was not controlled.
EPS are colloidal materials that contain construction materials for microbial aggregates
and a wide range of organics such as polysaccharides, proteins, lipids, amino-sugars, and
nucleic acids, other polymeric compounds, low molecular weight acids and neutrals.
Similarly, SMP are soluble materials that are released during cell lysis and they diffuse
through the cell membrane, but are lost during synthesis or are excreted for some reason
(Laspidou and Rittmann, 2002; Li et al., 2005). SMP contains a wide range of organics
similar to EPS and have moderate formula molecular weights and are biodegradable
(Laspidou and Rittmann, 2002). Laspidou and Rittmann (2002) stated that the SMP is also
responsible for the formation of effluent chemical oxygen demand (COD) and biochemical
oxygen demand (BOD) in the biological wastewater treatment process.
In literature, SMP and EPS have been analysed in many different ways. These include the
use of High pressure liquid chromatography (HPLC) with a SEC (Size Exclusion
Chromatorgraphy) column as a tool to measure molecular weight distribution (MWD) of
organics, Excitation Emission Matrix (EEM) and Liquid Chromatography - Organic
Carbon Detector (LC-OCD) and Ultraviolet (UV) Spectrophotometry (Jang et al., 2006;
Kimura et al., 2009; Aryal et al., 2011). The use of non-conventional methods such as
HPLC, LC-OCD and EEM provide more relevant information on the organic
characterization than conventional methods such as the phenol–sulfuric acid method and
the Lowry method. For example, Kimura et al. (2009) stated that conventional methods
employed in the analysis of SMP were not appropriate for investigatng of membrane
fouling in MBRs compared to non-conventional methods such as EEM analysis.
The above literature discussed the use of MBR in treating wastewater. In this section a
detailed review of the application of MBR in treating saline water is undertaken. It is
evident that the MBR system can also been effectively applied for treating saline water
such as seawater. A number of studies have been conducted on the classification of
microorganisms in saline water. Lay et al. (2010) reported that true halophilic
microorganisms or halophiles grow in saline environment and require a certain minimum
level of salt for continued existence. Halotolerant micro-organisms grow better in
freshwater environments but they can also live in saline environments. Halotolerant micro-
organisms are also able to tolerate high salt concentrations. Woolard and Irvine (1995)
stated that non-halophilic bacteria grow well in a medium containing 1% salt and are the
primary organisms in freshwater and terrestrial ecosystems. The microorganisms can be
classified into four classes depending on the salt concentration, these being: (i) Non-
halophilic, <10 g/L NaCl, (ii) Marine or slightly halophilic, 10 – 30 g/L NaCl, (iii)
Moderately halophilic, 30 – 150 g/L NaCl and (iv) Extremely halophilic >150 g/L NaCl
(Lay et al., 2010). Bassin et al., (2011) observed a more pronounced change in microbial
communities at shock salt loading compared to the gradual loading of salt. They employed
a combination of denaturing gradient gel electrophoresis (DGGE) and sequence analysis of
polymerase chain reaction (PCR)- amplified 16S ribosomal RNA (rRNA) gene fragments
and fluorescent in-situ hybridization (FISH) to validate the PCR-based results and to
observe the dominant bacterial populations. They found nearly 27 bands of
microorganisms which belonged to several phyla, these being , and Proteobacteria,
Bacteroidetes, Chloroflexi, Firmicutes, and Actinobacteria. Furthermore, BAssin et al.,
(2007) reported that protozoa, nematodes, rotifers and filamentous bacteria could not
withstand high salt concentrations.
From the classification of microorganisms it is evident that a number of them can tolerate
high concentrations of salt. These groups of microorganisms contained many aerobic
heterotrophs that are able to biodegrade organic carbon matter from saline water (Lay et
al., 2010). Many aerobic treatments had been employed with different salt concentrations
ranging between 10 - 150 g/L (Lefebvre and Moletta, 2006). Lefebvre and Moletta (2006)
stated that biological treatment of carbonaceous, nitrogenous and phosphorous pollution
has proved to be feasible at high salt concentrations but its efficiency depends on proper
adaptation of the biomass or use of halophilic organisms. Therefore MBR can be a helpful
pretreatment for seawater or brackish water reverse osmosis (RO).
Many studies had been conducted on the use of MBR to treat saline water. A study
conducted by Visvanathan et al., (2002) indicated that a fluidized bed biological granular
activated carbon, at 15 min empty bed contact time effectively removed nearly 100%
biodegradable DOC from seawater. They also investigated the effects of MBR on
biodegradable organic content removal and biofouling control. Their results show that
MBR succeefully removed 78% of DOC and that effluent from the MBR increased the
permeate flux of RO by 300% more than untreated seawater. Artiga et al., (2008) studied
the performance of hybrid MBR for the treatment of saline wastewater from a fish canning
factory (salt concentration was up to 73-83 g/L). They achieved 77% to 92% removal of
COD operating at organic loading rates of 0.3 upto 1.4 kg COD/m3d. Dan et al. (2002)
achieved 80% - 90% removal of COD employing two different types of MBR (yeast MBR
and biological MBR) operated at a high salt concentration of 32 g/L. Conversely, a change
in salinity may also affect the removal of BOD and COD due to an alteration in OLR
(Stewart et al., 1962, Kincannon and Gaudy, 1968). Yogalakshmi and Joseph (2010)
studied the effect of NaCl shock load on the removal efficiency of COD on a bench scale
aerobic submerged MBR. It operated at steady state OLR of 3.6 g-COD/L/d and hydraulic
retention time (HRT) of 8 h. They found almost 95% of COD was removed with a NaCl
shock loading of 5–30 g/L. The removal efficiencies of COD at NaCl shock loading of 50
and 60 g/L were 77% and 64% respectively.
The addition of salt may also have an effect on nutrient removal. Hong et al. (2007) studied
the effect of chloride concentration on the removal of nutrients by anaerobic/anoxic/oxic
(A2O) reactor. They found that a rise in chloride concentration from 150 to 5000 mg/L did
not influence the removal of ammonia nitrogen but it did affect the removal of phosphorus.
Another study by Artiga et al., (2008) showed that when the salt concentration was high
(up to 84 g/L) there was no nitrification eventuated but it did occur with a low salt
concentration of 15 g/L. Uygur (2006) observed a decreased removal efficiency of NH4-N
(from 3.0 to 0.80 mg NH4-N /gbiomass.h) and PO4-P (from 0.36 to 0.08 mg PO4-P
/gbiomass.h) when the salt content increased from 0% to 6% with experiments conducted
with SBR. Sharrer et al., (2007) found 91.8 ± 2.9% to 95.5 ± 0.6% removal of total
nitrogen at different salinity level of 0-32 g/L. They also discovered that the removal of
phosphate was affected by salinity. The removal efficiency of phosphate decreased from
96.1 ± 1.0% to 65.2 ± 5.4% when salinity rose from 0 to 32 g/L. They operated a MBR at a
HRT of 40.8 h, a solids retention time of 64 ± 8 days and food: micro-organism ratio of
0.029 day-1. Yogalakshmi and Joseph (2010) observed a decline in TKN removal
efficiency from 95% to 23% when NaCl shock loads increased from 5 to 60 g/L. They also
found that ammonia removal efficiency fell from 84% to 64% with NaCl shock loads of 5–
30 g/L, which further dropped to 13% at a 60 g/L shock load.
Fouling of the membrane surface is caused mainly by the deposition of organic and
inorganic matter during operation. Hence membrane fouling occurs due to the following
mechanisms (Meng et al., 2009):
• adsorption of solutes or colloids within/on membranes;
• deposition of sludge flocs onto the membrane surface;
• formation of a cake layer on the membrane surface;
• detachment of foulants attributed mainly to shear forces;
• the spatial and temporal changes of the foulant composition during the long-term
operation
Fouling mechanisms for MBR operated at constant flux are presented in Figure 2.3 below.
Figure 2.3. Fouling mechanisms for MBR operated at constant flux is presented in
(Source: Le-Clech et al., 2006)
Membrane fouling can be divided into three scales of fouling (Aryal et al., 2009). Firstly,
there is reversible and residual fouling caused by the deposition of mixed liquor particles
on to membrane surface and sludge build-up in between the fibres in membrane module.
Removable or reversible fouling can take place within very short time (10 min) and rate of
fouling is also high (0.1-1.0 mbar/min) (Table 2.4). Reversible fouling can be controlled by
back flushing or air scouring. Secondly, there is irreversible fouling which cannot be
removed by normal back flushing or air scouring. This can be minimized by means of
chemical cleaning. Thirdly, there is irrecoverable fouling which cannot be recovered by
any types of cleaning and it takes more than several years to achieve (Drews, 2010) and
rate of fouling is much lower than removable fouling (Table 2.4). Typical ranges of
different fouling rates occurring in full-scale MBR are presented in Figures 2.4 and 2.5 and
Table 2.4 (Drews, 2010)
Figure 2.4. Schematic representation of different fouling rates during long-term operation
of full-scale MBRs (Drews, 2010)
Figure 2.5. Schematic illustration of the formation and removal of removable and
irremovable fouling in MBRs (adapted from Meng et al., 2009).
Table 2.4. Typical ranges of the different fouling rates occurring at full-scale MBR
(Source: Drews, 2010)
Category Fouling rate in mbar/min Time frame
Reversible fouling (cake filtration) 0.1-1.0 10 min
Residual fouling 0.01-0.1 1-2 weeks
Irreversible fouling 0.001-0.01 6-12 months
Irrecoverable fouling 0.0001-0.01 Years
The treatment of wastewater by MBR mainly depends on its operating conditions (Le-
Clech et al., 2006) and these are as follows:
• solid retention time (SRT),
• hydraulic retention time (HRT),
• biomass concentration, and characteristics of biomass in-terms of soluble microbial
product (SMP) and extracellular microbial products (EPS),
• temperature,
• mode of operation
• filtration flux, and
• membrane morphology, etc.
The above mentioned factors are not only responsible for the treatment efficiency (in terms
of effluent quality), but are also the major factors causing membrane fouling. The
relationship between various fouling factors and membrane fouling is presented in Figure
2.6. A summary of membrane fouling with the factors influencing fouling is presented in
Table 2.5. Further interrelationship between permeability decline with biomass
concentration, floc size, sludge properties, concentration of EPS, SMP etc., is illustrated in
Figure 2.7. Furthermore, a comparison of relevant conditions and fouling results is shown
in Table 2.6.
Figure 2.6. Factors influencing membrane fouling in the MBR process (Adapted from
Chang et al., 2002)
Table 2.5. Relationship between various fouling factors and membrane fouling (Adapted
from Meng et al., 2009)
Condition Effect on membrane fouling
Sludge condition
MLSS Increase of MLSS concentration, decrease in normalized permeability and increased fouling potential and cake resistance
Viscosity Increase in viscosity, less membrane permeability and more membrane resistance
F/M Increase in F/M ration, increased fouling rate, protein concentration in foulant and increased removable and irremovable fouling
EPS Increase in polysaccharide concentration and fouling rate
SMP SMP is more important than MLSS and is most likely responsible for fouling; polysaccharide in SMP is a possible indicator of fouling
Filamentous bacteria Increase in filamentous bacteria, increase sludge viscosity. Bulking sludge could cause severe fouling to occur
Operating state
SRT Decrease of SRT results in higher membrane fouling. The optimum value of SRT is 20-50 days.
HRT Decrease of HRT increases fouling rate.
Aeration Increase in aeration rate will increase membrane permeability and decrease aeration rate but aggravate membrane fouling.
Permeate flux Sub-critical flux mitigates fouling
Figure 2.7. Inter-relationships between different operating factors and permeability loss in
lab scale MBRs (Source: Drews, 2010).
Table 2.6. Comparison of relevant conditions and fouling results (HF, hollow fibre; FS, flat sheet) (Source: Drews, 2006).
Module type
Nom. pore size
(μm)
Flux [L/(m2 h)]
TS [g/L]
SRT [d] HRT [h]
Total resistance (1012/m)
Fouling rate
[1010/(m d)]
Influence of increasing SRT
Correlation with PS
concentration
HF 0.1 0.8-5 5-26 - 18-110 <500 - - Yes
HF 0.1 13 7-27 10-30 3.3 1.3 - - Yes (with colloids)
HF 0.035 30.6 6.9-8.6 2-10 1.1-3.6 0.12-2.8 0.9-19 Fouling rate -
Tubulara 0.03 60 - - - <2.9 - - No
HF 0.1 17.5-20 3.2-8 8-40 4.5-12 1.4-7.7 4-40 Fouling rate ,
colloids contribution to Rtotal
No/Yesa
HF 0.1 19-21 7.1-14.1 8-14.8 11 2.3-6 3-17 PS concentration Yes
FS pilot 0.037 10 3.2-12 30
(irreg. wastage)
11 3-20 0-90 - No
FS 0.25 12.5 4.6-10 20-100 8 - - Less bound EPS,
spec. cake resistance
Yes (with bound EPS)
FS 0.22 10 3-8 30-120 12 <20 - - No (with SMP)
HF 0.05 12 12 23-40 9-12 - - Less bound EPS, filtration indexb No
FS lab 0.037 6-9 9 22-31 12-14 2.8 (average) 2-22 - No FS pilot 0.037 10 10 28-35 11 2.2-10 0-14 - No HF 0.4 25 2.7-7.1 17-102 8.5 - - Less SMP No
HF 0.04 12 2.5 10 6 - - - Yes (with SMP and protein/PS ratio)
The given maximum resistances are the ones when filtration was ceased before membrane cleanings were carried out. a Data from investigations in a non-aerated test cell (DFCm) instead of in the plant. b Determined in a dead-end stirred cell (150 kDa) at constant TMP (1 bar). Note: In this table the fouling rate is presented as the development of total membrane resistance per day (i.e. the rate of increase of membrane resistance per day)
A common strategy to minimise fouling and sludge accumulation on the membrane surface
is to provide aeration (air sparging) close to the membrane surface thereby inducing a local
shear stress which creates a favourable hydraulic distribution for mixing the sludge and
scouring the membrane surface (Bouhabila et al., 2001). However membrane aeration
forms a significant portion of the MBR operating cost (Cui et al., 2003; Judd, 2007).
Optimization of the aeration is always important in MBR operations. On the other hand
after a certain rate of aeration, the reduction of fouling is negligible i.e. sustainable flux
does not improve significantly. This can be explained by the important effect of induced
cross-flow that causes particles to be transported away from the membrane surface (Kim
and DiGiano, 2006). On the other hand, membrane fouling can also be minimized by
introducing a medium in suspension in the reactor. Thus, it may be possible to prevent
some fouling from occurring by 1) changing the operating parameters of the MBR under
specific conditions such as aeration, 2) using a adaptive membrane cleaning method and 3)
pre-treating the biomass suspension to limit its fouling propensity (through the addition of
adsorbent).
One of the most common strategies to reduce and control sludging/fouling is to provide
aeration (air sparging) close to the membrane surface thereby inducing local shear stress
which controls fouling and creates a favourable hydraulic distribution throughout the
fibre/sheet network (Bouhabila et al., 2001). However, aeration has high energy cost
which could be up to 70% of the total energy expenses (Drews, 2010). Thus, membrane
aeration forms an important part the operating cost of the MBR (Cui et al., 2003; Judd,
2007) and it is important to optimize the membrane aeration process. It is commonly
accepted that air bubbling close to the membrane is one of the most efficient means of
minimizing fouling and ensuring a sustainable operation (Wicaksana et al., 2006; Meng et
al., 2008).
Aeration is important in MBR both for biological oxidation of organic matter and
membrane defouling. Parameters related to aeration are the bubble size and shape, density
and viscosity of wastewater, internal circulation, temperature and presence of surface
active compounds (Malysa et. al., 2005). The air bubbles, after forming inside the
bioreactor tank, accelerate immediately and until it reaches the terminal velocity where the
forces acting on the bubble are balanced. Bubble dissolution in a reactor is a function of
the bubble size, liquid viscosity and its aeration rate (Baral, 2003). The presence of
contaminants such as surfactants can significantly change the bubbles dissolution rate,
shape and effective velocity (Lio and McLaughlin, 2000, Takemura, 2005). Recent studies
separate the submerged membrane reactor from the bioreactor tanks to achieve maximum
efficiency of aeration to remove foulants and to minimize the cost of aeration for biological
oxidation (Lebegue et al., 2007). In theory, small bubbles are better for biological
oxidation while large bubbles facilitate membrane defouling. Sofia et al. (2004) reported
that smallest bubble could produce the best performance, whereas Madec (2000) concluded
that the bubble’s size had no effect on membrane performance. A number of studies have
been done on the effect of aeration rate/amount of aeration required to reduce membrane
fouling (Ueda et al., 1997; Liu et al., 2003). They state that after a critical rate of aeration,
it has no effect on membrane flux. The critical aeration rates that have been reported were
in the range of 0.0048 – 0.010 m3m−2 s−1 (for MLSS concentrations varying form 2 – 10 g
L−1 and flux of 10 – 20 Lm−2 h−1). In addition, the air flow rate per membrane surface area
is reported to be 0.18 to 1.28 N m3/m2 h whereas the air flow rate per permeate flow
produced 10 to 65 m3/m3 (Drews, 2006). Some examples of MBRs are presented in Table
2.7 with their operating conditions and aeration rates.
Table 2.7. Aeration conditions for different full-scale MBRs (Source: EUROMBRA, 2006)
Membranes System Capacity Flux Aeration Conditions
(m3/day) (L.m-2.h-1) (m3.m-2.h-1)
Flat Sheet
Kubota 1.9 - 13 20 - 33 0.56 – 1.06
Brightwater 1.2 27 1.28
Toray 0.53 – 1.1 21.6 – 25 0.4 - 0.54
Huber 0.11 24 0.35
Colloide 0.29 25 0.5
Submerged plate 0.2-0.45 m
(240m²)
100 (sewage) 0.92
Hollow Fibre
Zenon 48 - 50 18 – 25 0.29 – 0.4
M. Rayon 0.38 10 0.65
USF Memcor 0.61 16 0.18
Asahi-kasei 0.9 16 0.24
KMS Puron 0.63 25 0.25
Submerged HF (0.2 m) 24 (product) 13.3 0.94
Because of the presence of irreversible interactions between soluble compounds or bacteria
and membrane material, membrane fouling cannot be controlled only by aeration. It is thus
important to define the procedure to reduce the concentration of these compounds at the
membrane surface and in solution. Incorporation of supporting media/adsorbents may be
relevant to scour some of the foulant on the membrane surface and capture some of the
fouling-causing organic substances prior to their contact with membrane material.
Thus, as an alternative to a higher aeration rate, membrane fouling could be minimized by
the use of a medium in suspension in the MBR. The use of a suspended medium (such as
activated carbon) could help to adsorb the organic matter and provide higher shearing
stress on the membrane surface. Many studies have been conducted on adsorbents in a
biological treatment tank to investigate their effect on the reduction of membrane fouling
(Guo et al., 2005; Lesage et al., 2008; Siembida et al., 2010; Xing et al., 2012; Jin et al.,
2013). A long-term (more than 600 days) pilot study by Siembida et al. (2010) with the
addition of granular material (Polypropylene) showed that the formation of fouling layers
on the membrane was reduced by the abrasion of granular material. They also reported that
the MBR process with the addition of granular medium can be operated at a 20% higher
flux than the conventional MBR process. A study by Pradhan et al. (2012) found that the
use of granular support medium in suspension in a submerged membrane reactor reduced
membrane fouling by around 85%. Akram and Stuckey (2008) observed that adding
powdered activated carbon (PAC) in the submerged anaerobic membrane bioreactor helped
to remove slowly biodegradable organics (both low and high molecular weight (MW)
residual COD). They also reported almost 4.5 times (filtration flux increased from 2 to 9
L/m2•h) higher filtration flux with the addition of PAC of 1.6 g/L than without any
addition.
Fang et al. (2006) had studied the effect of addition of activated carbon on the fouling of an
activated sludge filtration system. They found that the filtration resistance was reduced by
22% (from 6.4 ± 0.5×1012 m–1 to 5.0 ± 0.1×1012 m–1) with the addition of activated carbon.
Furthermore Li et al. (2005) reported that adding PAC in a SMBR helped to reduce
membrane resistance by 44%. Researchers also reported that the addition of adsorbent
help to remove a majority of soluble organic compounds that cause irreversible membrane
fouling (Guo et al., 2005; Chen et al., 2006; Lesage et al., 2008; Shanmuganathan et al.,
2015). Another study reported the adsorption of EPS on PAC on the operation of
submerged hybrid PAC-MBR (Kim and Lee, 2008). Studies by Sombatsompop et al.
(2006) and Guo et al. (2008) have shown a membrane-coupled moving bed biofilm reactor
(M-CMBBR) yielded a lower rate of biofouling than a conventional MBR. From the above
discussion it is evident that using the suspended medium in MBR could help to minimize
membrane fouling by: firstly adsorbing organic matters; and secondly providing extra
shearing stress on the membrane surface.
A periodic backwash during a membrane filtration process is an effective tool to control
reversible membrane fouling. Periodic backwashing help to reduce pressure build up and
thus help to prevent flux decline during membrane operation. However, it is important to
optimise the backwash for successful long-term operation of a membrane system (Smith et
al., 2006). The optimum backwash can be expressed as a function of the concentration of
the foulant, permeate flux and the operational temperature (Smith et al., 2006).
Optimisation of backwash is required to reduce energy requirement and to increase
membrane efficiency. Smith et al. (2006) reported that through an automated backwashing
system it is possible to control backwashing frequency, which reduce energy requirement
and increase productivity. Usually backflushing is done from the permeate side in the case
of hollow fibre modules and relaxation is applied for flatsheet modules for approximately
15–60 s every 3–12 min of filtration whereas frequent cleanings or maintenance cleanings
are conducted approximately every 2–7 d. The main cleanings are done once or twice a
year (Drews, 2010). On the other hand Le-Clech et al. (2006) recommended less frequent,
but longer backwashing (600 s filtration/45 s backwashing) which is more efficient than
frequent backwashing (200 s filtration/15 s backwashing). In addition to periodic
backwashing and relaxation chemical cleaning is also needed. According to extant
literature chemical cleaning may include the following phases (Le-Clech et al., 2006):
• Chemically enhanced backwash (on a daily basis),
• Maintenance cleaning with higher chemical concentration (weekly), and
• Intensive (or recovery) chemical cleaning (once or twice a year).
In this section a detailed literature review on the removal of nutrients using different types
of ion exchange resins/ adsorbed is carried out. It is sustainable to operate MBR at high
rate only to remove BOD and recover nutrients using a post-physico-chemical process such
as ion exchange or struvite process.
From the literature review it is evident that the MBR process can remove most organic
matter and nutrients from water but it fails when exposed to peak and variable loads
depending on the operating condition or state. Other disadvantages are the need for larger
reactor volumes, higher operating costs, and waste sludge production when phosphorus
removal is achieved by chemical precipitation. Phosphorus can also be removed by
biological processes and while they are environmentally friendly, the mechanism is
complex. Furthermore one cannot remove any phosphorus below a particular concentration
and consequently it is better to remove and recover nutrients by a post–physico-chemical
treatment such as ion-exchange.
Adsorption is a good technique for the removal of trace amounts of solute from aqueous
solution. Beler-Baykal and Guven (1997) used clinoptilite as an ion exchange for the
removal of ammonia from wastewater. Samatya et al. (2006) used selective ion-exchange
resin (purolite A520E) to remove nitrate from the water. They reported more than 95%
removal of nitrate at a resin dose of 0.2 - 0.5 g-resin/50ml (4 - 10 g resin per L). Nur et al.,
(2012) also reported 80% - 90% removal of nitrate by using ion exchange resin of purolite
A520E and A500P in the column study from synthetic wastewater. Other researchers also
reported good removal/uptake of nitrate (almost 90%) using different types of purolite
(purolite A100, A 300 and A520E) and with zero-valent iron (Fe0) (Bae et al., 2002;
Westerhoff, 2003; Primo et al., 2009; Bulgariu et al., 2010). Different ion exchange resins
and adsorbents have also been used to remove phosphorous from wastewater such as
purolite A500P, purolite A520E, Purolite FerrIX A33E, amberlite IRA910Cl (a strong
basic macroreticular anion exchange resin) and amberjet 1200Na (a strong acid cation
exchanger), aluminium oxide, iron oxide, zirconium oxide and hydrotalcite, (Chen et al.,
2002, Terry, 2009; Nur et al., 2012; Nur et al., 2014). Additionally, oxides of many
polyvalent metals, namely, Fe3+, Ti4+ and Zr4+ exhibit very favourable ligand sorption
properties for phosphorus through the formation of inner sphere complexes (Stumm and
Morgan, 1995; Dutta et al., 2004). Adsorbents such as aluminium oxide, iron oxide,
zirconium oxide, hydrotalcite etc are widely used for the removal of phosphate anions from
water and wastewater (Chen et al., 2002; Terry, 2009). Similarly zirconium salts (Zr4+)
have also been used for phosphorus recovery (Lee et al., 2007).
Ion-exchange processes using selective ion-exchange materials (such as purolite) are ideal
for reducing nitrate and phosphate to near-zero levels provided that the ion-exchange resin
is ammonia and/or phosphate selective, cost effective and amenable to efficient
regeneration and reuse. The ion exchange resins or adsorbent are used as filter media in a
filter-based system and after a period of usage they become saturated with phosphate and
nitrate and their removal efficiency declines. Once they are exhausted the sorbents can be
regenerated using a different regeneration reagent or a mixture of such reagent- such as
NaCl, NaOH, CaCl2, HCl, Na2CO3 etc (Loganathan et al., 2014). Phosphate so removed
can be recovered by precipitation with calcium/or magnesium salts and employed as
phosphate fertilizers. Alternatively they can be diluted with irrigation water for fertilizing
irrigated crops (Loganathan et al., 2014).
This study investigated the effect of organic loading rate (OLR), imposed flux, aeration
rate, and addition of suspended media on the removal of organic matters and nutrients and
on the membrane fouling in membrane bio-reactor (MBR). In addition, the effect of shock
and gradual loading of salt on MBR performance was also investigated. A set of
experiment was also carried out on the removal and recovery of nutrients by ion exchange
from high rate MBR effluent. The water characteristics, experimental and analytical
procedures are described in this chapter.
The MBR was initially seeded with mixed liquor obtained from a domestic sewage
treatment plant in Sydney, Australia. The sludge having a mixed liquor suspended solid
concentration of 3 - 4g/L was used during start-up of each experiment. The diffused
aeration system was used for mixing the sludge and supplying the air. The activated sludge
was acclimatized in a continuous MBR process for 40 days or more (based on the
operating condition). The concentration of MLSS, MLVSS, DOC, COD, ammonia and
nitrate was also monitored regularly. The pH was maintained at 6-8. A predetermined
amount of sludge was withdrawn to maintain the SRT of 20 or 40 days.
The bioreactor was continuously fed with a synthetic feed consisting of ethanol, beef
extract and peptone as an organic carbon source which is easily degradable and mineral
salts containing nitrogen and phosphorus as source of nutrients in the ratio of COD: N: P
ratio equal to 150 : 5 : 1. Ethanol (analytical reagent) was used as sources of organic
carbon which is easily bio-degradable. Analytical reagent inorganic salts were used as
source of N (NH4Cl) and P (KH2PO4). Analytical reagent NaHCO3 was used to maintain
the pH. The beef extract and peptone were used as source of protein. The beef extract and
peptone used in this study was also analytical reagent. The composition of synthetic
wastewater is given in Tables 3.3-3.4 for different operating conditions. The synthetic feed
was prepared using analytical grade (pure) chemicals where the amount of TSS is very
low. Moreover, the synthetic feed contained mostly biodegradable dissolved COD which is
easily biodegradable. On the other hand the real feed contains wide range of physico-
chemical and microbiological pollutants. The suspended solid concentration in real feed
varies significantly and also contains both particulate and non-biodegradable COD.
However, the synthetic feed has advantages over real feed as its physico-chemical and
microbiological characteristics can be kept constant. Further, it is also easy to vary the
influent concentration of the substrate in the synthetic feed by changing the quantities of
the ingredients.
The schematic diagram of submerged membrane bioreactor (SMBR) with and without
suspended media used in this study is shown in Figure 3.1. The suspended medium used in
this study was granular activated carbon (GAC: 150 - 300, 300 - 600 and 600 - 1200 μm).
Coal-based premium grade GAC (MDW4050CB) was supplied by James Cumming &
Sons Pty Ltd (GAC, particle size of 300 - 600 μm; average pore diameter was 30 Å;
surface area was ~1000 m2/g. This study targets the use of granular activated carbon
(GAC) to improve the process efficiency by increasing the permeate flux without the need
to increase aeration rate to reduce the irreversible membrane fouling. Further, GAC as a
suspended medium was rarely (or not) used in previous studies. Most of the previous
studies concentrated on the addition of powder activated carbon (PAC) (Guo et al., 2005;
Akram and Stuckey, 2008) mainly to achieve additional removal of organic matter thereby
also reducing the membrane fouling. A flat sheet membrane module with an area of 0.2 m2
made of polyvinylidene fluoride (PVDF) was used. In this study the flat sheet membrane
was used for experiments conducted with and without suspended media. This is because of
its sustainable operation than hollow fibre membrane as suspended media will tend to stack
on the top part (potted area) of hollow fibre module and may break the membrane fibre.
The details of membrane characteristics and operating conditions are presented Table 3.1
and Table 3.2 respectively.
Figure 3.1. Laboratory scale membrane bioreactor (membrane area = 0.2 m2, pore size =
0.14 m, volume of reactor = 10 L)
Table 3.1. Membrane characteristics used in this study
Item Unit Value
Company - A3, Germany
Material - polyvinylidene fluoride PVDF
Membrane area m2 0.2
Membrane pore size μm 0.14
Membrane configuration - Flat sheet
Dimension of this membrane module cm 11.5 (width) x10.5 (length) x22.5 (height).
