ISTANBUL TECHNICAL UNIVERSITY  GRADUATE SCHOOL OF SCIENCE

ENGINEERING AND TECHNOLOGY

DELFT UNIVERSITY OF TECHNOLOGY  GRADUATE SCHOOL OF CIVIL
ENGINEERING AND GEOSCIENCES

ANAEROBIC MEMBRANE BIOREACTORS

FOR COST-EFFECTIVE
MUNICIPAL WATER REUSE

Ph.D. THESIS

Hale ÖZGÜN

Department of Environmental Engineering

Environmental Sciences and Engineering Programme

OCTOBER 2015
ISTANBUL TECHNICAL UNIVERSITY  GRADUATE SCHOOL OF SCIENCE
ENGINEERING AND TECHNOLOGY
DELFT UNIVERSITY OF TECHNOLOGY  GRADUATE SCHOOL OF CIVIL
ENGINEERING AND GEOSCIENCES

ANAEROBIC MEMBRANE BIOREACTORS

FOR COST-EFFECTIVE
MUNICIPAL WATER REUSE

Ph.D. THESIS

Hale ÖZGÜN
(501072702)

Department of Environmental Engineering

Environmental Sciences and Engineering Programme

Thesis Advisor: Prof. Dr. Cumali KINACI

ThesisAnabilim DalıProf.
Co-Advisor: : Herhangi
Dr. JulesMühendislik,
B. van LIERBilim
Programı : Herhangi Program

OCTOBER 2015
 İSTANBUL TEKNİK ÜNİVERSİTESİ  FEN BİLİMLERİ ENSTİTÜSÜ
DELFT TEKNOLOJİ ÜNİVERSİTESİ  İNŞAAT MÜHENDİSLİĞİ VE YER
BİLİMLERİ ENSTİTÜSÜ

EVSEL ATIKSULARIN ANAEROBİK MEMBRAN BİYOREAKTÖRLER İLE

MALİYET ETKİN OLARAK YENİDEN KULLANIMI

DOKTORA TEZİ

Hale ÖZGÜN
(501072702)

Çevre Mühendisliği Anabilim Dalı

Çevre Bilimleri ve Mühendisliği Programı

Tez Danışmanı: Prof. Dr. Cumali KINACI

Anabilim Dalı
Eş Danışman: : Dr.
Prof. Herhangi
Jules B.Mühendislik,
van LIER Bilim
Programı : Herhangi Program

EKİM 2015
Hale ÖZGÜN, a Ph.D. student of ITU Graduate School of Science Engineering and
Technology and Delft University of Technology Graduate School of Civil
Engineering and Geosciences student ID 501072702, successfully defended the
dissertation entitled “ANAEROBIC MEMBRANE BIOREACTORS FOR COST-
EFFECTIVE MUNICIPAL WATER REUSE”, which she prepared after fulfilling
the requirements specified in the associated legislations, before the jury whose
signatures are below.

Thesis Advisor : Prof. Dr. Cumali KINACI ..............................

İstanbul Technical University

Co-advisor : Prof.Dr. Jules B. van LIER ..............................

Delft University of Technology

Jury Members : Prof. Dr. İzzet ÖZTÜRK .............................

İstanbul Technical University

Prof. Dr. İsmail KOYUNCU ..............................

İstanbul Technical University

Prof. Dr. Grietje ZEEMAN ..............................

Wageningen University

Prof. Dr. Louis Cornelis RIETVELD ..............................

Delft University of Technology

Assoc. Prof. Dr. Henri L.M.F. SPANJERS.............................

Delft University of Technology

Date of Submission : 08 September 2015

Date of Defence : 20 October 2015

v
vi
 To my father, Hamdi ÖZGÜN..

Canım Babama..

vii
viii
FOREWORD

This section is the most difficult part of this thesis for me to write since there are lots
of things that I would like to express and unfortunately, I am not so good at
expressing my deepest feelings by words. Apparently, for me writing this section
will be more tough than writing discussion parts of the thesis. But anyway, I will try
my best..
First of all, I would like to thank my two promoters, Prof. Cumali Kınacı and Prof.
Jules van Lier. Prof. Kınacı, thank you for the continuous support and guidance you
provided me throughout my time as your student since my undergraduate studies.
From the beginning, your positive outlook and confidence in my research have
inspired me and gave me confidence. Your support, patience, and knowledge that
you provided me were highly commendable. You are always more than a professor
for me. Jules, thank you for your guidance, understanding and teaching me how to do
research. I started my research journey with your most frequent and difficult
question: “What’s your research question?”. Soon after though, I realized how useful
and important this question was for performing a scientific research. During my PhD
study, I understand that not only have you been there as an expert to discuss the
research-related topics, but more importantly, you have been there as a friend
wishing me strength and giving me support about the unjustness of life for the most
difficult times, which I never forgot. I feel like I still need to learn a lot from you.
My sincere gratitude also goes to my co-promoter Dr. Henri Spanjers for his kind
supervision and support. The ideas you gave me as an input (e.g. “marble” idea for
my UASB) in our weekly discussions, whenever I had a hard time with the
experiments were of my moral support. I have learned a lot from you, from
theoretical knowledge to data presentation. It was an honor and pleasure to work
under your guidance.
I would like to express my sincere appreciation and thanks to Prof. İzzet Öztürk and
Prof. İsmail Koyuncu. Prof. Öztürk, thank you for giving me the opportunity of
learning many things from you, your continuous support, and for the many great
ideas and important comments that have helped me during my time as your student.
Prof. Koyuncu, thank you for introducing me the field of research on membranes,
sharing your knowledge, and continuous support you provided me during my
research.
Sincere thanks also goes to Tonny Schuit for his valuable inputs, discussions, and
helping me to find my way in the lab. Furthermore, my compliments go to Jennifer
Duiverman, Mieke Hubert and Patrick Andeweg for their support in administrative
staff and lab. Also, many thanks go to Peter and Robert from Carya Automatisering
for their valuable contribution during the software establishment, to Leon and Rob
from the TU Delft workshop for their help during the construction of my
experimental setups, and to Saskia from Deltares for her support. Also thanks to
Biothane that made their laboratory facilities available for the part of my
experimental work.
Many thanks to the other research member of the A-Racer project: Geo Smith. The
meetings we had urged me to sort out the pile of results and provided fresh insights.

ix
Words of special thanks go to Jan Pereboom for welcoming us at our first arrival to
Delft.
I also would like to express my gratitude to Turkish Academy of Sciences (TUBA)
for providing me the PhD Fellowship award to support my living and research in
Delft. I acknowledge A-Racer project and AgentschapNL for financial support to the
research. This thesis is produced as part of the A-Racer project with Pentair, Saxion,
TU Delft and Water Board Regge & Dinkel as partners. The project (IWA10007) is
partly funded by the Dutch Government via AgentschapNL under the InnoWator
program.
During the course of my research, I received great help from my colleagues and
friends: Juan B. Gimenéz, Jiang Li, Mirjam van der Spek, Yu Tao and Zhongbo
Zhou. I want to express my sincere appreciation for your valuable advices and
comprehensive explanations. Juan, coupling membrane to a UASB would not be
possible without your cooperation, inspiring discussions and your contribution. I am
deeply grateful to Mirjam who helped me with the countless experimental work.
Jiang, thanks for accompanying my lab journey at TU Delft on my first arrival. Yu,
thanks for introducing me to the fascinating world of environmental microbiology.
And Zhongbo, your positive energy is fabulous. I thank you very much for your
contributions to my thesis.
In Delft, I also would like to thank all friends for accompanying my life and PhD
journey. It was great to meet and know all of you. Xuedong, thank you for all the
time that we spent together sharing joy and happiness and for your continuous help;
Dara, thank you for always sharing laughs and making my days better not only in lab
but also in Delft; Jianmei, thank you for your everlasting smile and unforgetable
surprises and Annelies, thank you for the nice chats that we always have. Then,
there’s the huge, but probably still incomplete list of my other dear friends at TU
Delft: Annelies, Rui, Jingyi, Ran, Maria, Mostafa, Julian, Jixiang, Haoyu, Roslinda,
Fietha, Pawel, Faiz, Mohammed, Joraslie, Feifei, Marisa, Peter, Guido, Cuijie and
Peng. A warm thank you to Mehmet Abi, Fatma Abla, Meltem and Mustafa: I felt
more familiar and cozy living in Delft after I found Mehmet Abi at his Turkish
grocery. I would especially like to thank Kaan, not only for his significant
contributions in workload, but also for all the fun we had after working hours. It has
been great to share our Delft days with you. Many thanks go to you all for being
present in my life: Burcu, Burcu, Ayşegül&Turgut, Neslihan&Şafak, Zeynep, Özgül,
Başak, Seda, Tuba, Eelco, Emine&Kaan, Ömer and Emel. It is with you guys that I
have shared some of the most inspiring moments in my life, without you it would
never be the same.
I would like to thank Evren for being always there to support me and helping on
nearly every step of this research and sharing me the same way through life.
Last but certainly not least: thank you my parents Hamdi and Hatice Özgün and my
sister Hande, my brother in law Oğuzhan, my niece and nephew Dila and Bora for
your lifetime support and love at all times.. The laughter of my sweeties Dila and
Bora have always provided me joyful environment at home. My sister, thank you for
your continuous love and enthusiasm. My mother Hatice Özgün, thank you for your
endless devotion to my life. My father Hamdi Özgün, thank you for showing me that
in life only a few things matter, and for everything that I can not express in words
now. I dedicate this thesis to my father. I know he’s always with me..

September 2015 Hale ÖZGÜN

x
TABLE OF CONTENTS

Page

FOREWORD............................................................................................................. ix
TABLE OF CONTENTS.......................................................................................... xi
ABBREVIATIONS .................................................................................................. xv
SYMBOLS ..............................................................................................................xvii
LIST OF TABLES .................................................................................................. xix
LIST OF FIGURES ................................................................................................ xxi
SUMMARY ...........................................................................................................xxiii
ÖZET.....................................................................................................................xxvii
1. INTRODUCTION.................................................................................................. 1
1.1 Background ........................................................................................................ 1
1.2 Problem Statement, Objectives and Structure of the Thesis .............................. 2
1.2.1 Problem statement and research approach .................................................. 2
1.2.2 Research objectives..................................................................................... 5
1.2.3 Thesis outline .............................................................................................. 6
2. ANAEROBIC MEMBRANE BIOREACTORS FOR MUNICIPAL
WASTEWATER TREATMENT: INTEGRATION OPTIONS, LIMITATIONS
AND EXPECTATIONS .......................................................................................... 11
2.1 Introduction ...................................................................................................... 11
2.2 Integration Possibilities of Membranes with Different Types of Anaerobic
Reactors for Municipal Wastewater Treatment ..................................................... 14
2.2.1 Completely stirred tank reactor................................................................. 15
2.2.2 High-Rate anaerobic reactors.................................................................... 20
2.2.2.1 Upflow anaerobic sludge blanket reactor........................................... 20
2.2.2.2 Expanded granular sludge bed reactor ............................................... 22
2.2.2.3 Other reactor types ............................................................................. 23
2.3 Alternative Integration Possibilities of AnMBRs in Municipal Wastewater
Treatment Flow Schemes ....................................................................................... 24
2.4 Factors Affecting the Treatment Performance of AnMBRs for Municipal
Wastewater Treatment............................................................................................ 28
2.4.1 Operational conditions .............................................................................. 28
2.4.1.1 Temperature ....................................................................................... 28
2.4.1.2 Organic loading rate........................................................................... 30
2.4.1.3 Hydraulic retention time .................................................................... 30
2.4.1.4 Upflow velocity.................................................................................. 31
2.4.2 Sludge characteristics................................................................................ 31
2.4.3 Addition of adsorbents .............................................................................. 32
2.5 Factors Affecting the Membrane Performance of AnMBRs for Municipal
Wastewater Treatment............................................................................................ 33
2.5.1 Membrane characteristics ......................................................................... 33
2.5.1.1 Material .............................................................................................. 33
2.5.1.2 Module type and configuration .......................................................... 34

xi
 2.5.2 Operational conditions .............................................................................. 35
2.5.2.1 Shear rate............................................................................................ 35
2.5.2.2 Flux..................................................................................................... 36
2.5.2.3 Operation mode .................................................................................. 36
2.5.2.4 Temperature ....................................................................................... 37
2.5.2.5 Upflow velocity.................................................................................. 37
2.5.2.6 Solids retention time........................................................................... 38
2.5.2.7 Hydraulic retention time..................................................................... 39
2.5.3 Sludge characteristics................................................................................ 39
2.5.4 Addition of flux enhancers ........................................................................ 40
2.6 Cleaning Methods............................................................................................. 41
2.7 Economic Feasibility of AnMBRs for Municipal Wastewater Treatment ....... 43
2.8 Problems Encountered and Future Perspectives............................................... 45
2.9 Conclusion........................................................................................................ 48
3. EFFECT OF UPFLOW VELOCITY ON THE EFFLUENT MEMBRANE
FOULING POTENTIAL IN MEMBRANE COUPLED UPFLOW
ANAEROBIC SLUDGE BLANKET REACTORS .............................................. 49
3.1 Introduction ...................................................................................................... 49
3.2 Materials and Methods ..................................................................................... 51
3.2.1 Wastewater source..................................................................................... 51
3.2.2 Seed sludge................................................................................................ 52
3.2.3 Experimental setup.................................................................................... 53
3.2.4 Experimental procedure ............................................................................ 53
3.2.5 Analytical methods.................................................................................... 54
3.2.6 Filterability and reversibility test .............................................................. 55
3.3 Results and Discussion ..................................................................................... 57
3.3.1 Effect of upflow velocity on system performance .................................... 57
3.3.2 Filterability and reversibility at different upflow velocities...................... 64
3.4 Conclusion........................................................................................................ 65
4. IMPACT OF MEMBRANE ADDITION FOR EFFLUENT EXTRACTION
ON THE PERFORMANCE AND SLUDGE CHARACTERISTICS OF
UPFLOW ANAEROBIC SLUDGE BLANKET REACTORS............................ 67
4.1 Introduction ...................................................................................................... 67
4.2 Materials and Methods ..................................................................................... 70
4.2.1 Wastewater source..................................................................................... 70
4.2.2 Seed sludge................................................................................................ 71
4.2.3 Experimental setup.................................................................................... 71
4.2.4 Experimental procedure ............................................................................ 72
4.2.5 Analytical methods.................................................................................... 73
4.2.5.1 Experimental methods applied to effluent and permeate samples ..... 73
4.2.5.2 Experimental methods applied to sludge samples.............................. 73
4.3 Results and Discussion ..................................................................................... 76
4.3.1 Impact of membrane incorporation on UASB effluent ............................. 76
4.3.2 Impact of membrane incorporation on the overall system performance... 80
4.3.3 Impact of membrane incorporation on sludge properties.......................... 82
4.4 Conclusion........................................................................................................ 90
5. IMPACT OF TEMPERATURE ON FEED-FLOW CHARACTERISTICS
AND FILTRATION PERFORMANCE OF AN UPFLOW ANAEROBIC
SLUDGE BLANKET COUPLED ULTRAFILTRATION MEMBRANE......... 91
5.1 Introduction ...................................................................................................... 91

xii
 5.2 Materials and Methods ..................................................................................... 93
5.2.1 Wastewater source .................................................................................... 93
5.2.2 Seed sludge ............................................................................................... 94
5.2.3 Experimental setup.................................................................................... 94
5.2.4 Experimental procedure ............................................................................ 96
5.2.5 Experimental methods............................................................................... 97
5.2.5.1 Analysis.............................................................................................. 97
5.2.5.2 Critical flux determination ................................................................. 97
5.2.5.3 Filtration resistance ............................................................................ 97
5.2.5.4 Microbial community analysis ........................................................... 98
5.3 Results and Discussion..................................................................................... 99
5.3.1 Effluent characteristics.............................................................................. 99
5.3.2 Microbial community.............................................................................. 104
5.3.3 Critical flux tests ..................................................................................... 108
5.3.4 Filtration performance............................................................................. 109
5.3.5 Cleaning tests .......................................................................................... 111
5.4 Conclusion...................................................................................................... 113
6. COMPARATIVE EVALUATION OF THE SLUDGE CHARACTERISTICS
ALONG THE HEIGHT OF UPFLOW ANAEROBIC SLUDGE BLANKET
COUPLED ULTRAFILTRATION SYSTEMS .................................................. 115
6.1 Introduction .................................................................................................... 115
6.2 Materials and Methods ................................................................................... 117
6.2.1 Wastewater source .................................................................................. 117
6.2.2 Seed sludge ............................................................................................. 118
6.2.3 Experimental setup.................................................................................. 118
6.2.4 Experimental procedure .......................................................................... 119
6.2.5 Experimental methods............................................................................. 119
6.3 Results and Discussion................................................................................... 120
6.4 Conclusion...................................................................................................... 130
7. AN INTEGRATED ANAEROBIC MEMBRANE BIOREACTOR-
DIGESTER SYSTEM TO SUPPORT LOW TEMPERATURE ANAEROBIC
TREATMENT OF MUNICIPAL WASTEWATER .......................................... 133
7.1 Introduction .................................................................................................... 133
7.2 Materials and Methods ................................................................................... 139
7.2.1 Wastewater source .................................................................................. 139
7.2.2 Seed sludge ............................................................................................. 139
7.2.3 Experimental setup.................................................................................. 139
7.2.4 Experimental procedure .......................................................................... 141
7.2.5 Experimental methods............................................................................. 142
7.2.5.1 Analysis............................................................................................ 142
7.2.5.2 Filtration resistance .......................................................................... 143
7.2.5.3 Microbial community analysis ......................................................... 143
7.3 Results and Discussion................................................................................... 143
7.3.1 Treatment performance ........................................................................... 143
7.3.2 Filtration performance............................................................................. 148
7.3.2.1 Flux and TMP .................................................................................. 148
7.3.2.2 TSS and turbidity ............................................................................. 149
7.3.2.3 Particle size distribution................................................................... 150
7.3.2.4 Filtration resistances and cleaning tests ........................................... 152
7.3.3 Microbial community.............................................................................. 153

xiii
 7.4 Conclusion...................................................................................................... 158
8. CONCLUSIONS AND RECOMMENDATIONS ........................................... 161
8.1 Conclusions from the Various Research Steps............................................... 161
8.2 Main Outcomes and Evaluation ..................................................................... 165
8.3 Future Perspectives and Recommendations for Further Research ................. 168
REFERENCES ....................................................................................................... 171
APPENDICES ........................................................................................................ 193
APPENDIX A ...................................................................................................... 194
APPENDIX B....................................................................................................... 196
APPENDIX C....................................................................................................... 199
APPENDIX D ...................................................................................................... 201
CURRICULUM VITAE ........................................................................................ 203

xiv
ABBREVIATIONS

AD : Anaerobic Digestion
AMPTSII : Automated Methane Potential Test System
AnMBR : Anaerobic Membrane Bioreactor
BOD : Biochemical Oxygen Demand
COD : Chemical Oxygen Demand
CSTR : Completely Stirred Tank Reactor
DNA : Deoxyribonucleic Acid
DOC : Dissolved Organic Carbon
EGSB : Expanded Granular Sludge Bed
EPS : Extracellular Polymeric Substances
F/M : Food To Microorganism Ratio
FB : Fluidized Bed
FO : Forward Osmosis
GAC : Granular Activated Carbon
GLS : Gas-Liquid-Solids
HRT : Hydraulic Retention Time
MBR : Membrane Bioreactor
MF : Microfiltration
MLSS : Mixed Liquor Suspended Solids
OLR : Organic Loading Rate
OSN : Observed Species Number
OTU : Operational Taxonomic Unit
P/C : Protein/Carbohydrate Ratio
PAC : Powdered Activated Carbon
PBS : Phosphate Buffered Saline
PEI : Polyetherimide
PES : Polyethersulfone
PSD : Particle Size Distribution
PTFE : Polytetrafluoroethylene
PVDF : Polyvinylidene Fluoride
PVP : Polyvinylpyrrolidone
QIIME : Quantitative Insights Into Microbial Ecology
rDNA : Recombinant Deoxyribonucleic Acid
RDP : Ribosomal Database Project
RO : Reverse Osmosis
SMA : Specific Methanogenic Activity
SMP : Soluble Microbial Products
SRT : Solids Retention Time
SS : Suspended Solids
SUR : Specific Ultrafiltration Resistance
TMP : Transmembrane Pressure
TN : Total Nitrogen
TOC : Total Organic Carbon

xv
TP : Total Phosphorus
TS : Total Solids
TSS : Total Suspended Solids
UASB : Upflow Anaerobic Sludge Blanket
UF : Ultrafiltration
VFA : Volatile Fatty Acid
VS : Volatile Solids
VSEP : Vibratory Shear Enhanced Processing
VSS : Volatile Suspended Solids

xvi
SYMBOLS

CO2 : Carbon Dioxide

D50 : Median Particle Size
J : Permeate Flux
NaCl : Sodium Chloride
NaHCO3 : Sodium Bicarbonate
NaOCl : Sodium Hypochlorite
NH4-N : Ammonium Nitrogen
P : Membrane Permeability
PO43- : Phosphate
Rintrinsic : Intrinsic Resistance
Rirrecoverable : Irrecoverable Resistance
Rirreversible : Irreversible Resistance
Rremovable : Removable Resistance
RT : Total Filtration Resistance
η : Dynamic Viscosity Of Water

xvii
xviii
LIST OF TABLES

Page

Table 2.1 : Treatment performance of membrane coupled conventional anaerobic

treatment processes for municipal wastewater treatment........................ 16
Table 2.2 : Membrane performance of membrane coupled conventional anaerobic
treatment processes for municipal wastewater treatment........................ 18
Table 3.1 : Composition of the concentrated synthetic municipal wastewater....... 52
Table 3.2 : Characterization of the synthetic municipal wastewater at the UASB
inlet.......................................................................................................... 52
Table 3.3 : Characteristics of the seed sludge ......................................................... 53
Table 4.1 : VFA concentrations in the effluent and permeate before and after
membrane incorporation ......................................................................... 80
Table 4.2 : SMP and EPS amounts of the sludge in the UASB reactor .................. 85
Table 4.3 : SMA and stability of sludge in the UASB reactor before and after
membrane incorporation ......................................................................... 86
Table 4.4 : Microbial diversity index...................................................................... 87
Table 5.1 : Specifications of the membrane............................................................ 96
Table 5.2 : Total filtration resistances................................................................... 112
Table 6.1 : SMP and EPS contents of the sludge samples along the membrane
coupled UASB reactor height ............................................................... 124
Table 7.1 : Operational conditions of the AMBR-Digester system ...................... 142
Table 7.2 : SMA and stability of the AnMBR and the digester sludge ................ 146
Table 7.3 : Average biogas production rate in the UASB and the digester .......... 147
Table 7.4 : Influent and permeate characteristics in terms of COD, TSS, TN, NH4+
and TP during anaerobic municipal wastewater treatment in the single
stage AnMBR and the AnMBR-Digester system ................................. 148
Table 7.5 : Total filtration resistances................................................................... 152
Table A.1 : The dominant bacterial species detected by pyrosequencing in the
inoculum and sludge samples before and after membrane addition ..... 194
Table A.2 : The all archaea detected by pyrosequencing in the inoculum and sludge
samples before and after membrane addition........................................ 195
Table B.1 : The dominant bacterial species detected by pyrosequencing in the
inoculum and sludge samples at different operational temperatures .... 196
Table B.2 : The all archaea detected by pyrosequencing in the inoculum and sludge
samples at different operational temperatures....................................... 198
Table C.1 : The dominant bacterial species detected by pyrosequencing in the
sludge samples along the UASB reactor ............................................... 199
Table C.2 : The all archaea detected by pyrosequencing in the sludge samples along
the UASB reactor. ................................................................................. 200
Table D.1 : The dominant bacterial species detected by pyrosequencing in the
sludge samples from the UASB and digester during AnMBR, AnMBR-
Digester (UASB Sampling Point: S3), AnMBR-Digester (UASB
Sampling Point: S2) operation .............................................................. 201

xix
Table D.2 : The all archaea detected by pyrosequencing in the sludge samples from
the UASB and digester during AnMBR, AnMBR-Digester (UASB
Sampling Point: S3), AnMBR-Digester (UASB Sampling Point: S2)
operation ................................................................................................ 202

xx
LIST OF FIGURES

Page

Figure 1.1 : The schematic representation of the AnMBR-Digester system.............. 3

Figure 1.2 : Schematic representation of the research approach and thesis outline ... 9
Figure 2.1 : Membrane filtration used as a post-treatment step after UASB reactor22
Figure 2.2 : Alternative integrated flow schemes of AnMBRs ................................ 27
Figure 3.1 : Schematic diagram of the UASB reactor.............................................. 54
Figure 3.2 : A three-phase separator......................................................................... 54
Figure 3.3 : Schematic diagram of the SUR measurement experimental setup ....... 56
Figure 3.4 : Total and soluble COD concentrations in the effluent.......................... 57
Figure 3.5 : Biogas production at different upflow velocities.................................. 58
Figure 3.6 : TSS concentration in the supernatant and effluent ............................... 59
Figure 3.7 : Turbidity of the effluent........................................................................ 59
Figure 3.8 : TN and NH4-N concentrations in the effluent ...................................... 60
Figure 3.9 : TP concentrations in the effluent .......................................................... 60
Figure 3.10: SMP concentrations in the effluent ...................................................... 61
Figure 3.11: PSD at different upflow velocities (a) v=1.2 m/h, (b) v=0.6 m/h, (c)
v=1.2 m/h ................................................................................................ 63
Figure 3.12:SUR and ΔR/Δt values at different upflow velocities
................................................................................................................. 64
Figure 4.1 : The layouts of the experimental setups (a) UASB reactor and (b)
AnMBR system ....................................................................................... 72
Figure 4.2 : AMPTSII............................................................................................... 74
Figure 4.3 : Total and soluble COD concentrations in the UASB effluent and final
permeate .................................................................................................. 77
Figure 4.4 : TSS concentrations in the supernatant and effluent.............................. 78
Figure 4.5 : SMP concentrations in the UASB effluent and final permeate ............ 79
Figure 4.6 : Filtration performance of the AnMBR system ..................................... 82
Figure 4.7 : PSD of the seed and UASB reactor sludge (a) before and (b) after
membrane incorporation ......................................................................... 84
Figure 4.8 : Sludge settleability before and after membrane incorporation ............. 85
Figure 4.9 : Relative abundance of major bacterial genera ...................................... 88
Figure 4.10: Relative abundance of all archaeal species .......................................... 89
Figure 5.1 : Layout of the AnMBR, consisting of a UASB coupled UF unit .......... 95
Figure 5.2 : Images of the UF membrane (a) Cross-sectional (b) Permeate outlet .. 96
Figure 5.3 : Total and soluble COD concentrations in the UASB effluent and
permeate ................................................................................................ 100
Figure 5.4 : TSS concentration and turbidity in the effluent .................................. 101
Figure 5.5 : PSD of the UASB effluent at different sampling days ....................... 102
Figure 5.6 : SMP concentrations in the effluent and permeate .............................. 103
Figure 5.7 : Microbial diversity index.................................................................... 104
Figure 5.8 : Relative abundance of bacterial phyla ................................................ 105
Figure 5.9 : Relative abundance of major bacterial species ................................... 107

xxi
Figure 5.10: Relative abundance of all archaeal species......................................... 108
Figure 5.11: Relative abundance of all archaeal species......................................... 109
Figure 5.12: Filtration performance of the AnMBR system ................................... 111
Figure 5.13: Contribution of different resistances to RTs at the two operational
temperatures .......................................................................................... 113
Figure 6.1 : Schematic overview of the membrane coupled-UASB AnMBR and the
sludge sampling ports S1-S4 over the height of the UASB reactor ...... 119
Figure 6.2 : TS and VS concentrations and VS/TS along the height of the membrane
coupled UASB reactor........................................................................... 121
Figure 6.3 : PSD of the sludge samples in terms of (a) particle number and (b)
particle volume along the membrane coupled UASB reactor ............... 122
Figure 6.4 : Morphological changes of sludge samples along the UASB reactor.. 123
Figure 6.5 : SMA and stability of sludge samples along the height of the membrane
coupled UASB reactor........................................................................... 125
Figure 6.6 : Alpha-diversity index of each sludge sample along the UASB reactor
............................................................................................................... 126
Figure 6.7 : Lorenz distribution curves of each sludge sample along the UASB
reactor.................................................................................................... 127
Figure 6.8 : Relative abundance of major bacterial genera .................................... 128
Figure 6.9 : Soluble COD concentration along the height of the membrane coupled
UASB reactor ........................................................................................ 129
Figure 6.10: Relative abundance of all archaeal species......................................... 130
Figure 7.1 : Schematic diagram of the AnMBR-Digester system.......................... 140
Figure 7.2 : Total and soluble COD concentrations in the UASB effluent and final
permeate during three experimental stages ........................................... 144
Figure 7.3 : Filtration performance of a single stage AnMBR and an AnMBR-
Digester system ..................................................................................... 149
Figure 7.4 : TSS concentration and turbidity in the UASB effluent of a single stage
AnMBR and an AnMBR-Digester system ............................................ 150
Figure 7.5 : PSD of the UASB effluent during anaerobic municipal wastewater
treatment at different operational stages (a) AnMBR, (b) AnMBR-
Digester (UASB Sampling Point: S3), (c) AnMBR-Digester (UASB
Sampling Point: S2)............................................................................... 151
Figure 7.6 : Contribution of different resistances to RT at different operational stages
(a) AnMBR, (b) AnMBR-Digester (UASB Sampling Point: S3), (c)
AnMBR-Digester (UASB Sampling Point: S2). ................................... 153
Figure 7.7 : Alpha-diversity of the biomass samples collected from the UASB and
digester during AnMBR, AnMBR-Digester (UASB Sampling Point: S3),
AnMBR-Digester (UASB Sampling Point: S2) operation.................... 154
Figure 7.8 : Relative abundance of major bacterial species in the digester and UASB
during AnMBR, AnMBR-Digester (UASB Sampling Point: S3),
AnMBR-Digester (UASB Sampling Point: S2) operation.................... 157
Figure 7.9 : Relative abundance of all archaeal species in the digester and UASB
during AnMBR, AnMBR-Digester (UASB Sampling Point: S3),
AnMBR-Digester (UASB Sampling Point: S2) operation.................... 158

xxii
 ANAEROBIC MEMBRANE BIOREACTORS FOR COST-EFFECTIVE
MUNICIPAL WATER REUSE

SUMMARY

In recent years, anaerobic membrane bioreactor (AnMBR) technology has been

increasingly researched for municipal wastewater treatment as a means to produce
nutrient-rich, solids free effluents with low levels of pathogens, while occupying a
small footprint. An AnMBR can be used not only for on-site wastewater treatment,
but also for the generation of nutrient-rich irrigation water leading to reuse and
recycling possibility for agricultural applications as well. Furthermore, biogas
produced in the anaerobic process could potentially be used for minimising the
energy requirements of AnMBR operation. Despite the mentioned advantages, the
current state-of-the-art AnMBR technology for potential full-scale application raises
some concerns related to energy requirements and investment costs for membrane
fouling control, as well as the impact of high shear stress on biomass activity.
Municipal wastewater in many countries can be characterised as low strength and is
generated at high flows rates, related to the population served. However, ongoing
efforts in water conservation and the implementation of source separated sewer
systems may possibly change municipal wastewater characteristics in the near future.
Several factors involved in design and operation of AnMBR systems may influence
membrane fouling, such as treatment plant configuration and membrane
characteristics, feeding and biomass properties and operational conditions. Among
them, reactor design is proposed as an important factor that can change the AnMBR
sludge fouling propensity. To date, especially in the last decade, many studies have
been conducted in which various types of anaerobic reactors, including completely
stirred tank reactor (CSTR), upflow anaerobic sludge blanket (UASB) reactor,
expanded granular sludge bed reactor (EGSB), were used in combination with
various types of membranes. Among them, a membrane coupled UASB reactor may
be a promising approach to overcome problems related with fouling since the
membrane is only subjected to supernatant filtration and not by bulk sludge.
The purpose of this thesis was to investigate the applicability of an innovative
AnMBR configuration in order to produce a pathogen-free but nutrient-rich effluent
for use in agricultural irrigation, concomitantly enabling energy recovery. With the
realization of this study, flux enhancement by controlling the total solids (TS) load to
the membrane was the major starting point, bringing AnMBR technology for full-
scale sewage treatment one step closer to realisation.
The AnMBR-Digester system as developed in this study may offer a solution for the
mentioned challenges. The system consisted of a membrane integrated UASB reactor
coupled to a parallel operating digester that can be operated at any required process
temperature between 0-40 °C. The external cross-flow tubular membrane module
contained 28 membrane fibers with a very small internal tube diameter of 1.5 mm in
order to increase the filtration area for accommodating large sewage flows. Blocking
xxiii
of the membrane module lumens is minimised by the preceding UASB reactor,
which scavenges the large part of suspended solids (SS), having only the supernatant
subjected to membrane filtration. Besides that, the combination of a UASB reactor
and a sludge digester, the so-called UASB-Digester system, has been shown to be
successful for mutual sewage treatment at low temperature and sludge stabilization.
Therefore, membrane integration to the UASB-Digester system can be attractive for
producing high quality and nutrient-rich effluents for reuse purposes under moderate
climate conditions.
Based on this general aim, the optimum upflow velocity that will result in an effluent
with good filterability is determined in a laboratory-scale UASB reactor for the case
of membrane coupled UASB reactor systems. Filterability tests were carried out in
order to assess the effect of upflow velocity on subsequent membrane performance.
Results indicated a significant impact of upflow velocity on both biological
performance and physicochemical effluent characteristics. Operation at a higher
upflow velocity caused the washout of colloidal particles. Effluent characterization
results coincided with filterability tests. Results showed that filterability of the
effluent during the operation at 0.6 m/h was better than that during the operation at
1.2 m/h. The observed differences in protein/carbohydrate ratio (P/C) and particle
size distribution (PSD), which play important roles in membrane fouling, lead to the
hypothesis that upflow velocity is a critical parameter for effluent filterability in
membrane coupled UASB reactors.
In subsequent experiments, the membrane was integrated into the UASB reactor and
the system was operated as an AnMBR system with an upflow velocity of 0.6 m/h in
order to understand the impact of membrane addition on both the biological
performance and sludge characteristics. Membrane incorporation induced an
accumulation of fine particles and a decrease in extracellular polymeric substances
(EPS), resulting in a decrease in PSD and thus, a drop in sludge settleability.
Deterioration of sludge settleability led to an increase in sludge washout, with a
resultant increase in chemical oxygen demand (COD) and total suspended solids
(TSS) concentrations in the UASB effluent. However, SS-free permeate with an
average COD of 42 mg/L was obtained and despite the sludge bed deterioration, the
average transmembrane pressure (TMP) value was 85 mbar during reactor
operation, indicating that no severe membrane fouling occurred in the AnMBR.
Following that investigation, the AnMBR was operated at 15 °C in order to
investigate whether the membrane coupled UASB is a technically feasible alternative
in the treatment of municipal wastewater at lower temperatures. The results showed
that membrane fouling at 15 °C was more severe than at 25 °C. Increased COD and
soluble microbial products (SMP) concentrations, reduced particle’s diameter, and
higher turbidity in the UASB reactor effluent at lower temperature aggravated
membrane fouling. However, treatment performance was not considerably affected
by temperature possibly due to the retention capacity of membrane. Cake resistance
was found responsible for over 40% of the total fouling at both temperatures.
However, an increase was observed in the contribution of irreversible fouling
resistance to the total filtration resistance (R T) at lower temperature, related to the
larger amount of fine particles in the UASB reactor effluent.
During the operation at 15 °C, differences were observed in sludge characteristics at
different heights along the UASB reactor. The best location in the sludge bed for
conveying the sludge from the UASB reactor to the digester needed to be

xxiv
determined. Analysis over the height of the reactor with time showed that TS,
volatile solids (VS), TSS and volatile suspended solids (VSS) concentrations in the
reactor decreased with height, and highest COD concentration of 46 g/L was
observed at the bottom of the reactor. The active biomass remained near the inlet of
the reactor; whereas, non-active biomass consisted of loose, suspended particles and
flocculents moved towards the top. This was confirmed by the high specific COD
consumption rate near the inlet and the poor specific COD biodegradation in the
remaining portions of the bioreactor. Apparently, the assumption of a completely
mixed sludge bed behavior for the UASB reactor, being part of an AnMBR system,
does not hold for this type of reactor systems even at low temperatures, which makes
the location in the sludge bed from where the sludge is to be conveyed to the digester
of operational importance. Considering the observed sludge bed stratification with
regard to sludge stability and solids concentration, the sludge to be recirculated from
the UASB reactor to the digester is recommended to be extracted from the sampling
point where the sludge has the lowest stability. A low value of stability coincides
with a high amount of anaerobic biodegradable organic compounds present in the
sludge.
Finally, the impact of digester coupling on removal efficiency and filtration
performance of the AnMBR was investigated. Digester incorporation remarkably
influenced the characteristics of the UASB reactor effluent, resulting in a decrease in
turbidity, soluble and total COD and an increase in median particle size (D50), which
led to a substantial decrease in the RT. Improved stability and specific methanogenic
activity (SMA) of the sludge were achieved in the UASB reactor with an increase in
total biogas production of the AnMBR-Digester system. The sludge recirculation
from the sampling point having the lowest stability in the AnMBR-Digester system
improved both the solids physical removal and the conversion, which confirms the
importance of sludge transfer point selection, especially for the filtration
performance of AnMBR-Digester systems.
Overall, it can be concluded that the UASB reactor is a suitable alternative for
coupling membranes in AnMBR systems at 25 °C. Despite the sludge bed
deterioration, SS-free permeate with an average COD of 42 mg/L was obtained and
the average TMP of 85 mbar was maintained during the operation period, indicating
no severe membrane fouling. However, this configuration is not found technically
feasible at 15 °C, considering the deterioration of the filtration performance, which
would be a bottleneck to the practical engineering application of AnMBRs. Thus, the
filtration performance of the investigated single stage AnMBR was limited by the
low temperature. In that situation, the results revealed the high potential of the new
AnMBR-Digester configuration for treating municipal wastewater at 15 °C, as it
couples wastewater treatment and sludge stabilization. Under optimized sludge
recirculation conditions, the integrated AnMBR-Digester system represents an
efficient technology to mitigate rapid membrane fouling for the low temperature
applications of membrane coupled UASB reactors.