Number of vertical sheets - 8
Gap between two adjacent vertical
sheets (approximately) cm 1.1
Table 3.2. Operating conditions of laboratory scale SMBR with and without suspended
media
Parameters Unit Values
Volume of reactor L 10
Flux L/m2.h 5-40
HRT h 10-1.7
SRT Days 20
Aeration rate m3/m2membrane area.h 0.5-1.5
pH - 6.5-7.6
MLSS g/L 5-6
MLVSS g/L 4.5-4.8
Temperature o C 22±3
DO mg/L 2.8-4.6
GAC dose g/L (of volume of reactor) 0.5-2.0
At the beginning, the MBR was seeded with 3 L of mixed liquor (sludge) obtained from a
domestic sewage treatment plant. After seeding, the bioreactor was continuously fed with a
synthetic feed consisting of ethanol (as an organic source) and mineral salts containing
nitrogen and phosphorus (as nutrients) in the ratio of COD : N : P ratio equal to 150 : 5 : 1
with an organic load of 1.5 kg CODm-3.d-1. The concentration of the feed was modified
for different imposed flux to keep a constant organic loading rate (OLR) based on the
equation 3.1. In this study, the SRT was kept at 20 days as the aim of these experiments
was to investigate the effect of GAC and flux on membrane fouling reduction. Further,
based on recent experimental studies, lower MLSS concentrations was used in order to
avoid sludge build up in the membrane module (Jang et al., 2006; Kawasaki et al., 2011).
In addition, higher MLSS concentrations can cause operational problems like rapid
clogging of the membrane. Navaratna and Jegatheesan (2011) observed that lower MLSS
concentration of 4–7 g/L was effective in controlling and reducing fouling of the
membrane.
VQCC s
v*
=
where, Cv = organic load (kgCOD/m3.d); Cs = COD (kg/m3), V = volume of reactor (m3)
and Q = feed rate (m3/d)
This means, with the decrease of Q (increase in HRT/ decrease of flux), COD of the feed
(or the additional organic source) have to be increased with increase in HRT/decreased of
flux in order to maintain the organic loading rate. The composition of synthetic wastewater
is given in Table 3.3.
Table 3.3. Composition of synthetic wastewater
Components Unit Values
Ethanol mL/L 0.06-0.37
NH4Cl mg/L 13-80
KH2PO4 mg/L 3-20
Beef extract mg/L 1.8
Peptone mg/L 2.7
NaHCO3 mg/L 600
3.1
The schematic diagram of the MBR (used to study the effect of OLR and salinity) is
presented in Figure 3.2. Here a hollow fibre membrane module with an effective
membrane area of 0.2 m2 was used. The details of membrane characteristics and operation
parameters are presented in Table 3.4 and 3.5 respectively.
Figure 3.2. Experimental set up of membrane bioreactor (MBR)
Table 3.4. Membrane characteristics used in this work
Item Unit Value
Company - MANN+HUMMEL, Singapore
Material - hydrophilic modified PAN
Membrane are m2 0.20
Membrane pore size μm 0.10
Membrane configuration - Hollow fibre
Outer diameter mm 2.1
Inner diameter mm 1.1
Table 3.5. Laboratory scale hollow fibre MBR operated at different OLRs and salinity
Parameters Unit Values
Volume of reactor L 4
Flux L/m2.h 2.5
HRT h 8
SRT Days 40
Aeration rate m3/m2membrane area.h 1.5
pH - 6.5-7.6
MLSS g/L 3.5-5.3
MLVSS g/L 3.0-4.3
Temperature o C 22±3
DO mg/L 2.8-4.2
The MBR was initially seeded with 3 L of mixed liquor obtained from a domestic sewage
treatment plant in Sydney, Australia. Following initial seeding, the MBR was continuously
fed with a synthetic substrate made up of tap water. The COD : N : P ratio of synthetic
wastewater was maintained at 150 : 5 : 1. Different OLRs in the range 0.5 - 3.0
kgCOD/m3.day were used. The composition of synthetic wastewater is presented in Table
3.6. The concentration of the feed was modified for different organic loading rate (OLR)
based on equation 3.1.
Table 3.6. Composition of synthetic wastewater
Components Unit Values
Ethanol mL/L 0.10-0.40
NH4Cl mg/L 21.2-127.35
KH2PO4 mg/L 4.8-29.26
Beef extract mg/L 1.8
Peptone mg/L 2.7
NaHCO3 mg/L 600
The objective on the study the effect of salt concentration on the MBR performance was to
understand the effect of continual increase of salt concentration in treating saline
wastewater (such as wastewater from cheese industry) in continuous membrane bioreactor
process. After acclimation the MBR experiment was conducted with activated sludge (with
and without salt) to investigate the effect of gradual loading of salt (from 0 to 35 g-
NaCl/L). Initially, the MBR was run (after the acclimation period of 45 days) for around
10 days without salt following which 0.5 g/L of NaCl was added and the MBR was run for
around 11 days. The main focus of this study was to assess the short term effect of gradual
increase of salt concentration i.e mixing/leaching of salt water or high saline water in
biological process such as MBR without acclimatized the MBR with saline seed. As such,
activated sludge was used as seed material. The salt concentration was thus increased in
sequence as showed in Figure 3.3 with an elapse time of 10 - 11 days for each
concentration of salt. The profile of gradual loading of salt in MBR is presented in Figure
3.3. The gradual increase of salt concentration was made to study the effect of salinity
without changing other operating parameters such as SRT. The choice of 10-11 days was
to see the short term effect (less than acclimatization period) without providing any
addition time for acclimatization between each gradual increase of salt concentration. The
solid retention time (SRT) and hydraulic retention time (HRT) was kept constant at 40
days and 8 h (corresponds to an operational flux of 2.5 L/m2.h) respectively. The COD : N
: P ratio of synthetic wastewater was maintained at 150 : 5 : 1 at an OLR of 0.25
kgCOD/m3.d. The aim of this study was to assess the effect of gradual increase of salt
concentration on the organic and ammonia removal by MBR. Thus, in this study the OLR
fixed at 0.25 kg COD/m3 d. The concentration of the feed was modified based on equation
3.1.
Although, the results of laboratory scale membrane bioreactor can be different from that of
a proto-scale MBR plant, a laboratory scale study is the only way to assess the system
performance because at least some variables of the operational parameters in the complex
network of the MBR design can be fixed at a constant value (Drews, 2010). Furthermore,
the presentiveness of the results from lab scale will help in the design of a pilot scale plant.
The flux was low to keep the HRT at 8 h as the volume of reactor used in this study was
only 4 L. The low flux value also can be useful to treat high strength industrial wastewater.
For instance, generally in a conventional MBR treatment system higher nitrification takes
places when the HRT is around 6 h or more (Viero et al., 2007). In addition, the
operational flux can be increased by using a reactor of larger volume. For example, instead
of using a reactor of 4 L capacity, the operational flux could be increased to 10 L/m2.h with
a 16 L capacity of reactor by keeping the HRT the same as in the present study (8 h).
To study of effect of OLR and salinity, a relatively higher SRT of 40 days was employed
as higher SRT will improve nitrification than shorter HRT. The OLR was increased almost
6 times and salinity was also increased from 0 to 35 g/L
Figure 3.3. Profile of gradual loading of salt in MBR
The schematic diagram of the MBR (used for organic carbon removal) followed by the
purolite ion-exchange column (used for nitrogen (N) and phosphorous (P) removal) is
presented in Figure 3.4. The purpose of this experiment is to remove and recover the
nitrogen and phosphate using ion exchange column as post treatment. A high rate MBR
experiment was carried out at a 4 hour HRT only to remove only organic carbon allowing
the nitrogen and phosphorus remaining in the MBR effluent for possible recovery in ion
exchange process. Details of membrane and synthetic water characteristic used in this
study are given in Tables 3.5 and 3.6 respectively.
0
5
10
15
20
25
30
35
40
0 10 20 30 40 50 60 70 80 90 100 110
Salt
conc
entr
atio
n (g
-NaC
l/L)
Time (days)
(b)
Figure 3.4. Experimental set up membrane bioreactor (MBR) followed by purolite
(A500P and A520E) ion-exchange column)
After the acclimation period, permeate from the MBR was collected and passed through an
ion-exchange column packed with purolite to a depth of 6 cm. This ion exchange column
had an internal diameter of 2 cm and the volume of purolite used was 18.9 cm3. An up-
flow mode of filtration was employed with a filtration velocity of 2.5 m/h. Commercially
available macroporous anion-exchange resins (purolite A500P and A520E) were used.
The characteristics of Purolite A-500P and A520E are given in Table 3.7. A shorter depth
of ion-exchange was used to observe the exhaustion period of ion-exchange in a short
period.
Purolite A500P is designed for use as an organic scavenger, e.g. for the removal of tannins,
fulvic and humic acids, from domestic effluents. It was found to have a good phosphate ion
exchange capacity. The Purolite A-520E is a macroporous strong base anion resin which is
specially designed for the removal of nitrates from water for potable purposes. The
macroporous matrix and special ion-exchange group functionality impart ideal nitrate
selectivity to Purolite A-520E making this resin particularly suitable for nitrate removal
even when moderate to high concentrations of sulphate ions are present.
Table 3.7. Typical chemical and physical characteristic of A-500P and A520E
Parameters A500P A520E
Polymer Matrix Structure Macroporous Styrene-Divinylbenzene
Macroporous Styrene-Divinylbenzene
Physical Form and Appearance Opaque Near-White Spheres
Opaque Cream Spherical Beads
Functional Groups R-(CH3)3N+ Quartenery Ammonium
Ionic Form (as shipped) Cl- Cl-
Screen Size Range (British Standard Screen) 14-52 mesh, wet 16-50 mesh, wet
Particle Size Range (microns) +1200 <5 %, -300 <1% +1200 μm <5%, -300 μm <1%
Moisture Retention, Cl- form 63-70% 50-60%
Reversible Swelling Cl- ® OH 15% Negligible
Specific Gravity, Moist Cl- Form 1.06 -
Total Exchange Capacity, Cl-
Form (wet, volumetric) 0.8 eq/l min 0.9 eq/l min
pH Range (Stability), Cl- Form 0-14 0-14
(Operating), Cl- Form 5-10 5-10
A schematic diagram showing the submerged MFAH system is presented in Figure 3.2. A
set of MFAH experiment was conducted with the addition of zirconium (IV) hydroxide as
an adsorbent for the removal of phosphorus from wastewater. The properties of membrane
used in this study are presented in Table 3.4.
The reactor tank of 4 L capacity was filled to 3 L with wastewater and the membrane
module was placed in the centre of a tank just above the aerator plate. Air bubbles were
continuously injected at a fixed rate of 1.8 m3/m2membrane area.h from the bottom of the tank
which was predetermined and found to be effective enough to keep the adsorbent in
suspension.
The submerged membrane microfiltration was operated both with and without the addition
of adsorbent (Zr hydroxide) in suspension. Zirconium (IV) hydroxide was obtained from
Sigma-Aldrich, USA. It had a particle size range of 0.1 - 35 μm and density of 1100-1300
(kg/m3). Predetermined quantities of adsorbent based on adsorption isotherm results were
added to the tank prior to the experiment’s commencement. The adsorbent doses used were
between 1 and 5 g/L of the volume of the reactor. In the initial experiments the adsorbent
was added to the tank only once, i.e. at the beginning of the filtration test. In the
subsequent experiments the adsorbent was repeatedly applied to continuously provide new
sites for phosphate adsorption. In these experiments, 10% of adsorbent (1.5 g) was added
with 50 mL water every 24 h and the same amounts of used adsorbent and effluent volume
were removed from the bottom outlet. This made it possible to keep constant the total
amount of adsorbent and volume of water in the reactor.
This experiment was performed with synthetic wastewater which represents all the
characteristics of the biologically treated sewage effluent (BTSE). The characteristics of
wastewater are given in Table 3.8. The synthetic wastewater had a relatively low
concentration of phosphate-P, nitrate-N and sulfate-S ( 1 mg/L). Consequently, this
synthetic wastewater was spiked with phosphate (2-20 mg-P/L), nitrate (10-30 mg-N/L)
and sulfate (10-30 mg-S/L) using analytical grade reagents such as K2HPO4, KNO3 and
Na2SO4. The objective was to study the adsorption of these anions on Zr hydroxide from
solutions containing varied concentrations of these anions.
Table 3.8. Concentrations of the constituents of synthetic wastewater
Compound Concentration (mg/L) Fraction of DOC
Beef extract 1.8 0.065
Peptone 2.7 0.138
Humic acid 4.2 0.082
Tannic acid 4.2 0.237
Sodium lignin sulfonate 2.4 0.067
Sodium lauryl sulfate 0.94 0.042
Arabic gum powder 4.7 0.213
Arabic acid(polysaccharide) 5.0 0.156
(NH4)2SO4 7.1 0
K2HPO4 7.0 0
NH4HCO3 19.8 0
MgSO4.7H2O 0.71 0
Equilibrium adsorption experiments were studied at various adsorbent doses (Zr
hydroxide) (0.1 to 3 g/L) and using a phosphate concentration of 10 mg-P/L. The
suspensions were shakenat a speed of 120 rpm for 72 h in a flat shaker. The adsorption
equilibrium studies were conducted at different temperatures (20 - 60 oC), pH (4 - 10) and
in the presence of coexisting ions (nitrate and suphate).
The amount of phosphate adsorption per unit weight of Zr hydroxide at equilibrium, qe
(mg/g), was calculated using Equation (1),
( )M
VCCq ee
−= 0
where C0 = initial concentration of adsorbate (mg/L); Ce = equilibrium concentration of
adsorbate (mg/L); V = volume of the solution (L) and M = mass of adsorbent used (g).
The experimental results were treated with the Langmuir isotherm model (Nur et al.,
2014b; Equation 3.3). Langmuir model fitted better than the Freundlich, as such only
Langmuir model is described in this study.
CKCKqq
eL
eL
e +=
1max
where Ce = the equilibrium concentration of adsorbate (mg/L), qe= the amount of adsorbate
adsorbed per unit mass of adsorbent (mg/g), qmax = the maximum amount of adsorbate per
unit weight of adsorbent (mg/g), and KL = Langmuir constant (l/mg)
Batch kinetic experiments were studied with Zr hydroxide (0.5 - 5.0 g/L) in glass flasks
containing 100 mL of phosphate solution of concentration 10 mg P/L. The suspensions
3.2
3.3
wereagitated in a flat shaker at a speed of 120 rpm at 22 ± 2.0 oC for 5 h. Samples were
taken at various time intervals and phosphate concentrations were measured. The amount
of P adsorption at time t, qt (mg/g) was calculated using Equation (3):
( )M
VCCq tt
−= 0
where C0 = initial concentration of adsorbate (mg/L); Ct = concentration of adsorbate at
time t (mg/L); V = volume of solution (L); and M = mass of dry adsorbent (g).
The adsorption kinetic data were analyzed using pseudo-first order and pseudo-second
order models (Nur et al., 2014). The equations for these models are as follows:
(i) Pseudo-first order model
( )qqk
dqte
t
dt−= 1
3.5
where qe= amount of phosphate adsorbed at equilibrium (mg/g); qt = amount of phosphate
adsorbed at time, t (min) (mg/g); and k1 = rate constant of pseudo-first order adsorption
model (1/min).
(ii) Pseudo-second order model
( )qqk
dqtedt
t −=2
2 3.6
where qe= amount of phosphate adsorbed at equilibrium (mg/g); qt= amount of phosphate
adsorbed at time, t (min) (mg/g); and k2 = rate constant of pseudo-second order model
(1/min).
3.4
The measurement of mixed liquor suspended solids (MLSS) and mixed liquor volatile
suspended solids (MLVSS) were conducted as follows: initially the mixed liquor was
filtered through 1.2 μm filter paper. After the filter paper was kept in the oven at 100 0C
for 3 h, it was then kept in a desiccator for 24 h before the measurement of suspended solid
(which is MLSS).
After the measurement of MLSS, the filter paper containing MLSS was then kept in the
furnace oven for 30 min. The temperature in the furnace oven was set at 550 0C. After 30
min, the filter paper was placed in the desiccator for 24 h. Afterward the filter paper was
taken out from the desiccator and was subjected to measurement of the volatile suspended
solids (which is MLVSS). In this study, we have used 1.2 μm filter paper mainly to collect
the suspended solids (which will also comprise of microbes). The suspended solid
concentration was measured twice a day and average values were reported. A filter paper
of pore size of 1.2 μm is used normally for suspended solid measurement (APHA, 1995).
A few experiments were conducted with GAC in suspension. The MLSS and MLVSS
concentration with GAC were determined as follows: a known volume of mixed liquor
with GAC was collected and was put in rotary shaker at 150 rpm for 30 min to detach the
biomass attached on the GAC medium. It was then filtered using a sieve with a pore size of
150 μm. As the GAC particles used were 150 μm or more, the GAC particles were retained
on the sieve. The sieve was also carefully washed several times using distilled water to
make sure that no biomass was attached on the sieve. The mixed liquor after sieving was
collected and was subjected to MLSS and MLVSS by filtering and drying as described
earlier in this section. In both cases (with and without GAC) the MLSS and MLVSS
measurement was triplicated and the average value was reported.
SVI was determined by standard methods (APHA, 1995). The measurement of SVI was
done with mixed liquor containing both the biomass and GAC. However, the presence of
GAC did not have any significant effect on the SVI measurement as most of the GAC
particles settled down within a short time (less than a minute). The validation of this
procedure was checked as follows: initially the SVI was measured for mixed liquor after an
acclimation period of 40 days before the addition of GAC. This was compared with the
SVI of mixed liquor with the addition of GAC for the same mixed liquor. The SVI values
in both cases were almost the same with a variation of less than 2%.
The SOUR was calculated as follows: firstly OUR (oxygen uptake rate) was calculated
using the YASI5300A biological oxygen monitoring system. Prior to OUR measurements,
the instrument was calibrated according to the procedure described in the operations
manual. After the calibration percentage of DO (dissolved oxygen) with time was recorded
and the amount of DO (mg-O2/L) was calculated. Several measurements were made until
DO levels dropped below 1 mg/L. From this calculation, a graph was plotted based on the
observed reading (mg-O2/L) vs time (min). The slope of the linear best fit plot was the
amount of oxygen consumed in mg/L/min. From the measurement values of MLVSS and
OUR, the SOUR was calculated in mg/g-MLVSS.h.
Samples of the mixed liquor were frequently collected and analysed for SMP and EPS
using the method described by Le-Clech et al. (2006) presented in Figure 3.5. Initially the
mixed liquor was centrifuged at 5000xg for 5 min. Here g is the acceleration due to gravity
(9.81 m/s2). After centrifugation, the supernatant was taken out and filtered through 1.2 μm
filter. The filtrate is referred as SMP. After the collection of SMP, same volume of
deionised water of that of supernatant was added to residue and mixed for 10 min using
ultra sonication. This was followed by heating at 80 0C for 10 min and centrifuged for 10
min at 7000 x g. The supernatant was then taken out and filtered through 1.2 μm filter
which was referred as EPS. EPS are usually colloidal materials that contain construction
materials for microbial aggregates and a wide range of organics such as polysaccharides,
proteins, lipids, aminosugars, nucleic acids, other polymeric compounds, low molecular
weight acids and neutrals (Le-Clech et al., 2006). However not all the colloidal matters
present in mixed liquor can be EPS as SMP also contain soluble macro-molecules, colloids
and slimes (Jang et al., 2006). More details of SMP and EPS are discussed in introduction
section.
Figure 3.5. Extraction procedure of SMP and EPS from mixed liquor samples
Nutrients measurement such as phosphate (PO43-P), nitrite (NO2-N), nitrate (NO3-N) and
ammonium (NH4+-N), were carried out using the cell test method (Spectroquant, Merck)
and a photometer (NOVA 60, Merck). For some samples phosphate, nitrate and sulphate
were analyzed using a Metrohm ion chromatograph (model 790 Personal IC). The IC was
equipped with an auto sampler and conductivity cell detector. Na2CO3 (3.2 mmol/L) and
NaHCO3 (1.0 mmol/L) were used as a mobile phase with 0.7 mL/min flow rate. The
chemical oxygen demand (COD) was measured using COD reagent and a photometer by
the EPA 410.4 method.
DOC concentration of the mixed liquor, effluent and foulant for different HRT’s was
measured after filtering through 0.45 m filter using Multi N/C 2000 analyzer (Analytik
Jena AG).
MWD measurement were conducted to classify the nature of organic matter (OM) in terms
of molecular weight (in the range less than a few hundred Daltons to >35 kDa). High
pressure size exclusion chromatography (HPSEC, Shimadzu Corp., Japan) with a SEC
column (Protein-pak 125, Waters Milford, USA) incorporated with UV detector (254 nm)
was used to identify the MW distributions of OM. Polystyrene sulfonates of different
molecular size (1000, 1800, 4600, 8000, and 18000, 35000 and 75000 daltons) was used as
standards for calibration.
Excitation emission matrices (EEMs) were obtained using a spectrofluorometer (Varian
Eclipse) with a wavelength range of 280-500 nm by increasing the wavelength by 5 nm for
excitation and emission. When experiments were run at different aeration rates, the
selected fluorescent intensities (excitation:emission) were picked up and compared for the
abundance of fouling substances.
Size exclusion liquid chromatography with carbon detector, (LC-OCD, liquid
chromatography-organic carbon detection) a TSK HW 50-(S) column and a 0.028 molL-1
phosphate buffer were used to measure the hydrophilic and hydrophobic fractions of the
organic matters. It provides quantitative information on organic matter as well as
qualitative information on molecular size distribution of organics present in the
wastewater.
X-ray diffraction (XRD) was measure using a XRD Shimadzu S6000 (Japan)
diffractometer. The X-ray diffraction unit (Theta/2Theta) had a Cu target operated at 40 kV
and 30 mA. This was set at 5–45° 2-theta, step time 2° min-1, 25 °C.
FTIR was measured using IRaffinity-1 (Shimadzu, Japan) with a Zn/Se cell at room
temperature.
Total membrane resistance (Rt) were measured using the following Darcy’s equation
JPRt μ
Δ=
Equation 3.7 can be written as follows
)( pbscm RRRPJ
++Δ=
μ
Where, Rt (R t = Rm+Rsc+Rpb) is the total resistance (1/m), which is the combination of the
three resistances of the intrinsic membrane resistance (Rm; 1/m), the resistance of the
sludge layer deposited on the membrane surface (Rsc; 1/m), and the pore blocking
resistance (Rpb; 1/m) caused by solute adsorption on the pores of the membrane. J is the
permeation flux (m/s), P is the TMP (Pa), μ is the viscosity of the permeate (Pa.s),
The Rm, Rsc, and Rpb were calculated from equation 3.3 using the following experimental
procedure. The membrane after each experiment was taken out and submerged in distilled
water and its total resistance (Rt = Rm+Rsc+Rpb) was calculated at different the flux. The
3.7
3.8
membrane was then cleaned with distilled water by gentle shaking that removed the
deposited sludge cake and placed in distilled water again and resistance (Rpb+Rm) was
noted at different flux. Finally, the membrane was cleaned with chemicals and the
membrane resistance (Rm) was evaluated.
Membrane foulant attached onto the membrane surface was extracted using 0.5% (w/v) of
NaOH solution. After each experiment, the membrane was put in a container containing 2
L of NaOH solution (concentration mentioned above). The container was then placed on a
horizontal shaker for 3 h at 100 rpm. The foulant extracted solution was then analysed after
filtering it through 1.2 μm filter paper.
After extraction of the foulant, the membrane was then put into citric acid solution for 3 h
and was shaken using a horizontal shaker at 100 rpm. 0.5% of citric acid was used to
prepare citric acid solution. Afterward the membrane was then kept in sodium hypochlorite
solution (200 ppm) solution for another 3 h and was shaken using horizontal shaker at 100
rpm. Citric acid mainly to remove inorganic fouling and sodium hypochlorite was used to
clean bio-foulant from the membrane surface. Once membrane cleaning was completed,
the measurement of the flux of the clean membrane was performed and it was compared to
that of a virgin membrane. The flux test of virgin membrane was carried out from 1 to 50
L/m2.h. The flux was kept constant using a peristaltic pump and pressure was recorded for
the corresponding flux. After membrane cleaning, the flux test of clean membrane was also
carried out using the same flux values between 1 to 50 L/m2.h and pressure development
was recorded. It was found that the differences between clean and virgin membrane flux
was minimal (less than 5%). Thus, it was assumed that the cleaning procedure completely
regenerated the membrane.
Cluster analysis was conducted (using IBM SPSS statistics 19). Ward’s method of
hierarchical algorithm using the square Euclidean distance as a similarity measure was
used to establish clusters (Poulton 1989).
Organic loading rate (OLR) plays an important role in the treatment of wastewater by a
MBR. An increase of OLR decreases the filterability of the MBR and increase the
membrane fouling. Thus, in this study, a detailed influence of organic loading rate (OLR)
on the performance of a membrane bio-reactor (MBR) was investigated. The MBR was
operated with 6 different OLRs between 0.5 - 3.0 kgCOD/m3.d. The hydrodynamic
parameters of the MBR were kept constant. The hydraulic retention time and sludge
retention time were kept at 8 h and 40 days respectively. A detailed organic matter
characterization of membrane foulant, soluble microbial product and extracellular
polymeric substances were carried out. The organic matters such as bio-polymers type
substances together with humic acid and lower molecular neutral and acids were also
measure to explain their variation with OLR. Experimental setup (Figure 3.2) and
experimental details are provided in chapter 3 (section 3.3.2.).
4.2.1.
In this study the MBR was operated at 6 different OLRs ranging from 0.5 to 3.0
kgCOD/m3.d. The MBR was operated at a constant flux of 2.5 L/m2.h corresponding to an
HRT of 8 h. Although, the flux was low due to small reactor size, the HRT was 8 h which
is within the recommended range. A HRT of 8 h was chosen in order to achieve good
removal of organics and ammonium nitrogen. The aim of this study was mainly to find out
the effect of organic loading rate on organic matter in the effluent and on MBR fouling
without changing any hydrodynamic parameters such as HRT, SRT, Flux, aeration rate.
The only parameter which was changed was influent COD concentration ranging from 165
to 990 mg/L. This is to represent COD values of domestic sewage to biodegradable
medium strength industrial wastewater. The solid retention time was kept at 40 days. The
SRT value of 40 days was chosen based on literature. For example, Adham and Gagliardo
(1998) suggested the use of higher SRT of greater than 30 days, while Grelier et al. (2006)
observed a good bio-degradation of organic and nutrients with a higher SRT of 40 days.
The MLSS and MLVSS concentration was 3.5 - 5.3 g/L and 3.0 - 4.3 g/L respectively.
Although in the earlier research and existing MBR plants, the MLSS concentration was
usually maintained high at 8-12 mg/L, in this study we used a lower MLSS concentration
was also used by Kawasaki et al. (2011) and Jang et al. (2006). The MLSS concentration
was kept low in order to minimize the sludge accumulation between membrane fibres in
the membrane module. In addition, higher MLSS concentrations can cause operational
problems like rapid clogging of the membrane. High MLSS concentrations increase the
sludge viscosity and may also affect the oxygen transfer efficiency (Germain and
Stephenson, 2005). An increase of MLSS concentration from 2.4 to 9.6 g/L led to an
increase of membrane cake resistances ranging from 9 to 22 x 1011 m-1 (Fang and Shi,
2005). Navaratna and Jegatheesan (2011) observed that lower MLSS concentration of 4 to
7 g/L was effective in controlling of fouling of membrane and reducing fouling of the
membrane. Similarly, the F/M ratio was varied from a normally used value of 0.12 to a
high value of 0.57 d-1 to represent a range from low strength domestic waste to high
strength biodegradable industrial wastewater. From Table 4.1 one could also noticed that
an increase of OLR (thus F/M ratio) resulted in a lower consumption of biodegradable
organic matter. This is inspite of an increase of MLSS from 3.5 to 5.3 g/L with an increase
of OLR from 0.5 to 3 KgCOD/m3.d
An increase of OLR resulted in decreased DOC removal efficiency. The DOC removal
efficiency reduced from 93 - 98% to around 45 - 60% when the OLR increased from
between 0.5 - 1.0 to 2.75 - 3.0 kgCOD/m3.d. The removal of NH4-N was high at 83 - 88%
for OLRs of 0.5 - 1.0 kgCOD/m3.d. In the MBR, the removal of NH3-N usually occurred
through conversion of ammonia into nitrite and nitrate by a nitrification process and
through nitrogen loss by denitrification. From the performance result of MBR, it is found
that the conversion of NH4-N into NO3-N was high (83 - 88%) with OLRs of 0.5 and 1.0
kgCOD/m3.d. The conversion of NH4-N into NO3-N decreased from 83 - 88% to less than
67% when the OLR was increased to 2.0 - 3.0 kgCOD/m3.d (Table 4.1). The conversion of
NH4-N into NO3-N can be explained in terms of the ratio between influent NH4-N to
production of NO3-N which is shown in Figure 4.1. From the Figure 4.1, it is found that
the ratio of NO3-N and NH4-N decreased with an increase of OLR. A higher value of NO3-
N/NH4-N indicates higher nitrification (i.e conversion of NH4-N into NO3-N is high). This
could be due to the effect of competition between heterotrophic and autotrophic (nitrifying
organisms), as more organic carbon was present in the reactor under higher OLRs. Other
researchers also reported lower removal of inorganic nutrients with higher OLRs and F/M
ratios (Khoshfetrat et al., 2011; Shen et al., 2011). Further, the assimilated amount of N is
high at higher OLR. This could be due to higher biomass concentration than lower OLR of
0.5 kg-COD/m3.d, In addition, the concentration of NO2-N did not vary with OLRs and it
was virtually within the same range of 0.01-0.03 mg/L. This is expected as the nitrite
production is the rate limiting step in the nitrification process. The dissolved oxygen
concentration was also stable for the entire range of OLRs tested from 3.5 - 4.2 mg/L. As
expected, the MBR was not effective in removing phosphorus. The removal efficiency of
phosphorus was about 40 - 55% (Table 4.1). The removal of phosphorus may be due to
adsorption onto the membrane surface and some consumption by micro-organisms when
new cells are grown.