xxv
xxvi
 EVSEL ATIKSULARIN ANAEROBİK MEMBRAN BİYOREAKTÖRLER
İLE MALİYET ETKİN OLARAK YENİDEN KULLANIMI

ÖZET

Evsel atıksudan anaerobik membran biyoreaktör teknolojisi kullanılarak yüksek besi

madde, düşük patojen içeriği olan ve katı madde içermeyen su eldesi günümüzde
yoğun olarak araştırılmaktadır. Anaerobik membran biyoreaktörler diğer
teknolojilere göre daha az yer kaplamakta olup, sadece atıksu arıtma amaçlı değil,
aynı zamanda besi madde içeriği yüksek su eldesi amacıyla da kullanılmaktadır.
Böylece evsel atıksuların tarımsal sulama amaçlı yeniden kullanımına olanak
sağlanmaktadır. Ayrıca anaerobik prosesler ile elde edilen biyogaz anaerobik
membran biyoreaktör işletiminde gerekli olan enerji ihtiyacını azaltmaktadır.
Belirtilen avantajlarının dışında, anaerobik membran biyoreaktör teknolojisi tam
ölçekli uygulamalarda membran tıkanma problemine yol açmaktadır. Tıkanmaya
bağlı olarak oluşan yüksek enerji gereksinimi ve yatırım maliyeti bu teknolojinin
dezavantajları arasında sıralanabilir. Ayrıca anaerobik membran biyoreaktörlerde
biyokütle üzerinde yüksek kesme kuvveti söz konusu olmaktadır. Evsel atıksular
birçok ülkede düşük kuvvetli ve hizmet edilen nüfusa bağlı olarak yüksek debide
oluşmaktadır. Ancak su korunumu ve ayrık kanalizasyon sistemlerinin kurulması ile
ilgili devam eden çalışmalar yakın gelecekte evsel atıksu karakterizasyonunda
değişimlerin olabileceğini göstermektedir.
Arıtma tesisi konfigürasyonu, membran özellikleri, çamur ve giriş atıksu
karakterizasyonu ve işletme koşulları gibi tasarım ve işletme ile ilgili birçok faktör
anaerobik membran biyoreaktörlerde tıkanma potansiyeline etki etmektedir. Bunların
arasında, reaktör dizaynı çamurun tıkama potansiyelini etkileyen en önemli
faktörlerden biridir. Özellikle son on yılda tam karışımlı reaktör, yukarı akışlı
havasız çamur yataklı reaktör ve genleşmiş granüler çamur yataklı reaktörler gibi
farklı birçok anaerobik reaktör tipinin farklı membran türleri ile birlikte kullanıldığı
pek çok çalışma yapılmıştır. Bu alternatifler arasında yukarı akışlı havasız çamur
yataklı reaktörlerde çamur yerine reaktör üst suyu membrana ulaşmaktadır.
Dolayısıyla, tıkanma ile ilgili problemleri azaltması açısından bu kombinasyon
avantajlı görülmektedir.
Bu çalışmanın amacı evsel atıksudan tarımsal sulamada kullanma amaçlı düşük
patojen yüksek besi madde içeriği olan su eldesini ve enerji geri kazanımını
sağlayacak yenilikçi bir anaerobik membran biyoreaktör konfigürasyonunun
uygulanabilirliğinin araştırılmasıdır. Bu çalışmanın gerçekleştirilmesi ile membrana
ulaşacak olan toplam katı madde yükü kontrol edilerek akı artışı sağlanacak ve
anaerobik membran biyoreaktörlerin tam ölçekli evsel atıksu arıtımında kullanımına
yönelik önemli bir adım atılmış olacaktır.

xxvii
Anaerobik membran biyoreaktör ve çürütücü içeren sistem evsel atıksu arıtımında
farklı sistemlerde karşılaşılabilecek birçok dezavantajı önleyebilecek potansiyeldedir.
Önerilen sistemde; yan akım membran ünitesi, yukarı akışlı havasız çamur yataklı
reaktör ve buna paralel olarak çalışacak çürütücü kullanılacaktır. Çürütücü 0 ile 40
°C arasında istenilen sıcaklıkta işletilebilir. Sistemde kullanılacak olan membran
modülü tübüler olarak dizayn edilmiş olup, modül içerisinde 28 adet ince fiber
membran bulunmaktadır. Membran fiberlerinin çapı yüksek debili evsel atıksular
için filtrasyon alanını arttırma amaçlı oldukça düşük olacak şekilde 1.5 mm
seçilmiştir. Membran öncesinde yukarı akışlı havasız çamur yataklı reaktör
kullanılması büyük katı maddelerin reaktörde tutularak sadece reaktör üst sıvısının
membrana ulaşmasını sağlayacak; böylece modül girişinde tıkanma önlenmiş
olacaktır. Bunun yanısıra, yukarı akışlı havasız çamur yataklı reaktör ve çürütücü
kombinasyonu düşük sıcaklıklarda evsel atıksu arıtımı ve çamur stabilizasyonu
açısından ideal bir sistem olarak öne sürülmektedir. Bu kombinasyona membran
ilavesinin yapılması ile özellikle ılıman iklimli bölgelerde evsel atıksuyun tarımsal
sulama amaçlı yeniden kullanımı için yüksek kalitede besi madde içeriği yüksek su
elde edilmesi hedeflenmektedir.
Bu genel amaç doğrultusunda tezin ilk aşamasında laboratuvar ölçekli yukarı akışlı
havasız çamur yataklı reaktörde çıkış akımında en yüksek filtre edilebilirliği
sağlayacak optimum yukarı akış hızının belirlenmesi hedeflenmiştir. Yukarı akışlı
havasız çamur yataklı reaktör farklı yukarı akış hızlarında işletilmiş, bu hızların çıkış
suyu üzerindeki etkileri araştırılmıştır. Yukarı akış hızının membran performansı
üzerindeki etkisinin incelenebilmesi için filtre edilebilirlik testleri
gerçekleştirilmiştir. Sonuçlar; yukarı akış hızının sistemin biyolojik performansı ve
çıkış suyunun fizikokimyasal özellikleri üzerinde oldukça büyük etkisinin olduğunu
göstermektedir. Yüksek yukarı akış hızı kolloidal maddelerin sistemden kaçısına
neden olmaktadır. Yukarı akışlı havasız çamur yataklı reaktörde çıkış suyu
karakteristiği filtre edilebilirlik testleri ile parallel sonuçlar vermiştir. Sonuçlar çıkış
suyunun filtre edilebilirliğinin 0.6 m/sa yukarı akış hızında, 1.2 m/sa yukarı akış
hızına göre daha yüksek olduğunu göstermektedir. Yukarı akışlı havasız çamur
yataklı reaktör çıkış suyunda farklı yukarı akış hızlarında protein/karbondidrat
oranında ve partikül boyut dağılımnda gözlenen büyük değişimler yukarı akış hızının
çıkış suyu filtre edilebilirliği açısından anaerobik membran biyoreaktörlerde oldukça
kritik bir parametre olacağını göstermektedir.
Çalışmanın ikinci aşamasında; membran yukarı akışlı havasız çamur yataklı reaktöre
entegre edilmiş ve sistem bir önceki çalışmada seçilen 0.6 m/sa’lik yukarı akış
hızında anaerobik membran biyoreaktör olarak işletilmiştir. Bu çalışmadaki amaç,
membran ilavesinin sistemin biyolojik performansı ve çamur karakterizasyonu
üzerindeki etkisini incelemektir. Membran ilavesi ile küçük partiküllerin sistemde
birikmesi ve hücredışı polimerik maddelerin azalması sonucu çamurun partikül boyut
dağılımında azalma gözlenmiştir. Çamurun partikül boyut dağılımında gözlenen
azalma çamurun çökelme özelliğinin bozulmasına neden olmuştur. Çamurun
çökelme özelliğinin bozulması ile üst suya biyokütle kaçısı gerçekleşmiş, buna bağlı
olarak yukarı akışlı havasız çamur yataklı reaktör çıkış suyunda kimyasal oksijen
ihtiyacı ve askıda katı madde konsantrasyonlarında artış gözlenmiştir. Çıkış suyunda
gözlenen bu artışa rağmen membran süzüntü akımında kimyasal oksijen ihtiyacı
konsantrasyonu 42 mg/L olup, süzüntü katı madde içermememektedir. Ayrıca
çamurun çökelme özelliğindeki bozulmaya rağmen, membran ilavesi sonrası

xxviii
anaerobik membran biyoreaktörde transmembran basınç değeri 85 mbar civarında
olmuş, ciddi bir membran tıkanma problemi ile karşılaşılmamıştır.
Üçüncü aşamada anaerobik membran biyoreaktör sisteminin düşük sıcaklıkta evsel
atıksu arıtımında uygulanabilirliği araştırılmıştır. Bu amaç doğrultusunda; sistem 15
°C’de işletilmeye başlanmıştır. Sonuçlar membran tıkanma potansiyelinde 15 °C’de,
25 °C’ye göre ciddi seviyede artış olduğunu göstermiştir. Düşük sıcaklıkta yukarı
akışlı havasız çamur yataklı reaktör çıkış suyunda artan kimyasal oksijen ihtiyacı ve
çözünmüş mikrobiyal ürün konsantrasyonu, düşük partikül boyutu ve yüksek
bulanıklık değerleri membran tıkanmasında artışa neden olmuştur. Filtrasyon
performansında gözlenen değişime rağmen, arıtma performansında düşük sıcaklığa
bağlı herhangi bir değişim gözlenmemiştir. Bu durum membranın alıkoyma
kapasitesi ile ilişkilendirilebilir. Her iki sıcaklıkta da keke bağlı membran direnci
toplam direncin %40’ından fazlasını oluşturmuştur. Ancak düşük sıcaklıkta yukarı
akışlı havasız çamur yataklı reaktör çıkış suyunda artan küçük partiküller gözenek
tıkanmasına bağlı oluşan direncin toplam filtrasyon direncine oranında artışa neden
olmuştur.
Anaerobik membran biyoreaktörün 15 °C’de işletilmesi sırasında yukarı akışlı
havasız çamur yataklı reaktör boyunca çamur karakterizasyonunda değişim
gözlenmiştir. Bu durum çürütücü ilave edildiğinde, reaktörler arası çamur
transferinin gerçekleştirileceği noktanın belirlenmesi açısından önemlidir. Reaktör
boyunca farklı yüksekliklerden alınan çamur numunelerinde gerçekleştirilen analizler
toplam ve uçucu katı madde ve toplam ve uçucu askıda katı madde
konsantrasyonlarının reaktör yüksekliğine bağlı azaldığını ve en yüksek kimyasal
oksijen ihtiyacı konsantrasyonunun 46 g/L olacak şekilde reaktörün en alt
seviyesinde ölçüldüğünü göstermektedir. Aktif biyokütle reaktör girişine yakın
bölgelerde yer alırken, aktivitesi düşük gevşek askıda ve flok yapıda partiküller
reaktörün üst kısmına doğru hareket etmiştir. Bu durum reaktör girişine yakın
yerlerdeki yüksek spesifik kimyasal oksijen ihtiyacı tüketim hızı ve reaktörün geri
kalan bölümlerindeki düşük kimyasal oksijen ihtiyacı parçalanma hızı ile
desteklenmektedir. Sonuçlar, düşük sıcaklıkta dahi membran ilavesi sonrası
gerçekleşen çamur çökelme özelliğinde gözlenen bozulmanın, yukarı akışlı havasız
çamur yataklı reaktörü tam karışımlı reaktöre dönüştürmediğini göstermektedir.
Düşük sıcaklık koşullarında reaktör boyunca farklı seviyeler arası değişim devam
etmiştir. Dolayısıyla; çürütücüye transfer edilecek olan çamurun yukarı akışlı havasız
çamur yataklı reaktörün hangi seviyesinden alınacağı işletim açısından kritik bir
karar haline gelmiştir. Çamur yatağı boyunca çamur stabilizasyonu ve katı madde
konsantrasyonuna bağlı analizler gerçekleştirilmiş ve yukarı akışlı havasız çamur
yataklı reaktörden çürütücüye transfer edilecek olan çamurun stabilite değerinin en
düşük olduğu noktadan transfer edilmesine karar verilmiştir. Düşük stabiliteye sahip
çamur yüksek miktarda biyolojik olarak parçalanabilir organik madde içermektedir.
Çalışmanın son aşamasında, çürütücü ilavesinin sistemin arıtma ve filtrasyon
performansı üzerindeki etkisi araştırılmıştır. Çürütücü ilavesi sonrası, yukarı akışlı
havasız çamur yataklı reaktör çıkış suyunda çözünmüş ve toplam kimyasal oksijen
ihtiyacı konsantrasyonlarında ve bulanıklık değerlerinde azalma, partikül boyut
dağılım sonuçlarında ise artış gözlenmiştir. Bu değişimler toplam filtrasyon direncini
önemli ölçüde azaltmıştır. Ayrıca çürütücü ilavesi sonrası yukarı akışlı havasız
çamur yataklı reaktörde çamurun stabilite ve spesifik metanojenik aktivite
değerlerinde artış gözlenmiştir. Sistemin toplam biyogaz miktarında artış
gerçekleşmiştir. En düşük stabiliteye sahip noktadan çamur transferinin yapılması

xxix
sonucu sistemde katı madde giderim ve parçalanma potansiyelinde gözlenen olumlu
yöndeki değişimler çamur transfer noktasının seçiminin özellikle filtrasyon
performansı açısından oldukça kritik olduğunu göstermektedir.
Genel olarak sonuçlar değerlendirildiğinde; yukarı akışlı havasız çamur yataklı
reaktörlerin membran ile entegrasyonunun 25 °C için oldukça uygun olduğu
görülmüştür. Çamur çökelme özelliğinde gözlenen bozulmaya rağmen, girişte
530±30 mg/L olan ortalama kimyasal oksijen ihtiyacı, membran çıkışında 42
mg/L’ye düşmüş ve katı madde içermeyen bir süzüntü elde edilmiştir. Ayrıca
sistemin ortalama transmembran basınç değeri 85 mbar olarak belirlenmiş, ciddi
seviyelerde membran tıkanma problemi gözlenmemiştir. Ancak, sistem 15 °C’de
işletildiğinde, arıtma performansı 25 °C’deki ile aynı olmakla birlikte, filtrasyon
performansında ciddi seviyelerde bozulma gözlenmiştir. Bu durum sistemin
mühendislik uygulamaları açısından kullanılmasını büyük ölçüde engellemektedir.
Özetle, 15 °C’de sistemin filtrasyon performansında düşük sıcaklığa bağlı
bozulmalar belirlenmiştir. Çürütücü ilavesi sonrası gerçekleşen olumlu değişimler
anaerobik membran biyoreaktör ve çürütücü konfigürasyonunun atıksu arıtımı ve
çamur stabilizasyonunu birlikte gerçekleştirerek, 15 °C’de dahi evsel atıksu arıtımı
için yüksek bir potansiyeli olduğunu göstermektedir. Anaerobik membran
biyoreaktör ve çürütücü konfigürasyonu optimum çamur transfer noktası seçilerek
işletildiğinde, düşük sıcaklıklarda yukarı akışlı havasız çamur yataklı reaktör ve
membran entegre sistemlerde gerçekleşen ani membran tıkanma problemi
önlenmektedir.

xxx
1. INTRODUCTION

1.1 Background

Demographic growth, urbanization, higher living standards and technological

advance have led to an unprecedented increase in water demand, not only for
municipal use but also for agricultural and industrial use. In many places fresh water
is not sufficient to meet the water demand. Therefore, reclamation and use of treated
wastewater for various purposes such as urban, irrigational and industrial purposes
are increasingly recognized as essential strategies, especially for those areas suffering
from deficiencies in the water balance, such as various European and Mediterranean
regions. Reuse projects also exist in Japan, USA, Israel, and Australia (Zanetti et al.,
2010; Agrafioti and Diamadopoulos, 2012; Bunani et al., 2013; Tang et al., 2013;
Xiao et al., 2013).

Wastewater treatment and use of effluents have two major advantages including the
reduction of the environment contamination, and hence the health risks, and
conserving the limited freshwater reserves. Taking into account that high water
consumption increases the volume of wastewater generated, municipal wastewater is
one of the main alternative sources for water reclamation and reuse. However, the
presence of chemicals and microbial contaminants in the municipal wastewater
threats ecosystem safety and human health (Saddoud et al., 2006). In consequence,
municipal wastewater should be subjected to appropriate treatment to reduce the
health and ecological risks of reclaimed water (Ma et al., 2013).

Globally a major demand exists for more efficient municipal wastewater treatment
with a high effluent quality and reduced energy consumption, which will lead to cost
effective treatment systems with effluent reuse for irrigation. Throughout different
alternatives, anaerobic treatment systems have major potentials to meet the market
demand with almost no energy requirement meanwhile biogas is produced for
potential use. Anaerobic treatment is an energy generating process, in contrast to
aerobic systems that generally demand a high energy input for aeration purposes. In

1
addition, the lower anaerobic waste biosolids production compared with aerobic
treatment, reduces the costs and difficulties associated with biosolids management.
However, anaerobic sewage treatment systems only partly remove the organic
pollutants, with COD removal efficiencies reaching 70-80% in full-scale reactors
(Chernicharo et al., 2015; van Lier et al, 2010). In addition, they have a limited
feasibility for the treatment of sewage at low temperatures (<15 °C), while nutrient
removal needs to be addressed in subsequent systems.

Among anaerobic treatment technologies, the anaerobic membrane bioreactor

(AnMBR) is a promising technology as a means to retain all the biomass in the
reactor more effectively and achieve highly efficient solids-liquid separation,
producing superior effluent qualities. In addition, AnMBRs appear suitable for the
treatment of wastewaters with high organic suspended solids (SS) content, because
particles are confined inside the reactor, allowing their degradation (Fuchs et al.,
2003; Hu and Stuckey, 2007; Lin et al., 2010; Lin et al., 2011). The AnMBR has
gained popularity in recent years, and more attention has been focused on the
development of this technology for the treatment of high-concentration wastewater,
mainly food and beverage industry wastewaters, or for the treatment of swine manure
(Fakhru’l-Razi and Noor, 1999; Fuchs et al., 2003; He et al., 2005; Zhang et al.,
2007; Meabe et al., 2013). With respect to domestic sewage treatment, the
production of solids-free, nutrient-rich effluents with a low content of pathogenic
organisms is mentioned as a feature of major interest applying AnMBRs (Liao et al.,
2006; Kocadagistan and Topcu, 2007; Ho and Sung, 2010; Gao et al., 2010; Liu et
al., 2012).

1.2 Problem Statement, Objectives and Structure of the Thesis

1.2.1 Problem statement and research approach

Despite continuous developments in the field of AnMBR technology, the state-of-

the-art of this technology is not yet appropriate for municipal wastewater treatment
to reach re-use quality effluents, due to high energy and investment costs and high
shear stress on the biomass. It has been mainly applied for industrial or high strength
wastewater treatment.

2
A radical different AnMBR design is required to facilitate further development and
optimization of AnMBR technology for municipal wastewater treatment. Thus, an
innovative configuration of AnMBR will be investigated in order to solve the
bottlenecks of AnMBRs and enable energy recovery and effluent use for irrigation
(Figure 1.1). The system consists of an upflow anaerobic sludge blanket (UASB)
reactor coupled with a parallel-operated digester, which can be operated at any
required process temperature. Membrane integration with the UASB-Digester system
can be attractive for producing high quality and nutrient-rich effluents for reuse
purposes under moderate climate conditions.

The UASB within this concept improves primary sedimentation, sludge thickening
and methanogenesis. A UASB reactor coupled to a membrane is also useful for
decreasing the SS concentration of the effluent being sent to the membrane because
the sludge bed would entrap 70-80% of the particulate matter by adsorption and
biodegradation (Kataoka et al., 1992; An et al., 2009a; Wu et al., 2009; Zhang et al.,
2011). Most of the sludge is kept in the reactor instead of circulated across the
membrane. Thus, the membrane flux may become less dependent on the reactor
mixed liquor suspended solids (MLSS) concentration, possibly leading to increased
membrane fluxes. Using a UASB instead of a completely stirred tank reactor (CSTR)
reduces both the shear on the biomass and the required membrane surface and thus
the costs. Most of the sludge will be retained in the sludge bed of the UASB reactor,
so there is no shear impact on the sludge mass due to membrane filtration. It is
anticipated to feed the membranes with a maximum of 1 g/L total solids (TS).

Figure 1.1 : The schematic representation of the AnMBR-Digester system.

3
Municipal wastewater in many countries is generated at high per capita flow rates.
For treating these high flows large membrane surface areas are required, increasing
the capital costs of the membrane system. External cross-flow tubular membrane
modules that contain membrane fibers with a very small internal tube diameter of
only 1.5 mm are one hand very compact and on the other hand characterised by a
large membrane surface area and thus in principle very suitable to treat the high
sewage flows. Moreover, if the filtration step is preceded by a UASB reactor, then
TS level in the membrane unit is drastically reduced, which will reduce the clogging
and fouling potential of the membrane feed. In fact, the reduced clogging potential
enables the use of smaller membrane diameters and a simpler membrane process.

Since municipal wastewaters are often characterised by low temperature, membrane

integrated UASB systems must be adapted to overcome these temperature
limitations. The performance of UASB reactors at low temperature climates (5-20
°C) is highly limited by the hydrolysis of entrapped solids that accumulate in the
sludge bed, reducing the overall solids retention time (SRT). A low SRT will result
in a low sludge stability and thus, a low specific methanogenic activity (SMA),
resulting in a poor soluble chemical oxygen demand (COD) removal. In the proposed
system, a digester will be coupled to the membrane integrated UASB system. The
solids that are entrapped in the sludge bed of the UASB reactor, are conveyed to the
digester, which is operated at optimal temperature conditions, i.e. 35 °C. The solids
are then further degraded, and the methanogens-enriched digested sludge is re-
circulated to the UASB reactor, thereby providing additional methanogenic capacity
to improve soluble COD removal at low temperatures (Mahmoud et al., 2004;
Mahmoud, 2008; Zhang et al., 2013). Thus, the digester within this concept is
expected to improve sludge stabilization by methanation, as well as the sludge
dewatering characteristics, resulting in energy production and a continuous
inoculation of the UASB reactor with methanogens-enriched biomass. The
characteristics of the sludge that needs to be recirculated are crucial to the viability of
the AnMBR-Digester system. Optimization of the sludge recirculation needs to be
done carefully in AnMBR-Digester systems in order to maintain complete
conversion of biodegradable dissolved COD (Mahmoud, 2002; Zhang et al., 2012).

Overall, the investigated system will be a cost-efective technology with reduced

investment and operational costs and energy consumption. No energy will be

4
required for aerating the bioreactor and this will be a major step towards a positive
net energy balance. Compared to conventional anaerobic treatment, the treated water
is free of SS, has a log 4-5 lower levels of pathogenic organisms and has low
concentrations of soluble COD. With ultrafiltration (UF), the effluent will be
virtually free of pathogens, so it can be reused for unrestricted irrigation purposes.
The envisaged technology is further characterised by extremely low energy
consumption rates and biogas production (Liao et al., 2006; Kocadagistan and
Topcu, 2007; Ho and Sung, 2010). The envisaged system could be further optimised
towards net energy producing wastewater treatment plant.

1.2.2 Research objectives

In this study, the applicability of an innovative configuration of AnMBR will be

investigated in order to produce pathogen-free and nutrient-rich effluents for use in
irrigation, meanwhile energy is recovered. With the realization of this study, flux
enhancement in AnMBR technology by controlling the total suspended solids (TSS)
load to the membrane unit will be the major output of the study, bringing AnMBR
technology for full-scale sewage treatment one step closer to realisation. Based on
this general aim, specific aims were identified as follows:

 Identify the optimum upflow velocity that will result in an effluent with good
filterability for the case of membrane coupled UASB systems

 Understand the impact of membrane addition on both the biological

performance and sludge characteristics of a UASB reactor in an AnMBR
system treating municipal wastewater

 Investigate the relationship between the characteristics of the UASB effluent

and the filtration performance of the AnMBR at different operational
temperatures

 Identify the best location in the sludge bed for conveying the sludge from a
low temperature UASB reactor to the digester

 Investigate the impact of the digester coupling on the removal efficiency and
filtration performance of the AnMBR system

5
1.2.3 Thesis outline

The objectives mentioned above have been addressed in eight chapters and the
chapters are structured as follows (Figure 1.2):

Chapter 2 critically evaluates the potential of AnMBR technology for municipal

wastewater treatment with a focus on different types of anaerobic reactors that are
coupled with membranes. Various types of membrane coupled anaerobic sludge bed
reactors, such as UASB and expanded granular sludge bed (EGSB), are discussed as
an alternative approach to the more conventional AnMBRs consisting of CSTR type
bioreactors with either internal or external membrane units. Current anaerobic
bioreactor membrane process integration alternatives reported in the literature are
evaluated with proposals for future applications. In addition, the chapter discusses
the impact of various factors on both biological and filtration performances of
AnMBRs including strengths and limitations. Finally, possible uses of treated
municipal wastewater are discussed.

Chapter 3 deals with the effect of upflow velocity on biological removal efficiency
and effluent filterability in a laboratory-scale UASB reactor. Upflow velocities of
1.2, 0.6 and 1.2 m/h are applied in three successive stages over a total operation
period of 116 days. The effect of upflow velocity on the effluent quality is assessed
not only on biological based parameters such as organic matter removal but also on
physical characteristics. Filterability tests are carried out during each stage in order to
assess the effect of upflow velocity on subsequent membrane performance. From the
obtained results, an optimum upflow velocity for the UASB unit was identified that
will result in an effluent with good filterability for subsequent membrane filtration.
The information obtained from this study is valuable for optimizing the hydraulic
conditions in membrane coupled UASB systems in order to achieve stable effluent
quality and to minimize membrane fouling potential.

Chapter 4 presents the impact of membrane incorporation for effluent extraction on

the performance of a UASB reactor in terms of system performance and sludge
characteristics. A laboratory-scale system was operated in two different operational
stages: (i) before membrane incorporation as a sole UASB reactor and (ii) after
membrane incorporation as an AnMBR system. The impact is evaluated by
monitoring and comparing several indicators including COD removal efficiency,

6
TSS and soluble microbial products (SMP) concentrations in the UASB effluent and
sludge particle size distribution (PSD), sludge SMA and microbial community
analysis before and after membrane addition. This study is supposed to provide a
fundamental understanding of how a UASB system is affected by the elimination of
selective sludge washout, as well as to give a clue on the operation of a UASB
reactor as part of an AnMBR system.

Chapter 5 focuses on the relationship between the effluent characteristics and

filtration performance of an AnMBR. For this purpose, UF membrane coupled to a
UASB reactor is operated continuously for nearly 126 days at two operational
temperatures (25 °C and 15 °C). The effect of temperature on the overall
performance of the AnMBR system is determined in order to investigate whether the
membrane coupled UASB is a technically feasible alternative in the treatment of
municipal wastewater at 15 °C. Pyrosequencing is used to compare the microbial
community structure including both archaeal and bacterial communities and the
relative abundance of microbial species at different temperatures. The results provide
valuable new insights into the fouling mechanisms and would consequently further
advance knowledge on the behavior of membrane coupled UASB system at low
temperatures.

Chapter 6 provides a comparison of the sludge characteristics along the height of a

UASB reactor in terms of sludge morphology, activity and stability. The main aim of
this study is to identify the best location (i.e. where sludge is of lowest stability
and/or highest concentration) in the sludge bed for conveying the sludge from the
low temperature UASB reactor to the digester. Within this concept, the sludge profile
was investigated by collecting sludge samples along the different heights of the
UASB-AnMBR treating municipal wastewater. The sludge characteristics were
comparatively evaluated in terms of solids content, PSD, sludge morphology, SMA
and stability. Besides, pyrosequencing was employed to analyze samples from each
location in order to compare the microbial community composition including both
archaeal and bacterial communities and the relative abundance of microbial species.
The microbial community compositions elucidated in this study provide new insights
into understanding anaerobic wastewater treatment and shed some light on the
application potentials of UASB reactors in AnMBR concepts. The overall
information obtained from this study is valuable for the selection of the location

7
through which sludge will be transferred between UASB and digester in membrane
coupled UASB systems.

Chapter 7 evaluates the applicability of the proposed configuration. In the study, the
anaerobic sewage treatment using an AnMBR-Digester system was investigated at a
sewage temperature of 15 °C. It provides the impact of digester incorporation on
treatment and filterability performance of the AnMBR. The UASB reactor was firstly
operated without incorporating the digester aiming at monitoring of the overall
AnMBR performance, i.e. (i) treatment performance, (ii) filtration performance and
(iii) sludge quality. This is of particular importance to have base line records to be
used as reference values to assess the achievements obtained from incorporating a
digester. Afterwards, the single-stage UASB reactor was modified to a UASB-
Digester system by incorporating a digester operated at 35 °C in order to assess the
performance of the AnMBR-Digester system. The characteristics of the sludge that is
recirculated are crucial to the viability of the AnMBR-Digester system. Therefore,
the optimum sampling location for sludge transfer is selected. The overall
information obtained from this study is valuable for the operation and improvement
of the AnMBR-Digester system treating municipal wastewater at low temperature.

Chapter 8 concludes the thesis with a general discussion and summarizes the
conclusions, along with directions for future research.

8
 Introduction
Problem Statement and
Research Approach
Chapter 1

Fundamentals
State-of-the-art of
AnMBR Technology for
Municipal Wastewater
Treatment
Chapter 2

Research approach
UASB operation UASB and AnMBR operation
Determination of the Optimum Investigation of the Impact of
Upflow Velocity Membrane Addition
Chapter 3 Chapter 4

AnMBR operation
Investigation of the effect of Low Temperature on the Removal Efficiency and
Effluent Filterability
Chapter 5

AnMBR operation
Determination of the Variation of Sludge Characteristics along the UASB reactor
Chapter 6

Digester coupled AnMBR operation

Determination of the Impact of Digester Addition on the Removal Efficiency and
Effluent Filterability
Chapter 7

Research objective
Investigation of the applicability of an innovative configuration of AnMBR in order
to produce pathogen-free, and nutrient-rich effluent for use in irrigation

Conclusion
Discussion, Conclusion and
Future Perspectives
Chapter 8

Figure 1.2 : Schematic representation of the research approach and thesis outline.

9
10
2. ANAEROBIC MEMBRANE BIOREACTORS FOR MUNICIPAL
WASTEWATER TREATMENT: INTEGRATION OPTIONS, LIMITATIONS
AND EXPECTATIONS(1)

2.1 Introduction

Municipal wastewater is the most abundant type of wastewater that falls into the
category of low-strength waste streams, characterized by low organic strength and
high particulate organic matter content (van Lier, 2008). If recoverable, municipal
wastewater treatment plants have the potential to become net producers of renewable
energy, converting the chemically bound energy content in the organic pollutants of
raw municipal wastewater to useful energy carrier (Shizas and Bagley, 2004).
Therefore, selection of an appropriate energy recovery technology that can convert
the inherent energy in wastewater into a renewable energy source has become more
important.

Anaerobic technology has drawn considerable attention for municipal wastewater

treatment. In addition to the energy that can be recovered from methane-rich biogas,
the application of anaerobic processes distinctly reduces the overall energy demand
for municipal wastewater treatment because no aeration energy is required for
mineralizing the organics. Moreover, anaerobic processes produce mineralized
nutrients in the form of ammonia and orthophosphate enabling direct agricultural use
of the effluent for ferti-irrigation. The most important concern for further improving
the energy sustainability and resource preservation is the identification of an
appropriate anaerobic technology and methods for full recovery of the solubilized
methane from treated effluents (Shizas and Bagley, 2004; Liao et al., 2006).

(1)
This chapter is based on:

Ozgun, H., Dereli, R.K., Ersahin, M.E., Kinaci, C., Spanjers, H., van Lier, J.B. (2013) A review of
anaerobic membrane bioreactors for municipal wastewater treatment: Integration options, limitations
and expectations. Separation and Purification Technology 118, 89-104.

11
The retention of slow growing anaerobic biomass was the most important challenge
in the earlier development of an appropriate reactor technology for the anaerobic
treatment of municipal wastewater. Treatment of municipal wastewater by different
types of anaerobic processes has drawn considerable attention of various researchers
(Singh and Viraraghavan, 1998; Lettinga et al., 2001; Chan et al., 2009). The
invention of the UASB reactor in early 1970s by Lettinga and his colleagues was a
milestone in anaerobic wastewater treatment (Lettinga et al., 1980). The major
success of the UASB reactor lies in its ability to retain a high concentration of
biomass due to formation of a thick dense sludge bed, which, in dependence to
wastewater characteristics, may consist of well settleable methanogenic sludge
granules. Formation of well settleable sludge allows the decoupling of hydraulic
retention time (HRT) and SRT so that efficient treatment can be carried out at high
organic loading rates (OLRs) with a significant decrease in reactor size (Singh and
Viraraghavan, 1998). Owing to the high influent flows and low organic matter
content, municipal wastewater is treated at ambient temperatures in both aerobic and
anaerobic applications worldwide (van Lier et al., 2010), especially in developing
countries with (semi) tropical climates such as Brazil, India and Colombia. However,
municipal wastewater treatment by anaerobic systems in more temperate climates is
still considered a challenge since municipal wastewater belongs to the complex
wastewater category due to the high fraction of particulate organic material,
moderate biodegradability and its low strength (Lettinga et al., 2001). Under low
temperature (<20 °C) conditions, hydrolysis of particulate matter into dissolved
molecules becomes the rate-limiting step, which results in the accumulation of SS in
the reactor and a decrease in organic matter conversion efficiency together with a
decrease in methanogenic activity (van Lier et al., 1997; Kalogo and Verstraete,
1999; Lettinga et al., 2001; Martinez-Sosa et al., 2011a). Moreover, due to low
substrate affinity of anaerobic biomass compared to aerobic bacteria it is practically
very hard to achieve low effluent COD concentrations and to fulfill the
environmental regulations for wastewater reclamation and reuse (Salazar-Pelaez et
al., 2011a).

With ever growing application experiences from aerobic membrane bioreactors

(MBRs) (Santos et al., 2011), AnMBR technology has also started to be researched
for municipal wastewater treatment as a possible alternative to the conventional

12
anaerobic treatment processes (Martin et al., 2011). In AnMBRs, biomass can be
effectively retained inside the reactor providing optimal conditions for organic matter
degradation without any carry-over of SS. By incorporating membranes to anaerobic
municipal wastewater treatment, superior effluent quality in terms of COD, SS and
pathogen counts can be achieved in comparison with conventional anaerobic
processes, and a stable treatment performance can be obtained to meet stringent
discharge standards (Liao et al., 2006; Kocadagistan and Topcu, 2007; An et al.,
2009a; Ho and Sung, 2010). It was reported that AnMBRs provide a possibility for
the agricultural use of the treated effluent for non-potable purposes in many regions
suffering from water shortage (Martinez-Sosa et al., 2011b). Agricultural use of
treated effluents generally demands extensive pathogen removal along with the
availability of macronutrients. Since macronutrients such as ammonium and
orthophosphates are not removed by anaerobic bioprocesses and pathogens can be
retained by the membrane unit (Saddoud et al., 2006; Ellouze et al., 2009), permeates
of AnMBRs are certainly of interest for agricultural use. Information about a selected
number of guidelines on irrigation water quality can be found in the study of Norton-
Brandao et al. (2013). In addition to achieving high effluent qualities, a shorter start-
up period is required for AnMBRs in comparison to UASB systems, which is one of
the major advantages in the treatment of especially low-strength wastewaters. Start-
up periods of 6 and 12 days have been reported in the study of Hu and Stuckey
(2006) and Lin et al. (2011), respectively, whereas this period was in the range of
one to several months for UASB systems (Álvarez et al., 2006a).

Despite the mentioned advantages, there are still critical obstacles such as low flux,
membrane fouling, high capital and operational costs that limit the extensive use of
AnMBRs (Kocadagistan and Topcu, 2007; Hu and Stuckey, 2007; Zhang et al.,
2010). AnMBRs are generally operated at higher biomass concentrations compared
to aerobic MBRs, impacting rheology and thus, reactor hydraulics and pumping.
High MLSS concentrations also result in a more rapid and dense cake layer build up
in comparison to aerobic MBRs, requiring frequent physical cleaning, interval
operation, and likely sub-critical flux operation, in order to sustain the flux (Chang et
al., 2002; Liao et al., 2006; Jeison and van Lier, 2006a; Visvanathan and
Abeynayaka, 2012). Operational costs related to energy requirements for gas/liquid
recirculation for membrane fouling control and chemical costs required for

13
membrane cleaning are still heavy burdens on the economic feasibility of AnMBRs.
However, membrane acquisition and/or replacement costs have decreased
significantly due to a decline in membrane module costs (Judd, 2008; Santos and
Judd, 2010; Santos et al., 2011).

Despite above constraints, AnMBR is identified as a potential anaerobic process of

interest for treating municipal wastewater, whereby, various studies indicate that
treatment performances are seemingly dependent on the chosen process
configuration (Liao et al., 2006). Different types of anaerobic bioreactors, including
CSTR, UASB, EGSB, etc., have been investigated in combination with various types
of membranes. However, the most optimal process configuration, i.e. anaerobic
bioreactor type and the coupling of the bioreactor with the membrane module, yet
needs to be determined. Recently, several more general review papers on AnMBR
systems have been published, discussing the AnMBR feasibility (Smith et al., 2012;
Stuckey, 2012, Skouteris et al., 2012; Lin et al., 2013). This chapter focuses on the
perspectives of the various types of membrane coupled anaerobic sludge bed
reactors, such as UASB and EGSB, as an alternative approach to the more
conventional AnMBRs consisting of CSTR type bioreactors with either internal or
external membrane separation devices. The review chapter evaluates the current
anaerobic bioreactor-membrane process integration alternatives reported in the
literature, with proposals for future applications. Besides, many aspects of current
AnMBR applications for the treatment of municipal wastewater are discussed in
detail including treatability and filterability. Moreover, limitations and difficulties
encountered in anaerobic treatment and the possible use of the treated municipal
wastewater are discussed together with reasonable solution perspectives.

2.2 Integration Possibilities of Membranes with Different Types of Anaerobic

Reactors for Municipal Wastewater Treatment

Membranes can be coupled to various anaerobic reactor types such as CSTRs, UASB
and EGSB reactors, in different configurations for the treatment of municipal
wastewater. Table 2.1 and Table 2.2 present the performance of different AnMBR
applications for the treatment of municipal wastewater with regard to both biological
performance and membrane aspects, respectively. However, information regarding
the membrane performance is quite limited in most of the studies. Section 2.2

14
discusses both the advantages and disadvantages of different integration possibilities
of membranes with several types of anaerobic reactors.

2.2.1 Completely stirred tank reactor

By far, CSTR is the most common anaerobic process researched in AnMBR systems
(Kocadagistan and Topcu, 2007; Hu and Stuckey, 2007; Grundestam and Hellstrom,
2007; Ho et al., 2007; Vyrides and Stuckey, 2009; Ho and Sung, 2010; Martinez-
Sosa et al., 2011a; Gimenez et al., 2011) in analogy to the type of bioreactors used in
aerobic MBRs. In general, CSTRs are operated at equal HRT and SRT without any
internal biomass retention device. In-reactor biomass concentration can be increased
by applying a secondary clarifier with return flow, resulting in an anaerobic contact
process, or a membrane separation device. Simple flow-through CSTRs, without
sludge separation, are characterized by very low loading rates, resulting in long
retention times and, consequently, large reactor volumes. However, in membrane
coupled CSTRs, the complete retention of solids in the reactor decouples SRT from
HRT, which leads to an increase in biomass concentration and thus increase in
conversion rates alleviating the rate limiting step, e.g. hydrolysis/solubilization of
solids and/or methanogenesis. Often CSTRs are coupled to external cross-flow
membranes, resulting in high bioreactor liquid turnover rates, leading to a well-
mixed flow regime. In addition, as a result of the prevailing high shear stress and
intense mixing, an AnMBR setup might even increase the biological methane
potential of the substrate (Liao et al., 2006).

Notwithstanding its advantages, CSTRs directly expose the membrane to bulk

sludge, which results in heavy membrane fouling and low fluxes since effluent solids
concentration of CSTRs is equal to the bulk solids concentration (Liao et al., 2006).
The high solids concentration subjected to membrane filtration exacerbates cake
deposition in CSTR configuration, either with pressure-driven or vacuum-driven
membranes. Moreover, sludge recirculation through the membrane feed pump,
especially for external cross-flow membranes, results in a substantial decrease in the
mean particle size (Choo and Lee, 1998). On one hand, disruption of particles may
positively impact hydrolysis, but on the other hand it may negatively impact the
juxtapositioning of acetogens and methanogens, limiting the required interspecies
hydrogen transfer for attaining a high SMA (Brockmann and Seyfried, 1997).

15
 Table 2.1 : Treatment performance of membrane coupled conventional anaerobic treatment processes for municipal wastewater treatment.
COD Removal
Reactor Type/ Wastewater (based on
Volume Temperature Influent HRT OLR SRT MLSS Biogas Production Reference
Module Source membrane
COD
Configuration effluent)
- (L) (°C) - (mg/L) (h) (kg COD/m3.day) (day) (%) (g/L) (L CH4/g CODremoved) -
Kiriyama et al.
CSTR/Side-stream 8850 25 Raw 200a 89-120 1.4-2b 144 84.5c - -
(1992)
Grundestam
CSTR/Side stream 850 22 Raw 637 14.4 0.94 ∞d 92e - 0.12 and Hellstrom
(2007)
500-
CSTR/Side-stream - 25 Synthetic - - ∞d 97 9.6-16 - Ho et al. (2007)
1000
3, 6, 12, Hu and Stuckey
CSTR/Submerged 3 35 Synthetic 460 - 100 90-95 4.3-4.8 -
24 (2007)
f g Vyrides and
CSTR/Submerged 3 35 Synthetic 465 8-20 - 250 99 2-3 -
Stuckey (2009)
Ho and Sung
CSTR/Side-stream 4 15-25 Synthetic 500 12 1 - >85 - -
(2010)
d Achilli et al.
CSTR/Submerged 6 - Synthetic 425 12 - ∞ 83 4.6 -
(2011)
Gimenez et al.
CSTR/Submerged 900 - Raw 445 6-21 - 70 87 6-22 0.069
(2011)
CSTR/Submerged 60 30 Raw 322 10 1 ∞d 88 6.4-9.3 0.24 Lin et al. (2011)
Municipal+ Martinez-Sosa
CSTR/Submerged 350 35 630 0.8 0.6-1.1 680 90 15 0.27
Glucose et al. (2011a)
Municipal+ Martinez-Sosa
CSTR/Submerged 350 20 630 0.8 0.6-1.1 680 82-90 19 0.23
Glucose et al. (2011a)
Gimenez et al.
CSTR/ Submerged 1300 33 Pre-treated 410 6-21 0.71 76 - - -
(2012)
Gimenez et al.
CSTR/ Submerged 1300 21 Pre-treated 720 6-21 0.64 74 - - -
(2012)
An et al.
Digester/Submerged 12.9 15-20 Raw 259.5 2.6 2.36 - 52-87 - -
(2009b)
Kataoka et al.
UASB/Side-stream 5400 Ambient Pre-hydrolyzed 490 - 2.8b - 83 - -
(1992)
An et al.
UASB/ Side-stream 34 Ambient Raw 185.6 5.5-10 0.3-0.9 ∞d 77-81 12-32 0.062-0.121
(2009a)
Synthetic Wu et al.
UASB/ Side-stream 15.1 35 150h 6 0.3i - - - -
municipal (2009)

16
 Table 2.1 (continued) : Treatment performance of membrane coupled conventional anaerobic treatment processes for municipal wastewater
treatment.
COD Removal
Wastewater (based on
Reactor Type/ Module Volume Temperature Influent HRT OLR SRT MLSS Biogas Production Reference
Source membrane
Configuration COD
effluent)
- (L) (°C) - (mg/L) (h) (kg COD/m3.day) (day) (%) (g/L) (L CH4/g CODremoved) -
5.9- Zhang et al.
UASB/Submerged 45 10-15 Raw 302.1 8 - - 56.6-57.7 -
19.8g (2010)
Gao et al.
UASB/Side-stream 10 30 Synthetic 500 24 5 50 96 - -
(2010)
Zhang et al.
UASB/ Submerged 45 Ambient Raw 298.4 8 - - 63.4 - -
(2011)
Upflow anaerobic d Lew et al.
180 25 Pre-settled 540 12, 6, 4.5 1.08, 2.16, 4.32 ∞ 88 14-80 -
reactor/Side-stream (2009)
3.5, 4.6, Chu et al.
EGSB/Submerged 4.7 11, 15, 20, 25 Synthetic 383-849 1.6-4.5 145 76-96 13-23 -
5.7 (2005)
Pre-hydrolized
Fluidized bed (FB) Kataoka et al.
660 - solids from raw - 1.1b - 90 - -
reactor/- 353 (1992)
sewage
Concentrated
Hydrolyzation reactor Kataoka et al.
500 30 solids from raw 353 5 - - - - -
/Side-stream (1992)
sewage
Hydrolyzation reactor-
membrane-FB
1000 26 Raw 2187 - - - 98 - - Kimura (1991)
bioreactor/
Side-stream
Hydrolyzation reactor
5400 26 Raw 1144 - - - 94 - - Kimura (1991)
UASB/Side-stream
Jet flow anaerobic
Saddoud et al.
bioreactor/ 50 37 Raw 685 15-60 0.23-2 140 88 0.5–10g -
(2007)
Side-stream
Hybrid upflow
97.5- Wen et al.
anaerobic 17.7 Ambient Raw 4-6 0.5-12.5 150 97 16-22.5 0.13-0.42j
2600 (1999)
bioreactor/Submerged
a
mg biochemical oxygen demand (BOD)/L; b kg BOD/m3.day; c Volatile suspended solids (VSS) removal; d No discharge except sampling; e Total organic carbon (TOC) removal; f Dissolved
organic carbon (DOC) removal; g g VSS/L; h mg TOC/L; i g TOC/L.day; j m3/m3.day.

17
 Table 2.2 : Membrane performance of membrane coupled conventional anaerobic treatment processes for municipal wastewater treatment.
Transmembrane
Reactor Type/Module Configuration/Membrane Pore Filtration Flux Cross-flow
Membrane Type Material pressure Gas Sparging Reference
Configuration Size Area velocity
(TMP)
- - - (µm) (m2) (L/m2.h) (kPa) (m/s) (L/min) -
Kiriyama et al.
CSTR/Side-stream/Capillary organic membrane - - - 7.6 - - - -
(1992)
CSTR/Side-stream/Vibratory shear enhanced Polytetrafluoroethylene Grundestam and
- 0.45 1.59 - - - -
processing (VSEP) (PTFE) Teflon Hellstrom (2007)
PTFE laminated non-
CSTR/Side-stream/Tubular - 10 0.015 5-12 3.45-20.7 0.1-0.2 - Ho et al. (2007)
woven filter
Microfiltration Hu and Stuckey
CSTR/Submerged/- - 0.4 - 10-20 up to 30 - -
(MF) (2007)
Vyrides and
CSTR/Submerged/Flat sheet - Polyethylene 0.4 0.1 5-10 30-45 - 5
Stuckey (2009)
Ho and Sung
CSTR/Side-Stream/Tubular MF PTFE 1 0.09 5 6.9-55.2 - -
(2010)
Achilli et al.
CSTR/Submerged/Flat sheet MF Polyolefine 0.4 0.1 10.5 - - 8
(2011)
Gimenez et al.
CSTR/Submerged/Hollow fiber - - 0.05 30 10 8 0.23a
(2011)
Polyvinylidene fluoride
CSTR/Submerged/Flat sheet - 140b 0.6 12 - - 5 Lin et al. (2011)
(PVDF)
Martinez-Sosa et
CSTR/Submerged/Flat sheet UF Polyethersulfone (PES) 0.038 3.5 7 17.7 0.026 62a
al. (2011a)
Gimenez et al.
CSTR/Submerged/Hollow fiber UF - 0.05 30 - - - -
(2012)
Non-woven fabric
Digester/Side-stream/Tubular - (Polythylene 0.64 0.98 5 up to 30 - - An et al. (2009b)
terephthalate)
Kataoka et al.
UASB/External/Capillary UF Polyvinylalcohol 15b 100 - - - -
(1992)
UASB/Side-stream/Tubular - Polyacrylonitrile - 0.2 10.5 <50 0.4-1.45 - An et al. (2009a)
UASB/Side-stream/Flat sheet MF PVDF 0.22 0.05 25 30 - - Wu et al. (2009)
Zhang et al.
UASB/Submerged/Flat type Dynamic Dacron mesh 61 - 65 - -
(2010)
PVDF coated with a
UASB/Side-stream/Flat sheet UF 100b 0.052 8-12 - - - Gao et al. (2010)
Polyether block amide
Dacron mesh (Dynamic Zhang et al.
UASB/Submerged/Flat sheet - 61 - 65 25 - -
membrane) (2011)

18
 Table 2.2 (continued) : Membrane performance of membrane coupled conventional anaerobic treatment processes for municipal wastewater
treatment.
Transmembrane
Reactor Type/Module Configuration/Membrane Pore Filtration Flux Cross-flow
Membrane Type Material pressure Gas Sparging Reference
Configuration Size Area velocity
(TMP)
- - - (µm) (m2) (L/m2.h) (kPa) (m/s) (L/min) -
3.75,
Upflow anaerobic reactor/Side-stream/Hollow
MF - 0.2 4 7.50, <19.6c - - Lew et al. (2009)
fibre
11.25
EGSB/Submerged/ Hollow fibre - Polyethylene 0.1 0.1 - - - - Chu et al. (2005)
Kataoka et al.
FB reactor/External-pressure/Hollow fiber MF Polyethylene 0.1 54 - - - -
(1992)
Kataoka et al.
Hydrolyzation reactor/Internal-pressure/Tubular UF Polyacrylonitrile 13b 0.94 - - - -
(1992)
Hydrolyzation reactor-membrane-FB
MF Polyethylene 0.1 54 24 108 1 - Kimura (1991)
bioreactor/Side-stream/Hollow fiber
Hydrolyzation reactor UASB/Side- Polyvinylalcohol-
- - 100 16 147 0.7 - Kimura (1991)
Stream/Capillary Polysulfone
Saddoud et al.
Jet flow anaerobic bioreactor/Side-stream/- UF - 100b 1 3.5-13 100-200 3 -
(2007)
Hybrid upflow anaerobic bioreactor
- Polyethylene 0.03 0.3 5-10 <70 - - Wen et al. (1999)
/Submerged/Hollow fiber
a 3 2 b c
m /m .h; kDa, molecular weight cut off; Driven by water level difference.

19
Martinez-Sosa et al. (2011a) studied an AnMBR system consisting of an external UF
membrane coupled to a CSTR for the treatment of municipal wastewater and
obtained an effluent that can be used for agricultural irrigation. An innovative
process, i.e. vibrating membranes coupled to a CSTR, was proposed by Grundestam
and Hellstrom (2007) for the treatment of municipal wastewater. They achieved high
TOC removal efficiency of 92%. Authors subsequently applied reverse osmosis (RO)
as post treatment in order to concentrate the nutrients for further use on agricultural
lands. Ho and Sung (2010) reported high COD removal efficiencies for membranes
coupled CSTRs treating synthetic municipal wastewater. Gimenez et al. (2011)
tested hollow fiber membranes in an AnMBR system consisting of a CSTR and
submerged membranes immersed in an external chamber and achieved a COD
removal efficiency of 90% with a flux of 10 L/m2.h.

2.2.2 High-Rate anaerobic reactors

In high-rate anaerobic reactors such as sludge bed systems and anaerobic filters,
biomass is retained either by the formation of granular and/or thick flocculent sludge
or by attachment to a support material. Effluent SS concentration is significantly
lower than the biomass concentration in the reactor, which makes them feasible for
high hydraulic loadings. For instance, sludge bed systems are characterized by TSS
concentrations ranging between 20-40 kg/m3 reactor volume (van Lier et al., 2008).
As in these reactors biomass is not directly subjected to membrane filtration, dense
cake layer formation and consolidation will be less apparent in comparison with
CSTRs when combining these reactors with a membrane module. Therefore, high
rate anaerobic reactors may offer a good opportunity for being combined with
membranes, especially if very low SS concentration is required in the effluent and
retention of biomass is required under presence of toxicity related to discharges of
industrial wastewater in the municipal sewer systems, or hydraulic overload events
(Liao et al., 2006). However, the impact of membrane application on sludge
immobilization and granule stability is not yet well understood.

2.2.2.1 Upflow anaerobic sludge blanket reactor

UASB reactors coupled to membrane separation can be a reasonable option to

decrease the SS concentration being sent to the membrane since the sludge bed
would entrap most of the particulate matter by adsorption and biodegradation

20
(Kataoka et al., 1992; An et al., 2009a; Wu et al., 2009; Zhang et al., 2011). All
biological processes take place inside the dense sludge bed in the bottom of the
UASB reactor (van Lier et al., 2008). The physical removal of particulate organics
by settling, adsorption and entrapment in the sludge bed of the UASB makes it
appropriate for the first treatment step of municipal wastewater, which is generally
characterized by a high particulate COD/soluble COD ratio. In fact, UASB reactors
can be upfront used as biofilters before membrane treatment, which prevents the
membrane from excessive exposure to high SS concentrations. For instance, while a
UASB reactor had a biomass concentration of 20-30 g/L, SS concentration in the
effluent was below 1 g/L (Kleerebezem and Macarie, 2003). An et al. (2009a)
reported TSS concentrations in the range of 11-32 g/L in a UASB reactor; whereas,
TS concentration in the effluent was less than 50 mg/L. Obviously, the efficiency of
solids entrapment determines the amount and properties of solids leaving the UASB
with the effluent. In this context, HRT and upflow velocity in membrane coupled
UASB reactors seem to be the critical parameters determining the efficiency and
effluent fouling propensity. High SRTs and OLRs can be maintained by applying
UASB reactors, without increasing the effluent solids concentration that will be
subjected to membrane filtration. Thus, membrane flux may become less dependent
on the reactor MLSS concentration, possibly leading to high membrane fluxes. There
are many studies that focus on the fouling potential in AnMBR systems including
UASB reactors coupled to membranes. Liao et al. (2006) suggested that membrane
coupled UASB systems may decrease the capital cost of UASB reactors by
eliminating the need for a gas-liquid-solids (GLS) separator in a UASB. Moreover,
organic fouling may be reduced by increasing the SRT which leads to high biomass
concentrations in the reactor and, through increased entrapment in the bed, lower
COD concentrations in the effluent of the UASB reactor (Liao et al., 2006).
Obviously, with increased biogas production the absence of a GLS separator will
likely lead to high sludge carry over to the effluent.

Despite the expected more efficient membrane filtration step, for the treatment of
municipal wastewater under sub-mesophilic conditions, hydrolysis of the retained
particulates likely becomes the rate-limiting step and particulate matter accumulation
in the sludge bed will occur, subsequently resulting in activity loss (Lettinga et al.,
2001). Furthermore, a thinner and less porous cake layer on the membrane surface

21
can be expected, possibly leading to more serious pore clogging problems in UASB
coupled AnMBRs, since the membrane would be exposed directly to only fine
particles instead of a range of particle sizes. According to the previous results
particularly the small (submicron) particles determined the operational flux (Jeison et
al., 2009a; Jeison et al., 2009b). Therefore, the degree of small particle retention by
the sludge bed will be of prime importance for the feasibility of coupled membrane
filtration.

In some studies, researchers used membrane filtration as a polishing step after UASB
systems without returning the concentrate stream to the bioreactor as it is applied in
AnMBRs (Herrera-Robledo et al., 2010; Salazar-Pelaez et al., 2011a; Herrera-
Robledo et al., 2011; Salazar-Pelaez et al., 2011b) (Figure 2.1). These configurations
can be regarded as tertiary filtration. This approach has the advantage of easier
control of hydraulics in the UASB reactor and the preservation of the dilution rate as
bacterial selection criterion. However, it may have the disadvantage of high SS
concentrations exposure to the membrane because SS would be concentrated in the
concentrate collection tank.

Figure 2.1 : Membrane filtration used as a post-treatment step after UASB reactors.

2.2.2.2 Expanded granular sludge bed reactor

UASB reactors treating sewage at low and moderate temperatures are sometimes
characterized by a poor mixing regime, which causes a decrease in soluble COD
treatment efficiency. To solve this problem, tall reactors with a small footprint and
external or internal effluent recirculation are being applied, i.e. the so-called EGSB
reactors (van der Last and Lettinga, 1992). Bench-scale and (semi-)pilot studies have

22
shown that an EGSB reactor is an attractive alternative, especially for the treatment
of low strength wastewaters at ambient temperatures, due to the efficient biomass-
substrate contact induced by the applied high upflow velocity (de Man et al., 1988;
Kato et al., 1997; Lettinga et al., 1999). Besides, higher OLRs can be applied in
EGSB reactors compared to UASB reactors. However, SS are limitedly removed
from the wastewater and sludge washout would occur in the system due to high
upflow velocity. Therefore, anaerobic treatment applying EGSB systems is restricted
to pre-settled sewage only. As an alternative to pre-settling, Chu et al. (2005), as the
sole example in literature, proposed a membrane unit that is capable of retaining the
SS inside the EGSB reactor. In that study, the membrane module is submerged in the
upper part of the EGSB reactor and the applied high upflow velocity was beneficial
in the reduction of membrane fouling due to the increased hydraulic shear stress on
the membrane surface (Chu et al., 2005). Nonetheless, cake resistance governed the
achievable flux over the hollow fiber membranes, whereas pore blocking only was of
minor importance. Based on the high removal efficiencies even at low temperatures,
the use of EGSB reactors in AnMBRs was suggested as a potential technology for
treating municipal wastewater at ambient temperature (Chu et al., 2005). Following
their reasoning, submerged membrane configurations are found more appropriate
than external configurations for the use of EGSB reactors in AnMBRs. However,
membrane integration eliminates the hydraulic selection pressure required for
granulation, by avoiding the washout of flocculent sludge with poor immobilization
characteristics. Therefore, no granulation is expected in EGSB reactors coupled to
membrane filtration, which would decrease the settleability of the biomass on the
long-term operation.

2.2.2.3 Other reactor types

In addition to the most popular high rate reactors, i.e. UASB and EGSB reactors,
other reactor types such as hybrid upflow anaerobic systems (Wen et al., 1999) and
jet flow anaerobic reactors (Saddoud et al., 2006) were also used for the treatment of
municipal wastewater at low temperatures.

Wen et al. (1999) operated a hybrid upflow anaerobic system coupled to a membrane
unit for 200 days. The hybrid system consisted of an anaerobic bioreactor column in
which the bottom part was used as a sludge bed zone and the upper filter part was

23
packed with fine fibres to prevent sludge washout. They obtained a high COD
removal efficiency (>97%) with an effluent COD concentration up to 20 mg/L. It
was also reported that high OLRs (0.5-12.5 kg COD/m3.d) could be applied
successfully.

A jet flow anaerobic reactor (Saddoud et al., 2006) was coupled to membrane
filtration for the treatment of municipal wastewater. Circulation of the liquid inside
this type of reactor by using an inner tube and a nozzle system provides a good
homogenization inside the reactor. By using an UF membrane coupled to the jet flow
anaerobic reactor, total removal of pathogens and high quality effluent were obtained
of interest for agricultural usage (Saddoud et al., 2006).

2.3 Alternative Integration Possibilities of AnMBRs in Municipal Wastewater

Treatment Flow Schemes

In view of cost-effective municipal wastewater treatment, which also aims at

recovery of resources (nutrients), production of energy, and the use of the treated
effluent, integrated system approach deserves more attention than solely the coupling
of a specific anaerobic reactor type to membrane units. By that approach, the
optimum way of closing nutrient and water cycles should be determined. Some
alternative integration options for AnMBRs in conventional wastewater treatment
schemes can be found in the literature. However, innovative possibilities should be
continuously developed for energy and nutrient recovery.

For municipal wastewater treatment at low (~15 °C) and/or fluctuating temperatures
(15-25 °C), a two-step system including a UASB reactor combined with an
additional sludge digester was shown advantageous to prevent activity loss in the
sludge bed due to slowly degradable particulate matter accumulation (Mahmoud et
al., 2004). In this system, the non-degraded particulate COD entrapped in the sludge
bed of the UASB reactor is further stabilized in a separate CSTR type digester
operated under optimal mesophilic conditions (30-35 °C). The UASB-Digester setup
is proposed to overcome the hydrolysis limitation induced by low temperatures and
to reduce non-stabilized particulate matter accumulation in the sludge bed. By
recirculating the digested solids to the UASB reactor, the sludge bed SMA is
increased resulting in enhanced removal of soluble organics (Mahmoud et al., 2004).
Membrane integration to the UASB-Digester system can be attractive for producing

24
high quality and nutrient-rich effluents for reuse purposes under moderate climate
conditions (Figure 2.2(a)).

Since anaerobic processes are particularly of interest for concentrated waste stream,
upconcentration of sewage using forward osmosis (FO) technology is currently
receiving increasing interest (Zhao et al., 2012). FO uses an osmotic pressure
gradient to drive water across a semi-permeable membrane from the low osmotic
pressure side (feed solution) to the high osmotic pressure side (draw solution).
Owing to the absence of pressure driven water transport, FO membranes are not
susceptible for the traditional sewage fouling propensities (Achilli et al., 2009; Zhao
et al., 2012). FO processes may be used as a pretreatment to concentrate municipal
wastewater in coastal areas where the sea water can be used as a draw solution
(Achilli et al., 2009; Cath et al., 2010; Zhao et al., 2012). Alternatively, closed loop
draw solutions can be used, however, at the expense of energy consumption for draw
solution regeneration. For the optimization of energy and other operational costs,
appropriate draw solutes and concomitant regeneration methods are being researched
(McGinnis and Elimelech, 2008; Elimelech and Phillip, 2011). A decrease in
wastewater flow is regarded as the benefit of the FO based concept resulting in small
treatment systems with anaerobic digestion (AD) as the main bio-stabilization
process (Holloway et al., 2007; Zhao et al., 2012). A possible integration of the FO
unit in an AnMBR treatment system is given in Figure 2.2(b). Interestingly, a current
Dutch project researches the possibility to extract fresh water from sewage using FO,
whereas the AD-recovered energy from the concentrated sewage is used for re-
concentrating the draw solution. The energy needed for the desalination part of the
total process could be obtained by digestion of the concentrated sewage. However,
further research should be done in order to clarify the sufficiency of the produced
energy to drive the whole integrated concept (Roest et al., 2010).

Due to the low COD concentrations and concomitant high influent flow rates, high
membrane surface areas and/or high fluxes are required for municipal wastewater
treatment, using membrane technology. Therefore, a pretreatment step such as
(advanced) solid/liquid separation in municipal wastewater is of interest when
AnMBR systems are considered for municipal wastewater treatment. Consequently,
AnMBRs may take the role of sludge digesters in municipal wastewater treatment
systems as presented in Figure 2.2(c). A similar flow scheme was proposed by Sutton

25
et al. (2011). The latter authors carried out a model based feasibility study with mass
balance calculations for a conceptual flow diagram of a municipal wastewater
treatment plant consisting of aerobic and anaerobic MBRs and innovative nutrient
recovery processes such as ammonia removal by using zeolite, and orthophosphate
removal by a reactive filtration system. They proposed to shuttle the largest fraction
of organic carbon in the wastewater into solids slurry and to further digest the
organic solids in an AnMBR. Modeling and simulation results were used to optimize
the design of the system with respect to energy usage and solids production and to
obtain cost savings. Minimization of residual solids production and maximization of
the energy produced resulted in cost reduction. However, results are theoretical and
emphasis should be given to membrane performance, validating the concept with
pilot and full-scale studies.

Another alternative for AnMBR integration as a sludge digester is depicted in Figure

2.2(d). The main difference between Figure 2.2(c) and 2.2(d) is the presence of a
membrane coupled hydrolysis reactor in Figure 2.2(d). The hydrolysis reactor can be
used for solubilization of particulate organic matter which may accumulate in the
UASB reactor. The generated volatile fatty acids (VFAs) can be further converted to
methane in a subsequent high rate anaerobic system such as membrane coupled
UASB reactor or can be used for other purposes. This approach was investigated at
pilot-scale by Kiriyama et al. (1992) and Kiriyama et al. (1994) who concentrated
raw sewage by a filter unit and fed the thickened sludge to a hydrolysis reactor
equipped with external cross-flow membranes. VSS reduction rate increased from
45% to over 80% by the addition of a membrane module to the hydrolysis reactor.
The pre-acidified wastewater was further treated in a UASB reactor. However, no
information about the membrane performance was given in the study. Kataoka et al.
(1992) operated a two stage system consisting of a hydrolysis reactor and a FB
reactor for the treatment of municipal wastewater. Both reactors were equipped with
external membrane units. The membrane coupled FB reactor was operated at an OLR
of 1.1 kg BOD/m3.d with 91% BOD removal.

26
Figure 2.2 : Alternative integrated flow schemes of AnMBRs.

27
In case nutrient removal is required instead of recovery for reuse, all proposed flow
schemes can be modified by adding innovative nutrient removal systems such as
Sharon/Anammox (van Dongen et al., 2001; van Kempen et al., 2001) and struvite
recovery (Marti et al., 2008; Pastor et al., 2010). Grundestam and Hellstrom (2007)
proposed to treat AnMBR effluents by RO membranes in order to achieve high
nutrient recovery and water reuse. The system was operated at 22 °C without any
heating and a high permeate quality was obtained. Removal efficiencies obtained by
the proposed system were 91%, 99% and >99% for Kjeldahl nitrogen, phosphorus
and TOC, respectively. Total energy consumption of the system was reported as 3-6
kWh/m3.

2.4 Factors Affecting the Treatment Performance of AnMBRs for Municipal

Wastewater Treatment

Optimization of AnMBR processes generally focus on either the improvement of

biological efficiencies (Visvanathan et al., 2000) or on enhancing the membrane
based separation process (Huang et al., 2008; Dereli et al., 2012). Few authors
research the interaction between both processes (Chu et al., 2005; Vyrides and
Stuckey, 2009). In subsequent paragraphs the impact of the various factors, including
the key operational parameters such as temperature, HRT, upflow velocity, OLR as
well as sludge characteristics, and addition of adsorbents, on both biological and
physical aspects are further discussed.

2.4.1 Operational conditions

2.4.1.1 Temperature

The temperature dependence of biological reaction rates carries vital importance for
the overall efficiency of a biological treatment process. Generally, the activity of
microorganisms in a biological process decreases when the temperature decreases,
which consequently results in a decrease in COD removal efficiencies. Besides its
effect on the metabolic activities of the microbial population, temperature has a
profound effect on factors such as biogas solubility, solubility of (in)organic
compounds, and the settling characteristics of the biological solids due to water
viscosity changes (Tchobanoglous and Burton, 2003).

28
Chu et al. (2005) investigated the effect of temperature on the performance of
membrane-coupled EGSB reactor at moderate and low temperatures. Temperature
strongly affected the rate of the anaerobic conversion process and a decrease in both
COD removal efficiency and activity was observed with the decrease in temperature
from 25 °C to 11 °C. However, the contribution of the membrane to the COD
removal efficiency increased from 8.8% to 14.2% with the decrease in temperature,
which demonstrated the advantage and robustness of membrane-coupled reactors in
maintaining high treatment efficiency even at low temperatures due to the complete
retention of particulate COD and biomass. This is in agreement with findings of Ho
and Sung (2010) who also investigated the effect of temperature in the treatment of
municipal wastewater by AnMBRs. Two identical AnMBRs were operated at 25 °C
and 15 °C in parallel for 112 days. They reported that the physical removal capacity
by the membrane compensated for the decreased SMA and biological removal rate at
15 °C by rejecting soluble organics. Wen et al. (1999) also observed that AnMBRs
treating municipal wastewater had a strong tolerance to temperature fluctuations
between 12 °C and 26 °C. They obtained 88% COD removal even at 12 °C.
Martinez-Sosa et al. (2011a) reported that bioreactor temperature impacted methane
recovery in an anaerobic externally submerged MBR treating municipal wastewater.
COD removal efficiencies close to 90% were achieved at both 35 °C and 20 °C. A
stable cake layer formation on the membrane surface may compensate differences in
biological organic matter degradation rate at different temperatures. Although
methane recovery at 20 °C is lower owing to the higher solubility of methane at
lower temperatures, interestingly, the methane content of the biogas was higher at
low temperatures due to the gas solubility difference of carbon dioxide (CO 2) and
methane at 20 °C compared to 35 °C. Lower solubility of methane in comparison to
CO2 at 20 °C resulted in a major proportion of CO2 in the liquid phase leaving the
reactor and an increase in the proportion of methane from 80% to 88% in the gas
phase. Gimenez et al. (2012) also investigated the effect of temperature on methane
recovery efficiency in an AnMBR system treating municipal wastewater. They
obtained slightly lower values at 20 °C in comparison to 33 °C due to a reduction in
the treatment efficiency and an increase in the gas solubility.