Table 4. 1. Effect of OLR on the removal of DOC and nutrients (nitrogen and phosphorous)
(HRT = 8 h; SRT = 40 days).OLR
(kgCOD/
m3.d)
MLSS
(g/L)
MLVSS
(g/L) F:M
DOC
(mg/L)
COD
(mg/L)
NH4-N
(mg/L)
NO3-N
(mg/L)
PO4-P
(mg/L)
0.5
Influent
3.7±0.3 3.0±0.1 0.12
28.06 165.0 7.8 0.2 4.4
Effluent 1.7±0.1 9±1 0.9±0.1 6.6±0.2 2.5±0.3
% removal 94±0.5 94±1 85±1 - 42±8
1.0
Influent
3.8-4.2 3.2-3.6 0.24
74.9 330.0 14.4 0.2 8.7
Effluent 1.5±0.1 16±7 2.2±0.11 11±0.2 5.6±0.3
% removal 98±1 95±2 83±1 - 34±4
2.0
Influent
4.4-5.2 3.7-4.4 0.42
166.8 660.0 31.2 0.1 13.5
Effluent 42±10 31±11 12±1 13±0.1 6.4±0.6
% removal 74±6 94±2 63±4 - 52±6
2.25
Influent
4.4-4.5 3.6-4.1 0.50
184.1 700.0 35.0 0.1 15.2
Effluent 57±3 36±11 11±0.2 13.5±0.4 8.6±0.1
% removal 68±1 93±2 67±4 - 43±1
2.75
Influent
4.6-4.9 3.8-4.1 0.56
228.2 907.0 42.0 0.1 18.5
Effluent 114±14 48±16 31±10 9.3±6 11.3±0.5
% removal 50±6 94±2 25±23 - 35.5±0.5
3.0
Influent
5.1-5.3 4.2-4.3 0.57
258.3 990 45.9 0.1 22.3
Effluent 86.5±0.5 71±10 30.5±10 8.5±0.7 10.5±0.5
% removal 66±1 92±1 34±24 - 52.5±2.5
Figure 4.1. Effect of OLR on the conversion of NH4-N into NO3-N (HRT = 8 h, SRT = 40
days, volume of the reactor = 4 L)
4.2.2.
The effect of OLR on the membrane fouling is presented in Figure 4.2. The rate of
development of TMP was very low with OLRs of 0.5 - 2.0 kgCOD/m3.d. The rate of
development of trans-membrane pressure (TMP) with these OLR was around 0.001-0.005
kPa/h. The increase in OLR from 0.5 - 2.0 to 2.25 - 3.0 kgCOD/m3.d resulted in higher
TMP development. From Figure 4.2, it is seen that the rate of development of TMP was
very high at about 0.1 - 0.24 kPa/h with OLRs of 2.75 and 3.0 kgCOD/m3.d, although at
the beginning of first 50 - 100 h, there was no significant development of TMP which was
less than 0.001 kPa/h within 3 days. It was also found that with an OLR of 2.25
NO
3-N/N
H4-N
Time (day)
0.5 OLR 1.0 OLR 2.0 OLR 2.25 OLR 2.75 OLR 3.0 OLR
kgCOD/m3.d, the rate of development of TMP was much less during the operational period
of around 10 days and only after that it started to increase.
The higher development of TMP with higher OLR may be due to (i) an increase of MLSS
concentration (Chang and Kim, 2005; Cicek et al., 1999), (ii) production of non-
flocculated micro-organisms which continuously attached onto the membrane surface as
F/M ratio increased (Ng and Hermanowicz, 2005) and (iii) the accumulation of hydrophilic
compounds onto the membrane surface which resulted in higher membrane fouling (Pan et
al., 2010). From Table 4.1 it is seen that the MLSS concentration increased from 3.5 - 4.0
g/L to 5.1 - 5.3 g/L when the OLR was increased from 0.5 to 3.0 kgCOD/m3.d. This can be
correlated with the membrane fouling presented in Figure 4.3. This confirms the higher
attachment of bio-solids onto the membrane surface at high OLR.
In addition, from the fractionation of organic matter, correlations were made between the
total membrane resistance (Rt) and the hydrophilic and hydrophobic fraction of organics
present in mixed liquor (measured as SMP and EPS) and membrane foulant (Figure 4.3 a-
c). From these correlation it was found that increased OLR increased the hydrophilic
fraction of organic matter in SMP and EPS resulting in higher membrane fouling (Figure
4.3 a, b) (here it should be noted that the concentration of organic in mixed liquor is equal
to the sum of the organic present in SMP and EPS). With higher OLRs, the concentration
of hydrophilic compounds increased more significantly in SMP than EPS. A detailed
explanation on this variation is discussed in Sections 4.2.3 and 4.2.4. The characterization
of foulant also showed higher hydrophilic substances with higher OLR (Figure 4.3 c).
These correlations clearly demonstrate the effect of hydrophilic compound on membrane
fouling (i.e total membrane resistance (Rt) increase). On the other hand, the concentration
of hydrophobic organics did not vary significantly with OLR. Furthermore at higher OLR,
micro-organisms have positive affinity towards the membrane surface and start to grow on
the membrane surface rather than staying in the reactor in suspension. This can be justified
from the findings of a previous study which suggested that the hydrophilic fraction of
mixed liquor could be the major cause for membrane fouling (Pan et al., 2010). The
fractionation of organic matter of mixed liquor (measured as SMP and EPS) and foulant
also match with the TMP data. Hence, the rapid accumulation of bio-floc onto the
membrane with higher OLR resulted in rapid fouling. Shen et al. (2010) also observed
more rapid membrane fouling as OLR increased. Rosenberger and Kraume (2002)
suggested that fouling potential can be reduced by lowering F/M ratios. This will avoid
unnecessarily high shear stress on the biomass. Our results also showed that the operation
of a MBR with lower OLR (low F/M) helped to reduce membrane fouling.
Figure 4.2. Effect of OLR on membrane fouling (HRT = 8 h, SRT = 40 days, volume of
the reactor = 4 L)
0
10
20
30
40
50
60
0.5 OLR1.0 OLR2.0 OLR
2.75 OLR
2.25 OLRTMP
(kPa
)
Time (h)
0.5 OLR 1.0 OLR 2.0 OLR 2.25 OLR 2.75 OLR 3.0 OLR
3.0 OLR
Figure 4.3. Correlation between OLR with membrane fouling and hydrophobic and
hydrophilic fraction of organic (HRT = 8 h, SRT = 40 days, volume of the reactor = 4 L)
0
50
100
150
200
250
300
350
400
9.90E+09
1.60E+11
3.10E+11
4.60E+11
0.5 1 1.5 2 2.5 3
Hyd
roph
ilic/
Hyd
roph
obic
(mg/
L)
Tota
l mem
bran
e re
sita
nce
(Rt,
m-1
)
OLR (kgCOD/m3.d)
(a) SMPTotal Membrane resistanceHydrophilicHydrophobic
0
50
100
150
200
250
300
350
400
9.90E+09
1.60E+11
3.10E+11
4.60E+11
0.5 1 1.5 2 2.5 3H
ydro
phili
c/H
ydro
phob
ic (m
g/L
)
Tota
l mem
bran
e re
sita
nce
(Rt,
m-1
)
OLR (kgCOD/m3.d)
(b) EPSTotal Membrane resistanceHydrophilicHydrophobic
0
50
100
150
200
250
300
350
400
9.90E+09
1.60E+11
3.10E+11
4.60E+11
0.5 1 1.5 2 2.5 3
Hyd
roph
ilic/
Hyd
roph
obic
(mg/
L)
Tota
l mem
bran
e re
sita
nce
(Rt,
m-1
)
OLR (kgCOD/m3.d)
(c) Membrane foulantTotal Membrane resistanceHydrophilicHydrophobic
4.2.3.
The different organic fractions present in the MBR effluent, SMP, EPS and foulant were
analysed by using LC-OCD. The concentration of hydrophobic compounds present in the
MBR effluent were 1.1-10.9 mg/L, in SMP they were 0.1 - 8.7 mg/L, in EPS they were 0.5
- 8.8 mg/L and in membrane foulant they were 4 - 54 mg/L at different OLRs of between
0.5 - 3.0 kgCOD/m3.d. On the other hand, the concentrations of hydrophilic compounds
increased with higher OLRs (in effluent they were 0.3 - 85 mg/L, in SMP they were 0.6 -
109 mg/L, in EPS they were 0.6 - 68.7 mg/L and in membrane foulant they were 18 - 370
mg/L). The hydrophilic compounds present in MBR effluent, SMP, EPS and membrane
foulant were bio-polymers molecular weight (MW) of >> 20000 g/mol, humic substances
MW of ~1000 g/mol, building blocks MW of around 300-500 g/mol, low molecular weight
(LMW) neutrals MW of < 350 g/mol and LMW acids MW of < 350 g/mol (Huber et al.,
2011).
The fraction of organics showed that the concentration of biopolymers present in MBR
effluent was much lower < 0.02mg/L for all the tested OLRs. The lower concentration of
bio-polymer of < 0.02 mg/L in the MBR effluent indicates that most of the bio-polymer
compounds were rejected by the membrane. This can be confirmed from the values of bio-
polymer concentration in the mixed liquor. The biopolymers concentration of SMP and
EPS present in mixed liquor was 0.23 - 2.98 mg/L and 0.23 - 5.9 mg/L respectively. The
rejection of bio-polymer by the membrane could be due to the retention by the membrane
due to its small pore size and attachment onto the membrane surface. It was also found that
the MBR effluent sample had a higher concentration of building blocks ranging from 0.1 to
3.3 mg/L and low molecular weight neutrals which was between 0.20 to 85 mg/L than bio-
polymer for all OLRs tested. The concentration of low molecular weight neutrals was at
much lower concentrations of 0.20 to 11.9 mg/L with OLRs of between 0.5 - 2.0
kgCOD/m3.d than for higher OLRs of between 2.25 - 3.0 kgCOD/m3.d where it was
between 28 to 85 mg/L. A similar trend was also observed in case of humic substances.
The removal of humic substances as well as building blocks and LMW neutrals was higher
with lower OLRs of between 0.5 - 2.0 kgCOD/m3.d than 2.25 - 3.0 kgCOD/m3.d. The
rejection of building blocks and LMW neutrals was low. This may be due to the larger
membrane pore size than the smaller MW size of these organics. The membrane used in
this study had pore size of 0.1 μm whereas building blocks and LMW neutral had MW of
< 350 g/mol. These results also imply that the bio-degradation of hydrophilic organics
decreased with larger OLRs resulting in increased F/M ratios (Table 4.1). The lower
degradation of organic matter at higher OLRs resulted in higher organic concentration in
MBR effluent. Moreover, it was also observed that with an increase of OLR, the
concentration of hydrophilic substances in EPS did not change significantly while the
concentration of hydrophilic substances in SMP increased (Figure 4.3 a, b). This implies
that most of the organics present in the MBR mixed liquor, at higher OLR of < 2.0
kgCOD/m3.d, was present in soluble form as SMP rather than in colloidal form such as
EPS.
From this fractionation it is evident that among these (bio-polymers, humics, building
blocks, LMW neutrals) different hydrophilic organic compounds, the MBR system was
much more effective in removing high molecular weight (HMW) bio-polymers types
substances such as polysaccharides, proteins, amino-sugars molecular weight (MW) of >>
20000 g/mol followed by relatively LMW substances such as humic acids MW of ~1000
g/mol than LMW substances such as building blocks, low molecular weight (LMW)
neutrals and acids MW of < 500 g/mol.
In addition, the membrane foulant was also subjected to LC-OCD analysis to determine the
major organics substances responsible for membrane fouling. From the LC-OCD results, it
was found that the concentration of bio-polymer in the foulant sample was at higher
concentrations of 5.9 to 73.2 mg/L than that in the SMP and EPS of mixed liquor. It was
also found that there was no direct correlation with the formation of bio-polymers with an
increase of OLR. The concentration of bio-polymers present in the membrane foulant with
different OLR of 0.5, 1.0, 2.0, 2.25, 2.75 and 3.0 kgCOD/m3.d were 17.5, 5.89, 8.3, 0.25,
76.83 and 73.2 mg/L respectively. The percentage of hydrophobic organic compounds
occurring with OLRs of between 0.5 and 1.0 kgCOD/m3.d was about 53.7-56.0% whereas
it was about 9.18, 16.4, 1.1 and 3.76% with OLRs of 2.0, 2.25, 2.75 and 3.0 kgCOD/m3.d
respectively. The percentages of humic substances, building blocks and LMW neutrals
were 7.1 - 40.78, 3.7 - 40.9 and 15 - 40.9% respectively with the different OLR of between
0.5 - 3.0 kgCOD/m3.d. From this result, it is evident that bio-polymers were the major
foulant together with humic and lower organic molecules. This can be validated from the
results presented in Section 4.2.4.
4.2.4.
For a detailed understanding of the fundamentals of membrane fouling a representative
chromatogram of LC-OCD analysis is presented in Figure 4.4. The size-exclusion
chromatography in combination with organic carbon detection (SEC-OCD) is a robust
technology to study fractionation of organic matter in water, wastewater, seawater and
membrane foulant (Huber et al., 2011). In LC-OCD chromatography, the peaks are
categorized as (i) bio-polymers: retention time 30 - 40 min, (ii) humic substances: retention
is time around 40 - 52 min, (iii) building blocks: retention time just after humic substances
which is around 52 min, (iv) low molecular-weight acids: retention time is around 55min
and (v) low molecular-weight neutrals: retention time is around 60 min to later (Huber et
al., 2011). From Figure 4.4, it was also found that the MBR effluent had a very low signal
intensity in the region of bio-polymers followed by SMP < EPS < foulant samples. This
indicated that bio-polymer was one of the major foulant responsible for membrane fouling.
This fouling phenomenon can be further justified from the results presented in Table 4.2.
In Table 4.2, results of two bio-reactors are presented. These reactors (volume of 2 and 24
L) were operated without membrane with an OLR of 1.0 kgCOD/m3.d. The effluent
sample was collected by letting the sludge settle down and filtering through 0.45 μm filter.
SMP and EPS were collected in a similar way to that of MBR as described in the materials
and method section. The results indicate that the supernatant of the bio-reactors operated
without any membrane had higher concentration of bio-polymers of 1.3 to 3.99 mg/L than
a bioreactor operated with a membrane where the bio-polymer concentration in the MBR
effluent was <0.02 mg/L operated with membrane. This confirmed that the most of the bio-
polymers were rejected by the membrane in MBR system causing membrane fouling
aggregated with other organic substances such as humic, building block and low molecular
weight neutrals and acids. According to Sun et al. (2011) a submerged MBR is basically an
enclosed system that concentrates organic foulants in the sludge suspension resulting in
transformation of SMP to biopolymer clusters causing cake deposition and serious
membrane fouling.
Figure 4.4. LC-OCD chromatogram of MBR effluent, SMP, EPS and foulant (OLR = 1.0
kgCOD/m3.d, HRT = 8 h, SRT = 40 days, volume of the reactor = 4 L)
20 30 40 50 60 70 80 90 1000.00.10.20.30.40.50.6
2.02.5
Humics
Rel
ativ
e si
gnal
resp
onse
, OC
DRetention time (min)
(a) Effluent
Bio-polymersBuilding block
LMW neutralsLMW acids
20 30 40 50 60 70 80 90 1000.00.10.20.30.40.50.6
2.02.5
LMW Acids
LMW neutrals
Building block
HumicsBio-polymers
Rel
ativ
e si
gnal
resp
onse
, OC
D
Retention time (min)
(b) SMP
20 30 40 50 60 70 80 90 1000.00.10.20.30.40.50.6
2.02.5
LMW Acids
LMW neutrals
Building blockHumics
Bio-polymers
Rel
ativ
e si
gnal
resp
onse
, OC
D
Retention time (min)
(c) EPS
20 30 40 50 60 70 80 90 1000.0
0.5
1.0
1.5
2.0
2.5LMW Acids
LMW neutrals
Building block
Humics
Bio-polymers
Rel
ativ
e si
gnal
resp
onse
, OC
D
Retention time (min)
(d) Foulant
Table 4.2. Fractionation of OM by LC-OCD of bio-reactor effluent, SMP and EPS operated without membrane (OLR = 1.0 kgCOD/m3.d)
OLR kgCOD/m3.d Description
Influent DOC
DOC HOC CDOC BIO- polymers
Humic Substances
Building Blocks
LMW Neutrals
Dissolved Hydrophobic Hydrophilic
mg/l,
%DOC
mg/l,
% DOC
mg/l,
% DOC
mg/l,
% DOC
mg/l,
%DOC
mg/l,
% DOC
mg/l,
% DOC
1
Effluent (2 L)
26.8
5.64
100%
1.91
34%
3.72
66%
1.34
23.8%
0.49
8.7%
1.32
23.4%
0.57
10.1%
SMP (2 L) 5.76
100%
2.3
39.9%
3.46
60.1%
1.15
20%
0.65
11.3%
1.0
17.4%
0.65
10.2%
EPS (2 L) 2.8
100%
1.86
78%
0.52
22%
0.2
8.6%
0.04
1.6%
0.17
7%
0.11
4.8%
Effluent (24 L)
97.3
22.5
100%
15.84
70.5%
6.62
29.5%
3.99
17.7%
1.02
4.5%
1.48
6.6%
0.12
0.6%
SMP (24 L) 20.62
100%
14.02
68%
6.6
32%
3.9
19.1%
0.12
5.8%
1.34
6.5%
0.12
6%
EPS (24 L) 4.35
100%
0.68
15.6%
3.67
84.4%
0.2
4.6%
1.18
27.1%
1.29
29.67%
1.0
23%
The imposed flux or hydraulic retention time is equally important parameter as OLR to
minimize the footprint of reactor and membrane area to reduce the capital cost and the
operation coast. This study focused on imposed flux and its influence on sludge properties
such as EPS and SMP content. The effect of imposed flux on the membrane fouling was
also studied. It could be noted that in a fixed volume of reactor, the imposed flux is related
to hydraulic retention time (HRT). In addition to EPS and SMP analysis, the study
compares the apparent molecular weight distribution of the contents of an MBR reactor
operated at different imposed fluxes and their effect on membrane fouling was also
studied. Detail experimental setup (Figure 3.1) and experimental conditions are provided in
chapter 3 (section 3.3).
4.3.1.
The mixed liquor in the MBR was acclimatized for 40 days before studying the effect of
imposed fluxes and aeration rates. MLSS and MLVSS concentrations in the mixed liquor
had almost similar concentration throughout all the experimental periods at different fluxes
and aeration rates. The concentration of MLSS and MLVSS was between 5 - 6 g/L and 4.5
- 4.8 g/L, respectively. The aim of this study was to understand the effect of imposed flux,
thus the removal of organic and nitrogen compounds by the MBR are presented (Table
4.3), indicated data are only given for an aeration rate of 1.2 m3/m2.membrane area.h. An
increase or decrease of aeration rate may affect the removal efficiency mainly if dissolved
oxygen becomes limiting, it was not the case in these experiments. The effect of aeration
rate was then studied to assess its effect on membrane fouling.
After acclimation for 40 days at a sludge retention time of 20 days, the operation time of
MBR at different imposed flux of 20, 25, 30 and 40 L/m2.h were 13, 7, 3 and 2 days
respectively before significant membrane fouling was observed. The organic matters and
nutrients were measured in duplicate on a regular basis. The difference between two
measurements was less than 5%. Thus the average of two was reported. The number of
samples collected during the operation period after acclimation was 26, 14, 6 and 4 at
different imposed flux of 20, 25, 30 and 40 L/m2.h respectively.
From Table 4.3, it can be seen that the removal of DOC was in the range of 60 to 95%
under all tested fluxes. As expected, the highest DOC removal of around 95% was
achieved when the MBR was operated at a lower flux of 20 L/m2.h. This removal
efficiency reduced to 58 - 66% when the operating flux was increased to 40 L/m2.h. The
organic removal is discussed later in section 4.2.3 along with molecular weight distribution
results. The dissolved oxygen (DO) level shows (Table 4.3) that most parts of the reactor
experienced aerobic condition. Therefore, most of the DOC removal could be attributed to
the aerobic heterotrophic microbial biomass present in the reactor.
Table 4.3 also shows the removal efficiency of nutrients, especially nitrogenous
compounds such as NH4-N. It was found that the removal of NH4-N was slightly higher in
the range of 48 - 50% with a lower flux of 20 L/m2.h compared to about 30 - 35% at a
higher flux of 40 L/m2.h. The ammonia removals could be achieved by the micro-
organism assimilation/biosynthesis and nitrification reaction in the MBR. A mass balance
calculation was made to estimate the amount of inorganic nitrogen that was nitrified and
assimilated. Experimental error and any contribution from nitrite can be neglected as the
value of nitrite was very low. Only significant readings were observed for nitrate and
ammonia. The experimental errors can be ± 0.5 mg/L for ammonia and ± 0.1 mg/L for
nitrate as per measurement manual. Thus total measurement error in total inorganic
nitrogen (TIN) could be ± 0.6 mg/L. The amount of total nitrogen is the sum of nitrate-
nitrogen, nitrite-nitrogen, and ammonia-nitrogen. From the calculation, it was found that
the amount of nitrogen nitrified was 2.2 - 2.5, 1.0 - 1.2, and 0.2 - 0.6 mg/L at
corresponding imposed flux of 20, 25 and 40 L/m2.h, respectively. The nitrified amount is
the sum of the differences between the effluent and influent nitrite nitrogen and nitrate
nitrogen. All these readings were more than 0.2 mg/L and hence show a significant level of
nitrification. From the calculation of loss of total inorganic nitrogen, it was found that the
amount of nitrogen assimilated was 1.5 - 1.9, 1.1 - 1.2, 0.8 - 1.1 mg/L at corresponding
imposed flux of 20, 25 and 40 L/m2.h respectively. The amount of assimilated nitrogen
was calculated from the differences between the total influent and effluent nitrogen. Except
for the imposed flux of 20 L/m2.h for all other fluxes the assimilated amount was less than
1.2 mg/L (double the error in TIN level). This shows that the difference is not significant to
indicate the assimilation. TIN drop should indicate assimilation as there was sufficient DO.
However, according to the low level of nitrogen in influent composition, the total nitrogen
concentration in the effluent was also low. The choice of the ratio 150 : 5 : 1 for COD : N :
P in the wastewater was made to avoid any deficiency in nutrients to insure biomass
activity. A significant amount of ammonium was used for the cell growth and the
remaining ammonium (in excess) was partially oxidized to nitrate. In addition, the F/M
ratio in this study was 0.28, 0.29, and 0.28 for corresponding imposed flux of 20, 25 and
40 L/m2.h respectively. This indicated that the F/M ratio was stable. Although there were
no significant changes in the F/M, the removal of ammonia was relatively low for high flux
of 40 L/m2.h compared to lower flux of 20 L/m2.h. Thus, working at lower flux or higher
HRT leads to a relatively good conversion of ammonium. A lower conversion to nitrate of
less than 35% was observed at higher flux or lowest HRT.
Under lower or moderate flux conditions (or higher HRTs) the highest DOC and COD
removals were also achieved. This level of removal efficiency shows the great capacity of
MBR to degrade organic matter even when working under moderate flux. Nevertheless,
when the flux is too high or HRT is too low, the organic carbon removal can only be
achieved partially. Even if the ammonium concentration in influent was low, a nitrification
was observed under lower HRT (2.5 h). In conventional bioreactor, nitrification is
generally observed when working with higher HRT, 6h and more (Viero et al., 2007,
2008). Previous researchers also reported relatively lower removal of total nitrogen ranging
from 31% to 68% by MBR due to partial nitrification following the hydrolysis of the
organic-bound nitrogen to ammonia (Nguyen et al., 2012). The removal efficiency of PO4-
P remained very low of about 10 - 30% under all operating conditions, as ammonium, this
part of removed phosphorus was probably used for biomass growth. The amount
phosphorus accumulated was not calculated due to short t operation period.
Table 4.3. Removal of organic matter and nutrients by MBR operated at different imposed
fluxes (membrane area = 0.2 m2; membrane pore size= 0.14 μm; volume of reactor = 10 L;
SRT = 20 days; OLR = 1.5 kg COD/m3.day; aeration rate = 1.2 m3/m2.membrane area.h)
Flux
(L/m2.h)
DOC
(mg/L)
COD
(mg/L)
NO3-N
(mg/L)
NH4-N
(mg/L)
NO2-N
(mg/L)
DO
(mg/L)
20
Influent 123.50 179 0.6 8.1 0.01
3.5-4.2 Effluent 6.5-7.1 7-11 2.8-3.1 4.1-4.2 0.03
% removal 94-95 93-96 - 48-50 -
25
Influent 90.07 143 0.3 6.4 0.01
3.1-4.2 Effluent 6.4-8.5 11-14 1.3-1.5 4.0-4.3 0.01
% removal 90-93 90-92 - 32-38 -
30
Influent 83.50 120 0.8 5.4 0.02
3.5-4.1 Effluent 11.5-12.3 20-27 0.6-1.2 3.1-3.5 0.01
% removal 85-86 77-83 - 35-42 -
40
Influent 35.00 90 0.4 4.0 0.01
3.0-4.2 Effluent 11.8-14.5 29-33 0.5-1.0 2.6-2.8 0.01
% removal 58-66 63-67 - 30-35 -
4.3.2.
The effect of imposed fluxes and aeration rates on membrane resistance with
corresponding cumulative filtered volume (m3) was studied at all four different imposed
fluxes of 20, 25, 30 and 40 L/m2.h and aeration rates of 0.3, 0.6 and 1.2 m3/m2.membrane
area.h. The membrane resistance was calculated as function of the rate of change of TMP
development with flux (TMP/μ.J; 1/m). Here μ is the viscosity (Pa.s) and J is the flux
(m/s).
Figure 4.5 (a) and (b) show that a higher flux of 40 and 30 L/m2.h induced a higher
membrane resistance (ie. higher membrane fouling indicates higher TMP) compared with
lower flux of 20 and 25 L/m2.h. From this figure it is found that the membrane resistance
was lower and stayed stable during the initial period but later a sudden increase (jump) was
noted. The time at which this jump occurred was defined as when membrane resistance or
TMP increased two times the base membrane resistance or TMP that occurred during the
stable period. The sudden jump in membrane resistance or TMP at a flux of 20 L/m2.h
occurred after a filtration volume of around 4.0 and 14.0 m3/m2membrane area when the
aeration rate was 0.6 and 1.2 m3/m2.membrane area.h respectively (Figure 4.5 a and 4.5 b). This
phenomenon confirms the positive influence of membrane aeration and need for air
scouring. The jump in TMP occurred earlier for flux of 30 and 40 L/m2.h compared with a
20 L/m2.h. For flux of 30 and 40 L/m2.h, the TMP jump occurred after a filtered volume of
0.6 and 0.1 m3/m2membrane area respectively, whereas the development of membrane
resistance was slightly lower for a flux of 25 L/m2.h. This observation also confirms the
determining role of permeate flux when working in conditions close to the critical
permeate flux value (close to 30 L/m2.h. in MBR). At higher flux this jump may be not
only due to change of local flux but also due to the change of cake layer structure (Zhang
et al., 2006). Other researchers also reported a similar trend of membrane resistance or
TMP jump at higher flux (Navaratna and Jegatheesan, 2011).
From Figure 4.5 c, it is found that aeration rate had a positive effect on membrane
filtration. Higher aeration rates prolong the operation of membrane filtration. A slow
increase of membrane resistance was observed at a higher aeration rate of 1.2
m3/m2.membrane area.h compared to a lower aeration rate of 0.3 and 0.6 m3/m2.membrane area.h.
Similar trend was also observed on the effect of aeration rates with other filtration fluxes of
25 to 40 L/m2.h. The lower fouling at a higher aeration rate could be due to higher shear
stress against the membrane surface produced by larger air bubble which was generated at
higher aeration rates (Yu et al., 2003).
The effect of imposed flux and aeration rate on membrane fouling can be explained by
linear correlations between the cumulative filtered volume before a sudden TMP jump
occurs and aeration rates (Figure 4.6 a) and with imposed flux (Figure 4.6 b). These points
describe the amount of water that could be produced before membrane cleaning is needed
at different flux and aeration rates. It was found that the volume filtered before the TMP
jump (or fouling) can be directly correlated with an imposed flux and aeration rate with R2
of 0.80 to 0.99. It implies that an increase of aeration rate increase the filtered volume by 3
to 20 times. In case of flux, the effect of lowering the flux in reducing membrane fouling
was more significant than aeration rates. The lower flux of 20 L/m2.h produces almost 75
to 95 times more filtrate than a higher flux of 40 L/m2.h under 0.6 and 1.2 m3/m2.membrane
area.h of aeration. Higher aeration rate of 1.2 m3/m2.membrane area.h led only to 3.6 times more
water than that with 0.6 m3/m2.membrane area.h at flux of 20 L/m2.h. This implies that
operating the MBR at a lower flux reduced membrane fouling more efficiently than by
increasing the aeration rates. This can be explained by the difference in suction pressure
produced on the membrane surface at different fluxes. As stated earlier, at higher flux this
jump may be not only due to the change of local flux but also may be due to the change of
cake layer structure (Zhang et al., 2006). Thus, at higher flux the deposition of the sludge
on to the membrane surface should be faster and the cake layer should be well-built than at
a lower flux. Therefore, at higher flux, due to well-built cake layer, even a higher aeration
rate was not capable of dragging/scouring the particles from the membrane surface.
Therefore, the operation of MBR at lower flux will help to minimize membrane fouling.
This in turn, will also help to reduce the chemical cleaning frequency of membrane which
results in an increase of membrane life.
Figure 4.5. Temporal variation of membrane resistance at different imposed flux and
aeration rates (membrane area = 0.2 m2; membrane pore size = 0.14 μm; volume of reactor
= 10 L; SRT = 20 days; OLR = 1.5 kg COD/m3.day).
0 5 10 15 20 25 30 350.0
2.0x1012
4.0x1012
6.0x1012
8.0x1012
1.0x1013
1.2x1013
1.4x1013
(C)
(B)
Tot
al m
embr
ane
resi
stan
ce (1
/m)
Filtered volume (m3/m2membrane area)
(A) 20 L/m2.h (B) 25 L/m2.h (C) 30 L/m2.h
(a) Effect of imposed flux(aeration rate 0.6 m3/m2
memb.h)
(A)
0 5 10 15 20 25 30 350.0
2.0x1012
4.0x1012
6.0x1012
8.0x1012
1.0x1013
1.2x1013
1.4x1013
(A)
(B)
Tot
al m
embr
ane
resi
stan
ce (1
/m)
Filtered volume (m3/m2membrane area)
(A) 20 L/m2.h (B) 30 L/m2.h (C) 40 L/m2.h
(b) Effect of imposed flux(aeration rate 1.2 m3/m2
memb.h)
(C)
0 5 10 15 20 25 30 350.0
2.0x1012
4.0x1012
6.0x1012
8.0x1012
1.0x1013
1.2x1013
1.4x1013
Tot
al m
embr
ane
resi
stan
ce (1
/m)
Filtered volume (m3/m2membrane area)
(A) 0.3 m3/m2memb.h
(B) 0.6 m3/m2memb.h
(C) 1.2 m3/m2memb.h
A
B
C
(c) Effect of aeration rate
Figure 4.6. Correlation between filtered volume before getting rapid TMP rise with
imposed flux and aeration rates (membrane area = 0.2 m2; membrane pore size = 0.14 μm;
volume of reactor = 10 L; SRT = 20 days; OLR = 1.5 kg COD/m3.day)
4.3.3.
The organics present in SMP and EPS in MBR operated at different flux of 20, 30 and 40
L/m2.h was examined through MWD analysis. The focus was on investigating major
representative organic such as biopolymers, humic and fulvic acid type substances and low
molecular acid and neutrals. From MWD analysis it was found that the organics presents in
R² = 0.9854R² = 0.9625R² = 0.9877R² = 0.9973
0
2
4
6
8
10
12
14
16
0 0.2 0.4 0.6 0.8 1 1.2Volu
me
filte
red
befo
re T
MP
ju
mp
(m3 /m
2 mem
bran
e ar
ea)
Aeration rate (m3/m2memb.h)
(a) Effect of aeration rate
20 L/m²h 25 L/m².h 30 L/m².h 40 L/m².h
R² = 0.9334R² = 0.8013
0
2
4
6
8
10
12
14
16
0 10 20 30 40 50Volu
me
filte
red
befo
re T
MP
jum
p (m
3 /m2 m
embr
ane
area
)
Flux (L/m².h)
(b) Effect of flux
aeration rate = 1.2 m³/m² memb.h
aeration rate = 0.6 m³/m²memb.h
SMP and EPS had wide range of organics molecular weight (MW) of around less than 450
Da to about 48 kDa for imposed fluxes used (Figure 4.7 a, b). The organic of MW of about
48 kDa may biopolymers consisting of significant amount of proteins and polysaccharides,
600-1500 Da are humic and fulvic acid type substances and less than 600 Da are low
molecular weight acids and neutrals (Shon et at., 2006; Villacorte et al., 2009). In addition,
Jang et al. (2006) also reported that most of the protein and carbohydrate in SMP and EPS
had a molecular weight (MW) more than 10 kDa. This implies that SMP and EPS had
different nature of organics present in SMP and EPS.
The MWD of organics of SMP showed higher UV response of 0.5 mV in the region of
high MW organics of about 48 kDa at high flux of 40 L/m2.h as compared to those at lower
fluxes of 20 and 30 L/m2.h (Figure 4.7 a). The later had an intensity of less than 0.1 mV.
At high flux of 40 L/m2.h, the UV intensity was also high in the rage of 5.8 to 15 mV for
low molecular weight of organics of 1200-680Da than that of 20 and 30 L/m2.h. The UV
intensity of low MW of organics at lower flux was 2.8 to 3.1 mV. The UV intensity for
organics of relatively low MW of about 450Da were around 5.8, 7.6 and 9.1 mV for
different flux of 20, 30 and 40 L/m2.h respectively. This indicated that the formation of
lower molecular weight organics were much higher at higher flux of 40 L/m2.h. At the
higher flux of 40 L/m2.h the formation of biopolymers such as polysaccharides and protein
was also slightly higher than the lower flux of 20 L/m2.h resulting in a relatively high UV
intensity at the molecular weight region of more than 45 kDa.
On the other hand, the MWD of EPS showed relatively high UV intensity of 0.75 to 1.5
mV for molecular weight organics of more than 45 kDa (Figure 4.7 b). The UV intensity of
organics with MW of more than 45 kDa in SMP was in the range of 0.1 to 0.5 mV. The
organics of small MW of around 680 Da present in EPS showed relatively lower UV
intensity about 6.0 mV. The corresponding organics in SMP had UV intensity of about 15
mV. However, the UV intensity (19.5-24.1 mV) for very low molecular weight organics of
around 150 Da to less than 150 Da in EPS was much higher than that of SMP. UV
intensity for this range of organics in SMP was approximately 1.5 MV. This indicates that
EPS had relatively high concentration of organic matter compared to SMP. Additionally,
the nature of UV response for a wide range of molecules present in EPS was not exactly
the same in fluxes. In case of high MW of organics of around 48 kDa, high flux of 40
L/m2.h showed high UV response of 1.5 mV than lower flux of 20 and 30 L/m2.h (UV
intensity was around 1.4 and 1.5 mV respectively). But in the case of low MW organics of
690 Da, the UV intensity of 6.0 mV which was slightly higher at lower flux of 20 L/m2.h
than 30 and 40 L/m2.h (UV intensity was around 2.8 and 3.8 mV respectively). However,
at the very low MW region of about 150 Da the UV response was again high around 22.5
mV for higher flux of 40 L/m2.h than lower flux of 20 L/m2.h (UV intensity was around
19.8 mV). Although no firm conclusion can be made, it is clear that there is a change in
organic composition with a variation in flux. The organics in EPS could be from the
organics present in the influent and compositions from partly biodegraded influents in
addition to SMPs.
The mixed liquor was measured in terms of SMP and EPS as DOC, proteins and
carbohydrates as have been presented in the previous study (Le-clech et al., 2006; Jang et
al., 2006). The concentration of DOC in SMP was between 8.0 - 15.5 mg/L, whereas the
concentration of DOC in EPS was higher than SMP in the range of 25.1 - 46.2 mg/L. The
carbohydrate concentration of SMP was between 0.87 to 13.5 mg/L (C6H12O6 equivalent)
while protein concentration was between 19.33 to 30 mg/L during the experimental period.
However, the concentration of carbohydrate and protein of EPS was between 6.6 to 13.6
mg/L and 64.1 to 107.3 mg/L respectively. From carbohydrate, protein concentration and
MWD distribution of SMP and EPS, it was found that EPS contains more biopolymer as
well as lower molecular weight organic compounds than SMP. In case of higher flux of 40
L/m2.h, the carbohydrate and protein concentration in SMP and EPS was also higher in the
range of 8 to 13.6 mg/L and 20 to 107.3 mg/L respectively than at a lower flux of 20
L/m2.h. The carbohydrate and protein concentration in SMP and EPS at a lower flux of 20
L/m2.h was 0.87 to 2.5 and 19 to 64.6 mg/L respectively. From the above information, it
was found that at higher flux of 40 L/m2.h the formation of biopolymers as well as lower
molecular weight organics in SMP and EPS was higher than at a lower flux of 20 L/m2.h.
This results in a higher membrane fouling (Figure 4.5 b). Previous researchers observed an
exponential decline in critical flux with the increase of SMP and MLSS concentrations
(Navaratna and Jegatheesan, 2011). They also found a linear decline of flux with increase
in total EPS and protein concentration in EPS.
Figure 4.7. MWD of SMP and EPS in MLSS at different fluxes (membrane area = 0.2 m2;
membrane pore size = 0.14 μm; aeration rate = 1.2 m3/m2.membrane area.h; volume of reactor =
10 L; SRT = 20 days; OLR = 1.5 kg COD/m3.day)
4.3.4.
Both SMP and EPS had wide range of organic molecules ranged from very high to low
molecular weight (MW) of between 45 kDa to less than 200 Da. According to literature,
EPS and SMP have a nature that changes during the operation of the MBR (Flemming et
al., 1997). These substances can form a highly hydrated gel matrix in which microbial cells
are embedded. Therefore, they can create a significant barrier to permeate flow in
membrane processes. This can be explained from the results on MWD distribution of MBR
effluent and membrane foulant.
In this study, we did not characterize effluent in terms of SMP and EPS as most of the
bioflocs were retained by membrane in MBR. The membrane effluent was collected and
subjected to molecular weight distribution analysis. A significant reduction of organics in
the effluent was observed for organics of large molecular weight. The reduction of low
molecular weight organics was less. The reduction of large MW organics of 48 kDa was
almost 95% result in very low UV intensity of less than 0.05 mV, whereas the reduction of
lower molecular weight organics of 1200 - 450 Da was between 30 - 60% (UV intensity
was 0.5 - 2.2 mV) (Figure 4.8 a). This shows that the MBR system can effectively remove
the larger molecule of organic compounds of more than 45 kDa such as polysaccharide
types and to some extent the small molecules of organic substances such as amphiphilic,
acids and neutral. The removal of organics by MBR could be due to adsorption of
organics on to membrane surface and due to biodegradation of organics by the bio-film
formed on the membrane surface. Previous researchers also reported that the membrane
bio-film played a secondary filtration barrier for both low and high molecular weight
organic matters (Kang et al., 2007). They also stated that low molecular weight organics of
less than 1 kDa was removed by the bio-film using easily degradable organic matters as
microbial carbon and energy sources during the filtration. The removal of organics by
MBR was supported by the analysis of the foulant deposition layer on membrane surface.
The membrane foulant was directly extracted using 0.5% (w/v) NaOH solution and filtered
through 1.2 μm filter for further MWD analysis. The foulant sample contained particulate
and soluble materials such as SMP, EPS, biomass floc, cells. The foulant contained both
high molecular weight organics of more than 45 kDa with UV intensity of about 0.8 mV as
well as low molecular weight organics of less than 1000 Da (Figure 4.8 b). The UV
intensity was relatively high (14.1 mV) for organics of MW of around 660 Da than lower
MW of organic of 450 Da (UV intensity was about 6.2 mV). From this, it could be inferred
that a wide range of organics of both high and low molecular weights play a significant
role in membrane fouling. Similar result has been reported earlier (Teychene et al., 2008).
They reported that organic colloids and some humic-like substances or building blocks
play a role in membrane fouling.
Figure 4.8. MWD of organic matter of filtrate and foulant (backwash water) (Flux = 20
L/m2.h; aeration rate = 1.2 m3/m2.membrane area.h; membrane area = 0.2 m2; membrane pore
size = 0.14 μm; volume of reactor = 10 L; SRT = 20 days; OLR = 1.5 kg COD/m3.day)
8 10 12 14 16 18 2002468
10121416
UV
inte
nsity
(mV
)
Retention time (min)
(a) Effluent
MW:~48kDa
MW: ~1200Da
MW:~680Da
MW: ~450Da
8 10 12 14 16 18 2002468
10121416
MW: ~1200Da
UV
inte
nsity
(mV
)
Retention time (min)
(b) Membrane foulant
MW: ~48kDa
MW: ~660Da
MW: ~440Da
MW:~140Da
4.3.5.
Experiments were undertaken on a high rate MBR. The results clearly demonstrate that a
decrease of flux has positive effects on membrane resistance or membrane fouling. Lower
flux of 20 L/m2.h produced 75 to 95 times more water under the same aeration condition
compared to a higher flux of 40 L/m2.h (Figure 4.6). A lower flux helped in achieving
lower membrane fouling. However, it is noted that a lower flux will mean increased space
requirement, and this could overtake the very advantage obtained from the MBR system.
In addition, the amount of water produced before fouling for each aeration rate studied
showed an increased efficiency and performance by increasing aeration and the advantage
obtained is mostly physical. However, increased aeration would mean an increased cost
which may eventually overtake the savings obtained by alleviating fouling. Further the
study shows that reducing flux encouraged nitrification, implying that there is a change in
microbial composition. This information will help in understanding the basic need of
aeration and to choose sustainable flux in designing a MBR for an actual prototype
operation.
In addition, the MWD distribution of SMP, EPS, MBR effluent and foulant provides
fundamental information of organic matter present in activated sludge at different imposed
fluxes and their role on membrane fouling.
In real wastewater the fouling propensity may be different than that of synthetic
wastewater and the result may vary from 20 – 50 %. The synthetic feed was prepared using
analytical grade (pure) chemicals where the amount of TSS is very low. Moreover, the
synthetic feed contained mostly biodegradable dissolved COD which is easily
biodegradable. On the other hand the real feed contains wide range of physico-chemical
and microbiological pollutants. The suspended solid concentration in real feed varies
significantly and also contains both particulate and non-biodegradable COD.
The application of MBR in treating saline water is also very significant. Many wastewater
(especially industrial wastewater such as cheese industry) and wastewater water from
coastal area (where seawater seep through ground) contain salinity. Thus, a detailed study
on MBR under saline condition is important. In this study, the effect of gradual increase of
salt (from 0 to 35 g/L) in a continuous MBR was investigated. The effect of salt was
investigated in terms of organic and ammonia removal, specific oxygen uptake rate
(SOUR) represent the viability of micro-organism. Furthermore, a detailed organic
characterization were employed to better explore the effect of salt along with cluster
analysis of the DOC concentration of MBR effluent and mixed liquor. This study provided
relative advantages to understand the effect of continual increase of salt concentration in
treating saline wastewater in continuous membrane bioreactor process. Experimental
details are provided in chapter 3 (section 3.3.2)
4.4.1.
The results of the effect of gradual increase of salt concentration on organic removal in
MBR are presented in Table 4.4. From Table 4.4, it is found that salt concentration up to
3.0 g/L showed a relatively good removal efficiency of DOC (around 72%). Salt
concentrations of 5.0 and 10 g-NaCl/L showed only 35 - 40% removal of DOC. When the
salt concentration exceeded 10 g/L the DOC removal efficiency decreased significantly.
The DOC removal efficiency with salt concentration of 15, 20, 25 and 35 g-NaCl/L was
35, 30, 26, 10 and 10% respectively. The increase of DOC concentration at different salt
loading indicated that the degradation of organics by micro-organism decreased due to the
addition of salt. From literature it is found that, high salt concentrations of > 1% (10 g/L)
cause plasmolysis and reduce activity of cells (Reid et al., 2006). In addition, each time the
salt concentration increased, the organic concentration increased initially within 2 days and
steady state organic removal was achieved before 11 days. Thus, the time was chosen to
increase the salt concentration in the MBR was of 11 days to assess the short term effect of
gradual increase of salt concentration on MBR performance.
The specific removal of DOC with different salt concentration is presented in Figure 4.9.
The specific removal of DOC decreased by around 90% (decreased from around 17.0 mg-
DOC/g-MLVSS.d to 1.8 mg-DOC/g-MLVSS.d) when the salt concentration increased
from 0 to 35 g-NaCl/L although the concentration of MLVSS decreased by only around 20
- 22 % (from 4.2 to 3.4 g/L) (Figure 4.10). Figure 4.10 also presents the concentration of
MLSS and the specific oxygen uptake rate (SOUR) with different salt concentration. The
SOUR decreased with the increase of salt concentration and it decreased significantly
(almost 97%) when the salt concentration reached to 35 g-NaCl/L. The SOUR was
measured twice a day throughout the experimental period and an average value is
presented. The measurement of SOUR was conducted to assess the viability of
microorganism as it is a good indicator for assessing the viability of micro-organism
(Hasar et al., 2004). These results also indicate that only the MLVSS concentration does
not represent the viability of micro-organism. The reason is that MLVSS does not only
include the live cells. This can also be produced from the dead cells as well as the old
biomass (Hasar et al., 2004). Thus, the lower removal of organics (DOC) with high salt
concentration could be due to the adverse effect of salt on microbial activity. Yogalakshmi
and Joseph (2010) stated that the metabolic activity of microorganism reduced due to
addition of salt, and plasmolysis of microorganism caused the release of intracellular
constituents and soluble microbial products. The decrease of specific removal of DOC can
also be explained due to the effect of osmotic stress. The osmotic pressure of the water
increases about 8 bars for every 10 g/L of NaCl (Lay et al., 2010). They also reported that
most of the micro-organisms present in activated sludge are non-halophilic. These micro-
organisms can survive under low salt concentration up to 10 g/L. Previous study also
reported that an increase of salt concentration resulted in a decrease of organic removal.
For example, Uygur (2006) observed that the removal efficiency of COD decreased from
42.5 mg-COD/g-biomass.h to 10.7 mg-COD/g-biomass.h when salt concentration were
increased from 1 to 6% in a sequencing batch reactor (SBR) consisting of anaerobic, oxic,
anoxic and oxic phases.
Table 4.4. Removal of DOC and NH4-N by MBR with and without salt (gradual addition)
concentration (HRT = 8 h)
Salt
concentration
(g/L)
DOC
(mg/L)
(removal, %)
NH4-N
(mg/L)
(removal, %)
Salt
concentration
(g/L)
DOC
(mg/L)
(removal, %)
NH4-N
(mg/L)
(removal, %)
Influent 7.0 3.0 7.0 3.0
0 1.6±0.2
(77±3)
<0.5
(93±1) 10.0
4.6±0.2
(35±2)
<0.8
(76±2)
0.5 1.9±0.2
(72±2)
<0.5
(90±2) 15.0
4.9±0.1
(30±1)
1.7±0.1
(41±2)
1.0 1.8±0.5
(71±5)
<0.5
(90±2) 20.0
5.3±0.3
(26±4)
2.0±0.3
(33±4)
3.0 2.2±0.6
(68±3)
<0.5
(90±2) 25.0
6.3±0.1
(10±1)
3.7±0.3
(-)
5.0 4.2±0.2
(40±1)
<0.5
(88±1) 35.0
6.5±0.5
(10±1)
4.2±0.5
(-)
Figure 4.9. Profile of specific removal (average) of organic (DOC) and NH4-N with
different salt concentrations.
Figure 4.10. Profile of MLSS, MLVSS, and SOUR (average) with different salt
concentrations.
0
5
10
15
20
0 0.5 1 3 5 10 15 20 25 35
mg/
g-M
LVSS
.d
Salt concentration (g-NaCl/L)
mg-DOC/g-MLVSS.dmg-NH -N/g-MLVSS.d
012345678910
0
1
2
3
4
5
6
0 0.5 1 3 5 10 15 20 25 35
mg-
O2/g
-MLV
SS.h
mg/
L
Salt concentration (g-NaCl/L)
MLSS MLVSS MLVSS/MLSS SOUR
4.4.2.
The removal of NH4-N by the MBR is presented in Table 4.4. The removal of NH4-N was
more than 90% when no salt was added. An increase of salt concentration from 0.5 to 10 g-
NaCl/L showed a good removal of ammonia of 76 to 90%. A further increase in salt
concentration from 15 to 35 g-NaCl/L decreased ammonia removal (0 - 46%). The specific
NH4-N removal efficiency decreased by almost 100% (from 8.2 mg-NH4-N/g-MLVSS.d to
0 mg-NH4-N/g-MLVSS.d) when salt concentration reached to 35 g-NaCl/L (Figure 4.9).
These results clearly demonstrate the effect of high salt concentration on the nitrification
process. This finding is accordance with the findings of previous study. Sharrer et al.
(2007) also found that nitrification rate decreased linearly when salt concentration
increased from 0 to 60 g/L of NaCl and the nitrification rates was almost six times less at
higher salinity than freshwater. Yogalakshmi and Joseph (2010) also reported lower
removal of ammonia at high salt concentration. They found that ammonia removal of 84 to
64% with a NaCl loading of 5–30 g/L, which further dropped to 13% at a shock loading of
60 g/L. Vendramel et al. (2011) observed almost 90% of nitrification efficiency with Cl-
concentration of 0.05-6 g/L which decreased dramatically when Cl- concentration reached
to 12 g/L. Decrease in removal of NH4-N at high salt concentration could be due to the
effect of plasmolysis or lower availability of saline-resistant nitrifiers (Chen et al., 2003).
Uygur (2006) also reported that due to plasmolysis of the activated sludge organisms at
high salt content, specific nutrient removal rates decreased with increasing salt content.
Thus, the addition of salt results in an adverse effect on ammonia-oxidizing bacteria (AOB,
Nitrosomonas group) and nitrite oxidizing bacteria (NOB, Nitrobacte). Both of these
bacteria could have been reduced due to the addition of salt. Ye et al. (2009) reported that
the number of NOB strongly decreases when the salinity was above 1% (10 g/L). They
also reported that the feasibility of presence of NOB is less than 1% when salinity is higher
than 2% (w/v) or 20 g/L. Yogalakshmi and Joseph (2010) also reported that both nitrite
and ammonia oxidizers are more sensitive to short and long-term salt stress resulting in a
lower removal of nitrogen. Chen et al. (2003) reported that nitrification was good up to Cl-
concentration of 2.5 g/L and beyond that the nitrification rate started to decrease. They also
mentioned that Nitrobacter disappeared when Cl- concentration was more than 18 g/L.
Thus from the above findings it can be concluded that the salt tolerance ability of fresh
nitrifiers could be up to 15 g/L of NaCl. Furthermore, nitrification rate may improve by the
addition of salt tolerant culture such as halobacter halobium to the biomass or by providing
longer acclimation period to the biomass (Lay et al., 2010). Initially the raw activated
sludge had PO4-P concentration of 2.5 ± 0.2 mg/L. When the salt concentration was
between 0 - 5 g-NaCl/L, the PO4-P concentration was 1.9 - 2.1 mg/L. With the high salt
concentrations of 10-35 g-NaCl/L it was 2.3 - 2.5 mg/L. This indicated no or marginal
phosphate removal only.
In summary, the aim of this section of the study was to assess the effect of gradual increase
of salt concentration on the organic and ammonia removal by MBR. As such the DOC and
NH4-N concentrations were kept low at 7 and 3 mg/L respectively. From the Figures 4.9
and 4.10 it is evident that the removal of DOC and ammonia decreased with smaller values
of SOUR when the loading of salt concentration increased. As discussed before, when the
salt concentration increased, the viability of microorganism decreased. In this study we
have mainly focused on the organic and ammonia removal under different salt
concentrations which was increased gradually. Thus, we have measured the microbial
activity/mass in terms of SOUR and MLVSS concentration with the increase in gradual
salt concentrations. There are many previous studies which focused on the effect of salt on
microorganism activity (Wu et al., 2008; Cui et al., 2009; Ye et al., 2009; Lay et al., 2010;
Bassin et al., 2011). For example Ye et al. (2009) reported that the population of ammonia
oxidation bacteria decreased from 28.712 to 19.979 (log MPN/g-MLVSS; where MPN is
most probable number) whereas nitrite oxidation bacteria decreased from 19.73 to 11.92
(log MPN/g-MLVSS) when the salinity increased from 0 to 3 % (w/v). Similarly,
halophilic microorganisms or halophiles also required a certain minimum level of salt for
continued existence (Lay et al., 2010).
In addition, the MBR experiment with and without different concentration of salt was
conducted for 11 days. The salt was increased gradually as shown in Figure 3.3. The time
employed for each salt addition could be short for the microbes to become adapted with the
environment. However, its performance can be improved with the optimized operational
conditions such as acclimation for long time or by adapting the microorganisms naturally.
Furthermore, the nitrification rate may be improved by the addition of salt tolerant culture
such as halobacter halobium to the biomass or by providing a longer acclimation period to
the biomass (Lay et al., 2010). For example, Uygur (2006) used Halobacter in activated
sludge and found that the COD removal efficiency with the incorporation of Halobacter
increased from 42.6 to 63.3 mg COD g biomass-1 h-1 when the salt content increased from
0 to 1%. Furthermore Lefebvre and Moletta (2010) suggested to the use of a mixture of
halophilic organisms such as salterns will help to improve the pollutant removal efficiency.
In another study, Kubo et al. (2001) used Staphylococcus sp. and Bucillus cereus in
wastewater treatment of pickled plums, containing 15% of NaCl. They have reported a
removal efficiency of COD of 70% (in batch test) and 90% in a pilot plant when
Staphylococcus sp. and Bucillus cereus was used. Furthermore, Lay et al. (2010) reported
that a salt-in strategy to cope with osmotic pressure can be used to improve MBR
performance with salt water which could be bio-energetically less expensive, but requires
intracellular enzymatic systems. They have also suggested a compatible solute strategy
which is widely used but more expensive and could improve MBR performance with salt
water. Furthermore, SRT can also play an important role in treating wastewater with a high
salt concentration. Rene et al. (2008) reported lower organics and nitrogen removals at
SRT of 20 d compared with a SRT of more than 40 days (80% removal). In addition,
longer SRT may increase the concentration factor and increase the salt concentration (Lay
et al., 2010).
4.4.3.
The characterization of organic matter with different salt adoption strategies was
conducted to investigate how the major components of organics were affected by the
addition of salt in wastewater. The characterization of organics (hydrophobic and
hydrophilic) was conducted by LC-OCD. The hydrophilic organic compounds are bio-
polymers (consisting of protein and polysaccharides), humic acids, building blocks, low
molecular weight (LMW) acids and neutrals (Hubber, 2010).
A detailed characterization of MBR effluent, mixed liquor and membrane foulant was done
for better understanding of the gradual loading of salt on organic composition (Tables 4.5
and 4.6). The concentration of hydrophobic and hydrophilic compounds in feed water was
0.9 - 1.1 and 6.0 - 6.4 mg/L respectively. An increase of salt concentration increased the
effluent organic concentration by around 81%. The DOC values presented in this section
was measured on a daily basis. Each time the salt concentration increased, the organic
concentration increased initially within 2 days and afterwards steady state organic removal
was achieved. However, the variation of DOC concentration was minimal (less than 10%)
throughout the experimental period of 11 days which is presented as “±” (Tables 4.5 and
4.6). The increase of DOC was relatively bit higher during the first 2 days although the
variation was less than 10%. A significant change was observed with hydrophilic organic
concentration than hydrophobic organics. The concentration of hydrophilic organics in
MBR effluent increased from 0.9 - 1.6 mg/L (with 0 - 3 g-NaCl/L of salt) to 3.7 - 5.5 mg/L
when salt concentration reached to 5.0 - 35.0 g/L. Similarly, in mixed liquor it increased
from 1.6 - 4.9 (with 0-3 g-NaCl/L of salt) to 6.4 - 21.2 mg/L (with 5.0 - 35.0 g-NaCl/L of
salt). The amount of biopolymers increased by almost 95% (from 0.07 to 1.0 mg/L in MBR
effluent and 0.4 to 15.5 mg/L in mixed liquor) when salt concentration reached from 0 to
35 g-NaCl/L. Previous study reported 22 – 66% increase of soluble extracellular polymeric
substances due to addition of salt (Yogalakshmi and Joseph, 2010). Thus, it can be
assumed that a higher production of biopolymers could be due to the biosynthesis. As
stated earlier, plasmolysis of micro-organism caused the release of intracellular
constituents and soluble microbial products (Yogalakshmi and Joseph, 2010). A higher
production of biopolymers thus could resulted from plasmolysis due to changes in osmotic
pressure as the osmotic pressure of the water increases about 8 bar for every 10 g /L of
NaCl (Lay et al., 2010). Due to the higher osmotic pressure, the outer cell of
microorganism gets damaged which releases soluble microbial products as well as
extracellular polymeric substances. Literature shows that microorganisms respond to a salt
shock by aggregation of individual cells and acceleration of endogenous respiration,
accompanied by the release of organic cellular constituents such as soluble microbial
products and extracellular polymeric substances and cells autolysis (Reid et al., 2006).
However, the amount of humic type substances increased by around 80 - 85%. Moreover
building blocks and low molecular weight acid types substances were increased by around
60-90%. For all salt concentration (0 - 35 g-NaCl/L), the membrane foulant contained
biopolymers (7.5 - 18.3 mg/L), humics (8.8 - 17.6 mg/L), building blocks (7.3 - 13.1 mg/L)
and LMW neutral (5.6 - 16.0 mg/L) type substances. The results obtained from membrane
effluent and membrane foulant showed that more than 90% of the biopolymers substances
are rejected by the membrane. Whereas low molecular weight substances such as building
blocks and neutral could pass through the membrane.
In summary, in this study we did not measure the cellular carbon as sugar and protein
separately. However, we measured the biopolymers both in the MBR effluent as well as
the mixed liquor (Tables 4.5 and 4.6). Biopolymers contain polysaccharides with some
contribution from nitrogen-containing material such as proteins or amino sugars (Huber et
al., 2011). The results of Tables 4.5 and 4.6 clearly represent the formation or release of
biopolymers with increased salt concentration. Furthermore, literature shows that the
concentrations of protein and carbohydrate increase with higher salt concentration (Reid et
al., 2006). They have observed a rapid increase of protein and carbohydrate concentrations
when the salt concentration increased from 0 to 1 g/L NaCl, and further the changes were
slow. They also reported that the concentration of carbohydrate was higher than that of
protein concentration. In another study Yogalakshmi and Joseph (2010) found that the
amount of soluble extracellular polymeric substances increased by 22–66% after a shock
loading of NaCl (5-60 g/L). They also reported that both protein and carbohydrate
concentrations in soluble extracellular polymeric substances increased significantly
(protein concentration increased from 11-17 to 22-32 mg/gVSS and carbohydrate
concentration increased from 6-9 to 13–23 mg/gVSS).
Along with organic characterization, the dissolved organic nitrogen (DON) concentrations
present in biopolymers in MBR mixed liquor at different salt concentrations were also
measured (Figure 4.11). The DON in biopolymers was also measured by using LC-OCD.
Figure 4.11 clearly shows that the amount of DON present in biopolymer increased with
higher salt concentration. Due to the cell lyses resulting from salt addition higher
biopolymers were produced. In addition, the higher concentration of DON present in
biopolymers indicated that high concentration of protein in biopolymers as organic
nitrogen in biopolymers originates from proteins (Villacorte et al., 2006). Previous study
also showed an increase of protein concentration in soluble microbial products due to
addition of salt (Yogalakshmi and Joseph, 2010; Reid et al., 2006). Furthermore, the DON
present in MBR effluent was comparatively very low than mixed liquor in the range of 0
(with 0 to 5 g-NaCl/L) to around 0.12 mg/L (with 10 - 35 g-NaCl/L).