29
2.4.1.2 Organic loading rate

AnMBR processes have the advantage of tolerating changes in organic loading

similar to its tolerance to the fluctuations in temperature. OLRs ranging from 0.3 to
12.5 kg COD/m3.day have been applied in AnMBRs treating municipal wastewater
(Table 2.1). Wen et al. (1999) confirmed that an excellent effluent quality could be
obtained by the combination of an anaerobic bioreactor with a membrane unit in
spite of high organic loading fluctuations between 0.5 and 12.5 kg COD/m 3.day,
which generally would cause perturbations in conventional anaerobic reactors.
Accordingly, Lin et al. (2011) observed relative stability in the permeate COD
regardless of the fluctuations in the influent COD. The quality of the effluent was
also stable in the study of An et al. (2009a) in spite of the fluctuations in OLR. They
stated that AnMBR could achieve more stable effluent quality in comparison to a
conventional anaerobic process. Moreover, it was shown that the biogas yield
increased linearly with an increase in the organic loading.

2.4.1.3 Hydraulic retention time

HRT is an important parameter from an economic perspective as it has a strong

influence on capital costs, considering that shorter HRTs allow smaller reactors
(Salazar-Pelaez et al., 2011a). Therefore, the influence of HRT on the biological
removal efficiency of AnMBRs treating municipal wastewater has been investigated
by various researchers. Among them, Hu and Stuckey (2007) studied an AnMBR
operated at 35 °C and concluded that COD concentrations both in the reactor and
permeate increased slightly with a decrease in HRT due to an increased organic load.
Chu et al. (2005) also studied the effect of HRT at various temperatures on the
treatment performance of a membrane-coupled EGSB reactor. They found that COD
removal efficiency was independent of HRT at temperatures higher than 15 °C.
However, an increase in COD removal efficiency was observed with an increase in
HRT at 11 °C, indicating the significance of HRT at low temperatures. HRT was
controlled as an independent parameter from upflow velocity in the study of Chu et
al. (2005) by introducing effluent recirculation to the reactor. However, for upflow
reactors without recirculation, effects of HRT and upflow velocity can be evaluated
dependently as they are inversely proportional to each other as in the study of An et
al. (2009a). They reported an increase in the TOC removal efficiency when the HRT

30
of a membrane coupled UASB reactor was decreased from 10 h to 5.5 h. They
explained this as a result of improved wastewater distribution in the sludge bed and
better contact between biomass and substrate at higher upflow velocity. Based on the
results, it can be hypothesized that there is an optimum HRT for each case, which is
determined by many factors such as system hydraulics, wastewater characteristics
and sludge properties, to provide both efficient biological removal and filtration
performance.

2.4.1.4 Upflow velocity

Upflow velocity is an important parameter having two opposing effects on the

biological removal efficiency in upflow reactors. An increase in upflow velocity may
enhance mixing providing better substrate-biomass contact. On the other hand,
increasing upflow velocity may deteriorate the removal efficiency by exceeding the
settling velocity of particles, resulting in the detachment of the captured solids due to
high hydraulic shear force (Mahmoud et al., 2003). Different upflow velocities were
applied in Chu et al. (2005)’s study by using effluent recirculation in a membrane-
coupled EGSB reactor and a better COD removal performance was achieved at
higher upflow velocities which supported the former hypothesis of Mahmoud et al.
(2003). Besides, a significant increase in COD removal efficiency was observed with
an increase in upflow velocity at 11 °C in comparison with a slight increase at 25 °C,
which demonstrated the importance of adequate hydraulic mixing at lower
temperatures.

2.4.2 Sludge characteristics

Biomass characteristics, bacterial flora, e.g. the fraction of slow-growing bacteria,

and nutritional requirements, are mainly dependent on the type and operational
conditions of bioreactors (Kataoka et al, 1992). There are several reports asserting an
activity loss of biomass, especially of propionate degraders, in AnMBRs
(Brockmann and Seyfried, 1997; Padmasiri et al., 2007). This may be due to lysis of
the cells under high shear or disruption of juxta-positioning of the hydrogen
producing bacteria and hydrogenotrophic methanogens, enlarging the interspecies
hydrogen transfer distance. On the other hand, Jeison et al. (2009b) preserved both
methanogenic and acetogenic activities in a cross-flow AnMBR system applying
liquid superficial velocities of 1-1.5 m/s and gas slug upflow velocities of 0.1 m/s.

31
SMAs of the cross-flow AnMBR sludge using propionate as the substrate were even
higher than those of a parallel operating UASB system. Other studies focus on
differences in microbial species composition or activity profiles between the fouling
layer and biomass suspension of AnMBRs treating domestic sewage (Ho and Sung,
2010; Gao et al., 2010). Ho and Sung (2010) found that SMA of the biomass
attached to the membrane surface was lower compared to the bulk sludge in an
AnMBR treating municipal wastewater. Thus, the attached sludge did not play a
significant role as a biofilm for biological removal compared to suspended sludge.
Zhang et al. (2011) also investigated the characteristics of the fouling layer and the
bulk sludge in a dynamic AnMBR system in which separation is achieved by a
dynamic cake layer that is formed by bulk sludge. Microorganism communities were
found to be different from the bulk sludge, and activity of microorganisms in the
dynamic layer was found lower than bulk sludge due to the compactness of cake
layer which resulted in suppressed mass transfer.

Lin et al. (2011) investigated the effect of sludge concentration on biological

treatment of AnMBRs and observed a relative stability in the permeate COD
regardless of the fluctuations in sludge concentrations between 6.4 and 9.3 g
MLSS/L.

2.4.3 Addition of adsorbents

Powdered activated carbon (PAC) and zeolites can be added into AnMBRs in order
to adsorb soluble organic compounds. By this way, these materials reduce organic
fouling, enhance membrane flux and also affect biological treatment performance
(Akram and Stuckey, 2008). Hu and Stuckey (2007) investigated the effect of both
PAC and granular activated carbon (GAC) addition on the performance of AnMBRs.
The addition of PAC resulted in a significant increase in COD removal efficiency,
whereas only limited effect was observed after GAC addition. This was attributed to
the larger surface of PAC per unit mass compared to GAC, thus more colloids and
macromolecules could attach to PAC from the bulk solution. VFAs were only
limitedly absorbed to the activated carbon. Interestingly, the SMA of the sludge
growing in the AnMBR without activated carbon addition was lower compared to
that from the AnMBR with activated carbon addition. This finding was explained by
the support surface provided by activated carbon in order to protect the biomass from

32
high shear conditions. Besides, Vyrides and Stuckey (2009) observed an increase in
DOC removal after the addition of PAC due to the adsorption of high molecular
weight compounds.

2.5 Factors Affecting the Membrane Performance of AnMBRs for Municipal

Wastewater Treatment

Flux decline, especially in long-term operation, seems to be the most critical

constraint for the applicability and feasibility of AnMBRs for municipal wastewater,
since it plays a major role in the determination of the required membrane area. A
number of factors are associated with flux decline. Therefore, the research of
membrane fouling behavior and mechanisms requires the understanding of several
factors such as membrane characteristics, operational conditions and sludge
properties. So far, membrane fouling in AnMBRs has not been fully understood due
to the complex nature of membrane foulants and diversity of operational conditions,
membrane materials, configurations and wastewaters in different studies. Multiple
and complex interactions occur between physical and biological factors not only in
real plants but also in controlled lab-systems. Therefore, fouling phenomena should
be investigated under well-defined combinations of parameters and conditions which
individually might even present opposing effects (Drews, 2010). Having outlined the
key parameters affecting biological treatment performance, major factors that affect
filterability in AnMBRs for municipal wastewater treatment will now be discussed.

2.5.1 Membrane characteristics

2.5.1.1 Material

Membrane material characteristics may affect the degree of fouling in AnMBRs; e.g.
organic and inorganic membranes may show different fouling behaviors. Kang et al.
(2002) reported that cake layer formation was the main mechanism for fouling of
organic membranes, whereas inorganic precipitation, mainly struvite, played the key
role in the fouling of inorganic membranes.

Gao et al. (2010) observed differences in fouling rates and foulant layer composition
even for two different organic membrane materials. They found that fouling of
uncoated polyetherimide (PEI) UF membranes occurred faster than PVDF UF

33
membranes coated with polyether block amide, reaffirming the significance of
membrane material on fouling. Furthermore, they determined significant differences
in bacterial composition in the fouling layers of different membrane materials.
Although Bacteroidetes were absent in the fouling layer of PVDF membrane, they
were present on PEI membrane. Observations of Gao et al. (2010) showed that
membrane material may affect the interactions between different microbial species
and membrane surface, and thus the fouling phenomena, and support the value of
integrating membrane material science with deep understanding of the adhesion and
biofilm-formation potential of different microbial species.

2.5.1.2 Module type and configuration

Various membrane configurations such as flat sheet, hollow fiber, and tubular
membranes are applied in AnMBRs, using different types of module configurations
such as submerged/immersed and external cross-flow systems (Liao et al., 2006).
Submerged AnMBRs can also be used in different configurations including directly
immersed into the bioreactor or immersed in a separate membrane tank. Generally,
for submerged AnMBRs flat sheet and hollow fiber membranes are preferred,
whereas external cross-flow AnMBRs can be configured with tubular membranes as
well (Liao et al., 2006). Different hydrodynamic conditions in submerged and
external cross-flow AnMBRs have different effects on the bulk sludge properties,
attainable flux and membrane fouling due to the extent of applicable shear rate
(Berube et al., 2006). AnMBRs configured with external cross-flow membranes may
provide a higher flux (van Lier and Jeison, 2010) and need a lower membrane area in
comparison to their submerged counterparts. However, energy required for the cross-
flow pumps would be high due to the high flow to be pumped for providing enough
hydraulic shear force. Martin-Garcia et al. (2011) compared specific energy demands
for different configurations of AnMBR treating municipal wastewater and found that
0.3 kWh/m3 permeate and 3.7 kWh/m3 permeate were needed for submerged and
external cross-flow configurations, respectively. The amount of pumping energy may
be decreased by using small diameter tubular membranes since the packing area of
the membrane module can be increased. On the other hand, high hydraulic shear
force may also disrupt anaerobic biomass and generate small particles, which result
in significant membrane fouling. Besides, biological activity of anaerobic biomass
may decrease due to the high hydraulic shear force (Lin et al., 2010).

34
An et al. (2009a) investigated the effect of different tubular membrane diameters, i.e.
3.0, 1.9, 1.2 mm, on the filtration performance of an external cross-flow AnMBR
treating municipal wastewater. The results showed that the variations of TMP were
reciprocally correlated with the variations in tube diameter. This phenomenon was
related to the particle accumulation that occurred in the lumen. A smaller diameter
tube was simply blocked by large particles, which caused serious membrane
clogging in comparison to the tubes with larger diameter. The blockage of small
diameter tubes resulted in an uneven flux distribution along the membrane module
and an increase in the local flux, which finally caused more serious membrane
clogging in comparison to the tube with larger diameter.

2.5.2 Operational conditions

2.5.2.1 Shear rate

Removing the deposited cake layer from the membrane surface is very important to
achieve a steady operation in AnMBRs. Cross-flow and biogas sparging may be
regarded as principle mechanisms to provide shear over the membrane surface in
order to limit the deposition of particles and restrict their interaction with the
membrane (Liao et al., 2006). The cake layer formation rate was found to be
positively correlated with the fraction of small size particles and flux, whereas it was
negatively correlated with shear rate (Lin et al., 2010). Although increasing shear
rate can help to reduce fouling and cake resistance, and improve flux, there appears
to be a certain limit for both cross-flow velocity and gas sparging rate beyond which
it would practically have little or no added benefit. Choo et al. (2000) indicated that
cake layer resistance can be decreased by increasing cross-flow velocity. However, at
a Reynolds number of 2000, a plateau was reached and no further reduction in the
resistance could be obtained. The same is true for the gas sparging rate. Xie et al.
(2010) reported that the critical flux of a submerged AnMBR increased and the
fouling rate decreased when the biogas sparging rate was increased from 10 to 25
L/m2.min. However, they indicated that there was a practical limit above which
further increasing the biogas sparging rate provided little added benefit. Similar
results, also for submerged AnMBRs, were presented by Jeison and van Lier
(2006a).

35
Moreover, high shear rates may also stimulate the break-down of microbial flocs and
increase the cake layer resistance due to the selective deposition of fine particles in
the cake layer and membrane pores during long-term operation. This results in the
formation of a dense consolidated cake layer that is very hard to remove. This
phenomenon has been called as the shear rate dilemma (Jeison et al., 2009a). An et
al. (2009a) investigated the effect of cross-flow velocity on fouling in a UASB
reactor equipped with tubular membranes. They concluded that a low TMP can be
maintained for a long time at high cross-flow velocities. However, this operation
condition led to a high irreversible fouling due to the deposition of small sized
particles on the membrane surface and inside the pores during long-term operation.

Duration and frequency of applied shear rate may also have an effect on filterability.
Vyrides and Stuckey (2009) reported that switching from continuous biogas sparging
to intermittent mode (10 min ON/5 min OFF) resulted in a small increase in TMP.
However, it improved the DOC removal of submerged AnMBR due to the formation
of a thicker cake layer on the membrane surface. Under these conditions a higher rate
of biodegradation was observed because the substrate was in contact with a thicker
biofilm and more of the higher molecular weight solutes were retained until they
could be degraded.

2.5.2.2 Flux

Choice of the operational flux is very important in terms of fouling management.

Operation below the critical flux, which depends on membrane characteristics,
operating conditions and sludge characteristics, is an effective approach to avoid
severe fouling within a filtration system (Jeison and van Lier, 2006b; Meng et al.,
2009). Martinez-Sosa et al. (2011b) reported that an AnMBR treating municipal
wastewater could be operated stable for a long time at a critical flux of 7 L/m 2.h with
a biomass concentration of 14.8 g TSS/L and a gas sparging velocity of 62 m/h.
Increasing flux to 10 L/m2.h and 12 L/m2.h resulted in an unstable operation due to
the high fouling rate that could not be controlled even at higher gas sparging
velocities.

2.5.2.3 Operation mode

The operation mode of membrane systems is another important factor influencing

fouling in AnMBRs. Generally, backwashing and/or relaxation are used as strategies

36
to decrease fouling. Moreover, optimization of backwash and relaxation duration
and/or frequency plays an important role in flux enhancement. Lew et al. (2009)
emphasized the necessity of backwash frequency optimization for fouling reduction.
They obtained similar fouling rates for 15 and 30 minutes backwash frequencies;
however, higher fouling rate was observed at a higher backwash interval of 60 min.
Chu et al. (2005) also dealt with finding optimal operating modes and observed an
increase in permeability and enhancement in the permeate flux (J) recovery with
increasing relaxation time. During the relaxation period, efficient cake layer removal
from the membrane surface was achieved. Gimenez et al. (2011) operated a pilot-
scale AnMBR equipped with hollow fiber membranes for the treatment of municipal
wastewater and obtained a flux of 10 L/m2.h at an MLSS concentration of 22 g/L.
They prevented fouling with the help of a rather complex membrane operation mode
including backwash, relaxation and degasification cycles. An et al. (2009a) compared
backwash, relaxation and continuous filtration modes and maintained low TMP for a
longer time in backwash mode in comparison to relaxation or continuous filtration
modes.

2.5.2.4 Temperature

Temperature affects not only the rate of the biodegradation process but also the
viscosity of the filtered liquid and the solubility of various compounds and gases.
Since under practical conditions sewage temperatures cannot be altered, the
operational temperature is of significant importance for municipal wastewater
treatment using AnMBRs. Martinez-Sosa et al. (2011a) stated that the operational
temperature of an AnMBR is related to the observed membrane fouling. Although
they operated the system at a flux of 7 L/m2.h at 35 °C, they observed an increase in
the fouling rate at 20 °C, possibly due to TSS and soluble COD accumulation and a
higher viscosity in the bioreactor. Decrease in temperature and increase in solids
content were proposed as possible explanations for viscosity increase. The fouling
rate was reported as 0.14 mbar/d at 35 °C, whereas it was 2.61 mbar/d at 20 °C.

2.5.2.5 Upflow velocity

The upflow velocity in (submerged) AnMBRs generally has a positive impact on

filterability mainly due to increased shear stress. Chu et al. (2005) operated an EGSB
reactor equipped with a submerged membrane module and observed an increase in

37
permeability with the increase in upflow velocity likely due to the shear effect of
higher upflow velocities. However, on the long-term, a flux decline was observed
indicating a decreased contribution of upflow velocity to fouling control, suggesting
the inadequacy of shear force induced by high upflow velocity to combat cake layer
formation. Possibly, the observed phenomenon can be attributed to strong adhesion
of the foulants to the membrane surface and an increased thickness of the cake layer.

2.5.2.6 Solids retention time

SRT has been identified as the main parameter influencing flux with values typically
dropping at higher SRTs. Operation of anaerobic systems at ambient temperatures
becomes feasible only if SRT is approximately twice as high as that under
mesophilic conditions (Kashyap et al., 2003). As a rule of thumb, SRT should be at
least three times the doubling of the slowest growing organism responsible for
bioconversion (van Lier et al., 2008). High capability of solids retention in
membrane systems makes membranes ideally suited for anaerobic treatment of
municipal wastewaters especially at low temperatures when the degradation rate of
SS and colloidal materials is the rate limiting step (Zhang et al., 2010). Since the
particulate organics would also be retained in the reactor, they can eventually be
further hydrolyzed and degraded (Zhang et al., 2010). However, although membrane
processes result in solids retention independently of temperature (Liao et al., 2006),
the activity limitation of anaerobic microorganisms at low temperature might yield
high colloidal and soluble solids in anaerobic effluents, resulting in an increase in
membrane fouling propensity (Herrera-Robledo et al., 2011). SRT of more than 140
days was reported as a possible cause of severe membrane fouling and flux decline
(Saddoud et al., 2007). Huang et al. (2008) reported that longer SRTs (from 30 to
infinite days) produced higher protein/carbohydrate ratio (P/C) in extracellular
polymeric substances (EPS) and lower P/C in SMP, which resulted in serious
fouling. However, Herrera-Robledo et al. (2010) investigated the effect of SRT for
both short and long-term operations for the system that used UF membrane as a
polishing step after a UASB reactor. It was found that fouling rate and effluent
quality were not dependent on SRT. Their observations were based on trials for SRT
of 60 and 100 days in short-term operation. Sudden changes in TMP and flux were
observed for both SRTs. During long-term operation (500 h), they observed that
sudden increases in TMP and decreases in flux occurred in shorter filtration period

38
(e.g. 140 hours) with SRT of 60 days than with SRT of 100 days (e.g. 175 hours).
This leads to the assumption that a more rigid and strong fouling layer structure was
developed in the system having long SRTs.

2.5.2.7 Hydraulic retention time

Variations in HRT may change MBR fouling propensity. A few researchers have
reported on the effect of HRT in membrane coupled sludge bed reactor types treating
municipal wastewaters. Among them, An et al. (2009a) reported that a decrease in
HRT from 10 h to 5.5 h resulted in a decrease in solids removal efficiency of the
bioreactor. However, owing to the membrane separation, the reactor performance
was quite stable. Besides, Lew et al. (2009) observed that membrane fouling was
positively correlated to the particulate matter concentration reaching the membrane.

2.5.3 Sludge characteristics

Currently EPS, both bound and soluble, is often mentioned as the sludge factor of
prime importance in relation to membrane fouling. Soluble EPS is often also referred
to as SMP (Meng et al., 2009). Membrane inner pore accumulation and sorption of
EPS and SMP favor biomass attachment and cake layer formation, possibly leading
to severe fouling (Zhang et al., 2010; Herrera-Robledo et al., 2011). Operational
parameters such as SRT, OLR, temperature, pH and shear rate are the most important
factors influencing both the concentration and composition of SMP and EPS.

EPS excreted from the microbial cells was considered to have a significant effect on
the fouling due to both an increase in viscosity of the mixed liquor and an increase in
the filtration resistance (Nagaoka et al., 1996). Chu et al. (2005) measured the
amount of EPS from both granules in an AnMBR and sludge on the membrane
surface and claimed that EPS affected the cake resistance by filling the void spaces
between the particles in the cake layer resulting in a drastic reduction in the flux.
Viscosity of the liquor in the anaerobic reactor, which was equal to the water
viscosity, did not change with operating time. Apparently, the impact of EPS is only
on filtration resistance and not on liquid viscosity. An et al. (2009b) observed that
EPS extracted from the cake layer consisted of mainly protein-like and humic acid-
like substances. Moreover, Gao et al. (2010) found that EPS mainly consisted of
proteins and was the main reason of fouling in AnMBRs treating municipal
wastewater. An et al. (2009b) reported that the support layer (nonwoven fabric)

39
surface was covered with a rough and dense layer consisting of mainly protein, clay
materials and inorganic elements such as magnesium, aluminium, calcium, silisium,
and iron.

SMP was grouped into two predominant fractions including high and low molecular
weight SMP in Herrera-Robledo et al. (2011)’s study. High molecular weight SMP
was associated with long SRT, incomplete organic matter hydrolysis and rate
limitations of anaerobic microorganisms at low temperatures, e.g. 20 °C. Low
molecular weight SMP have the propensity to pass freely through membrane pores.
However, it can also adsorb onto the membrane surface leading to an intermediate
blocking or deposit inside the pores resulting in standard blocking.

Besides, microbial community composition is effective on fouling. Gao et al. (2010)

found differences in community composition between the cake layer and biomass
suspension for a membrane coupled UASB reactor. They detected a bacterium within
phylogenetic division OP11 at higher abundance in the cake layer compared to the
suspension. Moreover, some bacteria such as Bacteroidetes were not found in the
fouling layer although they existed in the suspension. These results may indicate that
some species play a direct role in fouling, e.g. by attaching to the membrane surface,
while others, including some that likely play a major role in the metabolism of
influent organics, play a less important or indirect role. When this hypothesis is true,
it means that cells are selectively incorporated within the fouling layer. Then,
isolation and characterization of representatives of these species may provide useful
information for biofouling control.

2.5.4 Addition of flux enhancers

The interest in flux enhancers that can act through a number of different phenomena
such as adsorption of SMP, coagulation, crosslinking between flocs and SMP has
greatly increased (Drews, 2010). Various additives such as activated carbon,
polyelectrolytes, coagulants and flocculants can be used to improve the flux and to
reduce fouling in MBRs (Hu and Stuckey, 2007; Wu et al., 2009).

Wu et al. (2009) investigated poly-aluminum chloride as a flux enhancer and found

that addition of 10 mg/L poly-aluminum chloride decreased fouling significantly in
comparison to other tested adsorbents/coagulants including PAC, zeolite, and
polyamide. In the study of Hu and Stuckey (2007), the effects of PAC and GAC

40
addition on TMP and flux were investigated in an AnMBR. In terms of both TMP
values and variations, better performances were achieved with activated carbon that
constantly scoured deposited particles from the membrane surface. Besides, a shift to
a relatively higher range in PSD was observed in an AnMBR with activated carbon
addition. On the other hand, PSD shifted to a lower size range in the AnMBR
without activated carbon addition, which probably led to pore clogging. Maintaining
stable TMP and reducing fouling are more pronounced with PAC than with GAC due
to the higher surface area of PAC. Moreover, rigid PAC particles may also make the
cake layer more porous since they are larger than the bioflocs. Besides, Vyrides and
Stuckey (2009) observed a reduction in biofilm (gel layer) resistance and thus, TMP
reduction after the addition of PAC. However, there appears to be an optimum dose
for PAC addition. Extensive addition can decrease the flux due to increased viscosity
of the sludge (Akram and Stuckey, 2008). Akram and Stuckey (2008) performed
flux experiments at a constant TMP and observed a significant flux improvement
from 2 L/m2.h to 9 L/m2.h with PAC addition of 1.67 g/L due to the adsorption of
fine colloids and dissolved organics, and the formation of a thin cake layer on the
membrane surface. However, flux decreased with the addition of 3.4 g/L PAC as a
result of the viscosity increase. Due to the significant effects of additives on mass
transfer properties and biochemical environment and concerns related to cost
effectiveness, the optimum dosage of additives should be determined above which no
further or even adverse impacts can occur (Drews, 2010). Most of the studies about
the effect of flux enhancers were done in short-term experiments so far. Therefore,
the feasibility of continuous addition of these chemicals during long-term operation
of AnMBRs treating municipal wastewater still needs to be assessed.

2.6 Cleaning Methods

The level and degree of membrane fouling depend on many factors such as
membrane operation, reactor type, membrane configuration, substrate type and
sludge characteristics. The cures and cleaning methods can be various depending on
the nature of fouling. Membrane cleaning can either be done physically or
chemically.

Physical cleaning is closely related to membrane operation such as regular backwash,

relaxation or short-term increase in shear rate to remove the cake layer accumulated

41
on the membrane surface. In addition, membranes can also be physically removed
from the membrane tank to be cleaned ex-situ, e.g. by applying water jets. Cake layer
formation on the membrane surface was found to play the major role in the increase
of membrane resistance and decrease of flux in AnMBRs (Jeison and van Lier,
2006a). As a matter of fact, Martinez-Sosa et al. (2011b) reported that the efficiency
of physical cleaning in an AnMBR was close to 100% which indicated that
irreversible fouling was not notable or even did not exist.

In most literature, cleaning is generally understood as ‘cleaning in place’, a chemical

cleaning procedure done with alkali, oxidants or acid cleaning agents. The cleaning
in place efficiency depends on many factors, such as foulant type, substrate type,
membrane material, cleaning procedure, i.e. type of chemical, chemical
concentration, soaking/flushing duration, temperature, and frequency.

Generally, an alkali cleaning solution such as sodium hydroxide and oxidants such as
hydrogen peroxide and sodium hypochlorite (NaOCl) are effective in removing
organic foulants such as organic matter with carboxylic and phenolic groups,
proteins and polysaccharides, whereas acid solutions (hydrochloric acid, sulfuric
acid, citric acid) have been extensively used to remove inorganic foulants such as
metal hydroxides and divalent cations (Liao et al., 2006; Porcelli and Judd, 2010).
Investigating municipal wastewater, Chu et al. (2005) applied chemical cleaning for
permeability recovery and compared two cleaning agents including only NaOCl and
the combination of NaOCl and sulfuric acid in terms of permeability recovery. No
significant difference was observed between two washing methods. Chemical
cleaning with NaOCl and backwashing methods were insufficient to remove
membrane biofouling in a membrane coupled UASB reactor (Calderon et al., 2011).
Especially methanogenic Archaea remained undisturbed by the chemical cleaning.
Instead of these conventional cleaning methods, some advanced methods including
nitric oxide cleaning, enzymatic disruption of exopolymers, bacteriophages, etc. were
proposed as effective cleaning techniques (Gao et al., 2010; Calderon et al., 2011).
Moreover, membrane material characteristics play a critical role on the application of
different chemicals for cleaning, i.e. PP membranes are not chlorine tolerant or
PVDF membranes cannot resist high pHs over 11 (Porcelli and Judd, 2010).

The latest experience gained from aerobic MBRs treating municipal wastewater
suggests that frequent maintenance cleaning (weekly or monthly) at mild chemical

42
concentrations can prevent permeability loss during long-term operation. That is,
long-term operation without any chemical cleaning results in residual membrane
fouling that is difficult to remove with chemicals (Zhang et al., 2007; Ayala et al.,
2011). Moreover, frequent cleaning with dilute solutions consumes ~30% less
reagents than infrequent cleaning that is applied i.e. quarterly or biennially using
high-strength reagents (Ayala et al., 2011).

One of the greatest concerns about using chemicals for membrane cleaning is the
effect of cleaning agents on membrane lifetime and integrity. Ayala et al. (2011)
suggested that a membrane lifetime of 6-7 years can be easily sustained in aerobic
MBRs without a significant loss of permeability and integrity if regular chemical
cleaning is practiced. Inorganic fouling that causes irreversible fouling during long-
term operation can decrease the membrane lifetime in AnMBR applications. Kang et
al. (2002) identified struvite as the most important inorganic foulant especially for
inorganic membranes. This would imply that inorganic fouling should not be
underestimated in AnMBRs. However, despite the fact that there are many
investigations aimed at understanding the fouling mechanisms in membrane systems,
the research on chemical cleaning is very limited to either qualitative measurements,
such as characterization of membrane foulants and their cleaning removal efficiency,
or permeability recovery observed after cleaning (Porcelli and Judd, 2010). Thus far,
chemical cleaning is neither fully understood nor given the required importance,
which is critical for the development of better cleaning agents and protocols, i.e.
cleaning conditions, sequence and method (Porcelli and Judd, 2010). More research
should be conducted on membrane cleaning in order to fully understand the
interaction between chemicals, foulants and membranes. This would definitely help
to develop better cleaning practices and procedures for AnMBRs.

2.7 Economic Feasibility of AnMBRs for Municipal Wastewater Treatment

Economic efficiency is a key decisive criterion for the selection of a process among
competing technologies. High membrane surface areas are required for the treatment
of municipal wastewater since influent flow rates are generally high. Therefore,
AnMBRs are particularly of interest if the (agricultural) use of the treated wastewater
is considered. Within this concept, a feasibility study should be performed based on
the energy consumption of pumps and extra costs of additional membrane system

43
devices and cleaning agents (Kocadagistan and Topcu, 2007). However, limited
information is available in the literature about the economic efficiency of AnMBRs
for municipal wastewater treatment.

The reduced flux is known as the main factor determining the economic feasibility of
membrane processes (Lee et al., 2001). This unavoidable disadvantage of AnMBR
technology is associated with an increase in operational costs, which arises from the
requirement of higher suction pressure, more intensive biogas recycling, larger
membrane surface area and more frequent membrane cleaning and replacement
(Meng et al., 2009; Herrera-Robledo et al., 2010).

Lin et al. (2011) assessed the economic feasibility of a submerged AnMBR system
for municipal wastewater treatment based on overall costs. Overall costs are
represented by the sum of capital costs including the costs of membranes, tanks and
plant equipment and operational costs which mainly comprise the costs of power,
sludge disposal and chemicals. It was reported that membrane costs account for the
largest fraction (72%) followed by the costs of tank construction and screens among
capital costs. Since membrane costs are linearly correlated to the applicable flux, the
observed low fluxes remain the obstacle for the potential application of submerged
AnMBRs in municipal wastewater treatment. For instance, similar to aerobic MBRs,
coarse screening followed by a fine screen should be considered in full-scale systems
in order to protect the membranes from damage and maximize membrane life (Lin et
al., 2011). Gas scouring energy represents the most significant operational cost.
Operational costs for aerobic MBRs were found three times higher than those for
AnMBRs due to the increased blower energy and sludge disposal costs. Moreover,
the operational costs for the AnMBRs can be partly offset by the use of produced
methane as an energy source (Liao et al., 2006). Martin et al. (2011) also reported
that energy requirement associated with fouling control in AnMBRs was 2-3 times
lower in comparison to the energy demands of aerobic MBRs. However, due to the
lower fluxes reported in literature for AnMBRs, the capital costs associated with
membranes might be higher compared to aerobic MBRs. They reported a wide
variation in energy demand (0.03 to 3.57 kWh/m3) in submerged AnMBRs. Achilli et
al. (2011) compared the operational costs of an aerobic MBR and AnMBR treating
municipal wastewater and found that the operational costs of the AnMBR were lower
than that of the aerobic MBR due to excess sludge management in the aerobic MBR.

44
However, the AnMBR system required a longer acclimation time for stable operation
than the aerobic MBR. Lin et al. (2011) also performed a sensitivity analysis
including the effect of variations in several aspects such as HRT, flux, membrane
price, flow, membrane lifetime, interest and specific gas demand per unit of
membrane, and reported that influent flow exerted the largest impact on the total life
cycle costs since it determined the capacity and footprint of the system. Besides,
costs are very sensitive to changes in the applicable flux, membrane price and
membrane lifetime followed by the moderate effect of interest. In contrast, HRT and
specific gas demand per unit of membrane were found to influence the costs to a
lesser extent. With the rapid development and implementation of AnMBR
technology especially in full-scale systems, more research on economic analysis is
required.

2.8 Problems Encountered and Future Perspectives

Upgrading of existing anaerobic treatment processes for municipal wastewater with

membranes can be of vital importance, especially when high effluent quality and/or
use of treated effluents are considered. AnMBR effluents are pathogen-free, which
makes them suitable for agricultural use (Saddoud et al., 2006; Ellouze et al., 2009).
However, the wide application of the AnMBRs in domestic utilities, especially in
full-scale systems, has not yet become a reality, which is partly because of system
novelty and reluctance owing to membrane fouling problems (Chu et al., 2005).
Therefore, AnMBR systems will likely benefit from the development of efficient
technologies to prevent fouling.

Inorganic fouling by the precipitation of struvite, potassium ammonium phosphate

and/or calcium carbonate may be one of the major concerns in fouling due to the
release of ammonia and phosphate (PO43-) from organic nitrogen and phosphorus
during AD and pH increases as a result of changes in CO2 partial pressure and
alkalinity generation in AnMBRs (Liao et al., 2006). Salazar-Pelaez et al. (2011b)
estimated that struvite would be undersaturated in municipal wastewaters with lower
concentrations of ammonium nitrogen (NH4-N), PO43- and magnesium compared to
industrial wastewaters and struvite precipitation would unlikely occur. However, not
only the concentration of these ions but also the membrane properties can play a
significant role in struvite precipitation (Kang et al., 2002). Therefore, this

45
phenomenon which may cause irreversible fouling should be further investigated
with long-term studies in AnMBRs treating municipal wastewater.

Other major concerns about using treated municipal wastewater in agriculture are
related with the removal of toxicity and endocrine disrupting chemicals. Industrial
discharges to municipal sewer systems can cause serious problems such as toxicity
and organic shock loads in the end-of-pipe wastewater treatment plants. Saddoud et
al. (2006) using an AnMBR equipped with a cross-flow UF module investigated the
treatability of municipal wastewater comprising industrial discharges containing
toxic substances. The AnMBR process was found to be inefficient, in terms of
unstable biogas production rate and composition, due to the high fluctuations in the
industrial toxicants. Besides, Ellouze et al. (2009) observed a residual toxicity in
wastewaters treated with AnMBRs due to the toxic soluble compounds from
industrial discharges into the sewer, which passed through the membrane. However,
the permeate obtained from the AnMBR was found to be significantly less toxic than
the effluent of conventional treatment processes, such as aerated lagoon and
activated sludge, based on micro-toxicity and phyto-toxicity analysis. Various studies
address the removal of endocrine disrupting chemicals in municipal wastewater
treatment by MBR systems (Cases et al., 2011; Tadkaew et al., 2011; Boonyaroj et
al., 2012). In aerobic MBRs, phenolic compounds, phthalates and estrogens can be
more effectively removed in comparison to conventional activated sludge systems
through biodegradation, adsorption and membrane rejection mechanisms (Cases et
al., 2011). However, there exists very little information about their fate and
biodegradation in AnMBRs (Ho et al., 2007). Obviously, more research should be
conducted on the fate of these priority pollutants in AnMBR systems.