Table 4.5. Characterization of organic matter in MBR effluent operated at different
gradual loading of salt
Salt (g/L)
DOC
Dissolved
approximate molecular weights (g/mol) HOC CDOC >>20,000 ~1000 300-500 <350
BIO- Humic Building LMW Hydrophobic Hydrophilic polymers Subst. Blocks Neutrals
ppm-C ppm-C ppm-C ppm-C ppm-C ppm-C ppm-C
Membrane effluent
0 1.6±0.2 0.6±0.1 1.0±0.2 0.07±0.01 0.4±0.03 0.3±0.02 0.2±0.02
0.5 1.9±0.2 0.9±0.1 0.9±0.1 0.09±0.03 0.3±0.04 0.4±0.01 0.2±0.05
1 1.8±0.5 0.18±0.1 1.6±0.2 0.11±0.02 0.4±0.04 0.45±0.05 0.6±0.08
3 2.2±0.6 0.6±0.05 1.6±0.3 0.04±0.01 0.5±0.1 0.6±0.1 0.35±0.05
5 4.2±0.2 0.3±0.2 3.8±0.3 0.11±0.01 2.1±0.6 0.6±0.4 0.9±0.1
10 4.6±0.2 1.0±0.2 3.7±0.1 0.19±0.04 1.7±0.1 0.95±0.05 0.8±0.1
15 4.9±0.1 0.7±0.1 4.0±0.3 0.3±0.05 2.1±0.2 0.7±0.07 0.9±1.1
20 5.3±0.3 0.8±0.3 4.3±0.3 0.4±0.08 2.1±0.2 0.8±0.07 1.1±0.04
25 6.3±0.1 0.7±0.1 5.5±0.5 1.0±0.5 2.3±0.05 1.1±0.2 1.1±0.08
35 6.5±0.5 0.5±0.1 5.0±0.3 1.0±0.2 2.1±0.1 1.0±0.2 1.0±0.1
Table 4.6. Characterization of organic matter in MBR mixed liquor operated at different
gradual loading of salt
Salt (g/L)
DOC
Dissolved
approximate molecular weights (g/mol) HOC CDOC >>20,000 ~1000 300-500 <350
BIO- Humic Building LMW Hydrophobic Hydrophilic polymers Subst. Blocks Neutrals
ppm-C ppm-C ppm-C ppm-C ppm-C ppm-C ppm-C
Mixed liquor
0 2.0±0.2 0.4±0.01 1.6±0.2 0.4±0.02 0.6±0.07 0.3±0.09 0.3±0.04
0.5 4.0±0.3 1.1±0.2 3.0±0.3 0.3±0.1 1.0±0.1 0.7±0.1 0.8±0.05
1 4.1±0.2 0.9±0.04 3.2±0.2 0.6±0.07 0.7±0.05 0.6±0.01 1.4±0.06
3 5.5±0.7 0.6±0.05 4.9±0.5 1.1±0.2 1.9±0.3 0.9±0.1 0.8±0.09
5 7.0±0.6 0.5±0.06 6.4±0.7 1.5±0.5 3.3±0.5 0.5±0.2 1.0±0.2
10 8.6-0.4 0.7-0.2 7.8-0.8 2.5-0.3 2.4-0.05 1.4-0.4 1.5-0.08
15 11.7±0.6 0.7±0.01 11.0±0.6 5.0±0.3 3.0±0.1 1.3±0.2 1.8±0.1
20 19.4±2.0 0.6±0.3 18.7±1.5 10.2±1.0 3.6±0.2 1.2±0.4 3.7±0.5
25 22.1±1.1 0.9±0.1 21.2±1.2 14.0±0.2 3.2±0.7 2.0±0.1 2.8±0.2
35 25.0±1.5 1.5±0.3 23.0±1.5 15.5±1.3 4.0±0.3 1.5±0.1 2.6±0.2
Figure 4.11. Dissolved organic nitrogen (DON) concentration present in biopolymer at
different salt concentration in MBR mixed liquor.
4.4.4.
The TMP development in MBR under different salt concentration is presented in Figure
4.12. Figure 4.12 a presents TMP values for salt concentrations between 0-10 g-NaCl/L,
whereas Figure 4.12 b is 15-35 g-NaCl/L. From this figure it is found that TMP
development was not significant. The lower development of TMP could be due to lower
operating flux (2.5 L/m2.h). An air scouring of 1 m3/m2membrane area.h which was effective
enough to scour sludge particle away from the membrane surface. The TMP development
at different salt concentration of 0, 0.5, 1.0, 3.0, 5.0, 10.0, 15.0, 20.0, 25.0 and 35.0 g-
NaCl/L was around 2.75, 2.58, 1.91, 4.54, 3.21, 2.4, 8.72, 8.52, 6.14 and 10.53 mbar
respectively. From the previous study it is found that the treatment of saline sewage led
higher a TMP development (0.43 kPa/day) than that of fresh water sewage (TMP
development was 0.29 kPa/day) (Tam et al., 2006). A lower development of TMP
observed in this study is due to the operation of MBR at relatively low flux of 2.5 L/m2.h.
02468
1012141618
0 5 10 15 20 25 35
mg/
L
Salt concentration (g/L)
Biopolymer
DON in biopolymer
The development of TMP also depends on aeration rate, biomass concentration, etc. An
air scouring of 1 m3/m2 membrane area.h was used in the MBR to provide shearing stress
on the membrane surface. This air flow is high enough to effectively prevent the deposition
of sludge particles onto the membrane surface. The higher development of TMP with high
salt concentration (15-35 g-NaCl/L) could be due to higher production of organics as
discussed in section 4.4.3. Additionally in organic matter, especially the increase of
biopolymers could result in a lower membrane permeability of the process which showed
higher TMP development (Figure 4.12 a, b) i.e. higher membrane fouling. However, the
production of biopolymers may not be significant over a long-term operation. Lay et al.
(2010) reported that higher salt concentration in mixed liquor increased the osmotic
pressure and thus required a higher driving force to operate the MBR. High salt
concentration may also increase the viscosity. Thus, the increase of salt concentration can
increase the fouling by forming a more densely packed cake layer on the membrane
surface Lay et al. (2010). Furthermore, the fouling of a membrane is not only due to
organics alone but also due to the deposition of sludge particle on to membrane surface.
Figure 4.12. TMP development with time in MBR at different salt concentration (Flux 2.5
L/m2.h; 1 m3/m2membrane area.h)
0 50 100 150 200 2500
5
10
15
20
25 (a)
TM
P (m
bar)
Time (h)
0.0 g-NaCl/L 0.5 g-NaCl/L 1.0 g-NaCl/L 3.0 g-NaCl/L 5.0 g-NaCl/L 10.0 g-NaCl/L
0 50 100 150 200 2500
5
10
15
20
25
TM
P (m
bar)
Time (h)
0.0 g-NaCl/L 15.0 g-NaCl/L 20.0 g-NaCl/L 25.0 g-NaCl/L 35.0 g-NaCl/L
(b)
4.4.5.
Cluster analysis was conducted (using IBM SPSS statistics 19) to find out the similarities
between the performance under different salt concentration (Figure 4.13 a, b). The cluster
analysis was made using the DOC concentration of MBR effluent and mixed liquor under
different salt concentrations. In both cases (MBR effluent and mixed liquor) three distinct
clusters (i) 0 - 3.0, (ii) 5 - 15 and (iii) 20 - 35 g-NaCl were found. Each cluster indicates the
similarity with each other and different from those in other clusters. For example, the
removal of DOC under 0 to 3 g-NaCl/L in cluster 1 was almost similar and this was
different to those of cluster 2. From the results of ammonia and organic removal along with
the characterization of organic matter presented in section 4.4.2 and 4.4.3. It is also found
that the removal of ammonia and organics follows a trend which is represented in Figure
4.13 a, b.
Figure 4.13. Cluster analysis of DOC concentration at different load of salt (a) MBR
effluent and (b) mixed liquor (S.C = salt concentration)
From the experimental investigation on the effect of OLR, imposed flux and gradual
increase of salt on MBR operation following conclusions can be made.
The removal efficiency organic (in terms of DOC) and NH4-N was 93-98% and 83-88%
respectively with low OLRs of 0.5 - 1 kgCOD/m3.d. Further increase of OLR resulted
lower removal of organic and ammonia. The higher organic loading rates have negative
impact of membrane fouling. The TMP development was 0.1 - 0.24 kPa/h at higher OLRs
of 2.75 - 3.0 kgCOD/m3.d. From organic characterization it was found that bio-polymer
was the major foulant followed by humic substances, building blocks, lower molecular
weight neutrals and acid.
It was found that both imposed flux and aeration rates had a strong influence on membrane
fouling. However, the effect of flux on membrane fouling reduction was much higher than
that of aeration rate. Lower flux of 20 L/m2.h produced almost 75 - 90 times more water
than higher flux of 40 L/m2.h with an aeration rate of 0.3-1.2 m3/m2.membrane area.h. Further,
high flux (lower HRT) also had a significant effect on the composition of organics present
in the SMP and EPS. At higher flux or low HRT, both SMP and EPS had organics of high
molecular weight of around 48 kDa and as well as lower molecular weight organic of less
than 200 Da.
Additionally, from the effect of gradual increase of salt concentration it was found that a
gradual increment of salt concentration above 5 g-NaCl showed inhibitory effect on
organic and ammonia removal efficiency by MBR. The DOC and NH4-N removal reduced
from 77 and 93% to 10 and 0% when salt concentration reached to 35 g-NaCl/L.
Furthermore, at high concentration of salt (35 g/L) due to plasmolysis the production of
organics such as biopolymers, humics was also high which resulted in higher membrane
fouling. From the above findings it could be concluded that the MBR process could be
useful to treat saline water under low salt concentration. However, its performance can be
improved by acclimation for sufficient time or by acclimatizing halophiles the
microorganisms naturally.
Industrial applications of MBR prove its interest in wastewater treatment, due to its ability
in completely removing solids (micro-organism included), its superior removal of nutrient
and organic matter, high loading rate capabilities, low sludge production and small
footprint. This makes the MBR particularly suitable when water reuse is envisaged.
The major challenge in the membrane filtration systems is the control of membrane fouling
and its minimization during operation. There is a pressing need to minimize the fouling
potential and/or develop a simple method to measure and predict the fouling potential of
wastewater. Thus, in this study SMBR was operated with and without the addition of
support media in suspension. The SMBR was compared in terms of membrane fouling
with and without the addition of suspended medium in the membrane reactor operated at
different filtration flux. The suspended medium used in this study was granular activated
carbon (GAC; particle size 300 – 600 μm) at air scouring (aeration) rates of 0.5-1.5 m3/m2
membrane area.h. Experimental set up (Figure 3.1) and experimental details are provided in
chapter 3 (section 3.3).
5.2.1.
To study the effect of filtration flux and aeration (air scouring) rate on membrane fouling
(i.e membrane resistance), experiments were conducted with different flux of 5, 10, 20, 25
and 30 L/m2.h (which correspond to HRT of 10, 5, 2.5, 2 and 1.7 h respectively) at aeration
rates of 1.5 and 1 m3/(m2membrane area.h). The membrane resistances were measured after 6-7
days of operation (after acclimation for 45 days). From this experimental investigation, it
was found that at the higher aeration rate, the effect of flux on membrane resistance was
negligible (Figure 5.1 a). On the other hand, the reduction of aeration rate from 1.5 to 1
m3/(m2membrane area.h) resulted in a sudden rise of TMP (i.e. membrane resistance increased
suddenly; Figure 5.1 b) for a flux of 25 L/m2.h. This could be due to the accumulation of
sludge on to membrane surface. The total hydraulic resistance is twice higher and this
increase appears mainly due to the deposit accumulation (Rsc three times higher) but also to
irreversible interactions (Rp twice higher; Table 5.1).
Figure 5.1. Effect of filtration flux and aeration of membrane resistance (membrane area =
0.2 m2; reactor size = 10 L; SRT=20 days)
a) Effect of flux (operated at aeration rate of 1.5 m3/m2 membrane area.h)
Cumulative filtered volume (m3/m2. membrane area)
0 1 2 3 4
Mem
bran
e re
sist
ance
(1/m
)
1e+12
2e+12
3e+12
4e+12
(A) 5 L/m2.h; HRT=10h(B) 10 L/m2.h; HRT=5h(C) 20 L/m2.h; HRT=2.5h(D) 25 L/m2.h; HRT=2h
b) Effect of aeration (operated at flux of 20 and 25 L/m2.h)
Cumulative filtered volume (m3/m2. membrane area)
0 1 2 3 4
Mem
bran
e re
sist
ance
(1/m
)
1e+12
2e+12
3e+12
4e+12
(A) 1.5m3/m2 membrane area; 20 L/m2.h; HRT=2.5h(B) 1.0m3/m2 membrane area; 20 L/m2.h; HRT=2.5h(C) 1.0m3/m2 membrane area; 25 L/m2.h; HRT=2h(D) 1.5m3/m2 membrane area; 25 L/m2.h; HRT=2h
Table 5.1. Rsc; Rpb, Rm and Rt at different operating flux and aeration rate (membrane area
= 0.2 m2; reactor size = 10 L; SRT = 20 days)
Flux Aeration Rsc Rpb Rm Rt
(L/m2.h) (m3/m2.membrane area.
h)
x1011
(m-1)
(% of
Rt)
x1011
(m-1)
(% of
Rt)
x1011
(m-1)
(% of
Rt)
x1011
(m-1)
5 1.5 7.5 36.8 7.0 34.5 5.8 28.7 20
10 1.5 8.2 37.7 7.7 35.5 5.8 26.8 22
20 1.5 8.9 39.9 7.7 35.0 5.6 25.1 22
25
1.5 9.6 41.8 8.0 34.8 5.4 23.4 23
1 27.9 54.5 17.5 34.5 5.7 11.0 51
5.2.2.
From the experimental results with higher flux (25 L/m2.h) at lower aeration (1.0 m3/m2
membrane area.h), there was a sudden rise of membrane resistance. This sudden rise of
membrane resistance could be minimized using a medium in suspension in MBR. This
medium will scour the foulant deposited on the membrane surface by producing extra
shearing stress. In this study, granular activated carbon (GAC, particle size of 300-600 μm)
was used in suspension (0.5-2 g/L of volume of reactor). From the Figure 5.2 (a) it was
found that the use of suspended media prevented sudden rise of membrane resistance. This
may be due to the extra shearing effect on to membrane surface produced by suspended
media which prevent (minimize) the deposition of particles by scouring. This resulted in
lower value of Rt (reduced from 51x1011 to 20x1011 m-1) (Table 5.2). The addition of GAC
also will adsorb some of the organic matter prior to their entry to the membrane. At lower
aeration (0.5 m3/m2membrane area.h) and higher flux, membrane resistance increased (Figure
5.2 b, c and Table 5.2). Similarly, membrane resistance increased when the dose of
suspended media was decreased (Figure 5.2 d). This shows that, the amount of suspended
medium has major effect on fouling reduction. Thus it is important to select the suspended
medium at appropriate concentration according to aeration rate and flux imposed.
Figure 5.2. Effect of suspended media on membrane resistance (membrane area = 0.2 m2;
reactor size = 10 L; SRT = 20 days)
a) Effect of suspended media (operated at a flux of 25 L/m2.h with aeration rate of 1 m3/m2 membrane area.h; HRT=2h)
Cumulative filtered volume (m3/m2. membrane area)
0 1 2 3
Mem
bran
e re
sist
ance
(1/m
)
1e+12
2e+12
3e+12
4e+12
5e+12
6e+12 (A) without suspended media(B) with suspended media (GAC @ 2g/L)
b) Effect of aeration on suspended media (operated at a flux of 25 L/m2.h; HRT=2h)
Cumulative filtered volume (m3/m2. membrane area)
0 1 2 3
Mem
bran
e re
sist
ance
(1/m
)
1e+12
2e+12
3e+12
4e+12
5e+12
6e+12 (A) 0.5 m3/m2 membrane area.h(B) 1.0 m3/m2 membrane area.h(C) 1.5 m3/m2 membrane area.h
c) Effect of flux (operated at aeration rate of 1.0 m3/m2 membrane area.h with suspended media GAC @ 2g/L)
Cumulative filtered volume (m3/m2. membrane area)
0 1 2 3
Mem
bran
e re
sist
ance
(1/m
)
1e+12
2e+12
3e+12
4e+12
5e+12
6e+12 (A) 25 L/m2.h; HRT=2h(B) 30 L/m2.h; HRT=1.7h
d) Effect of concentration of suspended media (operated at 25 L/m2.h with an aeration rate of 1.0 m3/m2 membrane area.h;HRT=2h)
Cumulative filtered volume (m3/m2. membrane area)
0 1 2 3
Mem
bran
e re
sist
ance
(1/m
)
0
1e+12
2e+12
3e+12
4e+12
5e+12
6e+12 (A) 0.5 g/L(B) 1.0 g/L(C) 2.0 g/L
Table 5.2. Rsc; Rpb, Rm and Rt at operating flux of 25 L/m2.h with and without GAC in
suspension at different dose and aeration rates:
GAC concentration Flux Aeration Cumulative
filter volume
Rsc Rpb Rm Rt
(g/L of volume of reactor) (L/m2.h) (m3/m2.membrane
area. h) m3/m2.
membrane area
x1011
(m-1) (% of
Rt) x1011
(m-1) (% of
Rt) x1011
(m-1) (% of
Rt) x1012
(m-1)
0 25 1 1.88 27.9 54.5 17.5 34.3 5.7 11.2 5.1
2 25 1 2.6 7.4 37.5 6.9 34.7 5.5 27.8 2.0
2 25 0.5 2.5 22.1 57.6 10.6 27.6 5.6 14.8 3.8
2 25 1.5 3.0 7.9 38.8 6.98 34.2 5.5 27.0 2.0
2 30 1.0 2.3 26.7 69.0 7.7 20.3 5.6 14.7 4.0
Table 5.2 clearly shows the importance of adding of GAC (i) to maintain low values of Rpb
irrespective of aeration rates and permeate flux and (ii) to maintain Rsc as significantly low
values. Nevertheless it is important to maintain a sufficient aeration rate (in our case 1
m3/m2membrane area. h) and (iii) work at practically high permeate flux in order to maintain a
good functioning of MBR under peak flows and also to reduce MBR footprint.
5.2.3.
The removal efficiency of dissolved organic matter and NH4-N was high when the MBR was
operated at a lower flux of 5 L/m2.h (Table 5.3). Here NH4-N was oxidised into NO3-N
(nitrification). At lower flux, the micro-organisms will have sufficient retention time to
assimilate the organic molecules and nutrients. Other researchers also found higher rate of
nitrification at lower flux (i.e. at higher HRT) (Viero, and Sant’Anna, 2007, 2008). Results
show a decrease of dissolved organic carbon (DOC) removal was low when the MBR was
operated at HRT of 2 h (corresponding flux of 25 L/m2.h). Nitrification was negligible when
the HRT was lower than 5 h. In any case, there was no removal of phosphates which confirm
the continuous aerobic working conditions.
The aeration rate (at least within the range studied) did not have any major effect on the
degree of organic matter removal as well as on nitrification. Similarly, suspended media also
had less effect on nitrification but had an influence on organic removal (Table 5.4). As GAC
was used as suspended medium, it adsorbed some of organic molecules before mixed liquor
enter the membrane. The specific oxygen uptake rate (SOUR) and dissolved oxygen (DO)
were between 10-12 mgO2/VSS/h and 2.8-3.6 g/L respectively for all the conditions used.
Table 5.3. Effect of operating flux on organic matter removal and on nitrification at an
aeration rate of 1.5 m3/m2.membrane area. h (membrane area = 0.2 m2; reactor size = 10 L;
SRT = 20 days):
Flux
(L/m2.h)
MLSS
(mg/L)
MLVSS
(mg/L)
DOC
(mg/L)
PO4-P
(mg/L)
NO2-N
(mg/L)
NO3-N
(mg/L)
NH4-N
(mg/L)
5
(HRT=10h)
Influent
4.3 3.6
123.0 2.72 0.02 1 10.1
Effluent 2.9 2.87 0.02 5 1.3
10
(HRT=5h)
Influent
5.1 4.6
90.0 1.34 0.01 0.8 4.9
Effluent 0.87 1.34 0.02 6.1 2.3
20
(HRT=2.5h)
Influent
5.3 4.3
63.0 0.52 0.02 1.5 2.4
Effluent 0.623 0.28 0.03 1.7 1.1
25
(HRT=2h)
Influent
5.1 4.7
35.0 0.96 0.01 2.7 2.3
Effluent 5.25 0.97 0.02 1.6 2.2
Table 5.4. Effect of suspended media on nutrients removal operated at a aeration rate of 1.5
m3/m2.membrane area.h (membrane area = 0.2 m2; reactor size = 10 L; HRT = 2 h; SRT = 20
days):
Flux
L/m2.h
Suspended media
(g/L of volume
of reactor)
DOC
(mg/L)
PO4-P
(mg/L)
NO2-N
(mg/L)
NO3-N
(mg/L)
NH4-N
(mg/L)
25
(HRT=2h)
Influent - 35.0 0.96 0.01 2.7 2.3
Effluent 0 4.72 0.88 0.02 1.1 1.8
Effluent 2 1.27 0.81 0.01 1.1 1.8
5.2.4.
The MWD distribution of organic matter in the MBR effluent, SMP, EPS and foulant is
presented in Figure 5.3 (a-d) when operated at higher permeate flux of 25 L/m2.h with an
aeration rate of 1.0 m3/m2membrane area.h with and without suspended media (GAC). The figure
showed narrow peak (A) in the spectra (detention time of 15.7 min and MW of around 1000-
1200 Da) in all four cases. This corresponds to humic type substances.
Without the GAC addition, the intensity of the peaks was in the order of EPS > foulant >
SMP > Effluent. SMP showed small peaks besides ‘A’ reflecting a wider range of organic
molecules of molecular weight of 1200-180 Da (peak detection time of 15.7 to 18 min). On
the other hand, EPS showed a broader range of spectrum with strong intensity after peak ‘A’.
Those molecules correspond to molecular weight range of 1200-180 Da (peaks detected at
15.7 to 18 min) are composed of humic acids, low molecular fulvic acid and low molecular
weight of neutrals. Besides, the spectra showed a strong peak at 9 min which corresponds to
biopolymers (MW of 35-40 kDa). These results indicate that EPS contained a wide range of
organic substances from bio-polymers to low molecular weight of organics. On the other
hand, foulant showed only two peaks, a small biopolymer at 9.7 min and a peak ‘A’ at 15.7
min that corresponds to humic type substances. This implies that foulant does not compose of
wide range of organic matter but only few selective organics.
In the presence of the suspended medium, although the peak spectra ‘A’ as well as other
spectra appeared in effluent, SMP, EPS and foulant, but the intensity of peaks was very low
as compared to those without GAC. These results show that the use of GAC as suspended
medium helped to absorb some smaller molecular weight of organic matter (ranging from
1200-150 Da). Other researcher also found similar results when adsorbent was introduced in
suspension in the SMBR (Guo et al., 2005; Chen et al., 2006; Akram, and Stuckey, 2008;
Lesage et al., 2008).
Figure 5.3. MWD distribution of organic matter in the MBR A) effluent; B) SMP; C) EPS
and D) foulant with and without suspended medium (flow rate = 25 L/m2.h; aeration rate =
1.0 m3/m2membrane area. h; suspended medium (GAC) @ 2g/L of volume of reactor)
5.2.5.
Fluorescence spectroscopy (excitation emission matrix) has been recently used to study a
wide range of organics present in wastewater and membrane foulant (Aryal et al., 2008;
2009). Every excitation emission spectra would be useful when studying the chemical
properties of organics of various origins. Based on nature of organic matter and its origin, the
spectrum is generally divided into five groups (Chen et al., 2003). These groups are i)
aromatic proteins (Ex : Em 200 – 250 : 280 - 330); ii) amino acid substances (Ex : Em 200 –
250 : 330 - 380); iii) the peptides and proteins (microbial byproducts) (Ex : Em 250 – 340 :
280 - 380); iv) fulvic acids type substances (Ex : Em 200 – 250 : 380 - 500) and v) humic
acids type substances (Ex : Em 250 – 500 : 380 - 500)
Figure 5.4 (a-d) shows the EEM spectra of effluent, SMP, EPS and foulant of MBR without
suspended medium. Without the addition of GAC as suspended medium, the EEM figure
shows that effluent has negligible organics whereas SMP had organics of small molecular
weight (aminoacid type; Ex : Em 200 - 250 : 330 - 380) and some fulvic acid type substances
(Ex : Em 200 – 250 : 380 - 500). EPS contained a wide range of organics that included amino
acids, biopolymers, humics and fulvics. Compared to EPS, the foulant contained mainly
humics and fulvic, with small amount of biopolymers (Ex : Em 280 – 320 : 310 - 330) that
corresponded polymers of lower molecular weight. This is supported by chromatogram
(Figure 5.3 d) where we observed polymer peak near to 36 - 30 kDa.
Figure 5.4. EEM distribution of A) effluent; B) SMP; C) EPS and D) foulant (flow rate = 25
L/m2.h; membrane area = 0.2 m2; aeration rate = 1.0 m3/m2membrane area.h; reactor size = 10 L;
SRT = 20 days)
The EEM spectra of effluent, SMP, EPS and foulant of SMBR (in the presence of GAC as
suspended medium) are shown in Figure 5.5 (a - d). The SMBR effluent contained negligible
amount of organics. However, SMP contained amino acid peaks as well as small biopolymer
peak in the EEM spectra. EPS showed very similar spectral pattern to that of EEM without
suspended medium (GAC) but had lower intensity. This shows that suspended GAC acted as
an adsorbent for biopolymers as well as humic and fulvic acid types substances. The foulant
showed peaks only in humic and fulvic regions. This result shows that GAC helped to reduce
the overall organics deposit possibly by scouring the membrane surface as well as by
adsorbing organics. Besides, no biopolymer peaks appeared in the EEM spectra (Figure 5.5
d). The result shows that suspended GAC media inhibit formation of biopolymer. Aryal et al
(2009) reported that the biopolymer formation occurs possibly by conversion of lower
molecular weight amino acids. Thus one could conclude that GAC possibly adsorbed low
molecular weight organics deposited on the membrane surface during scouring and inhibited
the formation of biopolymer.
Figure 5.5. EEM distribution of MBR A) effluent; B) SMP; C) EPS and D) foulant (flow rate
= 25 L/m2.h; aeration rate = aeration rate = 1.0 m3/m2membrane area.h; with suspended media
GAC @ 2g/L of volume of reactor).
In this study the effect of different particle sizes of granular activated carbon (GAC) on the
performance of a submerged membrane bioreactor (SMBR) was investigated. The sizes of
GAC used were 150 - 300, 300 - 600 and 600 - 1200 μm. In this study, a SMBR was operated
with a flux of 20 L/m2.h which corresponded to a hydraulic retention time (HRT) of 2.5 h.
The sludge retention time (SRT) was maintained at 20 days by withdrawing a predetermined
quantity of sludge every day. A detailed organic matter characterization of membrane foulant,
soluble microbial product and extracellular polymeric substances were carried out to
investigate the organic matters such as bio-polymers type substances together with humic acid
and lower molecular neutral and acids.
5.3.1.
The concentration of biomass in terms of MLSS and MLVSS are presented in Table 5.5.
After the acclimation period, the concentration of MLSS and MLVSS did not vary (only
minor changes of 10%) with the addition of different sizes of GAC particles. The food to
microorganism (F: M) ratio was almost constant within the range of 0.21 - 0.23 d-1. The SVI
values (Table 5.5) were below 150 mL/g which indicate a good settlability of the sludge. The
use of GAC as suspended medium in the size ranges of 150-300 and 300-600 μm helped to
reduce the SVI by around 30 - 40 % (Table 5.5). This indicated that the use of GAC prevented
the bulking properties of sludge. Previous study also reported lower SVI with the addition of
powdered activated carbon (PAC) in MBR due to the role of PAC in incompressible floc
formation (Satyawali and Balakrishnan, 2009). Another study by Li et al. (2005) reported that
the sludge viscosity was reduced by 45% with the addition of PAC. In all cases, the
concentration of dissolved oxygen was more than 2 mg/L.
Table 5.5. Biomass concentration and sludge properties with and without the addition of
GAC in SMBR
GAC particle size
(μm)
MLSS
(g/L)
MLVSS
(g/L) F:M
SVI
(mL/g)
DO
(mg/L)
0 6.6±0.6 5.20±0.2 0.23 90±5
4.5-6.5
150-300 7.0±0.5 5.48±0.3 0.21 50±5
300-600 6.6±0.4 5.41±0.3 0.23 50±5
600-1200 6.8±0.1 5.85±0.5 0.23 65±5
5.3.2.
The removal of organic matter, ammonia and phosphate with and without the addition of
GAC of different particle sizes is presented in Table 5.6. The removal of DOC and COD
without the addition of GAC was 89.2 ± 0.9 and 84 ± 0.3% respectively. The removals of
DOC and COD with the addition of GAC of different particle sizes were 94 – 95% and 93 -
95% respectively. Thus, an additional 10% DOC and COD removal could be achieved by
incorporating GAC in the SMBR. The DOC concentration in the SMBR effluent with the
addition of GAC is almost half (5.9 - 7.5 mg/L) than that without the addition of GAC (DOC
was around 14.6 mg/L). The influent DOC and COD concentrations were 129.8 and 158.6
mg/L respectively. The COD concentration in SMBR effluent with the addition of GAC was
also almost 35 - 40% lower than that without GAC addition. This indicated that the addition
of GAC helped to further reduce organic matter. The higher removal of organic matter with
GAC could be attributed to the adsorption of organics by GAC. Among of the three different
sizes of GAC, the highest removal (95.6 ± 0.7%) of DOC was found with the smallest particle
of GAC (150 - 300 μm) which can be explained in terms of the higher surface area. The
removal of DOC with GAC particle sizes of 300 - 600 and 600 - 1200 μm was around 94%.
In addition, the consumption of DOC and COD with and without the addition of GAC of
different particle size were around 0.21 g-DOC/g-MLVSS•d and 0.25 g-COD/g-MLVSS•d
respectively.
The removal of ammonium nitrogen without the addition of GAC was 36.6 ± 1.4%. The
removal of ammonium nitrogen with the addition of GAC of different particle sizes was 35 -
45%. The lower removal of ammonium nitrogen with and without the addition of GAC could
be due to the low nitrification. Further, GAC could not adsorb ammonium ions. Another
possible reason of the lower removal of ammonium nitrogen or partial nitrification could be
due to lower hydraulic retention time (2.5 h) employed in the SMBR. This time was not
sufficient for the microorganism to assimilate the nutrients. From literature it is found that in a
conventional membrane bioreactor the nitrification occurs with higher HRT of 6 h or more
(Viero et al., 2007). The removal of nitrogen was also assessed in-terms of Total Kjeldahl
Nitrogen (TKN) as well as total nitrogen. The nitrogen source of the synthetic wastewater
mainly comes from the NH4Cl (7.1 mg-NH4+/L) and a very small amount (0.3 mg/L) as
dissolved organic nitrogen (DON) from beef extract. The concentration of NO3- and NO2
- in
synthetic wastewater was 0.5 and 0.01 mg/L respectively. The concentration of Total Kjeldahl
Nitrogen (TKN) in synthetic wastewater was 5.7 mg/L as N. TKN is the sum of ammonia-
nitrogen and dissolved organic nitrogen. The removal of TKN without and with the addition
of different particle sizes of GAC of 150 - 300, 300 - 600 and 600 - 1200 was between 22 -
36% (Table 5.6). On the other hand the removal efficiency of total nitrogen with and without
the addition of GAC of different sizes was relatively low and between 9.4 ± 3.4 to 21.3 ±
3.4%. Total nitrogen is the sum of nitrate-nitrogen, nitrite-nitrogen, ammonia-nitrogen and
dissolved organic nitrogen. The lower removal of nitrogen in-terms of total nitrogen was due
to the presence of higher concentration of NO3- in the effluent due to partial nitrification. A
mass balance calculation of nitrogen was made to estimate the amount of nitrogen that was
nitrified. The nitrified amount is the sum of the differences between the effluent and influent
nitrite nitrogen and nitrate nitrogen. However, the concentration of nitrite could be neglected
as the value of nitrite was very low (less than 0.03 mg/L). Only significant readings were
nitrate nitrogen, dissolved organic nitrogen and ammonia nitrogen. From the calculation it
was found that the amount of nitrogen nitrified was 0.8 - 1.0 mg/L with and without the
addition of GAC of different sizes. From the calculation of the assimilated amount of
nitrogen, it was found that the amount of nitrogen assimilated without the addition of GAC
was 0.75 ± 0.2 mg/L. The amount of nitrogen assimilated with the addition of different sizes
of GAC of 150 - 300, 300 - 600 and 600 -1200 μm were 0.55 ± 0.2, 1.25 ± 0.2 and 1.05 ± 0.2
mg/L respectively. The amount of assimilated nitrogen was calculated from the differences
between the total influent and effluent nitrogen. Thus, the nitrogen balance calculation along
with TKN and nitrogen removal efficiency indicated a lower removal efficiency of nitrogen.