Salinity can be another problem for the treatment of municipal wastewater,

especially for coastal residential areas with improper infrastructure which allows the
infiltration of sea water into sewer systems. Vyrides and Stuckey (2009) examined
the performance of a submerged AnMBR treating saline municipal wastewater with
high fluctuations in salinity up to 35 g sodium chloride (NaCl)/L. They reported 99%
removal of DOC at 35 g NaCl/L, however the removal efficiency inside the reactor
was very low (40-60% DOC). Their results indicated that the retention of SMPs and
colloidal COD in the reactor was due to membrane rejection. But in this situation

46
more attention should be given to fouling control in order to have a feasible
treatment technology.

When agricultural use of the treated effluent is not considered, a nutrient removal
step is generally required to comply with discharge standards. Many studies report
the effluent nutrient concentration in AnMBRs treating municipal wastewater. Lin et
al. (2011) did not observe any removal of nitrogen or total phosphorus (TP) in an
AnMBR treating municipal wastewater as expected. Herrera-Robledo et al. (2011)
also did not observe any decrease in ammonia concentration due to filtration.
Interestingly, Kocadagistan and Topcu (2007) observed phosphorus removal up to
81% in an AnMBR treating municipal wastewater by membrane retention. Similar
results with regard to phosphorus removal together with calcium, magnesium and
iron were reported by Herrera-Robledo et al. (2011). Moreover, Saddoud et al.
(2007) reported high nitrogen and phosphorus removal, such as 60% and 30%
respectively, in an AnMBR which may not be explained solely by the uptake of these
macronutrients by biomass growth. Removal of these elements with low molecular
sizes in comparison to the membrane nominal cut-off may be attributed to their
sorption by the biofouling layer or even chemical precipitation. These contradictory
results are mainly due to the differences in membrane rejection capabilities of the
systems in different studies.

More research needs to be directed towards the applicability of AnMBR to low-

temperature municipal wastewaters. A significant reduction in the hydrolysis rate of
solids is typical at ambient temperatures due to the low activity of anaerobic
microorganisms (Lettinga et al, 2001; Zhang et al., 2010). Methane loss is another
important issue of concern especially at low temperature processes since methane
solubility in the liquid phase increases with decreasing temperature. Gimenez et al.
(2012) proposed biogas-assisted mixing as a limitation for super-saturation and a
guarantee for a minimum concentration of dissolved methane in the effluent of
AnMBR systems.

Alternative membrane materials, reactor types, and membrane integration

possibilities and configurations for AnMBR technology in municipal wastewater
treatment schemes are of interest to further develop. Dynamic membrane technology
that focuses on the use of meshes or fabrics as a support material instead of real
membranes is gaining interest in the AnMBR applications for municipal wastewater

47
treatment. Application of dynamic membrane systems will significantly reduce
capital costs associated with the purchase and renewal of membranes (Zhang et al.,
2010; Ersahin et al., 2012). So far, there have been several attempts to utilize
dynamic membrane filtration in AnMBR systems and the results in terms of removal
efficiency were very promising and comparable to conventional membranes (Ho et
al., 2007; An et al, 2009b; Zhang et al, 2010; Zhang et al., 2011). However, more
research should be conducted to understand the formation mechanisms of dynamic
membrane (cake) layer and to effectively control it for beneficial use of filtration.

2.9 Conclusion

AnMBR permeate may be used in agriculture or the nutrients in the permeate may be
recovered in order to get full benefit from the advantages offered by this technology.
Combination of membranes with different types of anaerobic high-rate reactor
configurations such as UASB and EGSB reactors, which have already been defined
as appropriate technologies for the treatment of dilute wastewaters, should be further
investigated. So far, integration with UASB reactors seems to be promising since
UASB reactors provide a pre-elimination of SS by entrapment and biodegradation in
the sludge bed. This may reduce the SS load to the membrane and limit the
membrane fouling due to cake layer formation. On the other hand, SMP and colloids
can be more important in fouling of the membranes in this type of configuration. As
a possible alternative, applying solid-liquid separation as a pretreatment and
digesting the concentrated slurry in an AnMBR can significantly reduce the initial
investment cost for membranes and bioreactor. For especially full-scale applications,
long-term reliability and operability of AnMBRs in municipal wastewater treatment
need to be further investigated.

48
3. EFFECT OF UPFLOW VELOCITY ON THE EFFLUENT MEMBRANE
FOULING POTENTIAL IN MEMBRANE COUPLED UPFLOW
(2)
ANAEROBIC SLUDGE BLANKET REACTORS

3.1 Introduction

The AnMBR is a potential process for treatment of municipal wastewater (Hu and
Stuckey, 2007; Vyrides and Stuckey, 2009; Martin et al., 2011). Research results
indicate the technology is feasible, particularly in moderate climate regions when
high effluent quality and/or agricultural use of treated effluents is required. The
performance of AnMBRs is affected by many factors including operational
parameters, sludge characteristics, membrane location and bioreactor
type/configuration (Liao et al., 2006). Possible configurations of AnMBRs include
CSTR, UASB and EGSB reactors, etc., coupled with a membrane module (Liao et
al., 2006; Ozgun et al., 2013a).

Research on AnMBRs, including membrane coupled UASB reactors, has received

considerable interest recently, leading to several studies investigating their potential
for the treatment of municipal wastewaters (Gao et al., 2010; Liu et al., 2012).
Extensive research to investigate the microbial communities in the UASB reactors
(Qui et al., 2013) and the biofouling layer on membranes (Calderon et al., 2011) has
also been conducted and identification of the species was proposed as an important
factor for biofouling control in membrane coupled UASB systems. Calderon et al.
(2011) analyzed the structure and diversity of the prokaryotic communities in the
biofouling layer of a membrane coupled UASB reactor and pointed out the
importance of particular microbial groups such as Sphingomonadaceae bacteria and

(2)
This chapter is based on:

Ozgun, H., Ersahin, M.E., Tao, Y., Spanjers, H., van Lier, J.B. (2013) Effect of upflow velocity on the
effluent membrane fouling potential in membrane coupled upflow anaerobic sludge blanket reactors.
Bioresource Technology 147, 285-292.

49
methanogenic Archaea in the occurance of biofouling. Due to the resistance of these
microbial groups to standard chemical cleaning, use of alternative antifouling
strategies was suggested as a promising approach for improving biofouling control
and prevention.

In membrane coupled UASB systems, the sludge bed at the bottom of the UASB
reactor acts as a biofilter prior to the membrane unit by entrapping most of the
particulate matter via adsorption and biodegradation (Seghezzo et al., 1998; Ozgun et
al., 2013a). Higher fluxes between 10.5-65 L/m2.h could be achieved with UASB
reactor coupled systems in comparison to CSTR coupled membrane systems (Wu et
al., 2009; An et al., 2009a). The membrane unit can be located either just above the
sludge bed in a UASB reactor or in a subsequent stage. In a membrane coupled
UASB system, the membrane is only exposed to the supernatant of the sludge bed.
Therefore, the fouling propensity of the effluent plays a major role in the stable
operation of such reactor systems (Martin-Garcia et al., 2011).

One of the important indicators of the effluent quality for membrane filtration
processes is filterability, which determines the loss of filtration performance over
time due to membrane fouling (Janssen, 2011). Good filterability, a requirement for
good performance and a trouble-free operation of AnMBRs, will lead to only a
limited increase in membrane resistance, and thus, a long-term operation at low
operational costs. Studies on AnMBRs show high variability in filterability (Spagni
et al., 2010) because it is influenced by many factors including the properties of the
feed water (Kang et al., 2003), membrane material/cake layer (Yun et al., 2006),
sludge characteristics (Zhang et al., 2010) and operational conditions such as shear
rate, SRT, temperature, HRT, operation mode and upflow velocity (Drews et al.,
2008).

For membrane coupled UASB systems, the efficiency of solids entrapment

determines the amount and properties of the solids leaving the UASB reactor with
the effluent. In this context, upflow velocity in membrane coupled UASB reactors
would seem to be the critical variable determining the reactor efficiency and effluent
fouling propensity (Mahmoud et al., 2003; Ozgun et al., 2013a). The applied upflow
velocity can have two opposing effects on the biological removal efficiency in
upflow reactors. An increase in upflow velocity may result in an increase in the
biological removal efficiency due to the increased interaction between the substrate

50
and the sludge. Conversely, high hydraulic shear force applied at high upflow
velocity may also deteriorate the removal efficiency due to detachment of the
captured solids (Mahmoud et al., 2003). Gonçalves et al. (1994) reported that SS
removal efficiency decreased from 70% to 51% when upflow velocity increased
from 0.9 m/h to 3.4 m/h in an upflow anaerobic reactor treating sewage. Moreover,
TSS and VSS removal efficiencies were found inversely proportional to the upflow
velocity in a UASB reactor treating settled sewage at moderate temperatures (van
Haandel and Lettinga, 1994; Seghezzo et al., 2002; Souza et al., 2006). Since upflow
velocity has a significant impact on the stable operation of UASB reactors, an
optimum upflow velocity should be applied in UASB reactors enabling the proper
distribution of gas pockets gathered in the sludge bed and providing a good contact
between substrate and biomass.

Until now, a large number of related scientific studies have focused on the effect of
upflow velocity on solely biological performance of the upflow reactors. However,
for the cases of membrane coupled UASB systems, there is no information in the
literature about its effect on filterability, a parameter that is closely related to
membrane fouling potential of the effluent. This study investigated this influence of
upflow velocity applied to a UASB reactor used for the treatment of municipal
wastewater. In addition to this general objective, the specific aim is to identify the
optimum upflow velocity that will result in an effluent with good filterability values
for the case of membrane coupled UASB systems. The effect of upflow velocity on
the effluent quality was assessed not only on biological based parameters such as
organic matter removal but also on physical characteristics. Possible relationships
between physical characteristics and filterability of the effluent were determined. The
information obtained from this study is valuable for optimizing the hydraulic
conditions in membrane coupled UASB systems in order to achieve a stable effluent
quality and to minimize membrane fouling potential.

3.2 Materials and Methods

3.2.1 Wastewater source

Synthetic municipal wastewater was used as feed. Macronutrient and micronutrient

solution compositions were slightly modified from the recipes in the studies of Aiyuk
and Verstraete (2004) and Martin et al. (2010), respectively. The feed solution was

51
prepared in concentrated form and stored at 4 °C. It was always thoroughly mixed
and diluted with tap water before being fed into the reactor. The composition of the
concentrated substrate solution and the characterization of the synthetic municipal
wastewater at the UASB inlet are presented in Table 3.1 and Table 3.2, respectively.

Table 3.1 : Composition of the concentrated synthetic municipal wastewater.

Macronutrient Solution Micronutrient Solution
Compound Unit Value Compound Unit Value
Urea mg/L 1200 FeCl3.6H2O mg/L 1000
NH4Cl mg/L 2000 CoCl2.6H2O mg/L 1000
CH3COONa.3H20 mg/L 7400 MnCl2.4H2O mg/L 250
Ovalbumin mg/L 450 CuCl2.2H2O mg/L 15
MgSO4.7H2O mg/L 180 ZnCl2 mg/L 25
KH2PO4.3H2O mg/L 1400 H3BO3 mg/L 25
CaCl2 mg/L 264.9 (NH4)6Mo7O24.4H2O mg/L 45
Starch mg/L 6400 Na2SeO3.H2O mg/L 50
Milk Powder mg/L 1500 NiCl2.6H2O mg/L 25
Yeast Extract mg/L 600 EDTA mg/L 500
Sunflower Oil mg/L 1000 HCl 36% mL/L 0.5
Micronutrients mL/L 26.6 Resazurin Sodium Salt mg/L 250
Yeast Extract mg/L 1000

Table 3.2 : Characterization of the synthetic municipal wastewater at the UASB

inlet.
Parameter Unit Concentration
(Average±Range)
COD mg/L 530±30
Soluble COD mg/L 159±25
Total nitrogen (TN) mg/L 54±5.2
NH4-N mg/L 36±5.5
TP mg/L 12±0.8
Alkalinity mg CaCO3/L 300±40
TSS mg/L 230±25
Turbidity NTU 75±15
pH - 7.4±0.2

3.2.2 Seed sludge

The reactor was seeded with flocculent anaerobic sludge obtained from a pilot-scale
UASB reactor treating black water (Sneek, the Netherlands). The characteristics of
the seed sludge are presented in Table 3.3.

52
 Table 3.3 : Characteristics of the seed sludge.

Parameter Unit Value

TS mg/L 22000±300
Volatile Solids (VS) mg/L 16900±235
TSS mg/L 20200±75
VSS mg/L 16900±225
COD mg/L 27100±330
pH - 7.9
Capillary Suction Time s 285
SMA g CH4-COD/g VS.d 0.3

3.2.3 Experimental setup

The experiments were performed using a laboratory-scale UASB reactor (Figure 3.1)
with an effective volume of 7 L. The system was equipped with feed, recycle and
effluent pumps (Watson Marlow 120U/DV), pH and temperature sensors (Elscolab,
M300 ISM), and gas meter (Ritter, Milligas Counter MGC-1 PMMA). A three-phase
separator (Figure 3.2) was installed at the top part of the reactor to separate the
biogas from the mixed liquor as well as to retain suspended particles in the reactor.
The reactor was equipped with sampling ports at different heights for sample
withdrawal. During the experiments, pH was in the range of 6.8 to 7.1. The
temperature of the jacketed reactor was controlled at 25±2 °C by a thermostatic water
bath (Tamson Instruments, the Netherlands). The experimental system was
connected to a computer running LabView software in order to control all pumps and
collect pH, temperature and biogas flow data on-line.

3.2.4 Experimental procedure

Before operation, the seed sludge was acclimated in the UASB reactor at 0.6 m/h for
40 days to obtain steady biogas production and stable sludge bed height. During the
study, the UASB reactor was continuously fed for a total of 116 days at different
upflow velocity conditions. Upflow velocities of 1.2 m/h, 0.6 m/h, and 1.2 m/h were
applied successively for periods of 33, 52 and 31 days, respectively. The third
experimental period, where an upflow velocity of 1.2 m/h was reestablished after the
period with 0.6 m/h, was included to verify whether the effect of upflow velocity on
effluent quality in the laboratory-scale UASB reactor was reproducible. During entire
experiment, HRT was fixed at 6 h, and the corresponding OLR was 2 kg

53
COD/m3.day. Total COD, soluble COD, TSS, turbidity, TN, NH4+-N and TP
analyses were performed on the influent and the effluent samples. TSS analysis was
also performed on samples taken from the middle of the upper phase of the UASB
reactor over the sludge bed, which is referred to as supernatant sample. SMP, PSD
and filterability analyses were performed on the effluent samples.

Figure 3.1 : Schematic diagram of the UASB reactor.

Figure 3.2 : A three-phase separator.

3.2.5 Analytical methods

COD, NH4+-N, TN, TP, TSS, VSS, TS, VS and alkalinity were measured according
to Standard Methods (APHA, 2005). Each analysis was performed in triplicate. The
samples for soluble COD measurement were filtered through a 0.45 µm filter. The
turbidity and pH were measured by HACH 2100N Turbidimeter and by a WTW

54
multi720 pH meter, respectively. The PSD of the effluent was assessed by a
Mastersizer 2000 (Malvern Instruments, Hydro 2000 MU) which has a detection
range of 0.02-2000 µm. Laser diffraction technique was used in order to measure the
size of the particles. A volume of 5 ml effluent was freshly sampled, filtered through
a 0.45 µm filter and diluted 4 times with phosphate buffered saline (PBS) (pH=7.2)
for SMP analyses. The phenol-sulfuric acid method was used to quantify
polysaccharides (Dubois et al., 1956). The concentration of protein was measured
using the Bradford method (Bradford, 1976). The methane content in the biogas was
measured using a Varian 3800 gas chromatograph equipped with a flame ionization
detector (Varian, Palo Alto, CA). Statistical calculations were processed by using
Minitab 16 software (Minitab Inc., USA).

3.2.6 Filterability and reversibility test

The Specific Ultrafiltration Resistance (SUR) test, developed at Delft University of

Technology (Roorda, 2004), was used to test the filterability of the UASB effluent.
SUR is defined as the filtration resistance per unit of filtered feed water per m 2
membrane area and it is calculated from the slope of a filtration curve, i.e. ratio of
filtration time and filtered volume vs. total volume of the filtered water, that is
measured within 30 min of filtration over an UF membrane at a constant temperature
and at a constant TMP (Janssen et al., 2008). Low SUR values indicate relatively
high filterability, whereas high SUR values indicate the poor filterability.

Reversibility is described as the extent to which the filtration resistance is returned

back to the original value after applying a hydraulic cleaning. The reversibility of the
fouling layer is calculated from the ratio between the change in resistance and time
caused by the cleaning of the membrane (ΔR/Δt).

The equipment (Figure 3.3) for the filterability and reversibility tests consisted of
three pressurized buffer tanks (A, B and C) with a volume of 10 L. Pressurized buffer
tank A was filled with demineralized water and pressurized buffer tank B was filled
with UASB reactor effluent. These pressurized buffer tanks were connected to the
membrane module by means of a three-way valve in between. The membrane
module contained three capillary UF membranes (Norit X-Flow) prepared from PES
and polyvinylpyrrolidone (PVP) with internal diameters of 0.8 mm and MWCOs of
150–200 kDa. The permeate is collected in the pressurized buffer tank C. To

55
maintain a TMP of 0.5 bar through the membrane module, two pressure control
devices were installed to keep the pressures of buffer tank A and B under 1.5 bar and
the pressure of buffer tank C under 1.0 bar. A digital pressure difference measuring
device was placed to check and log the TMP over the membrane module. The
permeate flow was continuously measured by a mass flowmeter. A data acquisition
program was used to store and display the TMP, flux, temperature and resistance.

Figure 3.3 : Schematic diagram of the SUR measurement experimental setup (A:
demineralized water vessel; B: wastewater vessel; C: magnetic stirrer; D: three-way
valve; E, H, I, K: valves; F: flow meter; G: membrane module; J: computer with
measuring program; L: permeate vessel; P: pressure difference sensor).

56
3.3 Results and Discussion

3.3.1 Effect of upflow velocity on system performance

Total and soluble COD concentrations in the effluent at different upflow velocities
are shown in Figure 3.4. After the 40 day acclimation period at an operating upflow
velocity of 0.6 m/h, the measuring period started at Day 0 with an increased upflow
velocity of 1.2 m/h. Average total and soluble COD concentrations during this first
period were 215±10 mg/L and 63±9 mg/L, respectively. After the decrease in upflow
velocity from 1.2 m/h to 0.6 m/h in the second period, a gradual decrease was
observed in total COD concentrations to the range of 130-160 mg/L. Soluble COD
concentrations were found to be similar at both upflow velocities. Average total
COD removal efficiencies of 60% and 72% were obtained at upflow velocities of 1.2
m/h and 0.6 m/h, respectively.

Upflow velocity was set to 1.2 m/h again at the start period on Day 88, in order to
determine if the effect of upflow velocity on total COD concentration is
reproducible. During this period, an increase in total COD concentration was
observed again, similar to the first period of 1.2 m/h. The observed increase in total
COD concentrations at an upflow velocity of 1.2 m/h may be either attributed to the
release of colloidal and suspended particles from the sludge bed or the decrease of
hydrolysis rate.

Figure 3.4 : Total and soluble COD concentrations in the effluent.

57
The difference in biogas production (Figure 3.5) observed during each period with
different upflow velocities was found to be insignificant (p>0.05). Average methane
content of the biogas was 61±5.7% throughout the study, and methane yields of
0.13±0.02 L CH4/g CODremoved and 0.11±0.01 L CH4/g CODremoved were achieved at
upflow velocities of 1.2 m/h and 0.6 m/h, respectively. The almost stable biogas
production throughout the whole study in spite of the larger variations in total COD
removal efficiency indicated that upflow velocity influences the release of colloidal
and suspended particles from the sludge bed rather than the biodegradation of COD
in the reactor.

Figure 3.5 : Biogas production at different upflow velocities.

Operation at a higher upflow velocity may promote the release and washout of
colloidal or particulate organic matter that would reside in the sludge bed at lower
upflow velocity. In order to confirm this assumption, TSS and turbidity were
measured both in the supernatant and effluent of the UASB reactor. A considerable
decrease from the range of 230-270 mg/L to 90-110 mg/L was observed in the TSS
concentrations of the supernatant when upflow velocity was decreased from 1.2 m/h
to 0.6 m/h. However, there was no significant change (p>0.05) in TSS concentrations
in the effluent as a result of the upflow velocity change: Average TSS concentrations
in the effluent were 91±29 mg/L and 81±13 mg/L at 1.2 and 0.6 m/h upflow
velocities, respectively (Figure 3.6). This finding may be attributed to the effective
separation capability of the GLS separator at the top of the UASB reactor. With the
increase in upflow velocity, particles present in the supernatant will rise to the top

58
but are prevented from rinsing with the aid of GLS separator. The average turbidity
in the effluent was found to be 55±5 NTU at the upflow velocity of 1.2 m/h, whereas
it decreased to 32±4 NTU at the upflow velocity of 0.6 m/h (Figure 3.7). Variations
observed in effluent turbidity (p<0.05) at different upflow velocities seem
contradictory to TSS results which do not show significant differences between each
upflow velocity. Observed results are attributed to the washout of colloidal particles
from the system instead of suspended particles as turbidity is an indicator of both
suspended and/or colloidal particles.

Figure 3.6 : TSS concentration in the supernatant and effluent.

Figure 3.7 : Turbidity of the effluent.

59
NH4+-N, TN (Figure 3.8) and TP (Figure 3.9) concentrations were also measured in
the UASB reactor effluent at different upflow velocities. TN mainly i.e. >75%
consisted of ammonium. The ratio of NH4+-N/TN in the effluent was higher at the
upflow velocity of 0.6 m/h than the one obtained at 1.2 m/h, indicating again the
washout of colloidal protein-like materials at higher upflow velocity. TP
concentration was stable between 10.9-12.9 mg/L throughout the whole study period.

Figure 3.8 : TN and NH4-N concentrations in the effluent.

Figure 3.9 : TP concentrations in the effluent.

60
The presence of SMPs is considered an important factor which can play a role in
membrane fouling by causing cake or gel layer compacting or even pore clogging
(Jarusutthirak and Amy, 2006; Meng et al., 2009; Huang et al., 2011; Zhang et al,
2011). Aquino et al. (2006) found that SMP was the main contributor of pore
clogging in AnMBRs, hence, SMP is likely have a negative impact on the
filterability of the UASB effluent. Fouling due to SMP and colloids may be relatively
more important in membrane coupled UASB systems than in membrane coupled
CSTR systems because of selective accumulation of SMPs on the membrane in the
former system (Ozgun et al., 2013a). In order to investigate the impact of SMP carry
over at different upflow velocity, the concentrations of protein and polysaccharide
were measured in the effluent (Figure 3.10). The protein concentration significantly
decreased at the upflow velocity of 0.6 m/h and was under the detection limit (2.5
mg/L) in the samples obtained from Day 54 to Day 85, indicating that at low upflow
velocity protein may be better metabolized or retained inside the reactor than at
higher upflow velocity. This was supported by the observation that protein
concentration again increased after upflow velocity was raised back to 1.2 m/h.
Unlike the effluent protein concentration, effluent polysaccharide concentration was
almost unaffected by the upflow velocity.

Figure 3.10 : SMP concentrations in the effluent.

Meng et al. (2006) indicated that P/C is an important indicator of the fouling
potential of an effluent fed to a membrane. They showed that proteins adhere more
easily to a membrane surface and are more likely to induce membrane fouling as

61
they are more hydrophobic than carbohydrates. Therefore, a high P/C would indicate
the probability of more severe membrane fouling. Huang et al. (2011) also found a
direct relationship between P/C and fouling. Due to the decrease in protein
concentration at lower upflow velocity shown in this study, the P/C of the effluent
decreased and thus, less fouling may be expected in a membrane coupled UASB
reactor by applying at low upflow velocity in the sludge bed.

Figure 3.11 presents PSD curves in the effluent at different upflow velocities. The
effluent at the upflow velocity of 0.6 m/h was more homogenous in terms of PSD
showing more unevenly distributed peaks (Figure 3.11-b), whereas the PSD at an
upflow velocity of 1.2 m/h was more heterogeneous showing more and lower peaks.

At an upflow velocity of 1.2 m/h, there was a shift to the left part of the distribution
pattern and an apparent decrease in the median particle size (D50) was observed,
whereas at an upflow velocity of 0.6 m/h peaks shifted to the right part of the PSD
curve and an increase in D50 was observed. From these results, it was concluded that
at the higher upflow velocity, small sized particles were washed out more vigorously,
resulting in an effluent having a D50 of 4.9 μm, whereas a D50 of 108 µm was
measured after the operation period of 0.6 m/h.

62
Figure 3.11 : PSD at different upflow velocities (a) v=1.2 m/h, (b) v=0.6 m/h, (c) v=1.2 m/h.

63
3.3.2 Filterability and reversibility at different upflow velocities

In order to investigate the relationship between filterability and effluent

characteristics at different upflow velocities, the SUR of samples obtained at
different upflow velocities was measured (Figure 3.12). Figure 3.12 shows that the
SUR values varied between 50.1012 m-2 and 300.1012 m-2 over the whole
experimental period. Increased SUR values with an increasing trend were obtained
during the operation at 1.2 m/h, whereas a decreasing trend was observed during the
operation at 0.6 m/h, which was an indicator of better filterability at lower upflow
velocity. Filterability deterioration at high upflow velocity was related to the
existence of colloidal material and a heterogeneous PSD during the operation at 1.2
m/h. This means that at the upflow velocity of 0.6 m/h, the average filterability of the
UASB effluent was better and more appropriate for membrane filtration in
comparison to the effluent obtained at 1.2 m/h.

Figure 3.12 : SUR and ΔR/Δt values at different upflow velocities.

Reversibility of the UASB effluent was also measured at different upflow velocities
in order to determine the extent to which filtration resistance is returned back to the
original value after applying a hydraulic cleaning (Figure 3.12). A high value for
ΔR/Δt indicates a low reversibility, whereas a high reversibility is indicated by a low
value for ΔR/Δt. The reversibility and SUR values of the UASB effluent were
reciprocally correlated. This relation is also endorsed by Roorda (2004). The increase
in upflow velocity from 0.6 m/h to 1.2 m/h resulted in a decreased reversibility,

64
which can be an indicator of unsteady process performance of the membrane unit
after prolonged periods of operation.

3.4 Conclusion

Upflow velocity is an important variable which affects the effluent characteristics of

a UASB, especially the physical characteristics and PSD. Results suggest that when
applying a high sludge bed upflow velocity in a UASB coupled membrane system,
there will be an increased load of small particles to the membrane. Effluent
characterization results coincide with the filterability and reversibility tests. The
increased effluent particle size seen at the lower upflow velocity was associated with
a better filterability. Overall the results of this study clearly indicate that selection of
an appropriate upflow velocity plays a major role in controlling the fouling potential,
which is one of the most important challenges for stable operational process
performance of membrane coupled UASB reactors.

65
66
4. IMPACT OF MEMBRANE ADDITION FOR EFFLUENT EXTRACTION
ON THE PERFORMANCE AND SLUDGE CHARACTERISTICS OF
UPFLOW ANAEROBIC SLUDGE BLANKET REACTORS (3)

4.1 Introduction

Recently, there has been a growing interest in increasing cost-efficiency and overall
sustainability of municipal wastewater treatment systems in addition to their goal to
protect receiving surface water and human health. As a result, municipal wastewater
is more and more considered as a source of energy, nutrients and water (Mo and
Zhang, 2012). When water reclamation for agricultural reuse is targeted,
conventional aerobic and anaerobic treatment schemes are being upgraded to
mainstream processes that are preferably capable of producing nutrient-rich and
solids-free effluents with a high retention of pathogens (Saddoud et al., 2006).

With growing concerns over climate change associated with fossil-fuel utilization,
the anaerobic treatment of municipal wastewater is receiving increased attention.
Compared with the widely used aerobic biological process for municipal wastewater
treatment, the anaerobic biological process has significant advantages including no
energy input required for aeration and energy recovery through methane production,
lower sludge production and most nutrients (nitrogen and phosphorus) remaining in
the effluent suitable for agricultural and landscaping irrigation reuse. Despite these
advantages, the perceived notion is that an anaerobic biological process is
inappropriate for municipal wastewater treatment with relatively low concentration
of organics and a significant particulate fraction, which make it technically

(3)
This chapter is based on:

Ozgun, H., Gimenez, J.B., Ersahin, M.E., Tao, Y., Spanjers, H., van Lier, J.B. (2015) Impact of
membrane addition for effluent extraction on the performance and sludge characteristics of upflow
anaerobic sludge blanket reactors treating municipal wastewater. Journal of Membrane Science 479,
95-104.

67
challenging to achieve a high OLR with good effluent quality (Wei et al., 2014).
However, within the alternatives to consider, the combination of anaerobic biological
treatment with membranes stands out as a promising alternative that may overcome
the drawbacks of the conventional anaerobic processes for municipal wastewater
through the excellent sludge-water separation by a membrane, which results in high
biomass and high OLR and organics removal and high quality effluent with lower
amount of particles and pathogens. Municipal wastewater is found to be a potential
source of energy (McCarty et al., 2011). AD of wastewater can produce methane gas,
which can be used by methane-driven engine to generate electricity. Anaerobic
processes not only produce methane as a renewable source of bioenergy but also
consume less energy for operation than aerobic systems. Complete anaerobic
treatment of municipal wastewater has the potential to achieve net energy production
while meeting stringent effluent standards (Dutta et al., 2014). In addition, the lower
anaerobic waste biosolids production compared with aerobic treatment reduces the
costs and difficulties associated with biosolids management (Yoo et al., 2012). In
anaerobic sewage treatment systems, the plant macronutrients ammonium and
orthophosphates are solubilized from organic matter, reaching effluent
concentrations even higher than influent values. In addition, a superior effluent
quality, in terms of COD, SS and pathogen counts, can be achieved by incorporating
membranes to the anaerobic sewage treatment systems (Brindle and Stephenson,
1996; Liao et al., 2006). The main disadvantage of anaerobic biological treatment
with membranes is high energy usage to reduce the fouling problem. In order to
reduce energy costs for membrane fouling control, UASB reactor has been proposed.
Coupling of UASB with membrane reactor was found to potentially reduce the
membrane energy cost. To date, the UASB reactor is the most commonly applied
technology for anaerobically treating municipal sewage (Chernicharo, 2007; Noyola
et al., 2012). UASB effluent extraction via membrane skits offers interesting
potentials, since membrane separation is then challenged by solely sludge
supernatant, due to the entrapment of most of the particulate matter by adsorption
and biodegradation in the sludge bed. As a consequence, membrane fluxes may
become less dependent on the reactor MLSS concentration, possibly leading to high
membrane fluxes compared to AnMBR consisting of CSTRs (Ozgun et al., 2013a).
Besides that, Chu et al. (2005), as the sole example in literature, proposed a
membrane unit that is capable of retaining the SS inside the EGSB reactor. In that

68
study, the membrane module is submerged in the upper part of the EGSB reactor and
the applied high upflow velocity was beneficial in the reduction of membrane fouling
due to the increased hydraulic shear stress on the membrane surface (Chu et al.,
2005). Nonetheless, cake layer resistance governed the achievable flux over the
hollow fiber membranes, whereas pore blocking only was of minor importance.
Based on the high removal efficiencies even at low temperatures, the use of EGSB
reactors in AnMBRs was also suggested as a potential technology for treating
municipal wastewater at ambient temperature (Chu et al., 2005). However,
membrane integration eliminates the hydraulic selection pressure required for
granulation, by avoiding the washout of flocculent sludge with poor immobilization
characteristics. Therefore, no granulation is expected in EGSB reactors coupled to
membrane filtration, which would decrease the settleability of the biomass on the
long-term operation.

A number of studies have been published assessing the performance of membrane

coupled UASB reactors for the treatment of municipal wastewater (Gao et al., 2010;
Herrera-Robledo et al., 2010; Salazar-Pelaez et al., 2011a; Salazar-Pelaez et al.,
2011b; Liu et al., 2012). In these studies, membranes have been coupled with UASB
reactors in two different configurations: one employs membranes as a part of an
AnMBR system, and the other uses the membrane unit as a polishing step after the
UASB reactor. In most of the studies (Herrera-Robledo et al., 2010; Herrera-Robledo
et al., 2011; Salazar-Pelaez et al., 2011a; Salazar-Pelaez et al., 2011b), membranes
have been applied as a polishing step after UASB reactors. In such setup, the
concentrate stream does not return to the bioreactor, preserving the liquid dilution
rate as a bacterial selection pressure factor. In such setup, the hydraulics in the
UASB reactor are not altered. However, the down-side of the approach is SS
accumulation in the membrane tank located after the UASB reactor, leading to a
progressive increase in SS loading on the membrane unit. Only few studies (Gao et
al., 2010; Liu et al., 2012) have employed membranes as a part of an AnMBR system
with concentrate flow recycle to the UASB reactor. In this case, concentrate flow
recycle leads to nearly absolute biomass retention and allows for operation at high
SRTs. Jeison et al. (2008) demonstrated the feasibility of AnMBRs including CSTRs
for enhanced treatment of wastewaters containing particulate organic matter because
they offer the possibility of TS retention. However, for UASB-based AnMBR

69
systems, the membrane barrier eliminates the hydraulic selection pressure required
for a sludge bed with good settling properties in the UASB reactor. Additionally, the
usually high concentration of partly degradable particulate matter in municipal
wastewater, which can be entrapped in the sludge bed, may further hamper the
applicability of UASB reactors in an AnMBR system configuration.

Membranes in UASB-based AnMBR systems are either located as side-stream (Gao

et al., 2010) in which the concentrate is recycled back to the bioreactor or submerged
(Liu et al., 2012). Gao et al. (2010) operated an AnMBR system consisting of side
stream membranes coupled with a UASB reactor and focused on the relative
abundance comparison of microbial species on the cake layer and biomass
suspension. On the other hand, Liu et al. (2012) carried out a study in which the
impact of food to microorganism ratio (F/M) on system performance was evaluated
by using two AnMBR systems with membrane units placed at the top of the UASB
reactor. However, neither of these studies investigated the impact of membrane
addition on the performance of UASB reactor by comparing the system performance
before and after membrane addition.

The goal of this study is to investigate the impact of membrane addition on both the
system performance and sludge characteristics of a UASB reactor in an AnMBR
system treating municipal wastewater. This impact is evaluated by monitoring and
comparing several indicators including COD removal efficiency, TSS and SMP
concentrations in UASB effluent and sludge PSD, sludge activity and microbial
community analysis before and after membrane addition. This study is supposed to
provide a fundamental understanding of how a UASB system is affected by the
elimination of selective sludge washout, as well as to give a clue on the operation of
a UASB reactor as part of an AnMBR system.

4.2 Materials and Methods

4.2.1 Wastewater source

In the present study, synthetic municipal wastewater was used as feed. Wastewater is
the same as the one described in Chapter 3 for synthetic municipal wastewater. The
composition of the concentrated substrate solution and the characterization of the

70
synthetic municipal wastewater at the UASB inlet were already presented in Chapter
3 in Table 3.1 and Table 3.2, respectively.

4.2.2 Seed sludge

The reactor was seeded with flocculent anaerobic sludge obtained from a pilot-scale
UASB reactor treating black water (Sneek, the Netherlands). The characteristics of
the seed sludge were presented in Table 3.3 in Chapter 3.

4.2.3 Experimental setup

The experimental setup consisted of a laboratory-scale UASB reactor with an

effective volume of 7 L coupled with an external membrane module. The
experimental period was divided into two stages: (i) before membrane incorporation,
when the UASB reactor was operated as the sole process and (ii) after membrane
incorporation, when the system was operated as an AnMBR. Before membrane
incorporation (Figure 4.1(a)), the system was equipped with feed, recycle pump and
effluent pump (Watson Marlow 120U/DV). After membrane incorporation, the
AnMBR stage was initiated by directing the UASB effluent through an external
cross-flow tubular module containing 28 UF (0.03 µm pore size) membrane fibers
(Pentair X-Flow) (Figure 4.1(b)), and returning the concentrate to the UASB reactor.
Each membrane fiber had an internal diameter of 1.5 mm and a length of 80 cm, with
an effective filtration area of 0.0038 m2. PES UF membranes with a pore size of 30
nm were used in the AnMBR. Peristaltic pumps (Watson Marlow 120U/DV) were
used for circulating the flow between the UASB reactor and the membrane module,
and obtaining permeate through the membrane. Furthermore, recirculation over the
membrane surface was applied by a peristaltic pump (Watson Marlow 620UN/R) in
order to maintain a cross-flow velocity of 1 m/s across the membrane surface
independently of the circulation flow between the UASB reactor and the membrane
module. This concentrate circulation flow rate was adjusted enabling a stable upflow
velocity in the UASB reactor. Pressure sensors (AE Sensors, ATM -800/+600 mbar,
the Netherlands) were installed in the membrane feed, concentrate and permeate
lines in order to measure the TMP. The biogas was collected by means of a three-
phase separator installed at the top of the UASB reactor. Biogas flow rate was
measured with a gas meter (Ritter, Milligas Counter MGC-1 PMMA). Temperature
was controlled by means of a water bath (Tamson Instruments, the Netherlands).

71
Temperature and pH inside the bioreactor were monitored on-line with a probe
combined with a transmitter (Elscolab, M300 ISM). A computer accessible via
LabView software (LabVIEW 10.0.1, National Instruments) was used to control the
pumps and collect the data.

Figure 4.1 : The layouts of the experimental setups (a) UASB reactor and (b)
AnMBR system.

4.2.4 Experimental procedure

The laboratory-scale experimental study was conducted during 126 days: the UASB
operation stage before membrane incorporation lasted 63 days, and was followed by
the AnMBR operation stage after the addition of membrane, which also lasted 63
days. The system was operated at an average HRT of about 6 h and an average OLR
of 2 kg COD/m3.day at each stage. The upflow velocity of the UASB reactor was
chosen as 0.6 m/h based on the results of Chapter 3 (Ozgun et al., 2013b) and kept
constant during the whole experimental period. The average pH and temperature of
the system remained stable around 6.9 and 25 °C, respectively. Membrane operation
consisted of alternating between 3 minutes filtration and 20 seconds backwash, at a
membrane flux of 12.3 L/m2.h during the AnMBR operation stage. The cross-flow
velocity of the membrane was 1 m/s. Backwash was applied by reversing the flow
direction, keeping the flux at the same value as during filtration. Membrane cleaning
including physical, base and acid cleaning steps was also carried out to recover the
permeability. During membrane cleaning, the membranes were first physically

72
cleaned by flushing with tap water to remove cake layer. Following that, the
membranes were chemically cleaned by soaking sequentially in 200 ppm NaClO-
0.4% Divos solution for 30 minutes and 1% (w/v) citric acid solution for 1 hour.

4.2.5 Analytical methods

Frequent analyses were performed to assess the characteristics of the wastewater,

effluent, permeate and sludge.

4.2.5.1 Experimental methods applied to effluent and permeate samples

COD and TSS concentrations were determined according to Standard Methods

(APHA, 2005). Each analysis was performed twice a week in triplicate. The samples
for soluble COD measurement were filtered through a 0.45 µm filter before analysis.
VFA concentrations were measured weekly using a Focus GC (Thermo electron
corporation) connected with flame ionization detector. A 30 m-long column (Hewlett
Packard HP INNOWAX) with an internal diameter of 0.25 mm and a film thickness
of 0.25 µm were used to separate VFAs. For SMP analysis, a volume of 5 ml effluent
was freshly sampled weekly, filtered through a 0.45 µm filter and diluted 4 times
with PBS (pH=7.2). The phenol-sulfuric acid method was used to quantify
polysaccharides (Dubois et al., 1956). The concentration of protein was measured
using the Bradford method (Bradford, 1976).

4.2.5.2 Experimental methods applied to sludge samples

PSD of the sludge was determined monthly by a Mastersizer 2000 (Malvern

Instruments, Hydro 2000 MU), which has a detection range of 0.02-2000 µm. Laser
diffraction technique was used to measure the size of the particles. For assessing the
SMP content of sludge samples, a volume of 5 ml sludge was washed by PBS
(pH=7.2) and centrifuged at 7000g for 7 minutes at 4 °C. The supernatant was
filtered using a 0.45 µm filter and the filtrate was collected for SMP analysis. The
pellet was re-suspended and vigorously washed with 10 ml PBS and then
ultrasonicated at 40 kHz (Cole-Parmer Ultrasonic, the Netherlands) for 3 minutes. A
high speed centrifuge (17000g) was applied for 20 minutes under 4 °C and the
supernatant obtained was filtered using a 0.45 µm filter for EPS assessment. The
phenol-sulfuric acid method was used to quantify polysaccharides (Dubois et al.,

73
1956). The concentration of protein was determined using the Bradford method
(Bradford, 1976).

SMA of the sludge was determined monthly by using an Automated Methane

Potential Test System (AMPTSII, Bioprocess Control, Sweden) (Li et al., 2011a).
The Automated Methane Potential Test is characterized by a high sensitivity
allowing on-line measurement of very low biomethane flows during the SMA test.
This test system works with the same principles as conventional methane potential
tests. The advantage of Automated Methane Potential Test over the conventional
tests is that the gas volume measurements and data logging are fully automatic
during the methane potential test (Figure 4.2).

Figure 4.2 : AMPTSII (1- Data storing, 2- Glass reactors with mixing and water
bath, 3- CO2 fixation unit, 4- Flow cell with methane production).

SMA tests were performed at 35 °C. Both samples and blanks were analyzed in
triplicate. SMA tests were carried out in 500 mL serum bottles (with a working
volume of 400 mL), which were filled with sludge and growth medium. The bottles
were sparged with nitrogen gas to remove oxygen from the headspace. The growth
medium for the blanks consisted of a mixture of macronutrients (6 mL of the stock
solution, per liter of growth medium), trace elements (0.6 mL of the stock solution,
per liter of growth medium) and phosphate buffer solution (10 mM). The growth
medium for the samples additionally included 0.5 g COD/L as sodium acetate.

The nutrient stock solution consisted of (g/L): NH4Cl (170), CaCl2.2H2O (8),
MgSO4.7H2O (9) and the trace element stock solution contained (g/L): FeCl 3.4H2O
(2), CoCl2.6H2O (2), MnCl2.4H2O (0.5), CuCl2.2H2O (30), ZnCl2 (50), H3BO3 (50),
(NH4)6Mo7.O2.4H2O (90), Na2SeO3.5H2O (100), NiCl2.6H2O (50), EDTA (1), HCl

74
36% (1 mL/L), Resazurine (0.5). The pH buffer stock solution was composed of
K2HPO4.3H2O (45.65 g/L) and NaH2PO4.2H2O (31.20 g/L). The sludge and growth
medium amounts were determined by setting a sludge VS amount to substrate COD
amount ratio of 2:1.

Stability of sludge samples was tested monthly by obtaining the specific ultimate
methane production. Stability tests were carried out in serum bottles with a volume
of 120 mL. The anaerobic medium was prepared by dissolving 3.5 g/L sodium
bicarbonate (NaHCO3) in tap water without the addition of extra nutrients. The
bottles were sealed with butyl rubber stoppers and aluminum caps. The head space
was flushed with N2:CO2 (70:30%) mixture. The sludge was kept in the bottle at 35
°C without substrate addition in order to degrade the accumulated particulate matter
in the sludge. During the test, the anaerobically degradable compounds were
converted to biogas. Biogas production was monitored in terms of pressure increase
with a pressure transducer (Centre Point Electronics PSI-30). Experiments lasted till
the cumulative biogas production reached to a plateau. Methane content in the biogas
was measured. The stability of the sludge was calculated as the ultimate methane
COD produced divided by the sludge COD. Therefore, a high value indicates that a
sludge is poorly stabilized and a high amount of anaerobic biodegradable organic
compounds is still present and vice versa.

Microbial community structure was investigated using 454-pyrosequencing. UASB

sludge samples were taken before and after membrane addition. All samples (seed
sludge and UASB sludge) were washed twice with PBS and then centrifuged at
10000g for 3 minutes. The supernatant was removed in order to reduce biomass
decay during sludge storage. All samples were stored at -25 °C until
deoxyribonucleic acid (DNA) extraction.

16S recombinant deoxyribonucleic acid (rDNA) of sludge samples were extracted

using a MoBio UltraClean Microbial DNA isolation kit (MoBIO Laboratories, Inc.,
CA, USA) following the protocol suggested by the manufacturer. In order to increase
the efficiency in DNA isolation process, a combination of heat, detergent and
mechanical force was applied. A minor modification including twice bead-beating (5
minutes) and heating (65 °C, 5 minutes) was applied to the protocol in sequence in
order to enhance the lysis efficiency of microbial cells. DNA isolation was confirmed

75
by agarose gel electrophoresis and the concentration of DNA was measured using
Nanodrop 1000 equipment (Thermo Scientific, Waltham, MA, USA).

Pyrosequencing of the 16S rDNA gene was performed at Research and Testing
Laboratory (Lubbock, TX, USA) using a Roche454 GS-FLX system (454 Life
Science, Branford, CT, USA) with titanium chemistry. Universal primers U515F
(GTG CCA GCM GCC GCG GTA A) and U1071R (GAR CTG RCG RCR RCC
ATG CA) (Wang and Qian, 2009) were used. By testing on Ribosomal Database
Project (RDP) (Maidak et al., 1997), both forward and reverse primers target over
90% bacterial and archaeal 16S rDNA.

Post-processing and analysis of pyrosequencing results, including chimera removal,

taxonomic classification and microbial diversity calculation (Chao1, Shannon index
and observed species), were performed by combining different programs from the
Quantitative insights into microbial ecology (QIIME) pipeline, version 1.6.0
(Caporaso et al., 2010).

4.3 Results and Discussion

4.3.1 Impact of membrane incorporation on UASB effluent

Figure 4.3 shows the total and soluble COD concentrations in the effluent of the
UASB reactor before and after membrane incorporation. The average total COD
concentration in the effluent before membrane incorporation was 149±5.9 mg/L.
After the membrane was coupled with the UASB reactor, total COD concentration in
the effluent increased dramatically, reaching concentrations up to 2420 mg/L at the
initial AnMBR operation stage. Afterwards, the effluent total COD concentration
decreased and remained stable within the range of 750-1000 mg/L, indicating that a
steady state condition was achieved. The increase in total COD concentration after
membrane incorporation can be either attributed to the observed increased solids
washout, indicating that significant quantities of sludge could not be retained in the
sludge bed, or to the accumulation of intermediate products of the anaerobic
degradation process. The latter might result from a biological process imbalance,
allowing a part of influent COD to leave the UASB reactor before being completely
degraded to the end products.

76
An increase in soluble COD concentration from 49±6.8 mg/L to 122±9.8 mg/L in the
UASB effluent was also observed after membrane incorporation (Figure 4.3). The
soluble fraction of the effluent might contain residual degradable and non- or slowly
biodegradable influent substrate, SMP and intermediate products such as VFA.
Therefore, this increase could be either related to soluble products of microbial origin
or residual substrate and substrate degradation intermediates.

Figure 4.3 : Total and soluble COD concentrations in the UASB effluent and final
permeate.

TSS was also measured both in the supernatant and effluent of the UASB reactor
(Figure 4.4). Samples taken from the middle of the upper section above the sludge
bed of the UASB reactor are referred as supernatant (Figure 4.1). As a result of
membrane incorporation, average TSS concentrations in the effluent increased from
74±10 mg/L to 387±48 mg/L, which is in agreement with the total COD increase
observed after membrane addition (Figure 4.3). Similar trends observed in total COD
and TSS concentration profiles on the basis of a fixed COD/TSS ratio confirmed the
buildup of sludge and/or suspended particles inside the system due to a complete
retention of small and colloidal flocs that otherwise would have been evacuated from
a conventional UASB reactor. Accumulation of fine particles in the system resulted
from both the elimination of selective sludge washout in the system due to the
complete retention by the membrane leading to recirculation of the membrane rejects
into the system and abrasive effect of the recirculation pump used for maintaining the

77
cross-flow velocity over the membrane leading to particle breakage and
fragmentation. After steady state conditions were achieved in the AnMBR, the
difference between effluent and supernatant TSS concentrations decreased from 38%
to 6%, which is also an indicator of GLS separator shortcoming for sludge washout
prevention. However, despite the deterioration of the sludge bed, UASB reactor still
acts like a biofilter before membrane treatment, which prevents the membrane from
excessive exposure to high TSS concentrations. An average TSS concentration below
0.5 g/L was achieved in the UASB effluent after membrane incorporation, which is
quite low in comparison to the ones in AnMBRs consisting of CSTRs.

Figure 4.4 : TSS concentrations in the supernatant and effluent.

In order to measure the soluble compounds that resulted in this increase, SMP
(Figure 4.5) and VFA (Table 4.1) concentrations were measured both in the UASB
effluent and permeate. SMP consists of polysaccharide and protein. Protein
concentration was under the detection limit (2.5 mg/L) in the UASB effluent
obtained from Day 46 to Day 62 before membrane incorporation. Polysaccharide
concentration was almost stable, with an average value of 15.3±4.1 mg/L. As shown
in Figure 4.5, membrane incorporation initially led to an increase in SMP production.
Protein and polysaccharide concentrations in the effluent rapidly increased to 80
mg/L and 40 mg/L, respectively. Stuckey (2012) reported that under any type of
stress SMP production can increase dramatically in an anaerobic culture. Therefore,

78
the rapid increase in SMP concentrations could be correlated with incomplete
hydrolysis of solids in the UASB reactor that caused stress in microorganisms, given
the limitations of anaerobic metabolism for particulate matter accumulation. Besides,
this dramatic increase in protein and polysaccharide concentrations may be related
with biomass associated products which are produced during cell lysis resultant from
sludge recirculation. The recirculation pump used for maintaining the cross-flow
velocity over the membrane may have had an abrasive effect that caused cell lysis
and release of organic material from cell lysis may result in an increase in protein
and polysaccharide concentrations in the UASB effluent. Following this initial
increase, a decrease in protein and polysaccharide concentrations was observed,
which was in agreement with the trends for COD and TSS concentrations, and is
likely due to the adaptation to the new conditions or emergence of new microbial
populations. However, protein and polysaccharide concentrations were still higher
than those observed before membrane incorporation, with average values of 64±11.1
mg/L and 26±11.2 mg/L, respectively. The operational change in the system induced
by the membrane incorporation apparently affected the microbes, inducing a higher
SMP secretion. A decrease was observed in polysaccharide to protein in ratio after
membrane incorporation in the UASB effluent.

Figure 4.5 : SMP concentrations in the UASB effluent and final permeate.

In contrast to the significant increases observed in SMP concentrations after

membrane incorporation, only a slight increase was observed in VFA concentrations
from an average value of 2 mg/L before membrane incorporation to 13.5 mg/L after

79
membrane incorporation, which can be an indicator of soluble fraction increase
coming mostly from microbial origin rather than residual substrate or substrate
intermediate. Acetic acid represented the largest proportion of the VFA in the
effluent and permeate.

Table 4.1 : VFA concentrations in the effluent and permeate before and after
membrane incorporation.

Time Stage VFA

(mg/L)
Effluent Permeate
C2 C3 i- C4 i- C5 C6 Total C2 C3 i- C4 i- C5 C6 Total
C4 C5 C4 C5
Day-1 BMa 3 0 0 0 0 0 0 3
Day-16 BM 2 0 0 0 0 0 0 2
Day-31 BM 2 0 0 0 0 0 0 2
Day-46 BM 2 0 0 0 0 0 0 2
Day-61 BM 2 0 0 0 0 0 0 2
Day-66 AMb 14 0 0 0 0 0 0 14 12 0 0 0 0 0 0 12
Day-81 AM 16 1 1 1 1 9 0 29 13 1 1 1 1 9 0 26
Day-96 AM 9 2 1 1 1 14 0 28 8 2 1 1 1 14 0 27
Day-111 AM 13 0 1 1 1 1 0 17 12 0 0 1 1 1 0 15
Day-126 AM 8 1 0 0 0 1 0 10 7 1 0 0 0 1 0 9
a
Before membrane incorporation.
b
After membrane incorporation.

4.3.2 Impact of membrane incorporation on the overall system performance

Soluble COD concentration in the UASB effluent (122±9.8 mg/L) was found to be
higher than the permeate COD concentration value (42±4.4 mg/L). The pore size of
the filter used for obtaining the sample for soluble COD determination (0.45 µm)
was approximately 10 times larger than the average pore size of the membrane (0.03
µm). This indicated that an important fraction of soluble COD (i.e. <0.45 µm) was
retained by the UF membrane. VFA concentrations in the permeate were almost
equal to those of the effluent because of their very low molecular weight. As
expected, the membrane had no considerable effect on VFA removal. In contrast to
VFA, the membrane was able to retain SMP. The average protein and polysaccharide
concentrations in the permeate were 20 mg/L and less than 2 mg/L, respectively. The
retention of protein by the membrane (66%) turned out to be lower than the retention
of polysaccharide (95%). These results indicate that more soluble protein was
discharged from the system when it was operated as an AnMBR, whereas more
polysaccharide was observed in the effluent when the system was operated as a sole
UASB.

80
As a general comparative evaluation of system performance before and after
membrane incorporation, regardless of the increases in COD and TSS concentrations
in the UASB effluent, SS-free (non-detected) AnMBR permeate was obtained after
membrane incorporation with an average COD of 42 mg/L, standing for an overall
COD removal efficiency of 92% in the AnMBR system (Figure 4.3), whereas total
COD removal efficiency was approximately 72% before membrane incorporation.
The OLR was kept constant at 2 kg COD/m3.d during the whole period. However,
the COD in the effluent of the system, i.e., UASB effluent before the membrane
addition and permeate after membrane addition, decreased from 149±5.9 mg/L to
42±4.4 mg/L, resembling an increase in the COD removal efficiency of about 20%,
from 10.1±0.2 g COD/d to 12.9±0.1 g COD/d. The higher COD removal efficiency
achieved after membrane addition, compared with the COD removal efficiency
before membrane addition, can be attributed to the complete retention of all
particulate and colloidal matter and biomass inside the reactor by the membrane.
Average TN, NH4-N and TP concentrations in the permeate were 57 mg/L, 54 mg/L
and 11.8 mg/L. The results clearly demonstrate that AnMBRs are not sufficient for
removing nutrient from wastewater. Therefore, a nutrient removal, when necessary,
can be achieved in a separate post-treatment step. Average pHs of the overflow from
UASB before and after membrane addition were 7.22±0.27 and 7.08±0.11,
respectively. The rise in VFAs in the UASB effluent was accompanied by a drop in
pH to 7.08±0.11 after membrane addition. Permeate pH was 7.10±0.26 after
membrane addition.

The variations of TMP with operation time are shown in Figure 4.6. Membrane
cleaning was carried out both at the middle (Day 94) and end (Day 126) of the
AnMBR operation stage to recover the membrane permeability (P). The membrane
was cleaned only once after its incorporation as the reactor was operated for 63 days.
A slow increase was observed in TMP with operation time as membrane flux was
kept at about 12.3 L/m2h.

The sudden TMP rise observed at the initial stage (Day 79) might be related to the
sudden and sharp increase observed in both effluent COD and TSS concentrations
(Figure 4.3 and Figure 4.4) just after operation, which resulted in the accumulation of
particles in the lumen. The small diameter membrane tubes that were selected in
order to increase the filtration area were simply blocked by particles, especially by

81
the bigger ones and this would result in an uneven flux distribution across the bundle
of tubes and then an increase in TMP at Day 79, which eventually caused more
serious membrane fouling. After steady state conditions were achieved, low average
TMP values were maintained, which is in agreement with the results of stable UASB
effluent characteristics having relatively low average TSS concentrations below 0.5
g/L. TMP remained quite stable with an average value of 87 mbar during AnMBR
operation. Thus, membrane flux may become less dependent on the reactor MLSS
concentration, possibly leading to low TMPs.

Figure 4.6 : Filtration performance of the AnMBR system.

4.3.3 Impact of membrane incorporation on sludge properties

In order to investigate the mechanisms behind the enhancement in overall system

performance, sludge characteristics were also monitored during each operational
stage. Considerable changes in sludge physical properties were observed during the
experimental period, mainly characterized by a decrease in PSD, coinciding with the
period after membrane incorporation. The success of UASB technology lies in its
capability to establish a dense sludge bed (Seghezzo et al., 1998) and the settling
ability of the sludge bed is related to the selective washout of the dispersed sludge,
which results in an increased growth of retained sludge agglomerates. When
membranes are coupled with the UASB process, this selection pressure is eliminated.
Under conditions of high selection pressure, light and dispersed sludge will be

82
washed out while heavier sludge particles can be retained in the UASB reactor;
whereas under low selection pressure conditions as in this case, growth will take
place mainly as dispersed sludge, which leads to the formation of a bulking type of
sludge (Hulshoff Pol et al., 2004).

Figure 4.7 presents PSD curves of the seed sludge, and UASB sludge before and
after membrane incorporation. From Figure 4.7, it can be observed that the seed
sludge had a broader PSD profile with D50 of 141 µm compared to sludge taken
from the UASB under operation. Already after 1 day of operation, a notable washout
of small particles was experienced. An increase in D50 value to 465 μm was
observed before membrane incorporation, indicating that even small sized particles
were washed out (Figure 4.7(a)). To the contrary, PSD curves after membrane
addition shifted to the left and a decrease in the D50 was observed (Figure 4.7(b)).
During the operation of the system as an AnMBR, fine sludge particles such as non-
settling particles, small particles including colloidal substances and macromolecules
of SMP could not be washed out and were returned to the UASB reactor with the
concentrate flow, thereby decreasing the D50 of the sludge in the UASB reactor to
138 µm. Besides retention of non-settling particles and small particles, the
recirculation pump used for maintaining the cross-flow velocity over the membrane
may have had an abrasive effect that caused particle breakage and fragmentation.
The decrease in particle size resulted in a reduction of the settling capacity of the
sludge, supporting the assumption that the settling capacity depends on the PSD of
the sludge, i.e. larger particles yield better settling properties. The deterioration of the
settling capacity of the sludge promoted the sludge washout from the UASB reactor
(Figure 4.8), which increased COD and TSS concentrations in the UASB effluent.

SMP and EPS were also analyzed in the sludge bed (Table 4.2). For the SMP of the
sludge, the protein amount decreased from an average of 11.4 mg/g VSS to 5.1 mg/g
VSS after membrane addition, whereas the polysaccharide concentration increased
from an average of 1.3 mg/g VSS to 2.0 mg/g VSS. For the EPS, both protein and
polysaccharide amounts decreased by 81% and 97%, respectively. EPS are crucial
for integrating cells into flocs and granules. Therefore, the decrease in EPS amount
might indicate a higher possibility of looser structures of flocs and granules,
contributing to the decrease in sludge particle size and also to the increase in SMP in
the effluent. For SMP, the increase in polysaccharide amount in the sludge might be

83
related with the increase in polysaccharide concentration in the effluent (Figure 4.5)
after membrane incorporation. However, it is intriguing to find the simultaneous
decrease of protein in sludge along with the increase of protein in the effluent (Figure
4.5). Results indicate that a large amount of bound protein from the sludge leaked
into the effluent and rose to the top of the UASB reactor and even the decrease in
protein concentration of SMP led to an increase in protein concentration in the
effluent. All these results show that the membrane addition had obviously affected
the EPS and SMP production in sludge, as well as the floc and granule structures

Figure 4.7 : PSD of the seed and UASB reactor sludge (a) before and (b) after
membrane incorporation.

84
 Figure 4.8 : Sludge settleability before and after membrane incorporation.

Table 4.2 : SMP and EPS amounts of the sludge in the UASB reactor.
Time SMP EPS
Protein Polysaccharide Protein Polysaccharide
(mg/g VSS) (mg/g VSS) (mg/g VSS) (mg/g VSS)
Day-1 11.5 1.2 21.4 5.1
Day-31 11.2 1.3 17.6 4.5
Day-61 11.4 1.4 17.4 4.3
Day-81 4.8 1.3 3.7 0.3
Day-96 6.1 2.1 3.7 0.2
Day-126 4.6 2.5 3.4 1.0

The impact of membrane incorporation on the SMA and stability of the UASB
sludge is shown in Table 4.3. After membrane addition, the SMA of the sludge
decreased by 24%, from 0.70±0.06 to 0.53±0.02 g CH4-COD/g VS.day. However,
activity decrease was not expected due to the increases observed in both COD
removal efficiency and methane production. In the bio-assays, the substrate COD to
inoculum VS ratio was maintained at 2:1. The drop in the SMA can likely be

85
attributed to influent particulate matter accumulation within the system, due to the
presence of a physical barrier, i.e. membrane, retaining all non- or slowly degradable
polymers. Therefore, the same amount of VS used as inoculum would contain a
lower amount of organisms, which would show a lower overall activity regardless of
a decrease in the activity. In order to test this assumption, stability tests were
performed. The sludge stability decreased from 0.35 to 0.48 g CH4-COD/g COD
after membrane incorporation. Apparently, after membrane incorporation, more
biodegradable solids accumulated in the sludge bed.

Table 4.3 : SMA and stability of sludge in the UASB reactor before and after
membrane incorporation.
SMA Stability
Time
(g CH4-COD/g VS.day) (g CH4-COD/g COD)
Day-1 0.76±0.02 0.32±0.02
Day-31 0.64±0.05 0.38±0.01
Day-61 0.70±0.05 0.36±0.01
Day-81 0.52±0.03 0.45±0.003
Day-96 0.55±0.06 0.49±0.01
Day-126 0.53±0.02 0.49±0.01

Collected samples were used to pyrosequence the former region of 16S rDNA gene
using the 454 GS-FLX sequencer before and after membrane addition. A total of
11167 reads were obtained. The overall estimations of microbial diversity based on
phylogenetic diversity index, Shannon-Wiener index, Chao1 and Observed species
number (OSN) (Table 4.4) exhibited a great increase after membrane incorporation,
demonstrating that both richness and evenness of sludge bed microbial community
increased after membrane incorporation in UASB. From a microbial ecological
perspective, adding a membrane implies, directly or indirectly, a wider niche for
UASB microorganisms. Firstly, the membrane module functions as a direct retention
of potentially-easily-washed-out microorganisms. After addition of membrane, the
recycled concentrate brings the poorly settled microbial groups to the bottom section
of the UASB and mixes with the surrounding sludge. Thus, all species are kept
inside, regardless of their abundance. However, the membrane has also an indirect
effect to microbial community as it retains otherwise-washed-out particulates that are
substrates to some specific bacteria.

86
 Table 4.4 : Microbial diversity index.
UASB Sludge UASB Sludge
Biodiversity Index Inoculum Before After
Membrane Membrane
Phylogenetic Diversity Index 21.3 19.2 26.2
Shannon-Wiener Index 6.2 6.0 6.7
Chao1 Richness 422.3 540.3 803.1
OSN 326 312 396

Pyrosequencing analysis showed a drastic change in the bacterial community after

membrane incorporation (Figure 4.9). In the initial UASB seed sludge, Bacteroidetes
was the most abundant bacterial phylum (755 sequences, 36% of total bacteria),
followed by Firmicutes (695, 33%), Proteobacteria (240, 11.5%), Chloroflexi (136,
6.5%), Acidobacteria (107, 5.2%) and other phyla with percentages lower than 2%
(Table A.1, Appendix A). This means that the bacterial community at the starting
point was very diverse and sequences belonged to more than ten phyla of bacteria.
However, the sequences belonging to Firmicutes, Acidobacteria, Synergistetes,
Planctomycetes, Fusobacteria, Deferribacteres and Thermotogae decreased
dramatically after start-up period. This is reasonably to expect due to both substrate
composition changes and washout conditions. The inoculated sludge was previously
fed with black water in a pilot UASB reactor, which contained a wider variety of
organic compounds compared to the laboratory-scale UASB, serving a more diverse
microbial community. After membrane incorporation, an increase was again
observed in some phyla such as Firmicutes, Planctomycetes and Fusobacteria.

The genus Cytophaga was further enriched and occupied the predominant position,
followed by the genus Flavobacterium after membrane integration (Figure 4.9). The
change in Cytophaga and Flavobacterium might be related with the change in the
feed from black water to synthetic wastewater since similar increasing trends were
sustained after membrane addition. Cytophaga genus predominance could be
associated with the increase in small (organic) particles after membrane
incorporation, as evidenced by PSD results (Figure 4.7). Cytophaga has been
reported to grow anaerobically utilizing simple carbon sources such as glucose and
starch, with final products such as acetic, propionic and succinic acids, in
approximately equimolar quantities, plus a small amount of lactic acid (Bachmann,
1955). On the other hand, the slight increase in VFA (mainly acetic acid) observed
after membrane addition (Table 4.1) could also be related with the large dominancy

87
of genus Cytophaga (Oh et al., 2004; Chen and Dong, 2005). Another common
bacterial genus that are closely involved in hydrolysis and acidogenesis (Rademacher
et al., 2012; Zheng et al., 2013) -Clostridium- became the third largest population
after membrane addition, with a 141% increase in abundance (Figure 4.9). However,
Clostridium had a decreasing trend before membrane addition which was related
with the change in the feed. It has been reported that some members of the genus
Clostridium are also involved in acetogenesis and syntrophic acetate oxidation
(Westerholm et al., 2012). Some bacterial genus, which were dominant during the
first part of the experiment, such as Desulfomicrobium, Bellilinea, Aeromonas and
Dechloromonas lost their dominancy after membrane addition (Figure 4.9, Table
A.1, Appendix A). As to Bacteroides, a commonly reported protein/carbohydrates-
dependent genus (Yu et al., 2012; Li et al., 2013), its further decrease in abundance
after membrane addition could be either linked to the change in the feed or to the
observation that the EPS amount in the sludge bed decreased by 81% (protein) and
97% (polysaccharide) after membrane addition (Table 4.2).

Figure 4.9 : Relative abundance of major bacterial genera.

The archaeal community structure was also significantly affected by the membrane
addition. Eight different species were identified in the seed sludge. Members of the
genus Methanobrevibacter comprised 43.8% of the total genera, followed by

88
Methanobacterium, Methanolinea and Methanosaeta, which accounted for 14.6%
(Table A.2, Appendix A). However, after inoculating the laboratory-scale UASB
reactor, a clear shift in the archaeal community appeared. The number of species
decreased from eight to six, indicating that the archaeal community had a lower
richness during the UASB operation stage compared to the seed sludge.

Figure 4.10 : Relative abundance of all archaeal species.

The relative abundances of Methanobrevibacter and Methanolinea-like methanogens

diminished and the relative abundance of Methanobacterium-like methanogens
increased in the sludge. After membrane incorporation, the species belonging to the
genus Methanosaeta (acetoclastic methanogens) (Sousa et al., 2013) became
dominant (56.2% in relative abundance) while before it only accounted for less than
15% (Figure 4.10, Table A.2, Appendix A). The previously dominant
Methanobacterium species (hydrogenotrophs) decreased by 45%. It is well known
that acetate is an important substrate for methanogenesis in anaerobic reactors, and
the archaeal genus Methanosaeta is representative of acetoclastic methanogens. In
addition, it was reported that Methanosaeta sp. contributed to the stable performance
of the anaerobic bioreactors (Talbot et al., 2008) and they are believed to be
competitive in established methanogenic communities (Bandara et al., 2012).
Therefore, the predominance of Methanosaeta can be explained by means of the
VFA accumulation (mainly acetic acid) observed during the AnMBR operational

89
stage. Similar results have also been reported in other studies (Siggins et al., 2011;
Ziganshin et al., 2013). Additionally, the increase in methane production after
membrane incorporation could be attributed to the further degradation of retained
particulate organic matter.

4.4 Conclusion

Coupling a membrane to a UASB reactor resulted in significant changes in both

physical and biological aspects, mainly due to the elimination of hydraulic dilution as
bacterial selection pressure factor. After membrane addition, the sludge PSD
decreased, coinciding with the complete retention of colloidal and fine particles by
the membrane. In addition, the EPS content of the sludge decreased and the abrasive
effect of the membrane recirculation pump caused particle fragmentation. Decrease
in the sludge particle size deteriorated the sludge bed settleability, which led to a
higher sludge washout, resulting in an increase in COD, TSS and SMP
concentrations of the UASB effluent. However, enhancement in overall system
performance was observed in terms of higher COD removal efficiency and methane
production after membrane incorporation due to the complete retention of all
particulate and colloidal matter and biomass inside the reactor by the membrane.
Besides, TMP values remained quite stable during AnMBR operation indicating no
severe membrane fouling. Pyrosequencing analysis showed that membrane
incorporation induced population dynamics in both bacterial and archaeal species in
the sludge. A decrease was observed in the sludge activity after membrane
incorporation, in spite of the increases observed in both COD removal efficiency and
methane production. Stability tests confirmed the observation that the decrease in
SMA was related to the particulate matter accumulation. Overall, it is impossible to
get an effluent that fulfills the regulations for water reuse by using solely a UASB
reactor for treatment. However, a pathogen-free but nutrient-rich permeate was
achieved during the AnMBR stage with quite stable TMP values. Therefore, the
results obtained in this study clearly supported the applicability of a membrane-
coupled UASB reactor for municipal wastewater treatment, in spite of the sludge bed
deterioration.

90
5. IMPACT OF TEMPERATURE ON FEED-FLOW CHARACTERISTICS
AND FILTRATION PERFORMANCE OF AN UPFLOW ANAEROBIC
SLUDGE BLANKET COUPLED ULTRAFILTRATION MEMBRANE (4)

5.1 Introduction

With the growing interest to produce nutrient-rich and hygienically safe effluents
from municipal wastewater, there is an increasing interest to assess the feasibility of
AnMBRs as treatment alternative. Municipal wastewater is characterized by a low
organic strength, implying low biomass growth rate and a significant amount of
particulate organics that have to be hydrolyzed. These characteristics constrain the
effective hold-up of biomass in the bioreactor, especially in reactors in which
flocculent biomass can be easily washed out such as UASB and EGSB reactors
(Martinez-Sosa et al., 2011a). The combination of a membrane separation unit and an
anaerobic bioreactor allows efficient retention of biomass (Gao et al., 2009; Tao et
al., 2012), eventually resulting in efficient hydrolysis of particulate organics and
advanced sludge stabilization at long SRTs (An et al, 2009b; Zhang et al., 2011).
Therefore, AnMBR is proposed as a promising technology for the treatment of
municipal wastewaters, with the added benefits of low sludge production, high
quality effluents with a very low content of pathogenic organisms but a high
fertilizing value, absence of additional costs for aeration, generally associated with
the conventional activated sludge process (Lin et al, 2011). However, thus far, the
development of AnMBR systems for municipal sewage treatment is challenged by

(4)
This chapter is based on:

Ozgun, H., Tao, Y., Ersahin, M.E., Zhou, Z., Gimenez, J.B., Spanjers, H., van Lier, J.B. (2015)
Impact of temperature on feed-flow characteristics and filtration performance of an upflow anaerobic
sludge blanket coupled ultrafiltration membrane treating municipal wastewater. Water Research 83,
71-83.

91
membrane fouling that increases the energy costs, reduces the membrane lifetime,
and makes the system less efficient (Nguyen et al., 2013).