The removal of phosphate without the addition of GAC was 43.7 ± 6.0% whereas with the
addition of GAC of different particle sizes were in the range of 35 - 45%. Phosphorus present
in wastewater can be removed by precipitation and/or adsorption, or by luxury uptake. Only a
small amount of phosphorus is used for cell metabolism and growth (Radjenovi´c et al.,
2008). To calculate the amount of P adsorbed on the membrane surface, a filtration test (with
same membrane) with synthetic water spiked with only P was conducted. The amount of P
removed by the membrane was less than 5%. This indicated that the adsorption of P on the
membrane surface was minimal. Furthermore, the GAC used as suspended medium also
showed no adsorption of P. Thus, the P removal could only be due to cell metabolism and
growth. From the influent and effluent P concentrations presented in Table 5.6 a mass balance
of P removal was calculated. The total amount of P coming to the SMBR system was about
100.2 mg-P/day. The amount of P retained in the SMBR with and without the addition of
GAC of different particle sizes were in the range of 34 - 43.8 mg-P/d. The amount of mixed
liquor sludge discharged everyday was 0.5 L/d. Thus the amount of P retained in the SMBR
following sludge removal was between 32 - 41.5 mg-P/d. Therefore the amount of P utilized
for cell metabolism and growth during the SMBR operation with and without the addition of
GAC was 0.6 to 0.7 mg-P/g-MLSS•d and was relatively low and coincides with the removal
of P presented in Table 4.6. In addition, the COD : P ratio could also be used to assess the
performances of biological nutrient removal in the SMBR process (Galil et al., 2009). From
the COD and P removal (presented in Table 5.6) it was found that the COD-used : P-removed
ratio was between 291 to 412 with and without the addition of different sizes of GAC particle.
The higher COD-used : P-removed ratio indicated that the SMBR process was P limited
(Galil et al., 2009).
Table 5.6. Removal of organic, ammonia and phosphate with and without the addition of GAC in SMBR (all the concentrations are in mg/L)
GAC
particle size
(μm)
DOC COD NH4+ NO3
- NO2- DON TKN as N Total N PO4
3-
Influent 129.80 158.60 7.10 0.50 0.01 0.20 5.70 5.85 3.20
0
Effluent
14.6±1.2
(89.2±0.9)
25.3±0.5
(84.0±0.3)
4.5±0.1
(36.6±1.4)
5.01±.2
0.01 0.6±0.05
4.0±0.2
(29.8±3.8)
5.1±0.2
(12.8±3.4)
1.8±0.2
(43.7±6.0)
150-300 5.9±0.9
(95.6±0.7)
8.0±0.5
(95.0±0.3)
4.6±0.2
(35.2±2.8)4.1±0.1
0.010.8±0.01
4.4±0.2
(22.8±3.5)
5.3±0.2
(9.4±3.4)
1.9±0.3
(40.6±9.0)
300-600 7.3±0.5
(94.5±0.4)
10.5±1.0
(93.3±0.6)
4.1±0.2
(42.2±2.8)4.3±0.2
0.030.6±0.01
3.7±0.2
(35.0±3.5)
4.6±0.2
(21.3±3.4)
2.1±0.2
(34.3±6.2)
600-1200 7.5±1.5
(94.2±1.2)
10.5±0.5
(93.3±0.3)
3.9±0.5
(45.0±7.0)4.5±0.3
0.010.6±0.02
3.6±0.4
(36.8±7.0)
4.8±0.2
(17.9±3.4)
2.1±0.2
(34.3±6.2) Note: The values within the bracket ( ) is the % removal efficiency and values without bracket is the concentration in mg/L.
5.3.3.
A detailed organic characterization of SMBR effluent was made using LC-OCD to
investigate the changes in different types of organic matter with and without the addition of
different particle sizes of GAC. From the characteristics of organic matter present in SMBR
effluents, it is found that the removal of both hydrophobic and hydrophilic organic matter
increased with the addition of GAC (Table 5.7). Without the addition of GAC, the removal of
hydrophobic organic was 86% whereas, with the addition of GAC it was around 96%. On the
other hand, the concentration of hydrophilic organic matter in the SMBR effluent with and
without the addition of different particle sizes of GAC was 4.4 - 6.2 and 12.2 mg/L
respectively. This indicated that the use of GAC as suspended media also helped to reduce
hydrophilic organic matter by 50 - 60%. Nguyen et al. (2012) also reported that GAC can
effectively remove hydrophobic as well as hydrophilic organic matter. They have reported
that GAC can adsorb higher amounts of hydrophobic compounds, whereas the removal of
hydrophilic compounds is due to hydrophobicity-independent mechanisms such anion
exchange, surface complexation and hydrogen bonding which play significant roles in
sorption of organic/trace organics onto GAC. Among the three different particle sizes of
GAC, the smallest size (150 - 300 μm) of GAC removed relatively higher amount of
hydrophilic organic matters. Furthermore, the addition of GAC of different particle sizes as
suspended medium reduced the concentration of biopolymers by 20% (from 1.53 to 1.20
mg/L). The concentration of humic, building block and low molecular weight (LMW) neutral
and acids in SMBR effluent with the addition of different particle sizes of GAC was reduced
by 66 - 76%, 20 - 50%, 30 - 56% respectively. From these results it is found that the addition
of GAC helped to remove lower molecular weight of organics than that of higher molecular
weight organics (such as biopolymers). From the literature it is found that the lower removal
of biopolymers by GAC could be due to the high molecular weight of biopolymers which
prevented access to the internal pore structure of the GAC particles (Velten et al., 2011).
Table 5.7. Organic characteristics of SMBR effluent with and without the addition of GAC in SMBR (all the units are in mg/L) GAC particle size (μm)
DOC HOC (hydrophobic)
CDOC (Hydrophilic)
Biopolymers Humic Building blocks LMW neutralsand acids
Influent 129.8 26.80 102.0 5.04 56.2 21.5 19.82
0
Effluent
14.6±1.2 3.50±0.1 12.20±0.5 1.53±0.1 4.84±0.1 2.15±0.2 2.65±0.5
150-300 5.90±0.9 1.30±0.1 4.60±0.2 1.23±0.1 1.15±0.2 1.10±0.2 1.21±0.5
300-600 7.30±0.5 1.40±0.5 6.40±0.5 1.20±0.1 1.21±0.5 1.30±0.1 1.56±0.2
600-1200 7.50±1.5 1.40±0.5 6.80±0.3 1.20±0.1 1.61±0.7 1.76±0.1 1.90±0.5
5.3.4.
The TMP development profile with time with and without the addition of different particle
sizes of GAC is presented in Figure 5.6. The development of TMP was high (38.9 kPa)
without the addition of GAC. The development of TMP with different sizes of GAC of 150
- 300, 300 - 600 and 600 - 1200 μm were 21.3, 16.0 and 28.5 kPa. The lower development
of TMP with the addition of GAC is due to the combined effect of adsorption of organic
matter by GAC and extra mechanical scour on the membrane surface created by GAC used
as suspended medium. Previous studies reported a significant reduction of membrane
fouling by the addition of support/suspended media into MBR. For example Fang et al.
(2006) reported a 22% reduction of membrane cake resistance by the addition of PAC. Li
et al. (2005) reported a 44% reduction of membrane resistance with the addition of PAC.
Figure 5.6. Transmembrane pressure (TMP) development profile with time with and with
the addition of different particle sizes of GAC (A = without GAC; B = with GAC particle
size of 300 - 600 μm; C = with GAC particle size of 150 - 300 μm; D = with GAC particle
size of 600 - 1200 μm).
05
1015202530354045
0 20 40 60 80 100 120 140 160 180 200 220
TM
P de
velo
pmen
t (kP
a)
Time (h)
ABCD
The lowest TMP development (16 kPa) was achieved with GAC particle sizes of 300 - 600
μm. The second lowest TMP development was with GAC particle sizes of 150 - 300 μm of
21.3 kPa. The lower (around 25%) development of TMP with GAC particle size of 300 -
600 μm could be due to higher mechanical scour on to membrane surface. The higher
reduction of TMP development with GAC particle size of 300 - 600 μm can also be
validated from the membrane resistance results reported in Table 5.8. The values of total
membrane resistance (Rt), cake resistance (Rc) and pore blocking resistance (Rp) are
presented with and without the addition of GAC of different particle sizes as suspended
medium. The values indicated that with the addition of GAC the Rt reduced by 60% (from
35.00 to 13.00 x 1012 m-1). It is also found that Rc and Rp with the addition of GAC of
particle size of 300 - 600 μm was lower (Rc = 8.2 x 1012 and Rp = 4.5 x 1012 m-1) than that
with particle sizes of 150 - 300 and 600 - 1200 μm (Rc = 11.2 and 15.6 x 1012 m-1 and Rp =
6.9 and 10.2 x 1012 m-1 respectively). Some of the larger particles of GAC (600 - 1200 μm)
used as supported medium, was found to settle down after 5 days of operation, thus
resulting higher TMP development afterwards. Smaller particles of GAC of 150 - 300 μm
led to relatively higher Rt and Rc values than GAC particle sizes of 300 - 600 μm. This
may be because due to the fact that the smaller GAC particle was not able to prevent the
deposition of sludge particle on the membrane surface. This helped to building
comparatively compact cake layer resulting in a higher Rc value. Although smaller particle
size (150 - 300 μm) of GAC showed slightly higher organic removal (Table 5.6 and 5.7)
than larger particle size of GAC of 300 - 600 μm but showed slightly higher TMP
development. Fouling of the membrane was not only due to the deposition of organics onto
the membrane surface but also due to the deposition of sludge particles (higher Rc value)
onto membrane surface. Thus, it is very important to choose the correct size of the GAC
particles to optimize organic removal with lower TMP development. In addition, the
settling of larger particle size of GAC of 600 - 1200 μm could be minimized by using a
higher aeration rate. However, the use of higher aeration rate is not cost effective.
Table 5.8. Rt, Rc and Rp with and without the addition of different particle of GAC
(membrane resistance, Rm = 0.59 x 1012 m-1)
GAC particle size
(μm)
Rt Rc Rp
(x 1012
m-1
) (x 1012
m-1
) (% of Rt) (x 1012
m-1
) (% of Rt)
0 35.00 20.4 58.28 14.1 40.28
150-300 19.20 11.2 58.33 6.9 35.90
300-600 13.00 8.2 63.08 4.5 34.60
600-1200 25.90 15.6 60.23 10.2 39.38
The effect of GAC on the membrane surface after the addition of GAC was tested
physically by measuring the clean water flux and the turbidity of the MBR effluent. It was
found that the GAC particles did not have any adverse effect on the membrane surface as
the clean water flux was the same as that of a virgin membrane and as the filtered turbidity
was reasonably low (less than 0.2 NTU). This result is also in agreement with the findings
of others. For example Siembida et al. (2010) used granular polypropylene particle size of
2.2 - 3.0 mm in a submerged membrane bioreactor for more than 600 days. They examined
the membrane surface with SEM and found only brush marks on the membrane surface
which did not affect the membrane performance.
From the experimental investigation, it is evident that the use of GAC could help to reduce
the membrane fouling resulting from the adsorption of organic matter and by providing
extra scouring effect on the membrane surface. However, it is important to investigate the
long term effect of GAC on the adsorption of organic matter.
From the experimental investigation it can be concluded the sudden rise of membrane
resistance could be minimized by incorporating a suspended medium in suspension in
MBR. The use of granular activated carbon in suspension (0.5 - 2 g/L of volume of reactor)
prevented the sudden rise of TMP (i.e. reduced membrane resistance) by producing extra
shearing effect on to membrane surface and reducing deposition of particles on to
membrane surface by scouring. The dose of suspended medium (here GAC) also had a
major effect on fouling reduction. Thus, a suitable amount of suspended medium need to
be used depending on the flux and aeration (or air scour) rate used. Further, the size of
GAC particle also played a major role on membrane fouling reduction. Total membrane
resistance (Rt) reduced by 60% with GAC particle sizes of 300 - 600 μm than smaller (150
- 300 μm) and larger (600 - 1200 μm) particle size of GAC. Thus, it is very important to
find out the suitable size of the medium when used as suspension in the SMBR.
The use of suspended media helped to reduce membrane fouling. The removal of organics
with the addition GAC as suspended medium was high (95%) than without the use of GAC
as support media in suspension. Moreover, GAC used as suspended medium also helped to
adsorb low molecular weight organics deposited on the membrane surface during scouring
and inhibited the formation of biopolymer.
The traditionally used MBR cannot remove nutrients (especially phosphorus) to an
appropriate amount. Further, the removal mechanism of nutrients by MBR is also complex.
Thus, a cost effective, more reliable and suitable physico-chemical technologies needed to
be used as a post treatment to MBR for the removal of nitrate and phosphate. Among the
different physic-chemical technologies such as reverse osmosis, electro dialysis, chemical
precipitation and ion exchange/adsorption is the most suitable process for the removal of
nitrate and phophste due to its simplicity, effectiveness and relatively low cost ((Bhatnagar
and Sillanpää, 2011). Another major advantage of ion exchange/adsorption process is its
ability to handle shock loadings and its capacity in functioning over a wide range of
temperatures. Further, ion exchange resins can easily be regenerated for several cycles and
the regenerant can be used as fertilizer. Thus, in this study, various ion exchangers such as
zirconium hydroxide, hydrated ferric oxide (HFO) and purolite ion-exchange resins
(A500P and A520E) were used to remove nutrients.
First part of this study was conducted with zirconium hydroxide for the removal of
phosphate from the wastewater. Various aspects in the removal of phosphate by Zr
hydroxide were studied. They are (1) the phosphate adsorption capacity of Zr hydroxide as
influenced by pH, temperature, and co-existing anions; (2) the kinetics and
thermodynamics of adsorption; and (3) incorporation of this adsorbent into a membrane
filtration adsorption hybrid reactor (MAHF) system for the continuous removal of
phosphate from wastewater. This is to demonstrate the potential of Zr hydroxide in
continuously removing phosphate from wastewater with minimum daily replacement of
adsorbent in the membrane reactor. The experimental details are discussed in chapter 3
(sections 3.3.4, 3.3.5 and 3.3.6). In order to study the above effects in detail, a synthetic
wastewater with high phosphate concentration was used instead of MBR effluent which
has varying amount of P at varying concentration. The characteristic of the wastewater is
given in Table 3.8.
The second set of the experiment was conducted with an anthracite filter mixed with
hydrated ferric oxide (HFO) to remove PO43- from the SMBR effluent operated with the
addition of GAC (particle size of 300 – 600 μm; chapter 5 section 5.3.2). This was
conducted to investigate the efficiency of HFO in removing phosphorus from MBR
effluent.
The third set of the experiment was conducted with MBR-purolite ion exchange combined
system. Two main objectives of the treatment process are the organic carbon removal and
nutrient recovery. The treatment train comprises: (a) primary treatment (step 1): removal of
mainly organic carbon through optimizing of the MBR at the lowest feasible HRT. (b)
post-treatment (step 2): an ion-exchange process to remove nitrogen and phosphorus which
was later recovered when the ion-exchange was regenerated. Details of experimental set up
are explained in chapter 3 (section 3.3.3). This configuration is advantageous for the
following reasons: (1) This allows a smaller MBR reactor volume (due to a low HRT), and
a correspondingly lower capital cost, and a lower oxygen demand, and (2) also allows a
maximum recovery of nutrients in the sludge, and a greater reuse potential of carbon.
However the operation of MBR at short HRT will result in higher membrane fouling which
was controlled by air scouring and placing suspended particle in the reactor to create
surface scouring. This was discussed in chapter 5.
X-ray diffraction analysis showed that Zr hydroxide was a poorly crystalline material
(amorphous) having a broad diffraction peak at 2 values of 29-30° (Figure 6.1). This
pattern matches with that obtained for an amorphous Zr hydroxide by Chitrakar et al.
(2006).
Figure 6.1. XRD pattern of Zr hydroxide
The FTIR pattern for Zr hydroxide before and after the adsorption of phosphate is
presented in Figure 6.2. The presence of a intense and wide band in the 3324 cm−1 region
and a peak at 2920 cm−1 for Zr hydroxide before adsorption may be due to the stretching
vibration mode of lattice water and hydroxyl groups (Liu et al., 2008; Nur et al., 2014a). A
peak at 1547 cm−1 (O-H bending vibration) is due to coordinated water molecules, and that
at 1394 cm−1 (O-H bending vibration) indicated surface hydroxyl group on the metal oxide
surface (Liu et al., 2008).
Figure 6.2. FTIR spectrum of Zr hydroxide
The percentage phosphate removed increased when contact time and adsorbent dose were
increased (Figure 6.3). The adsorption capacity increased up to 60 min and remained
steady thereafter. This revealed that the rate of adsorption was fast and it required
relatively low contact time (around 60 min) to produce maximum adsorption of phosphate.
The increase in adsorption with adsorbent dose is due to more adsorption sites being
available for the adsorption. Phosphate removed from the 10 mg P/L phosphate solution
was > 90% after 1 h for the adsorbent doses of 1 - 5 g/L whereas for a dose of 0.5 g/L the
removal was <70% up to 5 h. These results imply that adsorbent doses of 1 - 3 g/L are
effective in removing most of the phosphate (more than 90%) from wastewater containing
10 mg P/L.
900 1400 1900 2400 2900 3400Wave number (cm-1)
After adsorptionBefore adsorption
Figure 6.3. Effect of contact time and adsorbent dose (doses are shown as legends within
the figure) on the removal of phosphate by Zr hydroxide (initial phosphate concentration
10 mg-P/L)
Figure 6.4 show that the adsorption of phosphate remained nearly constant between pH 3
and 4 and then decreased when pH rose from 4 to 11, reaching almost 50% removal at pHs
10 and 11 compared to pHs 3 and 4. Previous studies on synthetic waters also reported
decrease in adsorption capacity of phosphate with rise in solution pH using Zr hydroxide or
Zr oxide adsorbents (Chitrakar et al., 2006; Liu et al., 2008).
The dependence of Zr hydroxide of phosphate adsorption on pH can be related to both the
amphoteric properties of the Zr hydroxide surface and the polyprotonic nature of
phosphate. The oxygen atoms and hydroxyl groups are present on the Zr hydroxide, thus
mainly the hydroxyl groups are responsible for the chemistry and the reactivity of such
0
20
40
60
80
100
0 50 100 150 200 250 300
Rem
oval
eff
icie
ncy
(%)
Time (min)
0.5 g/ L1.0 g/L3.0 g/L5.0 g/ L
metal hydroxide surfaces (Liu et al., 2008). At high pH the reduction in adsorption is likely
attributed to the adsorbent surface carrying more negative charges as evidenced by the
increase in negative zeta potential of Zr hydroxide with increased pH (Figure 6.5). This is
expected to reject the negatively charged phosphate ions. This resulted in the lower
adsorption of phosphate at higher pH values (Long et al., 2011). On the other hand, at low
pH, especially when the pH is below the zero point of charge (ZPC, the pH at which the
net surface charge is zero) the surface hydroxyl is protonated and becomes positively
charged (Liu et al., 2008). Thus, at low pH the adsorbent surface is positively charged and
electrostatic attraction occurs with negatively charged phosphate anions, which led to more
phosphate being removed at low pH (Chitrakar et al., 2006). Figure 6.5 shows that the ZPC
of Zr hydroxide was at pH 4.2 and therefore Zr hydroxide had predominantly positive
charges below this pH. The ZPC of 4.2 obtained in this study is consistent with the value
of 4.9 reported for a mesoporous ZrO2 by Liu et al. (2008).
Another reason for the reduction in phosphate adsorption with rise in pH is that the
negative charge on the phosphate species in solution increases as the pH is increased
(Loganathan et al., 2014). As the pH rose, the phosphate species progressively changes
from H2PO4- to HPO4
2- to PO43- in accordance with the first, second, and third ionization
constants (Delaney et al., 2011). This will increase the repulsive force between the
phosphate ions and the negatively charged Zr hydroxide surface since the pH is increased
and this leads in turn to decreased adsorption.
Despite the repulsive forces between the negatively charged phosphate anions and the
negatively charged Zr hydroxide, significant amounts of phosphate were adsorbed even at
high pH levels (Figure 6.4). This is because the phosphate anions were also adsorbed by
specific adsorption forming inner-sphere complexes through the ligand exchange
mechanism which does not involve coulombic forces (Loganathan et al., 2014). This
mechanism in the adsorption process is confirmed by the zeta potential data which showed
that adding phosphate shifted the ZPC to a lower pH (4.2 to <3.0) (Figure 6.5). This
phenomenon has also been reported for other specifically adsorbing anions like fluoride
(Nur et al., 2014a). FTIR pattern of Zr hydroxide after phosphate adsorption supports the
ligand exchange mechanism of phosphate adsorption. The broad peak at 1042 cm-1 for Zr
hydroxide after phosphate adsorption was more intense than that before adsorption (Figure
6.2), suggesting that the surface –OH and H2O groups were replaced by phosphate during
adsorption. This was also confirmed by Su et al. (2013) for phosphate adsorption on Zr
oxide nanoparticles.
Figure 6.4. Effect of pH on phosphate adsorption by Zr hydroxide (Zr hydroxide dose 0.1
g/L)
0
5
10
15
20
25
30
35
40
45
3 4 5 6 7 8 9 10 11
Ads
orbe
d am
ount
(mg-
P/g)
Initial pH
10 mg-P/L5 mg-P/L
Figure 6.5. Equilibrium phosphate adsorption isotherms as influenced by (a) temperature,
(b) pH, and (c) co-existing anions and Langmuir adsorption model fitting.
The data for the equilibrium adsorption of phosphate on Zr hydroxide at different pH,
temperature and in the presence of co-existing anions fitted well to the Langmuir
adsorption model (R2 = 0.85-98, Figure 6.6, Table 6.1). The Langmuir adsorption maxima
at 22 ± 2 oC and pH 7.1 was 21.10 mg-P/g which is approximately the same as the value of
29.71 mg/g reported by Liu et al. (2008) for a mesoporous ZrO2 at pH 6.7 – 6.9. The
results showed that Zr hydroxide can potentially work very well as a phosphate adsorbent
for removing phosphate from wastewater effluents.
The effect of temperature on phosphate adsorption showed that the Langmuir adsorption
maxima (mg P/g) at pH 7.1 increased from 21.10 mg/g at 20 ± 2 oC to 40.60 mg/g at 40 ± 3
-50
-40
-30
-20
-10
0
10
20
2 3 4 5 6 7 8 9 10 11
Zeta
pot
entia
l (m
V)
Initial pH
DI water5 mg-P/L 10 mg-P/L
oC. Further, an increase in temperature to 60 ± 3oC led to an additional increase of
adsorption to 61.50 mg/g (Table 6.1, Figure 6.6 a). The adsorption process is therefore an
endothermic reaction. The increased adsorption at higher temperature could be due to a
change in pore size or and/or activation of the adsorbent surface (Yan et al., 2010).
Consistent with the results shown in Figure 6.4 for initial P concentration of 10 mg/L and
Zr hydroxide dose of 0.1 g/L, the Langmuir adsorption maxima was highest (32.90 mg-
P/g) at the lowest pH of 4 and then decreased to 21.10 mg-P/g at pH 7.1 and 21.80 mg-P/g
at pH 10.0 (Table 6.1, Figure 6.6 b). The reasons for this decrease in adsorption capacity
when pH increases have been discussed in section 6.1.3.
A study was also carried out to determine the effect of coexisting anions, nitrate (10 mg
N/L) and sulphate (10 mg S/L) on the adsorption of phosphate from a solution containing
10 mg P/L by Zr hydroxide. The results showed that the phosphate adsorption capacity
decreased in the presence of sulphate ion but not in the presence of nitrate anions (Figure
6.6 c, Table 6.1). Nitrate is non-specifically adsorbed (outer-sphere complexation) to
metal oxides and therefore it was unable to compete well with phosphate which is
specifically adsorbed (inner-sphere complexation) (Loganathan et al., 2014). Sulphate, on
the other hand, can be adsorbed specifically and non-specifically and therefore it competed
weakly with phosphate. These results showed that the presence of sulphate ions had a low
inhibitory effect on the adsorption of phosphate by Zr hydroxide. A batch equilibrium
adsorption experiment showed that sulphate had a Langmuir adsorption capacity of 3.28
mg-S/g. Chitrakar et al. (2006) also noted that amorphous Zr hydroxide preferred
phosphate ions over sulphate ions.
Figure 6.6. Equilibrium phosphate adsorption isotherms as influenced by (a) temperature,
(b) pH, and (c) co-existing anions and Langmuir adsorption model fitting.
0
5
10
15
20
25
30
35
40
45
0 1 2 3 4 5 6 7 8
Qe
(mg-
P/g)
Ce (mg/L)
(a)
Langmuir20±2 °C40±3 °C 60±3 °C
0
5
10
15
20
25
30
0 1 2 3 4 5 6 7 8 9
Qe
(mg-
P/g)
Ce (mg/L)
(b)
LangmuirpH 4.0±0.1pH 7.1±0.3pH 10.0±0.3
0
5
10
15
20
25
0 1 2 3 4 5 6 7 8 9 10
Qe
(mg-
P/g)
Ce (mg/L)
(c)
LangmuirPP+NP+S
Table 6.1. Langmuir model parameters for phosphate adsorption at different temperatures,
pHs, and in the presence of nitrate and sulphate
Langmuir
parameters
pH 7.1, P only 22oC, P only 22oC, pH 7.1
22oC 40oC 60oC pH 4.0 pH 7.1 pH 10 P only P+N P+S
Qm 21.10 41.60 61.50 32.90 21.10 21.80 21.10 23.50 15.30
KL 2.69 0.36 0.32 0.96 2.69 1.53 2.69 1.07 3.77
R2 0.93 0.95 0.98 0.85 0.93 0.94 0.93 0.98 0.94
The wastewater used for all the experiments had a dissolved organic carbon (DOC)
concentration of approximately 8.5 mg/L. In each experiment, it was found that Zr
hydroxide had no affinity towards DOC as there was no significant removal of DOC.
Therefore, DOC would not have competed with phosphate adsorption on Zr hydroxide.
This is similar to the observation by Chen et al. (2002) who reported that DOC yielded an
insignificant effect on the adsorption of phosphate by: Amberlite IRA910 Cl and secondly,
Amberjet 1200Na .
Results for the adsorption kinetics experiments showed that both the pseudo-first order and
pseudo-second order models fitted very well to the experimental data for all the adsorbent
doses with the coefficients of determination (R2) greater than or equal to 0.95 (Table 6.2).
The calculated values of qe from both kinetic models were also approximately matched
with the experimental values of qe. It can therefore be concluded that both these models do
satisfactorily describe the kinetics of phosphate adsorption on Zr hydroxide.
Table 6.2. Pseudo first- and second-order adsorption rate constants and calculated and
experimental qe values for different Zr hydroxide doses (initial phosphate concentration 10
mg P/ L)
Pseudo-first-order model Pseudo-second-order model
Adsorbent dose (g/L) qe,exp
(mg/g)
k1
(1/h)
qe,cal
(mg/g) R2
k2
(g/mg. h)
qe,cal
(mg/g) R2
0.5 16.62 1.39 16.8 0.95 0.07 20.01 0.95
1 13.85 1.97 13.79 0.97 0.17 15.35 0.97
3 2.70 3.17 2.64 0.97 1.64 2.85 0.98
5 1.97 5.92 1.97 0.99 7.06 2.03 0.99
An adsorption thermodynamics study was conducted to determine the different
thermodynamic parameters involved in the adsorption process. From the batch adsorption
isotherms at pH 7.1 and different temperatures (22, 40, and 60 oC), the thermodynamic
parameters, equilibrium constant (K0) (phosphate distribution between solid and liquid
phases), change in free energy ( G°), the change in enthalpy ( H°), and change in entropy
( S°) were calculated based on the procedure described by Deliyanni et al. (2007) and Yan
et al. (2010). In this procedure, K0 was determined by plotting ln (qe/Ce) vs qe and
extraplotting qe to zero (qe and Ce were defined in equation 3.3). G° was calculated from
equation 6.1 and H° and S° were obtained from the slope and intercept of the plot of ln
K0 vs 1/T, respectively. These involved using equation 6.2, where R is the gas constant and
T is the absolute temperature.
G° = -RT ln K0 6.1
ln K0 = S°/R - H°/RT 6.2
The thermodynamic parameters for the adsorption of phosphate on Zr hydroxide are
presented in Table 6.3. It was found that G° declined from -48.31 to -118.98 KJ mol-1
when the temperature increased from 293 to 333 K. The negative values specify that the
adsorption process is spontaneous. The positive value for H° indicates that the process is
endothermic. Since the value of 466.38 kJ mol−1 obtained for H° is much higher than the
range of H° values of 8.4–41.8 kJ/mol reported for physical adsorption (Faust and Aly
2009), the adsorption process is considered to be chemical in nature. The negative value of
S° implies greater order (less randomness) of reaction during the adsorption (Raji and
Anirudhan 1998).
Table 6.3. The thermodynamic parameters for the adsorption of phosphate on Zr hydroxide
T
(K) ln ko
R
(J mol-1 K-1)
G°
(kJ mol−1)
S°
(kJ mol−1 K−1)
H°
(kJ mol−1)
293 19.83
8.314
-48.31
-1.76 466.38 313 35.59 -92.62
333 42.98 -118.98
Figure 6.7 shows the concentration of phosphate in the submerged MFAH system effluent
with time for different concentrations of inlet P concentrations. It emerged that at the
beginning (within 30 min) the amount of phosphate removed was small but then increased
with time up to 2 h of operation. The low level of phosphate being removed at the
beginning was due to the lagging in contact time for adsorption which was also evident in
the batch kinetic adsorption results presented in Figure 6.3. It was also found that as time
passed the removal of phosphate decreased. This is because in the continuous membrane
reactor process the new feed is continuously added but the amount of adsorbent used was
fixed (and remaining in the solution from the beginning).