Many attempts have been made to prevent or reduce membrane fouling (Martin-
Garcia et al., 2011; Salazar-Pelaez et al., 2011b). Briefly, several factors involved in
design and operation may influence fouling such as setup configuration and
membrane characteristics, feeding and biomass properties and operational conditions
(Gao et al., 2010; Salazar-Pelaez et al., 2011b). Among them, reactor design is
proposed as an important factor that can change the AnMBR fouling propensity
(Ozgun et al., 2013b). Most of AnMBRs studied employ CSTR because of the ease
of construction and use (Liao et al., 2006). While CSTR configurations feature the
general AnMBR advantages, they are characterized by exposing the membrane to the
bulk MLSS, which easily leads to fouling (Liao et al., 2006). The MLSS of a CSTR
AnMBR reactor is commonly >10000 mg/L, while the effluent from a UASB
AnMBR is 300-550 mg TSS/L (Cho and Fane, 2002; Liao et al., 2006). Exposure to
higher solids concentrations in CSTR designs usually increases the rate of cake
formation in comparison to reactor designs that retain biomass. Rapid cake formation
requires more maintainance and leads to longer down times (i.e. periods of non-
filtration) and thus lower overall fluxes. Besides, more intimate exposure of the
biomass to the membranes may induce EPS production, enhancing membrane
fouling (Liao et al., 2006). Given these considerations, UASB reactors have recently
attracted more interest to become the anaerobic bioreactor as a part of AnMBRs (Gao
et al., 2010; Liu et al., 2012; Ozgun et al., 2015a). Membrane coupled UASB
reactors may be a promising approach to overcome problems related with excessive
fouling. High fluxes are expected in membrane coupled UASB configurations since
the membrane is only challenged by the supernatant and not by the bulk sludge.
Lower reversible fouling has been reported for a membrane coupled UASB reactor
system in comparison to membrane coupled CSTR system for the treatment of black
water (van Voorthuizen et al., 2008).

It is well known that anaerobic processes strongly depend on operational temperature

that has a substantial effect on the conversion rate of organic matter and
consequently on the quality of the effluent. Temperature can also impact the
solubility of various compounds and gases and viscosity of sludge that has a great
influence on the membrane filtration performance. In fact, several studies (Wen et

92
al., 1999; Ho and Sung, 2010; Martinez-Sosa et al., 2011a; Robles et al., 2013) have
focused on the effect of operational temperature on the performance of AnMBRs
treating municipal wastewater. Ho and Sung (2010) reported that AnMBRs operated
at 25 °C and 15 °C, treating synthetic municipal wastewater achieved more than 85%
COD removal due to the physical retention of pollutants by the membrane, although
the methanogenic activity was suppressed at lower temperature. Martinez-Sosa et al.
(2011a) also observed that a pilot-scale AnMBR treating municipal wastewater
achieved 90% COD removal during operation for 100 days under both mesophilic
and psychrophilic conditions showing a strong tolerance to temperature changes.
Particularly, under sub-mesophilic conditions, hydrolysis of the retained particulates
likely becomes rate-limiting and in UASB reactors particulate matter accumulation
in the sludge bed results in a sludge activity loss (Lettinga et al., 2001). Under these
conditions, the degree of small particle retention by the sludge bed will be of prime
importance for the feasibility of UASB coupled membrane filtration systems (Ozgun
et al., 2013a). However, regarding UASB-membrane configurations, research on
particle retention in the sludge bed at lower temperatures and the effect of
temperature on filtration performance are difficult to find in literature.

Therefore, in the present research, a UASB coupled membrane system treating

municipal wastewater was operated for nearly 126 days at 25 °C and 15 °C in order
to investigate the relationship between the characteristics of the UASB effluent and
the filtration performance of the AnMBR at different operational temperatures. The
effect of temperature on the overall performance of the AnMBR system was
determined in order to investigate whether the membrane coupled UASB is a
technically feasible alternative in the treatment of municipal wastewater at 15 °C. In
addition, the effect of temperature on the microbial communities was also
determined. The results reported here provide valuable new insights into the fouling
mechanisms and will consequently further advance knowledge on the feasibility of
membrane coupled UASB systems at low temperatures.

5.2 Materials and Methods

5.2.1 Wastewater source

In the present study, synthetic municipal wastewater was used as feed, similar to that
described in Chapter 3. The composition of the concentrated substrate solution and

93
the characterization of the synthetic municipal wastewater at the UASB inlet were
already presented in Chapter 3 in Table 3.1 and Table 3.2, respectively.

5.2.2 Seed sludge

The reactor was seeded with flocculent anaerobic sludge obtained from a pilot-scale
UASB reactor treating black water (Sneek, the Netherlands). The characteristics of
the seed sludge were presented in Table 3.3 in Chapter 3.

5.2.3 Experimental setup

The experimental setup consisted of a laboratory-scale UASB reactor with an

effective volume of 7 L coupled to an external membrane module. A schematic
diagram of the experimental setup is depicted in Figure 5.1. Peristaltic pumps
(Watson Marlow 120U/DV) were used for circulating the flow between the
membrane module and UASB reactor and obtaining permeate through the
membrane. Furthermore, recirculation over the membrane surface was applied by a
peristaltic pump (Watson Marlow 620UN/R) in order to maintain a cross-flow
velocity of 1 m/s across the membrane surface independently from the upflow
velocity in the UASB reactor. Concentrate recycle flow-rate was adjusted to keep the
upflow velocity stable in the UASB reactor. The biogas was collected by means of a
three-phase separator installed at the top of the reactor. Biogas flow-rate was
measured by a gas meter (Ritter, Milligas Counter MGC-1 PMMA). The operational
temperature was controlled by using water baths. High temperature (TC16, Tamson
Instruments, the Netherlands) and low temperature (TLC 15, Tamson Instruments,
the Netherlands) water baths were used in order to maintain operational temperatures
of 25 °C and 15 °C, respectively. Temperature and pH inside the bioreactor were
monitored on-line by a probe combined with a transmitter (Elscolab, M300 ISM).
TMP was monitored using high accuracy pressure sensors (AE Sensors, ATM -
800/+600 mbar, the Netherlands) installed in the membrane feed, concentrate and
permeate lines. A computer accessible via LabView software was used to control the
pumps and collect the data.

The external cross-flow tubular membrane module contained 28 membrane fibers

with an internal diameter of 1.5 mm and a length of 80 cm. Cross-sectional and
permeate outlet images of the membrane module are presented in Figure 5.2(a) and
Figure 5.2(b), respectively. Pentair X-Flow (Enschede, the Netherlands) PES UF

94
membranes with a pore size of 30 nm were used in the AnMBR. The detailed
characteristics of the membrane are listed in Table 5.1.

Figure 5.1 : Layout of the AnMBR, consisting of a UASB coupled UF unit.

95
 (a) (b)
Figure 5.2 : Images of the UF membrane (a) Cross-sectional (b) Permeate outlet.

Table 5.1 : Specifications of the membrane.

Specifications UFC M5 LE
Material PES
Pore Size 30 nm
Structure asymmetric/microporous
Membrane area 0.11 m2
pH range 2-12
Chlorine exposure 250000 ppm·h
Temperature 1 – 80 °C
TMP -300….+300 kPa

5.2.4 Experimental procedure

The experimental work was carried out over two successive periods. The reactor was
operated in a continuous mode at 25 °C and 15 °C for 126 days. To isolate the impact
of the temperature transients on the system performance, HRT (6 h), OLR (2
kg/m3.day) and upflow velocity (0.6 m/h) were maintained constant.

The AnMBR was operated in cycles consisting of 180 seconds of filtration, followed
by 20 seconds of backwash. Backwash was applied by reversing the flow direction,
keeping the flux at the same value as during filtration. The membrane flux was
maintained constant at 12.3 L/m2.h with a cross-flow velocity of 1 m/s.

96
5.2.5 Experimental methods

5.2.5.1 Analysis

COD and TSS were determined following Standard Methods (APHA, 2005). Each
analysis was performed in triplicate. The samples for soluble COD measurement
were filtered through a 0.45 µm filter before analysis. The turbidity was measured by
HACH 2100N Turbidimeter. The PSD of the effluent was assessed by a Mastersizer
2000 (Malvern Instruments, Hydro 2000 MU) that has a detection range of 0.02-
2000 µm. Laser diffraction technique was used to measure the size of the particles.
For SMP analysis, a volume of 5 ml effluent was freshly sampled, filtered through a
0.45 µm filter and diluted 4 times with PBS (pH=7.2). The phenol-sulfuric acid
method was used to quantify polysaccharides (Dubois et al., 1956). The
concentration of protein was measured using the Bradford method (Bradford, 1976).

5.2.5.2 Critical flux determination

The critical flux was determined by applying a flux step method (Le Clech et al.,
2003). The flux below which there is no flux decline and no fouling observed over
time is defined as critical flux (Field et al., 1995; Ersahin et al., 2013). Critical flux
determination consisted of successive filtration steps of 15 min, with flux increments
of 5 L/m2.h. One minute of backwash was included between each cycle. In order to
have similar membrane initial conditions, a cleaning protocol including physical,
base and acid cleaning steps was followed every time the critical flux was
determined or when new flux conditions were imposed. During each step, TMP was
recorded with 30 seconds intervals. The critical flux was assumed to be exceeded
when a clear discontinuity occurs in the slope of the TMP against time (Montgomery
et al., 2001).

5.2.5.3 Filtration resistance

After long-term reactor operation, the fouled membranes were cleaned with a
sequential cleaning procedure to calculate different fouling resistances and evaluate
the characteristics of membrane fouling. Total filtration resistance (R T) was assessed
as a function of the TMP that was measured by pressure sensors located on the
membrane feed, concentrate and permeate lines. RT is calculated as the slope of the

97
measured TMP at each step versus the applied J, divided by the dynamic viscosity of
water (η) by the equation (5.1):

PT
RT   Rint rinsic  Rremovable  Rirreversible  Rirre cov erable (5.1)
 .J

Then, the measured total membrane resistance represents the sum of various
contributions including: the intrinsic resistance (Rintrinsic) of the membrane itself, the
resistance due to cake layer formation that can be physically removed by flushing
with water: Removable resistance (Rremovable), the resistance due to pore-clogging that
can be removed by chemical cleaning: Irreversible resistance (Rirreversible), and the one
that cannot be easily reverted by neither physical nor chemical cleaning:
Irrecoverable Resistance (Rirrecoverable).

Rintrinsic test of the membrane was carried out by filtration of deionized water through
the virgin membrane at the beginning of the experimental study. RT determination
was performed by measuring the J of deionized water using the fouled membrane.
The membranes were first physically cleaned by flushing with tap water to remove
the cake layer. The observed resistance, when filtering deionized water using the
physically cleaned membrane, was taken as the sum of Rintrinsic of the membrane,
Rirreversible and Rirrecovarable; by subtracting the obtained value from RT, the Rremovable
was obtained. Following that, the membranes were chemically cleaned by soaking
sequentially in 200 ppm NaClO-0.4% Divos solution for 30 minutes and 1% (w/v)
citric acid solution for one hour in order to calculate Rirrecovarable. Rirreversible was
obtained by subtracting the sum of Rintrinsic of the membrane, Rremovable, Rirrecoverable
from the RT. After each step, the clean water permeability of the membranes was
measured to determine the permeability recovery.

P (L.h-1.m-2.bar-1) was calculated by the equation (5.2):

J [ L / m 2 .h]
P (5.2)
TMP[bar]

5.2.5.4 Microbial community analysis

The microbial community structure was investigated using 454-pyrosequencing.

UASB sludge samples were taken at different operational temperatures. Details

98
about sample preparation, DNA extraction, pyrosequencing and post analysis of
pyrosequencing data can be found in Chapter 4.

5.3 Results and Discussion

5.3.1 Effluent characteristics

In order to clarify the relationship between the UASB effluent and the filtration
performance of the AnMBR, the changes in the UASB effluent and AnMBR
permeate characteristics, such as COD, SMP, TSS, turbidity and PSD, were
evaluated at 25 and 15 °C.

Total and soluble COD concentrations in the effluent and permeate of the system
during two runs are shown in Figure 5.3. A steady operational period was considered
to be achieved once stable effluent COD concentrations and biogas production were
obtained. Under steady operational conditions, the average COD concentration in
the UASB effluent was 898 mg/L at 25 °C. The high COD concentrations in the
UASB effluent can be ascribed to the membrane coupling, re-directing the reject
water back to the UASB reactor, leading to a higher degree of mixing inside the
UASB bulk. Upon the temperature decrease from 25 °C to 15 °C, the average
effluent COD concentration increased to 1759 mg/L, possibly due to a less complete
hydrolysis and a higher amount of accumulating SS. An increase in average soluble
COD concentrations from 122 mg/L to 153 mg/L was also observed with the
decrease in temperature. Generally, biological reactions proceed much slower at low
temperatures (Lettinga et al., 2001). Furthermore, microbial communities performing
the various steps of the anaerobic degradation pathway exhibit different responses to
temperature changes, which can result in the accumulation of intermediary products
of the anaerobic degradation as a result of unbalanced growth.

However, total COD removal efficiency, based on influent and permeate COD, was
only slightly affected by the temperature change, with total COD removal efficiency
higher than 89% even at 15 °C. Average total COD removal efficiency at 25 °C was
92% with an average COD concentration of 42 mg/L in the permeate. At the low
temperature (15 °C), the permeate quality was slightly less (52 mg/L), with an
average total COD removal efficiency of 90%. The exceptionally high COD removal
efficiencies can likely be ascribed to the retention of biomass by the membrane unit.

99
Our results highlight the possible role of membranes in performance stability under
temperature fluctuations, overall contributing to an improved permeate quality. A
slight difference in the overall treatment performance of an AnMBR at lower
temperatures has also been observed by Ho and Sung (2010) and it was attributed to
the physical retention capacity of the membrane. Higher rejection of small particles
or macromolecules by the membrane surface at 15 °C might result in the formation
of cake/gel layer by internal and external membrane adhesion and interception.

Figure 5.3 : Total and soluble COD concentrations in the UASB effluent and
permeate.

Average soluble COD removal efficiencies, based on the soluble COD

concentrations in the UASB effluent and permeate, were 65% and 67% at 25 °C and
15 °C, respectively. Since soluble COD entails all compounds below 0.45 μm, the
observed soluble COD removal can be attributed to the nominal pore size of 0.03
μm, a further decreased porosity owing to cake layer formation, but also to possible
internal and external adhesion, as well as possible degradation of organic compounds
by the cake layer on the membrane surface.

In order to determine the effect of temperature on the release and washout of

colloidal and suspended particles, TSS concentration and turbidity were measured in
the effluent of the UASB reactor. A slight increase in UASB effluent TSS
concentrations was observed from the range of 312-459 mg/L to 440-552 mg/L when

100
temperature was decreased from 25 °C to 15 °C. However, there was a considerable
increase in UASB effluent turbidity, as a result of the temperature decrease: average
turbidity in the effluent were 233±27 NTU and 732±46 NTU at 25 °C and 15 °C,
respectively (Figure 5.4). Higher variations observed in effluent turbidity, which is
contradictory to the lower variations in TSS concentration, might be attributed to the
washout of colloidal particles from the UASB reactor instead of suspended particles
at low temperature, as turbidity is an indicator of both suspended and/or colloidal
particles.

Figure 5.4 : TSS concentration and turbidity in the effluent.

In addition to the solids content, also the PSD of the UASB effluent has been
measured during each operational period (Figure 5.5). Particle size is considered a
crucial factor affecting particle deposition, since it can strongly determine the
magnitude of back-transport mechanisms (Altmann and Ripperger, 1997; Jeison and
van Lier, 2007). The range of D50 that was measured in the UASB effluent at 25 °C,
was between 80 µm and 137 µm, whereas at 15 °C an apparent decrease in D50 was
observed from 88 µm on Day 66 to 25 µm on Day 126. In addition, peaks shifted to
the left part, i.e. smaller sizes, of the PSD curve when temperature was decreased to
15 °C. The distribution of volume percentage exhibited one-peak curves at 25 °C,
which indicated that most of the particles in the UASB effluent had unimodal
distribution. However, percent volume distribution at 15 °C showed two-peak curves
with the emerge of a second peak curve at the left part of the distribution pattern,
confirming the presence of smaller sized particles in the UASB effluent. Small

101
particles show a high tendency to deposit on the membrane surface induced by their
low back-transport. Mahmoud et al. (2003) reported the substantial effect of
temperature on the conversion of organic matter and the characteristics of the sludge
bed. According to Lawler et al. (1986), large particles are broken and small particles
are accumulated in the system, which results in a larger specific surface area when
digestion does not work well. On the contrary, when digestion works well, particles
of all sizes are reduced with a preferential removal of small particles (conversion into
gaseous form) and the overall specific surface area will be reduced. Therefore, the
observed change in the particle size at 15 °C might be related to the unbalanced
disintegration rate of large composite influent particles into fractions of
carbohydrates, proteins and lipids and the slow hydrolysis and solubilization rates of
these products at low temperature. We therefore postulate that colloidal
carbohydrates, proteins and lipids accumulated within the system, which contributed
to the increase in carbohydrates and proteins in the effluent. In fact, the mentioned
compounds are colloidal particles that reduced the D50, and increased the turbidity
and suspended COD.

Figure 5.5 : PSD of the UASB effluent at different sampling days.

Figure 5.6 presents polysaccharide and protein concentrations of SMP in the UASB
effluent and final permeate of the AnMBR system. Temperature decrease resulted in
an increase in SMP concentration in the effluent of the UASB reactor, which might

102
be related to the decrease in SMP degradation rate due to kinetic considerations.
Hence, the lower the temperature, the lower the removal rate of COD in the UASB
reactor, and the higher the concentration of SMP. Furthermore, when the temperature
decreased from 25 °C to 15 °C, although the release of SMP decreased due to the
decrease in the rate of cell lysis and endogenous decay, the degradation rate of SMP
also decreased with the temperature drop and resulted in an overall increase in SMP
concentration.

Figure 5.6 : SMP concentrations in the effluent and permeate.

Concentrations of both polysaccharide and protein were always higher in the UASB
effluent than in the permeate; thus the membrane acted as a barrier for these
compounds, part of which could be deposited on the membrane surface, contributing
to its fouling. From Figure 5.6, it becomes clear that the membrane is more effective
in retaining the polysaccharide-based SMP than the protein-based SMP. This result is
consistent with the findings of Drews et al. (2006) and Salazar-Pelaez et al. (2011a).

Removal efficiency of SMP components were found to be slightly influenced by

temperature. Rejection of SMP components increased from 66% to 70% for proteins
and from 95% to 97% for polysaccharides with the decrease in temperature. The
observed higher mass flux of SMP through the membrane at higher temperatures
may be related with the increased diffusivity of the molecules at elevated
temperatures enhanced by solubility increase and decrease in adhesive forces.

103
5.3.2 Microbial community

454-Pyrosequencing analysis revealed a slight effect of temperature on both the

bacterial and archaeal communities. The communities’ biodiversities were assessed
based on α-diversity (Figure 5.7), in which the richness was indicated by
Phylogenetic Diversity Index, Chao1 richness and OSN, while the evenness was
shown by Shannon-Wiener Index. OSN and phylogenetic diversity slightly decreased
when the temperature decreased from 25 °C to 15 °C, indicating that a decreased
temperature may result in a decreased overall microbial diversity of an AD
community (Figure 5.7). This is consistent with previous studies of Gao et al. (2011)
and Lu et al. (2012).

Figure 5.7 : Microbial diversity index.

The dominant bacterial phyla (Figure 5.8) and species (Figure 5.9) were apparently
altered due to temperature decrease. From the results presented in Figure 5.8, it can
be deduced that the phylum of Bacteroidetes was significantly more abundant at 25
°C than at 15 °C, showing relative abundances of 47% to 21%, respectively. In
contrast, the abundance of phyla Proteobacteria and Firmicutes was about 50% and
100% higher, respectively, after the temperature drop (Figure 5.8). Although most of
the microorganisms were retained in the sludge bed of the UASB and they were not
directly in contact with membranes, it can be expected that bacterial spores and
dispersed cells were lifted from the sludge bed and subsequently deposited on the

104
surface of the membrane. The pyrosequencing results showed that the Firmicutes
strains, of which most are capable of yielding spores under stress conditions
(Wunderlin et al., 2014), accounted for 50.1% in relative abundance in the top part of
sludge bed at 15 °C, while such value at the same position at 25 °C was 18.2%. This
suggests an upward movement of endospore-forming species to the top part of the
sludge bed following the decrease in temperature, which possibly led to a higher
biomass fraction in the UASB effluent. Those sub-micrometer-scale endospores are
prone to washout, possibly contributing to membrane fouling.

Figure 5.8 : Relative abundance of bacterial phyla.

Some mesophilic bacteria lost their pre-dominant position. For instance, the species
belonging to genera Cytophaga decreased from 31% to 6% in abundance (Figure 5.9,
Table B.1, Appendix B). These species are usually found in mesophilic (Li et al.,
2011b; Kampmann et al., 2012) or thermophilic anaerobic reactors (Rademacher et
al., 2012). Another commonly reported genera Flavobacterium decreased by more
than half (from 8% to 3%) under low temperature (Figure 5.9, Table B.1, Appendix
B). These genera are normally found to be dominant in mesophilic reactors
(Cardinali-Rezende et al., 2009; Shin et al., 2010). Since genera Cytophaga were
largely pre-dominant at 25 °C, the loss of dominancy of this genus in the UASB
reactor might promote the relative abundance of other bacteria at low temperature.

105
For example, the mesophilic lactic acid bacterium, Lactococcus raffinolactis, became
dominant with abundance increased from 0.1% to 12.8%. Lactococcus raffinolactis
is able to ferment galactosides (Boucher et al., 2003), which were available in the
used synthetic wastewater made from milk powder. Compared to other lactic acid
bacteria, Lactococcus raffinolactis is reported to grow faster when temperature
ranges between 4 and 10 °C in natural environments (Hagi et al., 2004). To date, the
only evidence of Lactococcus raffinolactis in a reactor treating wastewater was
reported by (Lu et al., 2012) in a single-chamber microbial electrolysis cell for
glucose fermentation at 4 °C, where 7/96 clones of Lactococcus raffinolactis were
confirmed. Another two species, Acinetobacter sp. and Dechloromonas sp., were
also found to increase in relative abundance by 4~20 fold at low temperature.

Like the bacterial community, also the archaeal community distinctly changed from
the seed sludge conditions to the 25 °C and subsequent 15 °C conditions (Figure
5.10). The hydrogenotrophic methanogens, (that produce methane from formate, H2
and CO2) Methanobrevibacter sp, distinctly dropped from the seed sludge to the 25
°C reactor and almost vanished from the community when the reactor was operated
at 15 °C (Figure 5.10, Table B.2, Appendix B). Apparently the prevailing conditions
in the 15 °C reactor hamper the proliferation of this mesophilic archaea. Another
hydrogenotrophic methanogen, Methanobacterium petrolearium, also lost its
prevalence from 15% at 25 °C to about 3% at 15 °C (Figure 5.10, Table B.2,
Appendix B). Previous studies have shown that the optimum growth temperature of
Methanobacterium petrolearium is at 35 °C (Mori and Harayama, 2011). The
observed decrease in the abundances of hydrogenotrophic methanogens at low
temperature might be related to an increased competition with acetogenic hydrogen
consumers at low temperatures (Nozhevnikova et al., 2001; Kotsyurbenko et al.,
2001), suggesting a drop in relative importance of hydrogenotrophic methanogens
with decreasing temperature. The latter contrasts to the hypothesis of Smith et al.
(2012), that suggests that at low temperatures hydrogenotrophic methanogenesis may
become more significant, because hydrogenotrophic methanogenesis is more
exergonic at low temperature due to the increased solubility of hydrogen (Lettinga et
al., 2001). Most of the hydrogen is generated during fermentation, which is limited
by reduced hydrolysis and solubilization at low temperatures, leading to low
hydrogen concentrations in the bulk liquid. Therefore, despite the enhanced

106
hydrogen solubility at low temperatures, the lower production of hydrogen might
result in a substrate-limited hydrogenotrophic methanogenic population. In addition,
owing to the higher hydrogen solubility, homo-acetogenesis may become more
important under psychrophilic conditions, increasing the amount of COD that is
channeled via acetate as precursor for methane (Nozhevnikova et al., 2001;
Kotsyurbenko et al., 2001). Therefore, it is intriguing to find that Methanosaeta sp
became even more dominant after the temperature decrease (from previously 56% to
84%). Methanosaeta is a typical genus of acetate-utilizing methanogens that is
predominant in anaerobic reactors for methane production (Veiga et al., 1997). The
role of these acetate-utilizing archaea, which utilize only acetate as the substrate for
methanogenesis, has been proven to be able to out-compete other methanogens under
low temperatures (Dhaked et al., 2010). For example, Methanosaeta sp. were found
to be most numerous at low operating temperatures (5 °C-18 °C) in a fixed-bed
reactor revealed by both qPCR and clone library (Zhang et al., 2012). In another
study, it was found that Methanosaeta sp. showed an abundance of over 60% under
psychrophilic temperature (15 °C) in a bench-scale AnMBR (Smith et al., 2013).

Figure 5.9 : Relative abundance of major bacterial species (bacterial species with
abundance over 1%).

107
 Figure 5.10 : Relative abundance of all archaeal species.

5.3.3 Critical flux tests

Flux and TMP trends obtained from the critical flux tests during the experimental
study are given in Figure 5.11. For each test, a clear critical flux could be
distinguished based on the sudden TMP increase at a certain flux. Following the
flux-step method, the critical fluxes during operation at 25 °C were found to be 69.5,
41.3 and 53.8 L/m2.h at Day 1, Day 31 and Day 61, whereas the critical fluxes
decreased to 41, 38 and 34.3 L/m2.h at Day 81, Day 96 and Day 126, respectively,
during operation at 15 °C. The critical fluxes are much higher than most of those
reported in the literature under anaerobic conditions using different reactor and
membrane configurations (Gimenez et al., 2011; Wei et al., 2011; Martinez-Sosa et
al., 2012; Fox and Stuckey, 2015), which likely can be attributed to the applied
reactor configuration, preventing exposure of the bulk liquid to the membranes. The
highest critical flux was obtained at 25 °C. The lower critical fluxes obtained at 15
°C in comparison to 25 °C, agreed with a concomitant higher turbidity, COD and
SMP concentrations in the UASB effluent at 15 °C, confirming the negative impact
of a low operational temperature on membrane fouling and thus reactor performance.

108
 Figure 5.11 : Critical fluxes at different operating temperatures.

5.3.4 Filtration performance

The variations in TMP with operation time are shown in Figure 5.12. The results
correlate the UASB effluent characteristics with the filtration performance of the
entire AnMBR system. It can be observed that the TMP increased slowly with
operation time, while keeping the membrane flux at about 12.3 L/m 2.h during the
first 16 days. During each experimental period, membrane cleaning was carried out
nearly every 30 days to recover the P, followed by another operation period. The
sudden TMP rise observed on Day 16 resulted in a higher TMP increasing rate at the
initial stage (Day 1-Day 31) in comparison to the second stage (Day 31-Day 63) of
operation at 25 °C. The sudden TMP increase coincides with the observed sudden
and sharp increase in both UASB effluent COD and TSS concentrations. In addition,
during the initial stage the UASB effluent was characterized by a wider PSD. The
observed non-stable effluent characteristics during the initial stage resulted in the
accumulation of particles in the lumen of the membrane unit. The small diameter
membrane tubes that were selected in order to increase the filtration area were simply
blocked by particles, especially by the bigger ones. In addition, particles with a lower
diameter could easily accumulate in the lumen due to the reduction in nominal tube
diameter. Such partial blockage would result in an uneven flux distribution across the
bundle of membrane tubes. Moreover, the sudden increase in local flux at Day 16,
could eventually cause more serious membrane fouling, in agreement with the
observations made by Cho and Fane (2002), who also related the sudden rises in

109
TMP to the local flux increases. However, after stable operational conditions were
achieved at 25 °C, it can be observed that the TMP increased only slowly with
operation time and low TMP values were maintained. This stable filtration period is
in agreement with the stable UASB effluent characteristics and selected flux that is
below the critical flux at the same operation period. However, a gradual increase in
the TMP was observed over time under subcritical flux conditions of 12.3 L/m 2.h
during operation at 15 °C. Although the system was operated at fluxes below the
critical flux for new or chemically-cleaned membranes, a rapid membrane fouling
was observed. The relatively high critical flux values presented in Figure 5.11 can be
attributed to short-term test operation. During the critical flux test, cross-flow
conditions were applied for only a couple of hours. Apparently, for long-term
operation periods, the predictive value of the assessed critical flux becomes
insignificant, meaning that only much lower fluxes can be admitted for attaining
stable membrane filtration. Nonetheless, the distinct difference between 25 °C and 15
°C suggests a decreased filtration performance at lower temperatures. Le Clech et al.
(2003) concluded that the critical flux resulting from the flux-step method cannot be
used to estimate the critical flux of systems operated continuously because this
method is based on a short-term run. The findings of our study at 15 °C also
confirmed this observation. Since TSS concentrations were very similar at both
temperatures, the higher fouling rate at lower temperature was probably not
associated with the TSS concentration. Particularly small particles, which were
abundantly measured at the low temperature, have a high fouling propensity, likely
related to the low back-transport (Martinez-Sosa et al., 2012). Presence of more SMP
at low temperature could also have increased the possibility of external and internal
pore clogging and biofilm growth on the membrane surface.

Our filtration performance results were obtained under laboratory-scale conditions.

Although the laboratory-scale studies provide valuable information about the
treatability and flux levels, the hydraulic conditions and pressure drops in a
laboratory-scale membrane module can be significantly different from pilot- and full-
scale membrane modules. Moreover, the applied shear rate and its effect on
anaerobic biomass can significantly differ in laboratory and full-scale reactors due to
the use of different pumps and valves. Therefore, laboratory-scale results should be
validated with pilot- and full-scale studies since the latter ones can provide more

110
representative data in terms of attainable flux, TMP and fouling. To date, AnMBR
application for domestic utilities in pilot- and full-scale systems has not yet become a
reality, which is partly because of system novelty and reluctance owing to membrane
fouling problems (Ozgun et al., 2013a). With the rapid development and
implementation of AnMBR technology especially in full-scale systems, long-term
reliability and operability of AnMBRs in municipal wastewater treatment need to be
further investigated.

Figure 5.12 : Filtration performance of the AnMBR system.

5.3.5 Cleaning tests

In order to further control membrane fouling, applying an effective strategy to extend

the stable filtration period, membrane filtration resistances were measured and
summarized in Table 5.2. The results clearly showed that the filtration resistances at
25 °C were significantly lower than the ones at 15 °C when operated under similar
hydrodynamic conditions. Indeed, RT after the cleaning procedure was similar to the
resistance of the virgin membrane at 25 °C. Higher contents of the foulants in UASB
effluent and more fine particles tended to form a more dense and non-porous cake
layer, giving rise to increased filtration resistance at 15 °C. Jeison and van Lier
(2006a) and Massé et al. (2006) also reported that a reduction in the particle size
resulted in a more compact cake layer structure and hence, a higher cake resistance.
The latter results confirm our observations: the cake layer resulting from the
material, which contains only small particles, showed a compact structure right from

111
the beginning of the filtration experiment. In contrast, the cake built up with bigger
particles was more porous and therefore allowed a higher permeate flow. The TMP
increase with time might be explained by cake consolidation, penetration of smaller
particles into the porous cake layer.

For both experimental periods, Rintrinsic of the membrane and the resistance caused by
pore-clogging were found smaller compared to the cake layer resistance. Cake layer
resistance accounted for most of the R T, representing 51% and 43% of the RT at 25
°C and 15 °C, respectively, which is consistent with the results obtained by Chu et al.
(2005) (Figure 5.13). By far most of the sludge was retained in the sludge bed of the
UASB module. However, some dispersed solids in the reactor apparently reached the
membrane module and adhered to the membrane surface. It was clear that a foulant
layer had built up on the surface, while the fouling of the membrane pores was
insignificant. Irreversible fouling was less significant since it only represented
around 20% and 35% of the total resistance at 25 °C and 15 °C, indicating the
feasibility and importance of the fouling control by minimizing the cake resistance in
AnMBRs. However, the importance of Rirreversible increased at low temperature. As
shown in Figure 5.13, an increase was observed in the contribution of the irreversible
fouling resistance to the RT of the system. The increase in irreversible fouling
indicated more pore blocking, which is a drawback of extending the membrane
operation period. Possibly, an increase in SMP resulted in the blocking of the
membrane pores.

Table 5.2 : Total filtration resistances.

Procedure Parameter Unit T=25 °C T=15 °C
Day-31 Day-61 Day-96 Day-126
Permeability L/(m2.h.bar) 723 751 779 679
Virgin membrane
Resistance 1012 m-1 0.55 0.53 0.48 0.56
Permeability L/(m2.h.bar) 187 214 119 100
Before cleaning
Resistance 1012 m-1 2.02 1.80 3.16 3.69
After physical Permeability L/(m2.h.bar) 385 434 194 184
cleaning Resistance 1012 m-1 0.98 0.89 1.9 2
After NaOCl-Divos Permeability L/(m2.h.bar) 546 550 339 264
solution cleaning Resistance 1012 m-1 0.69 0.7 1.11 1.4
After citric acid Permeability L/(m2.h.bar) 659 721 640 395
cleaning Resistance 1012 m-1 0.57 0.54 0.59 0.93

112
 Figure 5.13 : Contribution of different resistances to RTs at the two operational
temperatures.

5.4 Conclusion

Temperature affected the filtration performance of the membrane coupled UASB

reactor. Decreasing the temperature from 25 °C to 15 °C greatly influenced the
characteristics of the UASB effluent, resulting in an increased turbidity, COD and
SMP concentrations and a decreased D50 as a result of colloids accumulation at
lower temperatures. As a result, the RT increased significantly. The overall microbial
diversity also changed throughout the experimental periods, possibly invoked by the
drop in temperature. The applicable flux in the AnMBR at 15 °C was mainly limited
by higher SMP production and the decrease in D50. However, the overall treatment
performance of the AnMBR was not affected significantly, because the physical
removal by the membrane compensated for the decreased biological removal rate at
15 °C. The higher load of SMP and submicron particles to the membrane possibly
led to the formation of a compact gel layer on the membrane surface at the low
temperature. The results revealed more severe fouling at 15 °C in comparison to 25
°C, with cake layer representing the main fouling resistance during both
experimental periods. Results of critical flux tests also coincide with the R T tests of
the AnMBR. However, severe fouling at 15 °C despite the operation at sub-critical
flux conditions confirmed that critical flux is not a useful indicator to determine the
long-term filterability in the UASB–UF coupled AnMBR setup. Based on the
observed results, although the overall treatment performance is high, membrane
coupled UASB reactors are presently not yet considered a technically feasible
alternative in the treatment of municipal wastewater at 15 °C in terms of filtration
performance. Therefore, further studies should be done and other configurations,

113
including a UASB reactor combined with an additional sludge digester, need to be
evaluated in order to reduce non-stabilized particulate matter accumulation in the
sludge bed and mitigate rapid membrane fouling for the low temperature applications
of membrane coupled UASB reactors.

114
6. COMPARATIVE EVALUATION OF THE SLUDGE
CHARACTERISTICS ALONG THE HEIGHT OF UPFLOW ANAEROBIC
SLUDGE BLANKET COUPLED ULTRAFILTRATION SYSTEMS (5)

6.1 Introduction

Membrane coupled UASB reactors have been considered as an alternative of interest

for the treatment of municipal wastewater in AnMBR systems (Gao et al., 2010; Liu
et al., 2012; Ozgun et al., 2013a; Ozgun et al., 2015a). UASB reactors can be used
upfront as biofilters before membrane treatment, which prevents the membrane from
excessive exposure to high SS concentration since the sludge bed would entrap most
of the particulate matter by adsorption and biodegradation (Kataoka et al., 1992; An
et al., 2009a; Wu et al., 2009; Zhang et al., 2011; Ozgun et al., 2013b). Enhanced
interception of solids in the sludge bed limits the solids and colloidal load to the
membrane and thus, membrane flux may become less dependent on the reactor
MLSS concentration, leading to high membrane fluxes (van Voorthuizen et al., 2008;
Ozgun et al., 2013a; Martin Garcia et al., 2013).