The operation time for efficient removal of phosphate can be improved by increasing the
concentration of the adsorbent. Thus, the microfiltration/adsorption experiment was
repeated with different concentrations of the adsorbent (0-5 g/L of the volume of the
reactor). The results showed that when the dose of Zr hydroxide was increased, more
phosphate was removed (80%) and this was achieved using 5 h of operation (Figure 6.8).
Nonetheless the removal efficiency continued to decrease after a certain time for each
adsorbent dose.
Figure 6.7. Effect of time on P concentration in the effluent in MFAH system with
addition of Zr hydroxide (1 g/L) for different initial P concentrations
0
5
10
15
20
25
0 30 60 90 120 150 180 210 240 270 300 330 360
Eff
luen
t con
cent
ratio
n (m
g P/
L)
Time (h)
0 mg P/L10 mg P/L20 mg P/L
Figure 6.8. Effect of time on P removal efficiency in MFAH system with with addition of
different doses of Zr hydroxide (Inlet concentration 10 mg-P/L).
A set of long-term microfiltration adsorption experiments was then carried out replacing
10% of Zr hydroxide (1.5 g) every 24 h. The results showed (Figures 6.9 a, b) that after the
first 24 h of operation the phosphate removal efficiency decreased significantly to 6 and
20% for the feed concentrations of 10 and 5 mg/L, respectively (Figure 6.9 a). At this stage
when 10% of Zr hydroxide was added the phosphate removal efficiency increased to 70%
which was then reduced to 3-6% after the second 24 h of operation. Similarly, with another
replacement of 10% of Zr hydroxide after 48 h total time of operation the removal
efficiency increased to 60%. Thus, the replacement of the used adsorbent with fresh
additions helped to continuously maintain the high P removal efficiency. Similarly, a
decrease in filtration flux from 10 to 5 L/m2.h led to more phosphate being removed
(Figure 6.9 b). Furthermore, based on the mass balance calculation it was found that the
adsorption of phosphate was approximately 16.67 mg/g of Zr hydroxide which is around
0102030405060708090
100
0 30 60 90 120 150 180 210 240 270 300 330 360
Rem
oval
eff
icie
ncy
(%)
Time (min)
0 g/L 1 g/L 3 g/L 5 g/L
80% of Langmuir maximum adsorption capacity. The smaller percentage of phosphate
removed at higher inlet phosphate concentration and at higher flux is due to the higher
loading of phosphate in the reactor. Based on the two filtration fluxes (5 and 10 L/m2.h)
and 10% daily replacement of Zr hydroxide, the amount of Zr hydroxide required was
0.0625-0.125 g to treat 1 L of water, for relatively good removal of phosphate when the
phosphate concentration was around 10 mg/L. The optimum dose of replacing the
adsorbent will vary based on operational conditions as well as the inlet concentration of
phosphate.
Figure 6.9. Effect of repeated additions of Zr hydroxide (5 g/L) to MFAH system on the
phosphate removal at (a) 5 L/m2.h filtration flux and two inlet P concentrations and (b) 10
mg/L inlet P concentration and two filtration fluxes
0102030405060708090
0 10 20 30 40 50 60
Rem
oval
eff
icie
ncy
(%)
Time (h)
(a)5 mg P/L10 mg P/L
0102030405060708090
0 10 20 30 40 50 60
Rem
oval
eff
icie
ncy
(%)
Time (h)
(b)5 L/m².h10 L/m².h
The removal of PO43- by SMBR alone was not sufficient (Table 5.6). Thus, it is vital to
remove PO43- using an additional process. In this study, an anthracite filter mixed with
HFO was used to remove phosphate from SMBR effluent (chapter 5, section 5.3.2).
Adsorption column experiments (packed with anthracite and HFO) were conducted at a
low filtration velocity of 2.5 m/h. The filtration column was packed with 36 g of anthracite
(particle size of 0.6 – 1.18 mm) as an inert material with varying percentage of HFO by
mass (0%, 1%, 5%, and 10% of anthracite) in-filled in the anthracite medium to investigate
the effect of different amounts of HFO on the removal of phosphate.
The adsorption capacity of PO43- by HFO (doses of 0.1 - 7.0 g/L) was first evaluated using
a batch equilibrium study. From the batch equilibrium study, it was found that the removal
of PO43- increased from 11.3% (with 0.1 g/L of HFO) to 90.4% with 7.0 g/L of HFO.
Furthermore, the equilibrium (isotherm) data was successfully fitted with Langmuir
isotherm model (figure is not shown). From the Langmuir isotherm model, the maximum
adsorption capacity of PO43- was found to be 41.9 mg-PO4
3-/g-HFO. This result is in
agreement with the findings of previous study (Gupta et al., 2012). They reported the
maximum adsorption capacity of P by HFO of 14 mg-P/g (or 42.9 mg-PO43-/g). Gupta et
al. (2012) also have investigated a comparative removal of P by purolite ion exchange
resin and HFO. A 50% higher adsorption capacity of P by HFO (14 mg-P/g) than purolite
ion exchange resin (7 mg-P/g) was reported.
From the literature, it is also found that oxides of polyvalent metals such as Fe3+, Ti4+ and
Zr4+ exhibit very favourable ligand sorption properties for phosphate through the formation
of inner sphere complexes (Blaney et al., 2007). Moreover, a new phosphate-selective
sorbent, referred to as hybrid anion exchanger (HAIX) has been successfully used to
remove phosphorus. HAIX is essentially a polymeric anion exchanger within which
hydrated ferric oxide (HFO) nanoparticles have been dispersed irreversibly (Blaney et al.,
2007).
The results of the HFO/anthracite column experiment are presented in Figure 6.10. The
performance of the filtration column was analyzed in terms of phosphate removal. From
Figure 6.10, it is clear that the anthracite filter medium itself was not effective in removing
of PO43-. The incorporation of HFO helped to remove PO4
3-. The removal of PO43-
increased with high percentages of HFO. With 1% of HFO, the filtration column was
saturated (Ct/C0 = 1) within 125 bed volumes, whereas it took place after 350 bed volumes
with HFO of 10% (here, Ct is the effluent PO43- concentration with time (t) and C0 is
influent PO43-concentration; The No. of bed volume is defined as flow rate x time/volume
of filter medium). From the Figure 6.10 it is also found that at the beginning (within 10
min) of the operation the Ct/C0 values with 0, 1, 5 and 10% of HFO were 0.95, 0.70, 0.25
and 0.16, respectively. The lower values of Ct/C0 and longer bed volume filtered with 10%
of HFO could be due to the greater availability of adsorbing sites of HFO for the
adsorption PO43-. Thus HFO can be used as an absorbent/ion exchanger with other inert
materials such as anthracite or sand to remove PO43- from the effluent of a high rate
membrane bioreactor. The breakthrough adsorption capacity for different % HFO mixed
with anthracite in the column was calculated. The adsorption capacity of HFO was found
36-39 mg-PO43-/g-HFO which is slightly lower than that obtained from the batch
equilibrium study (41.9 mg-PO43-/g-HFO). This could be due to the change in operational
condition. In equilibrium study it is assumed that all the HFO particles dispersed properly
in water will have 100% availability of adsorption sites. On the other hand in the column
some HFO particles may interact with anthracite particles as well as among HFO particles
themselves resulting in minor losses of adsorption sites.
HFO is not suitable for long-term use in fixed-bed columns due to its lack of mechanical
strength although HFO has higher adsorptive capacity for phosphate (Blaney et al., 2007).
Thus the application of HFO on the P removal may not be cost effective as the recovery of
HFO material is difficult. Thus, purolite ion exchange resins were used as post-treatment to
remove and recover nutrients although it has lower ion exchange/ adsorption capacity than
HFO.
Figure 6.10. PO43- removal by HFO from SMBR effluent (0, 1, 5 and 10% of HFO by
mass with anthracite coal as inert material was used as filter medium; influent PO43-
oncentration to the post treatment HFO adsorption column was 2.2 mg/L)
0
0.2
0.4
0.6
0.8
1
0 100 200 300 400
Ct/C
0
No. of bed volume filtered
0 % HFO 1 % HFO
5 % HFO 10 % HFO
The removal efficiency of nutrients (PO4-P and NO3-N) from the MBR effluent with the
two selected purolite resins (purolite A520E and A500P) is presented in Figure 6.11 (a, b).
The purolite resins were chosen to remove and the recover nutrients from the MBR
effluent. The removal efficiency of phosphorous was observed to be higher for purolite
A500P, at 50-90%, compared with purolite A520E, at 10-30%, (Figure 6.11 a). Initially
the removal efficiency of PO4-P was low. This effect could be minimised by using higher
bed depth of ion exchange resin or lower filtration velocity. On the other hand, purolite
A520E showed a higher removal efficiency for NO3-N (almost 94%) than purolite A500P
(50-90%) during the operating period. As shown in Figure 6.11 b, the number of bed
volumes filtered was 1950. Here the number of bed volume is the volume filtered divided
by the purolite volume (No. of filter bed volume = flow rate x time/volume of ion-
exchange medium). The Purolite A-520E is a macroporous strong base anion resin which
is specially designed for the removal of nitrates from water for potable purposes. It showed
higher removal efficiency of nitrogen compared to phosphorous. On the other hand, A500P
is designed for use as an organic scavenger, e.g. for the removal of tannins, fulvic and
humic acids, from domestic sewage effluent. This resin also showed high removal
efficiency of phosphate. The concentration of nitrite (NO2-N) in the effluent was very low
at 0.01 - 0.02 mg/L for both these resins. The concentration of ammonium (NH4-N) after
purolite A520E and A500P was between 0.6 - 2.2 and 0.7 - 2.5 mg/L respectively.
Figure 6.11. Comparison between purolite A520E and A500P (bed height = 6 cm; velocity
= 2.5 m/h, the concentration of PO4-P and NO3-N of the MBR effluent was 3.1 and 11
mg/L respectively)
a) Removal efficiency of PO4-P with purolite A500P & A520E
No. of filter bed0 40 80 120 160 200 2401000 1500 2000
Rem
oval
eff
icie
ncy
(%)
0102030405060708090
100
purolite A500Ppurolite A520E
b) Removal efficiency of NO3-N with purolite A500P & A520E
No. of filter bed0 40 80 120 160 200 240 1000 1500 2000
Rem
oval
eff
icie
ncy
(%)
0102030405060708090
100
purolite A500Ppurolite A520E
From the single ion exchange column experiments, it was found that purolite A520E
enhances the removal of NO3-N over a longer period whereas, A500P provides superior
PO4-P removal. Based on these findings, an experiment was carried out with purolite
A520E and A500P ion-exchange columns in series. The first column was filled with
purolite A520E and the second with purolite A500P in order to improve the removal
efficiency of NO3-N and PO4-P. The result of this experiment is presented in Figure 6.12.
These results indicated that the two (purolite A520E and A500P) columns in series helped
to improve the removal efficiency of PO4-P and NO3-N and could be run for longer period.
Figure 6.12. Effect on nutrient removal of two types of purolite ion-exchange resin
columns in series (velocity = 2.5 h, the concentration of PO4-P and NO3-N of MBR
effluent was 4.18 and 9 mg/L respectively)
No. of filter bed0 40 80 120 160 200 240 1000 1500 2000
Rem
oval
eff
icie
ncy
(%)
0102030405060708090
100
PO4-PNO3-N
The recovery of phosphorus and nitrate helps both in keeping good quality of receving
water (with less or no eutrofication) and also extracting valuable phosphorus. The amount
of nutrients recovered from the MBR effluent by the ion-exchange column was calculated
using the following equation:
Recovery of nutrients (mg) = ( )
tCQ CC iini
iΔ
+− −
=
= 21
01
Where,
Q = flow rate (L/h)
C0 = influent concentration (mg/L)
Ci = effluent concentration after the ith time step of ion-exchange column operation (mg/L)
t = time difference (h)
The amount of nitrogen and phosphorous retained in the purolite ion-exchange media is
presented in Table 6.4. Although the amount of nitrate and phosphate were low in the
MBR effluent, an amount of 20 kg and 82 kg of PO4- and NO3
- respectively could be
recovered on a daily basis in a 10000 m3/d plant (calculated based on the results obtained
in this study).
6.3
Table 6. 4. Estimation of retention of nutrients in the purolite ion-exchange column
Time
(h)
Cumulative filtered
volume
(m3)
Cumulative recovery
of PO4-P
(in mg)
(initial concentration
= 4.18 mg/L)
Cumulative recovery of
NO3-N
(in mg)
(initial concentration =
11 mg/L)
0.5 0.000393 1.0 3.6
1 0.000785 1.9 7.5
2 0.001571 3.9 14.1
3 0.002356 6.5 21.4
4 0.003142 8.6 27.3
5 0.003927 10.9 34.6
6 0.004712 12.9 42.9
22 0.017279 32.3 141.7
46 0.036128 71.2 195.1
Ion-exchange processes using selective ion-exchange materials (such as purolite) are ideal
for reducing ammonia and phosphate to near-zero levels provided that the ion-exchange
resin is ammonia and/or phosphate selective, cost effective and amenable to efficient
regeneration and reuse. In this study, the regeneration of purolite was conducted using
NaCl and Na2SO4 solutions of 1, 2, 3 and 5% (w/v). The regeneration was conducted by
backwashing the resin with a velocity of 5 m/h. Over 95 - 97% phosphate recovery can be
obtained with 20 bed volumes of 1% NaCl or with 4 bed volumes of 2 - 3% NaCl.
Similarly 95 - 98% of nitrate recovery was obtained with only 20 bed volumes of 1% NaCl
or with 6 bed volumes of 2 - 3%. However, Na2SO4 with similar concentration can recover
only 40 - 50% of adsorbed NO3- and PO4
3-. Use of NaCl for regeneration is not practical
due to the disposal of saline NaCl solution into the receiving water body. However, this
solution can be mixed with (or diluted) with sewage effluent and discharged into receiving
water body. KCl also had a similar regeneration capacity and the KCl solution can be used
as soil fertilizer. Besides NaCl and Na2SO4, other commercially available mild alkaline
solutions such as Ca(OH)2 may be attempted as sodium salts at high concentrations may be
not suitable as fertilizer in agriculture. On the other hand the regeneration of HFO and
zirconium (IV) hydroxide was not as effective as pourolite A520E and A500P although
their capacity in phosphorus adsorption was high.
The removal of phosphate in the normal biological treatment processes such as bioreactor
does not exceed 30% (Deliyanni et al., 2007). Although the removal of phosphate by
biological process can be improved by using polyphosphate accumulative organisms, this
treatment process is very complex and slow, requiring large amount of infrastructural
investment. Further it is not efficient in treating wastewater with high phosphate
concentration (Long et al., 2011). On the other hand, removal of phosphate by chemical
precipitation is expensive and it is not easy to recover phosphorus from chemical sludge
(Long et al., 2011). In the adsorption process, the used adsorbent loaded with phosphate
can be profitably used in agriculture (Zhang et al., 2009).
The cost of Zr hydroxide is $ 150-200/kg (analytical grade). Whereas the price of purolite
and HFO is around $200 and $1400/kg ((the price quoted is based on small quantity.
However, in large quantity, the price will be very much less). The material cost for
removing 1 g of P using Zr hydroxide and purolite ranges from $ 8 – 12 and for HFO is
around $30. From literature, the unit cost (only material) for the removing of 1 g of P using
Titanium mesostructure, Zirconium mesostructure, Layered double hydroxides, Alumina,
and Ion exchange resins ranges from $33-1150 (Choi et al. 2011). This shows the ion
exchange resins used in this study was comparatively cheaper than other ion exchange
resins.
From the experimental investigation following conclusion can be made:
Amorphous Zr hydroxide used in this study proved to be effective in removing phosphate
in terms of speed and capacity of adsorption. The maximum adsorption of phosphate by Zr
hydroxide was 21.1 mg P/g at 22 oC and pH of 7.1. The Langmuir adsorption maximum
increased with temperature rising up to 60 oC but decreased when pH increased from 4 to
10. The FTIR pattern and zeta potential data before and after adsorption of phosphate
showed that phosphate was specifically adsorbed on Zr hydroxide through inner-sphere
complexation. Consistent with this adsorption mechanism, the non-specifically adsorbing
nitrate anion had no effect on the adsorption of phosphate. Sulphate, which can adsorb
specifically on metal hydroxides, reduced the adsorption of phosphate on Zr hydroxide.
The thermodynamic parameters, G° and H° were negative and positive, respectively,
showing that the adsorption process was spontaneous and endothermic. S° was negative
and this implies that there was a greater order of reaction (less randomness) during the
adsorption process.
A submerged membrane filtration Zr hydroxide adsorption hybrid reactor system where Zr
hydroxide was added only once at the start of the experiment showed that the removal of
phosphate declined after 5 h of operation. However when Zr hydroxide was repeatedly
added once every 24 h satisfactory removed of phosphate at a constant level. Long-term
effectiveness of the reactor in removing phosphate depends on the dose and frequency of
Zr hydroxide being added as well as the filtration flux.
Further, a simple treatment of 10 % HFO + 90 % anthracite mixed column showed a high
removal (90 %) of PO43- from the SMBR effluent. It could be operated for a long time
before the HFO became exhausted as the number of filter volume is more than 300 for
10 % of HFO.
In addition, purolite ion exchange resins removed almost 90% of phosphate and nitrate.
The removal efficiency of phosphate was observed to be higher (90%) for purolite A500P
as compared to purolite A520E (30%). However, purolite A520E showed a higher removal
efficiency of NO3-N (almost 94%) than purolite A500P (50-90%).
In conclusion, over 95 - 98% phosphate and nitrate recovery was obtained during the
regeneration of purolite columns with 1% NaCl of 20 bed volumes. However, the
regeneration of HFO and Zr hydroxide using NaCl and NaOH was not effective. Hence,
the use of high rate MBR with an ion-exchange system (which is easy to regenerate such as
purolite) as post treatment can be an efficient wastewater treatment with a useful nutrients
recovery and zero nutrient discharge.
The aim of this study was to operate the MBRs at high rate to remove organic carbon and
then remove and recover nitrogen and phosphate using ion exchange resins as post
treatment. The effect of organic loading rates, imposed fluxes and salt concentrations were
investigated. It was found that the removal efficiency of DOC and ammonium nitrogen
was more than 90% and 80% respectively when the MBRs were operated at lower organic
loading rates and lower flux (less than 10 L/m2.h). On the other hand the removal of DOC
and ammonium nitrogen decreased significantly with the increase of salt concentration
from 5 to 35 g-NaCl/L. In addition, operating MBRs at high flux (lower HRTs) led high
membrane fouling. To overcome this problem, the MBR was operated with the
incorporation of GAC in suspension. The incorporation of GAC in MBR helped to reduced
membrane fouling by more than 50%. This also helped to remove an addition amount of
DOC. The removal and recovery of nutrients from MBR effluent was studied with
different types of ion exchange resins such as purolite, hydrated ferric oxide and zirconium
hydroxide. The ion exchange processes effectively remove more that 90% of nitrogen and
phosphorus. The specific experimental findings at different operating conditions of MBR
are described below.
The removal efficiency of organic in terms of DOC decreased with the increase in organic
loading rate (OLR), imposed flux and salt concentration. The removal efficiency of DOC
was 93 - 98% with low OLRs of 0.5 - 1 kgCOD/m3.d. The removal of DOC reduced to 45 -
60% when the OLR was increased to 2.75 - 3.0 kgCOD/m3.d. The removal of NH4-N was
high for low OLRs of 0.5 - 1.0 kgCOD/m3.d of 83 - 88%. The removal of phosphorus with
different OLRs was ranging from 30-58%. The removal efficiency of organic and nutrients
was high (more than 90%) at lower flux of 5 L/m2.h. Further, the removal efficiency of
organic reduced to 58 – 66% when the operating flux was increased to 40 L/m2 h. The
removal of ammonia and phosphate at different imposed flux of 20 - 4 0 L/m2.h were 30 –
50 % and 10 – 30 % respectively.
The increase of salt concentration had negative impact on the removal of organics and
nutrients. The uptake rate of dissolved organic carbon and ammonia decreased from
around 17.0 mg-DOC/g-MLVSS.d to 1.8 mg-DOC/g-MLVSS.d and from 8.2 mg-NH4-
N/g-MLVSS.d to 0 mg-NH4-N/g-MLVSS.d respectively when salt concentration reached
to 35 g-NaCl/L. The lower removal of organics (DOC) and ammonia with high salt
concentration could be due to the adverse effect of salt on microbial activity. The amount
of biopolymers increased to 95% (from 0.07 to 1.0 mg/L in MBR effluent and 0.4 to 15.5
mg/L in mixed liquor) when salt concentration was increased from 0 to 35 g-NaCl/L. The
concentration of dissolved organic nitrogen (DON) in bio-polymer also increased from
0.05 to 3.31 mg/L when salt concentration was increased to 35 g-NaCl/L. The removal of
PO4-P at different salt concentration was relatively low (less than 30%). The study on the
gradual increase of salt concentration was tested to model the start-up of a system where
salt was gradually leached into the system to acclimatise the biomass. For example, in
coastal areas, salt water can gradually enter into MBR plant. Furthermore, many industrial
sectors such as food industries, petroleum and leather industries produce wastewater
containing high salinity and high organic matter. Thus, this study was made to understand
the effect of continue increase of salt concentration in membrane bioreactor process.
The increase of OLR resulted in higher membrane fouling. The TMP development was low
at 0.001-0.005 kPa/h with organic loading rates of 0.5 - 2.0 kgCOD/m3.d as compared to
higher OLRs of 2.75 - 3.0 kgCOD/m3.d. TMP development was closer to 0.1 - 0.24 kPa/h
at higher OLRs. Similar to OLR, imposed flux had a strong effect on membrane fouling.
Lower flux of 20 L/m2.h produced almost 75 – 90 times more water than higher flux of 40
L/m2h with an aeration rate of 0.3- 1.2 m3/m2.membrane area.h. The development of TMP
increased from 2.0 mbar (at 0 g-NaCl/L) to 10.5 mbar when salt concentration was
increased to 35 g-NaCl/L. The characterization of foulant showed that bio-polymer was the
major foulant followed by humic substances, building blocks, lower molecular weight
neutrals and acid along with the deposition of bio floc on the membrane surface.
At higher aeration rate, the effect of flux on membrane resistance was negligible but when
the aeration rate was reduced (from 1.5 to 1 m3/m2membrane area.h), a sudden rise of TMP
was observed. This could be due the accumulation of sludge on to membrane surface.
Thus, the use of suspended media in the membrane reactor can reduce this accumulation of
the sludge by mechanical scouring. The use of granular activated carbon (GAC, particle
size of 150 - 1200 μm) in suspension (0.5 - 2 g/L of volume of reactor) prevented the
sudden rise of TMP (i.e. reduced membrane resistance). It also helped to reduce the sludge
volume index (SVI) and TMP development by 30 - 40% and 58% respectively. This is due
to the combined effect of adsorption of organic matter by GAC and extra mechanical scour
on the membrane surface created by GAC used as suspended medium.
The amount of suspended medium also had a major effect on fouling reduction. In this
study, a suspended media concentration of 2 g/L of the volume of the reactor showed better
reduction of membrane fouling. It is also important to apply sufficient aeration to keep the
medium in suspension in the reactor. Further, GAC of particle size of 300 - 600 μm
showed 25 % higher reduction of membrane fouling than GAC of particle sizes of 150 -
300 and 600 - 1200 μm. Thus, a suitable amount and size of suspended medium need to be
used depending on the flux and aeration (or air scour) rate used. It was found that the GAC
particle did not have any adverse effect on membrane surface as the clean water flux was
the same as that of a virgin membrane and as the filtered turbidity was reasonably low (less
than 0.2 NTU; even after long term use of the membrane).
High rate MBR experiments were carried out at a short hydraulic retention time (HRT) to
remove only organic carbon allowing the nitrogen and phosphorus remaining in the MBR
effluent for their removal and recovery by an ion-exchange process. The ion exchange
experiments were conducted in batch, column and membrane hybrid adsorption systems.
The ion exchange resins used were purolite (A520E and A500P), HFO and zirconium
hydroxide. It was found that the removal efficiency of phosphate was observed to be
higher (90%) for purolite A500P as compared to purolite A520E (30%). The removal
efficiency of phosphate by HFO and Zr hydroxide was high (more than 90%). Further,
purolite A520E showed a higher removal efficiency of NO3-N (almost 94%) than purolite
A500P.
To make the ion exchange system sustainable, the regeneration of purolite was conducted
using NaCl solutions of 1, 2, 3 and 5% (w/v). Over 95 – 97% phosphate recovery can be
obtained with only 20 bed volumes of 1% NaCl or with 4 bed volumes of 2 – 3% NaCl.
Similarly 95 – 98% of nitrate recovery was obtained with 20 bed volumes of 1% NaCl or
with 4 bed volumes of 2 – 3%. However, NaCl may not be suitable to be disposed. As such
alternative regenerants such as Na2SO4 and Ca(OH)2 were tried but their efficiency in
regeneration was less (40 - 50%). The regeneration of HFO and zirconium hydroxide was
neither effective nor practical. Further, HFO is not suitable for long-term use in fixed-bed
columns due to its lack of mechanical strength although HFO has higher adsorptive
capacity for phosphate.
A post-treatment of ion exchange demonstrates that high rate MBR with an ion-exchange
system is an efficient wastewater treatment with a useful nutrients recovery. This
configuration is advantageous for the following reasons: (1) This allows a smaller MBR
reactor volume, and a correspondingly smaller footprint, and a lower oxygen demand, and
(2) also allows a maximum recovery of nutrients in the sludge, and a greater reuse potential
of carbon from the grey water. However the operation of MBR at short HRT resulted in
higher membrane fouling which has to be controlled by air scouring or by introducing
medium in suspension in the reactor.
The following recommendations are made further study:
• Future research is needed to explore highly efficient, low cost suspended media for
use as suspended media in the reactor. This should not also affect membrane life.
• A long term comparative study with different types of membrane such as flat sheet,
hollow fibre with different density/packing density and configuration with the
addition of suspended media need to be conducted. This is to establish a sustainable
high rate MBR.
• Study on the influence of salinity with the addition of suspended activated carbon
needed to explore.
• Low cost adsorbents with high adsorption/ion exchange capacity are needed to be
developed for the removal of nutrients.
• The MBR-ion exchange hybrid system developed in this study for the removal and
recovery of nutrient needs to be tested in pilot plant scale as it has the potential for
the removal and recovery of phosphate fertilisers from waste water.
• Suitable, cost effective regeneration process for the regeneration of ion exchange
resins need to be explored.
Adham, S. and Gagliardo, P. (1998). Membrane Bioreactors for Water Repurification-
Phase I. Desalination Research and Development Program Report No. 34, Project
No. 1425-97-FC-81-30006J; Bureau of Reclamation: Denver, Colorado.
Ahmed, Z., Cho, J., Lim, B-R., Song, K-G., and Ahn, K-H. (2007). Effects of sludge
retention time on membrane fouling and microbial community structure in a
membrane bioreactor. Journal of Membrane Science, 287, 211–218.
Akram, A. and Stuckey, D. C. (2008). Flux and performance improvement in a submerged
anaerobic membrane bioreactor (SAMBR) using powdered activated carbon
(PAC). Process Biochemistry, 43(1), 93-102.
Al-Halbouni, D., Traber, J., Lyko, S., Wintgens, T., Melin, T., Tacke, D. and Hollender, J.
(2008). Correlation of EPS content in activated sludge at different sludge retention
times with membrane fouling phenomena. Water Research, 42(6), 1475-1488.
Álvarez-Hornos, F.J., Volckaert, D., Heynderickx, P.M., Van Langenhove H. (2011).
Performance of a composite membrane bioreactor for the removal of ethyl acetate
from waste air. Bioresource Technology, 102, 8893–8898
APHA, (1995). American Public Health Association, Standard methods for the
examination of water and wastewater, 21st centennial edition. APHA, Washington.
Artiga, P., García-Toriello, G., Méndez, R. and Garrido, J. M. (2008). Use of a hybrid
membrane bioreactor for the treatment of saline wastewater from a fish canning
factory. Desalination, 221(1), 518-525.
Artiga, P., García-Toriello, G., Méndez, R., & Garrido, J. M. (2008). Use of a hybrid
membrane bioreactor for the treatment of saline wastewater from a fish canning
factory. Desalination, 221(1), 518-525.
Aryal, R., Legegue, J., Vigneswaran, S., Knadasamy, J., Heran, M. and Grasmick, A.
(2009). Identification and characterisation of biofilm formed in membrane
bioreactor. Separation and Purification Technology, 67(1), 86-94.
Bae, B., Jung, Y., Han, W., and Shin, H. (2002). Improved brine recycling during nitrate
removal using ion exchange. Water Research, 36, 3330 - 3340.
Baral, B. (2003), Numerical study on the bubbly plume for water purification system,
Master Thesis. University of Tokyo, Japan.
Bassin, J.P., Kleerebezem, R., Muyzer, G., Rosado, A.S., van Loosdrecht, M.C. and
Dezotti, M. (2012). Effect of different salt adaptation strategies on the microbial
diversity, activity, and settling of nitrifying sludge in sequencing batch reactors.
Applied Microbiology and Biotechnology, 93(3), 1281-1294.
Beler-Baykal, B., Oldenburg, M. and Sekoulov, I. (1996). The use of ion exchange in
ammonia removal under constant and variable loads. Environmental Technology,
17 (7), 717-726.
Bhatnagar, A. and Sillanpää, M. (2011). A review of emerging adsorbents for nitrate
removal from water. Chemical Engineering Journal, 168(2), 493-504.
Birima, A. H., Megat Mohd Noor, M. J., Mohammed, T. A., Idris, A., Muyibi, S. A.,
Nagaoka, H., and Abdul Ghani, L. A. (2009). The effects of SRT, OLR and feed
temperature on the performance of membrane bioreactor treating high strength
municipal wastewater. Desalination and Water Treatment, 7(1-3), 275-284.
Blaney, L.M., Cinar, S., and Sengupta, A.K. (2007). Hybrid anion exchanger for trace
phosphate removal from water and wastewater. Water Research, 41, 1603 - 1613.
Bouhabila, E. H., BenAïm, R. and Buisson, H. (2001). Fouling characterisation in
membrane bioreactors. Separation and Purification Technology, 22-23, 123-132.
Bulgariu, L., Ceica, A., Lazar, L., Cretescu, I., and Balasanian, I. (2010). Equilibrium and
Kinetics Study of Nitrate Removal from Water by Purolite A100 Resin. Revista de
Chimie, 61, 1136 - 1141.
Chang, I.S., Le Clech, P., Jefferson, B. andJudd, S. (2002). Membrane fouling in
membrane bioreactors for wastewater treatment. Journal of environmental
engineering, 128(11), 1018-1029.
Chang, I. S. and Kim, S.N. (2005). Wastewater treatment using membrane filtration effect
of biosolids concentration on cake resistance. Process Biochemistry, 40, 1307–
1314.
Chen, G.H., Wong, M.T., Okabe, S. and Watanabe, Y. (2003). Dynamic response of
nitrifying activated sludge batch culture to increased chloride concentration. Water
Research, 37(13), 3125-3135.
Chen, J.P., Chua, M.L. and Zhang, B. (2002). Effects of competitive ions, humic acid, and
pH on removal of ammonium and phosphorous from the synthetic industrial
effluent by ion exchange resins. Waste Management, 22(7), 711-719.
Choi, J.W., Lee, S.Y., Park, K.Y., Lee, K.B., Kim, D.J., Lee, S.H. (2011). Investigation of
phosphorous removal from wastewater through ion exchange of mesostructure
based on inorganic material. Desalination, 266(1), 281-285.
Chitrakar, R., Tezuka, S., Sonoda, A., Sakane, K., Ooi, K. and Hirotsu, T. (2006).
Selective adsorption of phosphate from seawater and wastewater by amorphous
zirconium hydroxide. Journal of Colloid and Interface Science, 297(2), 426-433.
Cicek, N., Franco, J.P., Suidan, M.T., Urbain, V. and Manem, J. (1999). Characterization
and comparison of a membrane bioreactor and a conventional activated sludge
system in the treatment of wastewater containing high-molecular weight
compounds. Water Environment Research, 71, 64–70.
Cicek, N., Macomber, J., Devel, J., Suidan, M. T., Audic, J. and Genestet, P. (2001). Effect
of Solids Retention Time on the Performance and Biological Characteristics of a
Membrane Bioreactor. Water Research, 30, 1771–1780.
Cui, Y., Peng, C., Peng, Y. and Ye, L. (2009). Effects of salt on microbial populations and
treatment performance in purifying saline sewage using the MUCT process.
CLEAN–Soil, Air, Water, 37(8), 649-656.
Cui, Z. F., Chang, S. and Fane, A. G. (2003). The use of gas bubbling to enhance
membrane processes. Journal of Membrane Science, 221, 1–35.
Dan, N.P., Visvanathan, C., Polprasert, C. and Aim, R.B. (2002). High salinity wastewater
treatment using yeast and bacterial membrane bioreactors. Water Science and
Technology, 46 (9), 201–209.
Dan, N.P., Visvanathan, C., Polprasert, C. and Aim, R.B. (2002). High salinity wastewater
treatment using yeast and bacterial membrane bioreactors. Water and Wastewater
Management for Developing Countries, 46(9), 201-209.
Delaney, P., McManamon, C., Hanrahan, J.P., Copley, M.P., Holmes, J.D. and Morris,
M.A. (2011). Development of chemically engineered porous metal oxides for
phosphate removal. Journal of Hazardous Materials, 185(1), 382-391.
Delgado S., Diaz F. Villarroel R., Vera L., Diaz R., Elmaleh S. (2002). Influence of
biologically treated wastewater quality on filtration through a hollow-fibre
membrane. Desalination, 146(1-3), 459-462.
Deliyanni, E.A., Peleka, E.N. and Lazaridis, N.K. (2007). Comparative study of
phosphates removal from aqueous solutions by nanocrystalline akaganéite and
hybrid surfactant-akaganéite. Separation and Purification Technology, 52(3), 478-
486.
Drews, A. (2010). Membrane fouling in membrane bioreactors—Characterisation,
contradictions, cause and cures. Journal of Membrane Science, 363, 1–28.
Dutta, P.K., Ray, A.K., Sharma, V.K. and Millero, F.J. (2004). Adsorption of arsenate and
arsenite on titanium dioxide suspensions. Journal of Colloid and Interface Science
278 (2), 270–275.
EUROMBRA, (2005). Membrane bioreactor technology (MBR) with an EU perspective
for advanced municipal wastewater treatment strategies for the 21st century,
Contract No. 018480 Deliverable Report – D5, INSA Toulouse, France.
Fane, A. G. (2002). Membrane bioreactors: design & operational options. Filtration &
Separation, 39(5), 26-29.
Fang, H.H.P., Shi, X. (2005). Pore fouling of microfiltration membranes by activated
sludge. Journal of Membrane Science, 264, 161–166.
Faust, S.D.and Aly, O.M. Adsorption process for water treatment. London: Butterworths,
2009, pp.108-113.
Flemming, H.C., Schaule, G., Griebe, T., Schmitt, J. and Tamachkiarowa, A. (1997).
Biofouling—the Achilles heel of membrane processes. Desalination, 113(2), 215-
225.
Galapate, R. P., Agustiani, E., Baes, A. U., Ito, K. and Okada, M. (1999). Effect of HRT
and MLSS on THM precursor removal in the activated sludge process. Water
Research, 33(1), 131-136.
Galil, N.I., Malachi, K.B.D. and Sheindorf, C. (2009). Biological nutrient removal in
membrane biological reactors. Environmental Engineering Science, 26(4), 817-824.
Germain, E., and Stephenson, T. (2005). Biomass characteristics, aeration and oxygen
transfer in membrane bioreactors: their interrelations explained by a review of
aerobic biological processes. Reviews in Environmental Science and
Bio/Technology, 4, 223–233.
Grelier, P., Rosenberger, S., Tazi-Pain, A. (2006). Influence of sludge retention time on
membrane bioreactor hydraulic performance. Desalination 192, 10–17.
Guo, W. S., Shim, W. G., Vigneswaran, S. and Ngo, H. H. (2005). Effect of operating
parameters in a submerged membrane adsorption hybrid system: experiment and
mathematical modeling. Journal of Membrane Science, 247, 65-74.
Guo, W., Vigneswaran, S., Ngo, H. H., Xing, W., and Goteti, P. (2008). Comparison of
the performance of submerged membrane bioreactor (SMBR) and submerged
membrane adsorption bioreactor (SMABR). Bioresource Technology, 99(5), 1012-
1017.
Gupta, M.D., Loganathan, P. and Vigneswaran, S. (2012). Adsorptive removal of nitrate
and phosphate from water by a purolite ion exchange resin and hydrous ferric oxide
columns in series. Separation Science and Technology, 47(12), 1785-1792.
Halling-Sørensen, B. and Jørgensen, S.E. (1993). The removal of nitrogen compounds
from wastewater. Elsevier, 443, 55-56.
Hasar, H., Kınacı, C. and Ünlü, A. (2004). Production of non-biodegradable compounds
based on biomass activity in a submerged ultrafiltration hollow fibre membrane
bioreactor treating raw whey. Process biochemistry, 39(11), 1631-1638.
Hermanowicz, S.W. (2011). Membrane Bioreactors: Past, Present and Future?. Water
Resources Center Archives.
Holler, S. and Trösch, W. (2001). Treatment of urban wastewater in a membrane
bioreactor at high organic loading rates. Journal of Biotechnology, 92, 95–101
Hong S. P., Bae T. H., Tak, T. M., Hong S., Randall, A. (2002), Fouling control in
activated sludge submerged hollow fiber membrane bioreactors. Desalination,
143(3), 219-228.
Hong, C.C., Chan, S.K. and Shim, H. (2007). Effect of chloride on biological nutrient
removal from wastewater. Journal of Applied Sciences in Environmental
Sanitation, 2(3), 85-92.
Hong, S., Rupak, A., Vigneswara, S., Johir, M.A.H., and Kandasamy, J. (2011). Influence
of hydraulic retention time on the nature of foulant organics in a high rate
membrane bioreactor. Desalination, 287, 116-122.
Horan, K. (2007). Membrane Fouling Studies in Suspended and Attached Growth
Membrane Bioreactor Systems. PhD thesis, Asian Institute of Technology, School
of Environment, Resources and Development, Thailand, May 2007.
Horan, N.J. (2003). Handbook of water and wastewater microbiology. Elsevier, 819p:
p150-151.
Huber, S. A., Balz, A., Abert, M. and Pronk, W. (2011). Characterisation of aquatic humic
and non-humic matter with size-exclusion chromatography–organic carbon
detection–organic nitrogen detection (LC-OCD-OND). Water Research, 45(2), 879-
885.
Hwang, B.K., Lee, W.N., Park, P.K., Lee, C.H. and Chang, I.S. (2007). Effect of
membrane fouling reducer on cake structure and membrane permeability in
membrane bioreactor. Journal of Membrane Science, 288, 149–156.
Jang, N., Ren, X., Kim, G., Ahn, C., Cho, J., Kim, IS. (2006). Characteristics of soluble
microbial products and extracellular polymeric substances in the membrane
bioreactor for water reuse. Desalination, 202, 90–98.
Jefferson, B., Brookes, A., Le-Clech, P. and Judd, S.J., (2004). Methods for understanding
organic fouling in MBRs. Water Science and Technology, 49, 237–244.
Jin, L., Ong, S.L., and Ng, H.Y. (2013). Fouling control mechanism by suspended biofilm
carriers addition in submerged ceramic membrane bioreactors. Journal of
Membrane Science, 427, 250-258.
Judd, S. (2007). Membrane bioreactor technology costs. Proceedings of International
Membrane Science and Technology Conference 2007, Nov., Sydney, Australia
Jung, Y., J. Koh, H. W., Shin, W. T. and Sung, N. C. (2006). A novel approach to an
advanced tertiary wastewater treatment: Combination of a membrane bioreactor
and an oyster-zeolite column. Desalination, 190, 243-255.
Kang, S., Lee, W., Chae, S. and Shin, H. (2007). Positive roles of biofilm during the
operation of membrane bioreactor for water reuse. Desalination, 202(1), 129-134.
Kawasaki, K., Maruoka, S., Katagami, R., Bhatta, C.P., Omori, D., Matsuda, A. (2011).
Effect of initial MLSS on operation of submerged membrane activated sludge
process. Desalination, 281, 334–339.
Ke, O. and Junxin, L. (2009). Effect of sludge retention time on sludge characteristics and
membranefouling of membrane bioreactor. Journal of Environmental Science, 21,
1329–1335.
Khoshfetrat, A.B., Nikakhtari, H., Sadeghifar, M. and Khatibi, M.S. (2011). Influence of
organic loading and aeration rates on performance of a lab-scale upflow aerated
submerged fixed-film bioreactor. Process Safety and Environmental Protection,89,
193–197.
Kim, J. and DiGiano, F.A. (2006). Defining critical flux in submerged membranes:
influence of length-distributed flux. Journal of Membrane Science, 280(1), 752-
761.
Kim, J.S., and Lee, C.H. (2003). Effect of powdered activated carbon on the performance
of an aerobic membrane bioreactor: comparison between cross flow and
submerged membrane systems. Water Environment Research, 75, 300-307.
Kimura, K., Naruse, T. and Watanabe, Y. (2009). Changes in characteristics of soluble
microbial products in membrane bioreactors associated with different solid
retention times: relation to membrane fouling. Water Research, 43, 1033–1039.
Kincannon, D.F. and Gaudy, A.F. (1968). Response of biological waste treatment systems
to changes in salt concentrations. Biotechnology and Bioengineering, 10(4), 483-
496.
Kornboonraksa, T. and Lee, S.H. (2009). Factors affecting the performance of membrane
bioreactor for piggery wastewater treatment. Bioresource Technology, 100, 2926–
2932.
Kubo, M., Hiroe, J., Murakami, M., Fukami, H. and Tachiki, T. (2001). Treament of
hypersaline-containing wastewater with salt-tolerant microorganisms. Journal of
Bioscience and Bioengineering, 91(2), 222-224.
Laera, G., Pollice, A., Saturno, D., Giordano, C. and Sandulli, R. (2009). Influence of
sludge retention time on biomass characteristics and cleaning requirements in a
membrane bioreactor for municipal wastewater treatment. Desalination, 236(1),
104-110.
Laspidou, C.S. and Rittmann, B.E. (2002). A unified theory for extracellular polymeric
substances, soluble microbial products, and active and inert biomass. Water
Research, 36, 2711–2720.
Lay, W.C., Liu, Y. and Fane, A.G. (2010). Impacts of salinity on the performance of high
retention membrane bioreactors for water reclamation: a review. Water Research,
44(1), 21-40.
Lebegue, J., Heran, M., and Grasmick, A. (2007), Proceedings of the Sixth International
Membrane Science and Technology Conference, November 5–9, Sydney,
Australia.
Le-Clech, P., Chen, V. and Fane, T.A.G. (2006). Fouling in membrane bioreactors used
in wastewater treatment. Journal of Membrane Science, 284(1-2), 17-53.
Lee, S.H., Lee, B.C., Lee, S.H., Choi, Y.S., Park, K.Y. and Iwamoto, M. (2007).
Phosphorus recovery by mesoporous structure materials from wastewater. Water
Science and Technology, 55(1-2), 169-176.
Lee, W., Kang, S. and Shin, H. (2003). Sludge characteristics and their contribution to
microfiltration in submerged membrane bioreactors. Journal of Membrane Science,
216(1-2), 217-227.
Lefebvre, O. and Moletta, R. (2006). Treatment of organic pollution in industrial saline
wastewater: a literature review. Water Research, 40(20), 3671-3682.
Lesage, N., Sperandio, M. and Cabassud, C. (2008). Study of a hybrid process: Adsorption
on activated carbon/membrane bioreactor for the treatment of an industrial
wastewater. Chemical Engineering and Processing: Process Intensification, 47(3),
303-307.
Li, H., Yang, M., Zhang, Y., Liu, X., Gao, M. and Kamagata Y. (2005). Comparison of
nitrification performance and microbial community between submerged
membrane bioreactor and conventional activated sludge system. Water Science and
Technology 51, 193–200.
Li, Y-Z., He, Y-L., Liu, Y-H., Yang, S-C. and Zhang, G-J. (2005). Comparison of
thefiltration characteristics between biological powdered activated carbon sludge
and activated sludge in submerged membrane bioreactors. Desalination, 174, 305-
314.
Liang, S., Liu, C., and Song, L. (2007). Soluble microbial products in membrane bioreactor
operation: Behaviors, characteristics, and fouling potential. Water Research, 41, 95
– 101.
Lio, Y. and McLaughlin, J.B. (2000). Bubble motion in aqueous surfactant solutions.
Journal of Colloid and Interface Science, 224 (2), 297-310.
Liu, H., Sun, X., Yin, C. and Hu, C. (2008). Removal of phosphate by mesoporous ZrO2.
Journal of Hazardous Materials, 151(2), 616-622.
Liu, S. X. (2007). Food and agricultural wastewater utilization and treatment, Blackwell
Publishing, 277p: p148.
Loganathan, P., Vigneswaran, S., Kandasamy, J. and Bolan, N.S. (2014). Removal and
recovery of phosphate from water using sorption. Critical Reviews in Environmental
Science and Technology, 44(8), 847-907.
Long, F., Gong, J.L., Zeng, G.M., Chen, L., Wang, X.Y., Deng, J.H. and Zhang, X.R.
(2011). Removal of phosphate from aqueous solution by magnetic Fe–Zr binary
oxide. Chemical Engineering Journal, 171(2), 448-455.
Madec A. (2000). Influence of two phase flow on the filtration through submerged
membrane processes, Ph.D thesis, INSA, Toulouse, France
Malysa, K., Krasowska, M. and Krzan, M. (2005). Influence of surface active substances
on bubble motion and collision with various interfaces. Advances in Colloid and
Interface Science, 114-115, 205-225.
Meng, F., Chae, S-R., Drews, A., Kraume, M., Shin, H-S., and Yang, F. (2009). Recent
advances in membrane bioreactors (MBRs): Membrane fouling and
membrane material. Water Research, 43, 1489-1512.
Meng, F., Shi, B., Yang, F. and Zhang, H. (2007). Effect of hydraulic retention time on
membrane fouling and biomass characteristics in submerged membrane
bioreactors. Bioprocess and Biosystems Engineering, 30, 359-367.
Meng, F., Yang, F., Shi, B. and Zhang, H. (2008). A comprehensive study on membrane
fouling in submerged membrane bioreactors operated under different aeration
intensities. Separation and Purification Technology, 59, 91–100.
Metzger, U., Le-Clech, P., Stuetz, R.M., Frimmel, F.H., and Chen, V. (2007).
Characterisation of polymeric fouling in membrane bioreactors and the effect
of different filtration modes. Journal of Membrane Science, 301, 180–189.
Navaratna, D. and Jegatheesan, V. (2011). Implications of short and long term critical flux
experiments for laboratory-scale MBR operations. Bioresource Technology, 102,
5361–5369.
Ng, H.Y. and Hermanowicz, S.W. (2005). Membrane bioreactor operation at short solids
retention times: performance and biomass characteristics. Water Research, 39,
981– 992.
Nguyen, L.N., Hai, F.I., Kang, J., Price, W.E. and Nghiem, L.D. (2012). Removal of trace
organic contaminants by a membrane bioreactor–granular activated carbon (MBR–
GAC) system. Bioresource technology, 113, 169-173.
Nur, T., Johir, M.A.H., Loganathan, P., Nguyen, T., Vigneswaran, S. and Kandasamy, J.
(2014). Phosphate removal from water using an iron oxide impregnated strong base
anion exchange resin. Journal of Industrial and Engineering Chemistry, 20(4),
1301-1307.
Nur, T., Johir, M.A.H., Loganathan, P., Vigneswaran, S. and Kandasamy, J. (2012).
Effectiveness of purolite A500PS and A520E ion exchange resins on the removal of
nitrate and phosphate from synthetic water. Desalination and Water Treatment,
47(1-3), 50-58.
Nur, T., Loganathan, P., Nguyen, T.C., Vigneswaran, S., Singh, G. and Kandasamy, J.
(2014a). Batch and column adsorption and desorption of fluoride using hydrous
ferric oxide: Solution chemistry and modeling. Chemical Engineering Journal, 247,
93-102.
Ognier, S., Wisniewski, C. and Grasmick, A. (2004). Membrane bioreactor fouling in sub-
critical filtration conditions: a local critical flux concept. Journal of Membrane
Science, 229, 171– 177.
Pan, J.R., Su, Y. and Huang, C. (2010). Characteristics of soluble microbial products in
membrane bioreactor and its effect on membrane fouling. Desalination, 250, 778–
780.
Poulton D.J. (1989). Statistical zonation of sediment samples using ratio matching and
cluster analysis. Environmental Monitoring and Assessment 13, 379–404.
doi:10.1007/BF00394241
Pradhan, M., Vigneswaran, S., Kandasamy, J. and Ben Aim, R. (2012). Combined effect
of air and mechanical scouring of membranes for fouling reduction in submerged
membrane reactor. Desalination, 288, 58–65.
Primo, O., Rivero, M., Urtiaga, A., and Ortiz, I. (2009). Nitrate removal from electro-
oxidized landfill leachate by ion exchange. Journal of Hazardous Materials, 164, 389
- 393.
Radjenovi , J., Matoši , M., Mijatovi , I., Petrovi , M. and Barceló, D. (2008). Membrane
Bioreactor (MBR) as an Advanced Wastewater Treatment Technology. Hdb. Env.
Chem. 5, Part S/2, 37–101.
Raji, C. and Anirudhan, T.S. (1998). Batch Cr (VI) removal by polyacrylamide-grafted
sawdust: kinetics and thermodynamics. Water Research, 32(12), 3772-3780.
Reid, E., Liu, X. and Judd, S.J. (2006). Effect of high salinity on activated sludge
characteristics and membrane permeability in an immersed membrane bioreactor.
Journal of Membrane Science, 283(1), 164-171.
Ren N, Chen Z, Wang X, Hu D, Wang, A. (2005). Optimized operational parameters of a
pilot scale membrane bioreactor for high-strength organic wastewater treatment.
International Biodeterioration and Biodegradation 56:216–223.
Ren, N., Chen, Z., Wang, A., Hu, D., (2004). Removal of organic pollutants and analysis
of MLSS-COD removal at different HRTs in a submerged membrane bioreactor.
International Biodeterior Biodegradation. 55, 279-284.
Ren, N., Chen, Z., Wang, X., Hu, D. and Wang, A. (2005). Optimized operational
parameters of a pilot scale membrane bioreactor for high-strength organic
wastewater treatment. International Biodeterioration and Biodegradation 56:216–
223.
Rene, E.R., Kim, S.J. and Park, H.S. (2008). Effect of COD/N ratio and salinity on the
performance of sequencing batch reactors. Bioresource Technology, 99(4), 839-
846.
Rosenberger, S. and Kraume, M. (2002). Filterability of activated sludge in membrane
bioreactors. Desalination, 151, 195-200
Samatya, S., Kabay, N., Yüksel, Ü., Arda, M. and Yüksel, M. (2006). Removal of nitrate
from aqueous solution by nitrate selective ion exchange resins. Reactive and
Functional Polymers, 66(11), 1206-1214.
Satyawali, Y. and Balakrishnan, M. (2009). Effect of PAC addition on sludge properties in
an MBR treating high strength wastewater. Water Research, 43(6), 1577-1588.
Schramm, A., De Beer, D., Wagner, M. and Amann. R. (1998). Identification and
activities in situ of Nitrosospira and Nitrospira spp. as dominant populations
in a nitrifying fluidized bed reactor. Applied and Environmental Microbiology ,
64(9), 3480.
Shanmuganathan, S., Johir, M.A., Nguyen, T.V., Kandasamy, J. and Vigneswaran, S.
(2015). Experimental evaluation of microfiltration–granular activated carbon (MF–
GAC)/nano filter hybrid system in high quality water reuse. Journal of Membrane
Science, 476, 1-9.
Sharrer, M.J., Tal, Y., Ferrier, D., Hankins, J.A. and Summerfelt, S.T. (2007). Membrane
biological reactor treatment of a saline backwash flow from a recirculating
aquaculture system. Aquacultural Engineering, 36(2), 159-176.
Shen, L., Zhou, Y., Mahendran, B., Bagley, D.M. and Liss, S.N. (2010). Membrane
fouling in a fermentative hydrogen producing membrane bioreactor at different
organic loading rates. Journal of Membrane Science, 360, 226–233.
Shon, H.K., Kim, S.H., Erdei, L. and Vigneswaran, S. (2006). Analytical methods of size
distribution for organic matter in water and wastewater. Korean Journal of
Chemical Engineering, 23(4), 581-591.
Siembida, B., Cornel, P., Krause, S. and Zimmermann, B. (2010). Effect of mechanical
cleaning with granular material on the permeability of submerged membranes in the
MBR process. Water Research, 44, 4037-4046.
Smith, P. J., Vigneswaran, S., Ngo, H. H., Ben-Aim, R., Nguyen, H. (2006). A new
approach to backwash initiation in membrane system. Journal of Membrane
Science, 278, (1-2), 381-389.
Sofia, A., Ng, W., Ong, S. L. (2004). Engineering design approaches for minimum fouling
in submerged MBR. Desalination, 160(1), 67-74.
Sombatsompop, K., Visvanathan, C., and BenAim, R. (2006). Evaluation of biofouling
phenomenon in suspended and attached growth membrane bioreactor. Desalination,
201, 138-149.
Stewart, M.J., Ludwig, H.F. and Kearns, W.H. (1962). Effects of varying salinity on the
extended aeration process. Journal (Water Pollution Control Federation), 1161-
1177.
Stumm, W. and Morgan, J.J. (1995). Aquatic Chemistry: Chemical Equilibria and Rates in
Natural Waters. Wiley, New York.
Su, Y., Cui, H., Li, Q., Gao, S. and Shang, J.K. (2013). Strong adsorption of phosphate by
amorphous zirconium oxide nanoparticles. Water Research, 47(14), 5018-5026.
Sun, F., Wang, X., Li, X. (2011). Change in the fouling propensity of sludge in membrane
bioreactors (MBR) in relation to the accumulation of biopolymer clusters.
Bioresource Technology, 102, 4718–4725
Takemura, F. (2005). Adsorption of surfactants onto the surface of a spherical rising
bubble and its effect on the terminal velocity of the bubble. Physics of Fluid, 17,
048104-1.
Terry, P. A. (2009). Removal of nitrates and phosphates by ion exchange with hydrotalcite.
Environmental Engineering Science, 26(3), 691-696.
Teychene, B., Guigui, C., Cabassud, C. and Amy, G. (2008). Toward a better identification
of foulant species in MBR processes. Desalination, 231(1), 27-34.
Trussell, R.S., Merlo, R.P., Hermanowicz, S.W. and Jenkins, D. (2006). The effect of
organic loading on process performance and membrane fouling in a
submerged membrane bioreactor treating municipal astewater. Water Research,
40, 2675–2683.
Ueda, T., Hata, K., Kikuoka, Y. and Seino, O. (1997). Effects of aeration on suction
pressure in a submerged membrane bioreactor. Water Research, 31(3), 489-494.
Uygur, A. (2006). Specific nutrient removal rates in saline wastewater treatment using
sequencing batch reactor. Process Biochemistry, 41(1), 61-66.
Velten, S., Knappe, D.R., Traber, J., Kaiser, H.P., von Gunten, U., Boller, M. and Meylan,
S. (2011). Characterization of natural organic matter adsorption in granular
activated carbon adsorbers. Water Research, 45, 3951-3959.
Vendramel, S., Dezotti, M., & Sant Anna, G. L. (2011). Nitrification of an industrial
wastewater in a moving bed biofilm reactor: effect of salt concentration.
Environmental technology, 32(8), 837-846.
Viero, A.F. and Sant’Anna, G.L. (2008). Is hydraulic retention time an essential
parameter for MBR performance?. Journal of Hazardous Materials, 150(1), 185-
186.
Viero, A.F., Sant’Anna, G.L. and Nobrega Jr., R. (2007). The use of polyetherimide
hollow fibres in a submerged membrane bioreactor operating with air
backwashing. Journal of Membrane Science, 302 (1–2), 127–135.
Villacorte, L.O., Kennedy, M.D., Amy, G.L. and Schippers, J.C. (2009). The fate of
transparent exopolymer particles (TEP) in integrated membrane systems: removal
through pre-treatment processes and deposition on reverse osmosis membranes.
Water Research, 43(20), 5039-5052.
Visvanathan, C., Boonthanon, N., Sathasivan, A., and Jegatheesan, V. (2003). Pretreatment
of seawater for biodegradable organic content removal using membrane bioreactor.
Desalination, 153(1), 133-140.
Westerhoff, P. (2003). Reduction of nitrate, bromate, and chlorate by zero valent iron
(Fe0). Journal of Environmental Engineering, 129, 10 - 16.
Wicaksana, F., Fane, A.G. and Chen, V. (2006). Fibre movement induced by bubbling
using submerged hollow fibre membranes. Journal of Membrane Science, 271,
186–195.
Woolard, C.R. and Irvine, R.L. (1995). Treatment of hypersaline wastewater in the
sequencing batch reactor. Water Research, 29(4), 1159-1168.
Wu, G., Guan, Y. and Zhan, X. (2008). Effect of salinity on the activity, settling and
microbial community of activated sludge in sequencing batch reactors treating
synthetic saline wastewater. Water Science and Technology. 58 (2008) 351-358
Wu, J., Chen, F., Huang, X., Geng, W. and Wen, X. (2006). Using inorganic coagulants to
control membrane fouling in a submerged membrane bioreactor. Desalination,
197 (1-3), 124-136.
Xing, W., Ngo, H. H., Guo, W. S., Listowski, A. and Cullum, P. (2012). Optimization of
an integrated sponge–granular activated carbon fluidized bed bioreactor as
pretreatment to microfiltration in wastewater reuse. Bioresource Technology, 113,
214-218.
Yan, L.G., Xu, Y.Y., Yu, H.Q., Xin, X.D., Wei, Q. and Du, B. (2010). Adsorption of
phosphate from aqueous solution by hydroxy-aluminum, hydroxy-iron and
hydroxy-iron–aluminum pillared bentonites. Journal of hazardous materials, 179(1),
244-250.
Yang, Z., Peng, X.F., Chen, M.Y., Lee, D.J., and Lai, J.Y. (2007). Intra-layer flow in
fouling layer on membranes. Journal of Membrane Science, 287, 280–286.
Ye, L., Peng, C.Y., Tang, B., Wang, S.Y., Zhao, K.F. and Peng, Y.Z. (2009).
Determination effect of influent salinity and inhibition time on partial nitrification
in a sequencing batch reactor treating saline sewage. Desalination, 246(1), 556-566.
Yogalakshmi, K.N. and Joseph, K. (2010). Effect of transient sodium chloride shock loads
on the performance of submerged membrane bioreactor. Bioresource technology,
101(18), 7054-7061.
Yoon, S-H., Kim, H-S, Yeom, I-T. (2004). The optimum operational condition of
membrane bioreactor (MBR): cost estimation of aeration and sludge treatment.
Water Research, 38, 37-46.
Yu, K., Wen, X., Bu, Q. and Xia, H. (2003). Critical flux enhancements with air sparging
in axial hollow fibers cross-flow microfiltration of biologically treated wastewater.
Journal of Membrane Science, 224(1), 69-79.
Yuzir, A., Chelliapan, S., Sallis, PJ. (2011), Influence of step increases in hydraulic
retention time on (RS)-MCPP degradation using an anaerobic membrane
bioreactor. Bioresource Technology, 102, 9456–9461.
Zhang, J., Chua, H.C., Zhou, J. and Fane, A.G. (2006). Factors affecting the membrane
performance in submerged membrane bioreactors. Journal of Membrane Science,
284(1), 54-66.
Zhang, G., Liu, H., Liu, R., Qu, J. (2009). Removal of phosphate from water by a Fe–Mn
binary oxide adsorbent. Journal of Colloid Interface Science 335(2), 168-174.
Zheng, X., Sun, P., Han, J., Song, Y., Hu, Z., Fan, H., and Lv, S. (2014). Inhibitory factors
affecting the process of enhanced biological phosphorus removal (EBPR)–A mini-
review. Process Biochemistry, 49(12), 2207-2213.