Results of the study in Chapter 4 elucidated the potentials of membrane coupled

UASB reactors for the treatment of municipal wastewater under mesophilic
conditions. However, under sub-mesophilic conditions, hydrolysis of the retained
particulates likely becomes the rate-limiting step and particulate matter accumulation
in the sludge bed will occur leading to activity loss (Lettinga et al., 2001; Ozgun et
al., 2015b). Low temperature based limitations in the growth of methanogens will
lead to further sludge bed deterioration resulting in a poor soluble COD removal,
which will lead to more severe pore clogging problems. The latter was confirmed

(5)
This chapter is based on:

Ozgun, H., Ersahin, M.E., Zhou, Z, Tao, Y., Spanjers, H., van Lier, J.B. Comparative evaluation of
the sludge characteristics along the height of upflow anaerobic sludge blanket coupled ultrafiltration
systems. under review.

115
with the results of the study in Chapter 5. Therefore, for municipal wastewater
treatment at low and/or fluctuating temperatures, membrane integration to a UASB-
Digester system (Mahmoud et al., 2004) can be an attractive option in order to
overcome the hydrolysis limitation induced by low temperature and to reduce non-
stabilised particulate matter accumulation in the sludge bed.

In the UASB-Digester system, the non-degraded particulate COD entrapped in the

sludge bed of the UASB reactor is further stabilized in a separate CSTR type digester
operated under optimal mesophilic conditions (30-35 °C) (Mahmoud et al., 2004).
By recirculating the digested solids to the UASB reactor, the sludge bed SMA is
increased resulting in enhanced removal of soluble organics. System performance is
affected by several factors related to the sludge recirculation such as sludge
characteristics, recirculation rate and sludge volume. Therefore, careful optimization
of the sludge recirculation in UASB-Digester systems is required in order to maintain
complete conversion of biodegradable dissolved COD.

Sludge characteristics generally change over the height of UASB reactors. Variation
of sludge concentration along the UASB reactor depends on several factors such as
gas production or COD load per unit area; upflow velocity, HRT and sludge settling
characteristics (Narnoli and Mehrotra, 1997). As a consequence, a number of studies
and research activities dealing with the assessment of the change in sludge
characteristics over the height of UASB reactors due to applied upflow velocity
and/or entrapped or rising gas bubbles have been carried out on bench, pilot and full-
scale UASB installations (Agrawal et al., 1997; Uemura and Harada, 2000; Sponza,
2001; Mahmoud et al., 2004; Mahadevaswamy et al. 2004; Elmitwalli and Otterpohl,
2007; Vadlani and Ramachandran, 2008). Agrawal et al. (1997) investigated the SS
and VSS concentrations along the reactor height and maximum SS concentration in
the sludge bed was found at 30 to 70 cm height from the bottom of the reactor having
a height of 180 cm. SS concentration was lower at the bottom of the reactor due to
the floating mode of sludge related with the occlusion of gas in the bed, especially
during the winter time. Operating a bench-scale UASB reactor treating raw sewage at
relatively low temperatures, Mahmoud et al. (2004) reported a declining trend in TS
and VS values with the increase in sludge bed height and observed a clear
stratification around 40% height. Reactor scale effects might have played a role in
the obtained results. However, VS/TS ratio was almost constant through the sludge

116
bed height indicating the equal stabilization of sludge along the bed height. Uemura
and Harada (2000), operating a UASB treating sewage, observed an increase in
soluble COD at the bottom levels of the sludge bed due to the liquefaction of
entrapped solid organics. However, the values decreased afterwards due to the
further progress of methanization.

All these studies about the sludge characteristics were investigated on solely UASB
reactors. However, following the results of the study in Chapter 4, significant
changes were observed in the sludge bed due to elimination of hydraulic shear as
bacterial selection pressure factor when membranes are coupled to UASB reactors.
Membrane incorporation induced an accumulation of fine particles and a decrease in
EPS concentration in the sludge, both partly responsible for decreasing the PSD and
thus, affecting sludge settleability. Therefore, a significant change in sludge
stratification along the UASB reactor height is expected when membrane units are
coupled to the effluent. However, so far, there is no information in the literature
available concerning this possible impact.

The main aim of this study is to identify the best location in the sludge bed, i.e.
agreeing with sludge having the lowest stability and/or highest concentration, for
conveying the sludge from the UASB reactor to the digester. Within this concept, the
sludge profile was investigated by collecting sludge samples along the different
heights of the UASB reactor in an AnMBR treating municipal wastewater. The
sludge characteristics were comparatively evaluated in terms of solids content, PSD,
sludge morphology, SMA and stability. Besides, pyrosequencing was carried out for
the samples from each location in order to compare the microbial community
composition including both archaeal and bacterial communities and the relative
abundance of microbial species.

6.2 Materials and Methods

6.2.1 Wastewater source

Synthetic municipal wastewater was used as feed, having the same composition as
the one described in Chapter 3. The composition of the concentrated substrate
solution and the characterization of the synthetic municipal wastewater at the UASB
inlet are presented in Chapter 3 in Table 3.1 and Table 3.2, respectively.

117
6.2.2 Seed sludge

The reactor was seeded with flocculent anaerobic sludge obtained from a pilot-scale
UASB reactor treating black water (Sneek, the Netherlands). The characteristics of
the seed sludge are presented in Table 3.3 in Chapter 3.

6.2.3 Experimental setup

The sludge samples were derived from a laboratory-scale AnMBR consisting of a

UASB reactor with an effective volume of 7 L coupled to an external membrane
module. A schematic diagram of the experimental setup is illustrated in Figure 5.1 of
Chapter 5.

Peristaltic pumps (Watson Marlow 120U/DV) were used for circulating the flow
between the UASB reactor and the membrane module, and obtaining permeate
through the membrane. Furthermore, recirculation over the membrane surface was
performed by a peristaltic pump (Watson Marlow 620UN/R) in order to maintain a
cross-flow velocity of 1 m/s across the membrane surface independently of the
circulation flow between the UASB reactor and the membrane module. This
concentrate circulation flow-rate was adjusted enabling a stable upflow velocity in
the UASB reactor. The biogas was collected by means of a three-phase separator
installed at the top part of the UASB reactor. Biogas flow rate was measured with a
gas meter (Ritter, Milligas Counter MGC-1 PMMA). Temperature was controlled by
means of a water bath (Tamson Instruments, the Netherlands). Temperature and pH
inside the bioreactor were monitored on-line by a probe combined with a transmitter
(Elscolab, M300 ISM). A computer accessible via LabView software was used to
control the pumps and collect the data.

The external cross-flow tubular membrane module contained 28 membrane fibers

with an internal diameter of 1.5 mm and a length of 80 cm. Each membrane fiber had
0.0038 m2 of total membrane surface area. Pentair X-Flow (Enschede, the
Netherlands) PES UF membranes with a pore size of 30 nm were used in the
AnMBR. The cross-sectional image and the detailed characteristics of the membrane
are presented in Figure 5.2 and Table 5.1 of Chapter 5. Pressure sensors (AE
Sensors, ATM -800/+600 mbar, the Netherlands) were installed in the membrane
feed, concentrate and permeate lines in order to measure the TMP.

118
6.2.4 Experimental procedure

Sludge samples were taken from different heights of the UASB reactor on Day 47
and on Day 63 after an operational period of 63 days, in order to study the degree of
sludge stratification over the reactor height. Four sampling ports were selected for
sludge sampling; the first one at 9 cm above the base of the UASB and the others at
16 cm to 24 cm intervals along the height of the reactor (Figure 6.1). Sludge samples
are denoted as S1, S2, S3 and S4, respectively.

The AnMBR system was operated in a continuous mode at an average HRT of about
6 h and an average OLR of 2 kg COD/m3.day. The upflow velocity of the UASB
reactor was set at 0.6 m/h based on results discussed in Chapter 3 and kept constant
during the whole experimental period. The average pH and temperature of the system
remained stable around 6.9 and 15 °C, respectively. Membrane operation consisted
of alternating between 3 minutes filtration and 20 seconds backwash, at a membrane
flux of 12.3 L/m2.h. The cross-flow velocity of the membrane was 1 m/s.

Figure 6.1 : Schematic overview of the membrane coupled-UASB AnMBR and the
sludge sampling ports S1-S4 over the height of the UASB reactor.

6.2.5 Experimental methods

COD, TN, TSS, VSS, TS and VS were measured following Standard Methods
(APHA, 2005). Each analysis was performed in triplicate. The samples for soluble
COD measurement were filtered through a 0.45 µm filter.

119
The PSD of the sludge samples was assessed by using direct image analysis. Image
analysis of particles with a size range of 50~5000 µm was obtained by fluorescence
stereo microscope (Leica M205 FA, Germany) with QWin image analyzing software
(QG2-32, version V 3.5.1). For assessing the SMP content of sludge samples, a
volume of 5 ml sludge was washed by PBS (pH=7.2) and centrifuged at 7000g for 7
minutes at 4 °C. The supernatant was filtered using a 0.45 µm filter and the filtrate
was collected for SMP analysis. The pellet was re-suspended and vigorously washed
with 10 ml PBS and then ultrasonicated at 40 kHz (Cole-Parmer Ultrasonic, the
Netherlands) for 3 minutes. A high speed centrifuge (17000g) was applied for 20
minutes under 4 °C and the supernatant obtained was filtered using a 0.45 µm filter
for EPS assessment. The phenol-sulfuric acid method was used to quantify
polysaccharides (Dubois et al., 1956). The concentration of protein was determined
using the Bradford method (Bradford, 1976).

SMA of the sludge samples was determined by using AMPTSII (Bioprocess Control,
Sweden) (Li et al., 2011a). Stability of sludge samples was tested by obtaining the
specific ultimate methane production. Detailed explanations about SMA and stability
tests can be found in Chapter 4.

Microbial community structure was investigated using 454-pyrosequencing. Details

about sample preparation, DNA extraction, pyrosequencing and post analysis of
pyrosequencing data can be found in Chapter 4.

6.3 Results and Discussion

Figure 6.2 present the profiles of sludge characteristics along the height of UASB
reactor in terms of TS and VS. VS/TS increased from 0.70±0.01 for S1 to 0.88±0.01
for S3, followed by a slight decrease to 0.87±0.02 for S4. The relatively low VS/TS
ratio observed at the bottom of the UASB reactor might be attributed to an inorganic
sedimentation layer being present there. The maximum concentration of TS in the
sludge bed was found to be 41 g/L at the bottom of the reactor, whereas a decrease
was observed at the higher part of the reactor, which is consistent with the results of
Mahmoud et al. (2004) treating sewage in a UASB and a UASB-Digester system.
The TS concentration decreased to 8 g/L in S4. The VS-TS as well as VSS-TSS
showed a declining trend in concentration from bottom to top of the UASB with
clear stratification at about 46% height of the bottom of the reactor. Highest TSS and

120
VSS concentrations were obtained at the bottom part of UASB reactor followed by a
decrease with height. Besides, a significant increase in the VSS/TSS ratio from
0.73±0.02 for S1 to 0.97±0.004 for S3 was observed, followed by a slight decrease to
0.93±0.01 for S4. The total COD concentrations decreased from 46 g/L for S1 to 9
g/L for S4, in agreement with the TS(S) and VS(S) concentrations.

Figure 6.2 : TS and VS concentrations and VS/TS along the height of the membrane
coupled UASB reactor.

The number and volume based PSD of each sludge sample are depicted in Figure
6.3(a) and Figure 6.3(b), respectively. A different size distribution was obtained for
each of the sludge samples from different heights. Figure 6.3(b) shows that, 85% to
90% of the total particle volume for S1 are bigger than 2 mm. The smaller particles
with diameters below 100 µm represented only 1% of the total volume of particles
for S1, but made up a large fraction of the total number in the sludge. An increase in
the height led to a shift in the size distribution, with an increase in the volume of
smaller sized particles. Fraction of smaller particles with diameters below 100 µm
increased to 24% of the total volume of particles for S4 and particles bigger than 1.6
mm are not present for S4. Overall, the general trend in size distributions was very
similar in both types of PSD graphs with the majority of bigger particles contributing
to 85-90% of the total volume at lower levels in comparison to upper levels. The
results of PSD analysis along the UASB height are in agreement with the results of
Sponza (2001), which also confirmed the decrease in particle size in the upper part of
the UASB reactor.

121
Morphological changes of the sludge were also visualized along the height of
membrane coupled UASB reactor (Figure 6.4). More circular flocs with smoother
surface and a larger diameter were observed at the lower levels, whereas smaller
flocs with rough surface existed at the upper levels, which supported the PSD results.

Figure 6.3 : PSD of the sludge samples in terms of (a) particle number and (b)
particle volume along the membrane coupled UASB reactor.

122
 S4

S3

S2

S1
Figure 6.4 : Morphological changes of sludge samples along the UASB reactor.

123
SMP and EPS concentrations of the sludge samples along the UASB reactor height
are presented in Table 6.1. These samples contained cells, cell flocs bound by EPS,
SMP, and un-degraded or partially degraded particulate matter. The protein contents
of SMP and EPS were found higher than their polysaccharide contents for all of the
sludge samples. Higher protein content may be attributed to the presence of exo-
enzymes in the sludge flocs and cell lysis compounds (Neyens et al., 2004). An
increasing trend was observed for SMP concentration with the increase in reactor
height. Drews et al. (2006) indicated that a drop in SMP content can occur due to
biodegradation in a MBR. Therefore, SMP increase at higher levels might result
from the lower degradation rates observed with the increase in height. Protein and
polysaccharide content in the SMP were 1.89 mg/g VSS and 1.01 mg/g VSS for S1,
which increased to 3.30 mg/g VSS and 2.62 mg/g VSS for S4, respectively.
However, a decreasing trend was observed for EPS compositions, with the highest
concentrations at the down part of the sludge bed, coinciding with the dominance of
large particles (Figure 6.3).

Table 6.1 : SMP and EPS contents of the sludge samples along the membrane
coupled UASB reactor height (Average±Standard Deviation).
Time SMP EPS
Protein Polysaccharide Protein Polysaccharide
(mg/g VSS) (mg/g VSS) (mg/g VSS) (mg/g VSS)
S1 1.89±0.05 1.01±0.09 3.79±0.10 2.10±0.14
S2 2.10±0.03 1.38±0.04 2.48±0.03 0.90±0.05
S3 3.28±0.19 1.52±0.25 2.07±0.07 0.80±0.11
S4 3.30±0.08 2.62±0.11 1.77±0.08 0.80±0.13

Figure 6.5 shows the SMA and stability along different heights of the UASB reactor.
SMA values of S1 and S2 are higher than those of S3 and S4, which indicates that
the biomass in the lower levels is more active than that in the upper level. Highest
SMA value of 0.64 g COD/g VSS.d was observed for the sludge collected from the
bottom. Following that, SMA decreased gradually to 0.44 g COD/g VSS.d for S2 and
then remained constant thereafter for S3 and S4. Large particle sizes may reduce the
SMA due to enhanced mass transfer limitations (Bhunia and Ghangrekar, 2007).
However, SMA values were found to be higher at the low part of the sludge bed that
was characterised by a large particle size. The highest value at 9 cm may be ascribed
to the contribution due to the effect of properly controlled wastewater flow, growth

124
of hydrolytic microorganisms capable of decomposing the solid organics
accumulated in the reactor bottom part and population shift. Stability tests were also
performed in order to determine whether the sludge was equally stabilized over the
bed or not. According to the results, the sludge stability decreased from 0.53 g CH 4-
COD/g COD for S1 to 0.64 g CH4-COD/g COD for S2. Following that decrease, an
increase in stability was observed for S3 and S4. So according to the sludge stability
tests, S2 contains more biodegradable solids accumulated in the sludge bed, which
makes it favourable as an extraction point of sludge for the coupled digester.

Figure 6.5 : SMA and stability of sludge samples along the height of the membrane
coupled UASB reactor.

16S rDNA was extracted from sludge taken from the four sampling points. A total of
10078 sequences were retrieved by 454-tag Pyrosequencing. The clone libraries
showed marked differences in microbial community composition at different reactor
heights. Remarkable differences in microbial community structure were found
between S1-S3 and S4, while such differences between S1, S2 and S3 were small.
Details of bacterial and archaeal abundance can be seen in Appendix C.

The peaks of alpha-diversity index (Shannon, Chao1 and OSN) appeared for S2
(Figure 6.6), indicating the most diverse microbial community at the second bottom
sampling port of the UASB. The index of S1 and S3 were slightly lower than S2. The
alpha-diversity index of S4 was lowest compared to the other three sludge samples,
and decreased by 42% (Chao1), 38% (OSN) and 20% (Shannon) compared to the

125
corresponding maximum. This means that the UASB biomass from the top part of
the sludge bed had a much less diversity than the middle and bottom biomass. Figure
6.7 presents the Lorenz distribution curves based on the abundance of each
operational taxonomic unit (OTU) that originated from pyrosequencing of 16S rDNA
of the communities of different heights of UASB. The 45 degree diagonal is the
theoretical perfect evenness line representing an absolutely even microbial
community. A curve further away from the diagonal has a lower evenness or higher
dominance since evenness quantifies how relative abundance is distributed among
species (Wilsey and Stirling, 2007). The Lorenz distribution curves showed that the
biomass community of S4 was less even than the other three positions, meaning that
several species greatly dominated the upper part of UASB microbial community. The
biomass community of S4, dominated by less species, can be considered less diverse
in comparison to the ones in which several different species have a similar
abundance.

Figure 6.6 : Alpha-diversity index of each sludge sample along the UASB reactor.

The abundances of Bacteria and Archaea were quantified along the height of the
reactor. Remarkable increasing/decreasing trends of several bacterial and archaeal
species were discovered along the height of UASB from bottom to top. However, the
compositions of the microbial community for S1, S2 and S3 were markedly different
from that of S4. For example, the number of Cytophaga clones decreased
dramatically from 19.9% for S1 to 2% for S4. The most abundant Clostridium-

126
related OTU obtained for S1, S2 and S3 was also observed for S4, but the number of
clones was significantly less. Instead, OTUs similar to Ferrimonas marina (28%), as
well as Novispirillum itersonii (14.3%) and Cloacibacterium normanense (6.4%)
appeared to be dominant for S4 (Figure 6.8, Table C.1, Appendix C).

Figure 6.7 : Lorenz distribution curves of each sludge sample along the UASB
reactor.

Likely, availability of specific substrates and/or physical selection had contributed to

the varied microbial communities along the UASB height. For instance, as typical
primary fermenting bacteria, Cytophaga sp. can grow competitively under nutrient
rich environments (Pankratov et al., 2011; Panichnumsin et al., 2012). The relative
abundance of Cytophaga sp. decreased by 90% from bottom to top. This can be
related to the gradient decrease of substrate bioavailability (i.e. COD, Figure 6.9).
The drop in Cytophaga sp. abundance in the upper part of the UASB might be
related to the lack of granules there. PSD results (Figure 6.3) and microscopic
pictures (Figure 6.4) showed much more granular sludge in the bottom part of the
UASB and nearly no granules were found in the middle part. Abundant growth of
acetogenic bacteria such as Cytophaga sp. is often related to granular biofilms and its
abundance decreased when the granules are deteriorated (McKeown et al., 2009).

127
 Figure 6.8 : Relative abundance of major bacterial genera.

Water quality profile for soluble COD is shown in Figure 6.9. Influent has an
average soluble COD concentration of 159 mg/L. This increased to 385 mg/L at a
height of 10 cm from the bottom, likely due to hydrolysis of entrapped solid organics
that accumulated in the lower portion of the reactor and then decreased along the
reactor height due to further progress of methanization. As suggested by the water
quality profiles, most of the organic matter in the wastewater was degraded in the
bottom part of the reactor, implying that the dominant phylotypes in the bottom part
of the UASB could be responsible for removing organic matter. Higher soluble COD
removal in the lower granular sludge bed zone was also observed in the study of Ahn
et al. (2001). The same authors also measured higher specific uptake rates near the
inlet compared to other volume fractions, indicating accumulation of active biomass
near the reactor inlet (Ahn et al., 2001). The most abundant OTU in the S1 belonged
to the genus Cytophaga (Figure 6.8). This OTU, that represented approximately 20%
of all of the clones from S1 (344/1721 clones), decreased with reactor height, i.e.
236/1798, 70/1662 and 43/2034 clones in S2, S3, S4, respectively. This drop
coincided with the observed soluble COD decrease along the height of the reactor
(Figure 6.9).

128
 Figure 6.9 : Soluble COD concentration along the height of the membrane coupled
UASB reactor.

The variation in sludge composition is also illustrated by the variation in archaeal

communities over the height of the sludge bed (Figure 6.10, Table C.2, Appendix C).
Species belonging to genus Methanosaeta were the pre-dominant archaeal species in
the lower parts (S1 and S2) of the UASB, but their abundance decreased to 1/2 for S3
and even down to 1/3 for S4. As a typical acetoclastic methanogen, the growth of
Methanosaeta sp. is highly related to the presence of acetate and its availability to
biomass (Penning et al., 2006; Ince et al., 2011). High abundance appeared in the
bottom part, coinciding with the likely presence of the methanogenic precursor
acetate that is generated by acidogens and acetogens converting influent organic
matter. In addition, Methanosaeta sp. are commonly identified as the predominant
acetate consuming methanogen in granular sludge, which was particularly present at
the bottom of the UASB reactor (Figure 6.3 and Figure 6.4) (El-Mamouni et al.,
1997; McKeown et al., 2009; Kovacik et al., 2010; Kobayashi et al., 2011;
Khemkhao et al., 2012).

129
 Figure 6.10 : Relative abundance of all archaeal species.

6.4 Conclusion

Irrespective the AnMBR operation mode, sludge bed stratification manifested in a

UASB reactor coupled to a UF membrane filtration unit. Results showed striking
differences in sludge characteristics at different reactor heights along the UASB
reactor. Even though membrane incorporation led to a deterioration in sludge bed
settleability, the reactor biomass became segregated into several zones. In the lower
zones, there was variation in biomass concentration with height; and in the upper
zones, the concentration of biomass was more or less uniform. The lower zone had
dense granulated biomass, whereas the upper zone had loose active and nonactive
biomass. Analysis over the height of the reactor showed that the TS, VS, TSS and
VSS concentrations in the reactor decreased with increased height, and highest COD
concentration of 46 g/L was found at the bottom of the reactor. Image analysis
showed significant differences in particle size between each sampling port. Results
clearly showed that the highest SMA was obtained with the sludge taken from the
bottom parts. The most active biomass remained near the inlet of the reactor;
whereas, non-active biomass consisting of loose, suspended particles and flocculents
was present at the top. Comparative sequence analysis at the phylum level revealed
changes in microbial community composition with reactor height. A high diversity of

130
microbial species were present in the lower part of the reactor, whereas less diversity
was observed in the middle and upper parts. The observed sludge bed stratification
contradicts the pre-assumption that the UASB sludge bed is fully mixed when
coupled to a membrane unit in an AnMBR setup. Our findings indicate that sludge
conveyance from the low temperature UASB reactor to the parallel operating
digester for sludge stabilisation is location specific. Considering the low sludge
stability and solids content stratification, the sludge to be recirculated from UASB
reactor to digester is recommended to be taken from 40-50% of the sludge bed
height.

131
132
7. AN INTEGRATED ANAEROBIC MEMBRANE BIOREACTOR-
DIGESTER SYSTEM TO SUPPORT LOW TEMPERATURE ANAEROBIC
TREATMENT OF MUNICIPAL WASTEWATER(6)

7.1 Introduction

Anaerobic processes are generally operated under mesophilic (35 °C) and sometimes
thermophilic (55 °C) temperature conditions. However, for wastewaters with a low
organic content and relatively low temperature (e.g., municipal wastewater), methane
production is by far insufficient to cover the heating requirement for distinctly
increasing the temperature. Therefore, anaerobic reactor systems treating low
concentrated wastewaters are generally opereated under ambient temperature
conditions (An et al., 2009a). This also explains why full-scale anaerobic treatment
of municipal wastewater has been restricted to tropical areas, where the temperature
of municipal wastewater allows for sufficiently fast hydrolysis of complex organics
and SS (Zhang et al., 2013). Laboratory-scale research has shown the feasibility of
low temperature (6–15 °C) application of anaerobic processes with easily
biodegradable substrates (such as VFAs, semi skimmed milk or nonfat dry milk) for
both high and low strength wastewaters (Rebac et al., 1998; Rebac et al., 1999;
Zhang et al., 2013). Although operation at low ambient temperatures appears
technically feasible, SRTs need to be lengthened compared to mesophilic operation
since hydrolysis of solids is restrained due to the lower temperature, leading to
accumulation of non-stabilised matter in the sludge bed (van Lier et al., 2008). Thus,
municipal wastewater treatment by anaerobic systems in more temperate climates is
still considered a challenge since municipal wastewater belongs to the complex

(6)
This chapter is based on:

Ozgun, H., Zhou, Z, Ersahin, M.E., Tao, Y., Spanjers, H., van Lier, J.B. An integrated anaerobic
membrane bioreactor-digester system to support low temperature anaerobic treatment of municipal
wastewater. under review.

133
wastewater category due to the high fraction of particulate organic material,
moderate biodegradability and its low strength. Under low temperature (<20 °C)
conditions, the growth of methanogens is very slow and hydrolysis of particulate
matter into dissolved molecules becomes the rate-limiting step, which results in the
accumulation of SS in the reactor (Ozgun et al., 2013a). The solids accumulation will
impose a more frequent sludge discharge. Consequently, at lower temperatures the
amount of excess sludge produced will increase leading to a shorter SRTs, which
might then limit the net yield of methanogens resulting in a poor organic matter
conversion efficiency and thus deterioration of sludge stability with a low SMA
(Mahmoud et al., 2004). The latter will result in a poor soluble COD removal and an
overall deterioration of the digestion process (Mahmoud, 2008). Therefore, the
anaerobic treatment efficiency is largely dependent on local ambient temperatures
(Lew et al., 2011). Due to solids accumulation, the SRT becomes too short,
furthermore, the excess sludge of the UASB reactor still requires stabilization before
appropriate use or final disposal (Zhang et al., 2012), and the liquid effluent needs
further treatment.

The current challenge in anaerobic sewage treatment technology development is to

amend the system to temperate climate conditions and/or to regions that are
periodically subjected to low temperatures. Various investigations have been carried
out to overcome the process constraints related to the low temperatures and it has
been stated that anaerobic sewage treatment is certainly not limited to regions of hot
climates (Mahmoud, 2002). Besides operational aspects such as sludge
characteristics and upflow velocities, the technical design of the overall process
influences the performance (Halalsheh, 2002; Mahmoud, 2002; Seghezzo, 2004).
One of the current solutions for the treatment of raw domestic sewage at low
temperature is a two-step UASB system where in the first step mainly the removal of
SS and partial hydrolysis and acidification occurs, while in the second step the
hydrolyzed organic matter is converted to biogas (Zeeman et al., 1997; Elmitwalli et
al., 2002; Halalsheh, 2002; Seghezzo, 2004). However, results show similar
performances of the two-step reactor in comparison to a one step system, due to low
removal efficiencies and low SRT in the second stage. Zeeman et al. (1997) and
Elmitwalli et al. (2002) proposed two-stage systems consisting of a high loaded first
stage entrapping the SS, followed by a methanogenic stage for soluble COD

134
removal. The produced sludge from the high loaded first stage is, by definition, not
stabilised and needs further stabilization in a separate digester. Typical problems of
high loaded UASB reactors, such as sludge flotation and washout of active biomass
were observed. Applications of membranes for enhanced sludge retention may
alleviate some of those challenges. For example, application of an AnMBR for
sewage treatment resulted in lower effluent COD values (<100 mg/L) and SS
concentrations (<1 mg/L) compared to conventional UASB reactors at 18–25 °C (An
et al., 2009a). However, based on the results of Chapter 5, although overall treatment
performance is high, membrane coupled UASB configuration is not found
technically feasible for the treatment of municipal wastewater at 15 °C considering
the deterioration of the filtration performance, which would be a bottleneck to the
practical engineering application of AnMBRs.

Lettinga and Hulshoff Pol (1991) proposed a novel technology consisting of an

integrated high loaded UASB reactor and digester for treating sewage at low
temperature. In the UASB-Digester system, sludge circulation occurs between a
UASB and a digester. Municipal sewage was treated in a UASB at low temperature.
The solids, which are entrapped in the UASB sludge bed, are conveyed to a digester
operated at optimal conditions where they can be stabilised and the digested sludge is
re-circulated to the UASB reactor, thereby providing additional methanogenic
biomass to convert soluble COD at low temperatures. In this system, only the
fraction of the municipal sewage that is transferred from a UASB reactor to a
digester needs to be heated. The UASB reactor of this system is operated under cold
conditions (8–20 °C), while designed for summer conditions. As the loading rate is
too high to allow for complete stabilization of entrapped SS in the low temperature
UASB reactor, these solids are transferred and stabilized in the sludge digester that
operates at higher temperature.

Several studies (Álvarez et al., 2004; Mahmoud et al., 2004; Mahmoud, 2008;
Mahmoud et al., 2008; Lew et al., 2011; Zhang et al, 2012) have focused on the
impact of digester coupling on the performance of UASB reactors. A previous
laboratory experiment indicated that the removal efficiencies are increased by about
5% applying a UASB-Digester system compared to single stage UASB reactor (Ruiz
et al., 1998). Inhibition of methanogenic bacteria appeared in the external digester,
limiting the performance of the system, but this problem could be solved by means of

135
an adequate start-up and feeding strategy (Álvarez et al., 2004). Álvarez et al. (2004)
investigated the performance of a UASB-Digester system for the treatment of low
temperature raw municipal wastewater and found stable anaerobic treatment and a
slight SMA increase. Mahmoud et al. (2004) also compared the one-stage UASB
reactor with a combined UASB-Digester, where the accumulated sludge from the
UASB reactor was directed to a CSTR-Digester, operating at 35 °C, for further
treatment. The obtained results were promising as compared with the one-stage
UASB reactor. The UASB-Digester had a better performance than a single stage
UASB reactor at 15 °C, with a COD removal of 66% versus 44%. This system was
investigated in the Netherlands under controlled temperature conditions. Following
that study, Mahmoud (2008) investigated the system in Palestine, where climate and
sewage characteristics are quite different in order to assess the system response to the
Mediterranean climate seasonal temperature fluctuations. UASB performance had
remarkably improved after incorporating a digester to the system and anaerobic
treatment of high strength sewage during a hot period in the UASB-Digester system
was found very promising. Mahmoud et al. (2004) achieved an average COD
removal efficiency of 66% at 15 °C with municipal sewage of a low soluble COD
fraction (19–24% of total COD). However, several authors have shown that the
average COD removal efficiency of the system decreases to only 37–46% when
treating municipal sewage with a considerably higher soluble COD fraction (33–
44%), mainly caused by insufficient methanogenic activity in the UASB reactor
(Álvarez et al., 2004; Zhang et al., 2012). Zhang et al. (2013) improved the
performance of UASB-Digester system for these types of municipal sewage by
adding co-substrate to the sludge digester. A pilot-scale UASB-Digester system was
applied to treat real municipal sewage, and glucose was added in the sludge digester
as a model co-substrate in order to produce additional methanogenic biomass, which
was continuously recycled to inoculate the UASB reactor. An increase was observed
in soluble sewage COD removal efficiency, SMA and methane production and co-
digestion is proposed as a suitable approach to support a UASB-Digester for
pretreatment of low temperature municipal sewage.

Optimization of sludge recirculation should be done carefully in UASB-Digester

systems in order to maintain complete conversion of biodegradable dissolved COD.
System performance is affected by several factors in sludge recirculation such as

136
sludge characteristics, recirculation rate and sludge volume. In previous studies of
Álvarez et al. (2004), Mahmoud et al. (2004) and Mahmoud (2008), the sludge
recirculation rate was regulated to maintain the sludge bed height at a certain level.
Besides, the location for sludge removal in the sludge bed of the UASB reactor was
not optimized. The amount of sludge that needs to be recirculated is crucial to the
viability of the UASB-Digester system, as it determines the required energy input to
heat the transferred sludge from 15 to 35 °C. Lew et al. (2011) optimized the sludge
recirculation process by determining the best location in the sludge bed for removing
sludge. During cooler periods below 20 °C, influent wastewater particulate matter
accumulates in the top part of the UASB reactor sludge and can be conveyed to the
digester. The digester can be designed with an operational volume 33% smaller than
the UASB. Zhang et al. (2012) also improved the sludge recirculation process of a
UASB-Digester system treating domestic sewage at low temperature by investigating
the effect of sludge recirculation rate on the performance. In this work, the effect of
the sludge recirculation rate in the UASB-Digester system on COD removal
efficiency, biogas production, the stability of the UASB sludge and the digester
sludge, and methanogenic activity of the UASB sludge at 15 °C, was investigated.
The results showed that the total COD removal efficiency increases with increasing
sludge recirculation rate. Based on the energy available in the wastewater, a
maximum sludge recirculation flow of 2-2.5% of the influent flow rate still allows
energy self-sufficiency. The results present practical information for the operation
and improvement of the UASB-Digester system treating domestic sewage at low
temperature.

All of these studies about the digester incorporation were investigated on solely
UASB reactors. However, with growing interest in cost-effective water reclamation
techniques, AnMBRs are becoming increasingly popular for the treatment of
municipal wastewater when high effluent quality and/or agricultural use of treated
effluents is required (Ho and Sung, 2010; Ozgun et al., 2013a; Prieto et al., 2013;
Wei et al., 2014). Membrane fouling, which causes a reduction in the flux throughout
the operation, remains an unavoidable drawback of the AnMBR and limits its
widespread application in wastewater treatment (Wang et al., 2013). Fouling
mechanisms depend on many factors such as membrane characteristics, sludge
characteristics, environmental conditions, reactor design and the operation strategy

137
(An et al., 2009a). Membrane coupled UASB reactors may be a promising approach
to reduce problems related to membrane fouling (Ozgun et al., 2013b). High fluxes
are expected in membrane coupled UASB configurations since the membrane is now
only challenged by the supernatant but not by the bulk sludge. The results presented
in Chapter 4 clearly supported the applicability of a membrane-coupled UASB
reactor for municipal wastewater treatment at 25 °C, in spite of the sludge bed
deterioration. A pathogen free but nutrient rich permeate quality was achieved during
the AnMBR stage with quite stable TMP values (Ozgun et al., 2015a). However, at
15 °C the system performance deteriorated (Chapter 5; Ozgun et al., 2015b). Further
studies should be done and other configurations need to be evaluated in order to
mitigate rapid membrane fouling for the low temperature applications of membrane-
coupled UASB reactors. Digester coupling can be an attractive alternative for
producing high quality and nutrient-rich effluents for reuse purposes under moderate
climate conditions. So far, there is no study in the literature about the impact of
digester incorporation to the membrane-coupled UASB systems treating low
temperature sewage.

The present work investigated an AnMBR-Digester system treating municipal

wastewater at a temperature of 15 °C. The AnMBR was firstly operated without
incorporating the digester aiming at monitoring of the overall AnMBR performance,
i.e. (i) treatment performance (ii) filtration performance and (iii) sludge quality. This
was done to obtain a reference to assess the achievements obtained from
incorporating a digester. Afterwards, the one-stage AnMBR was modified to an
AnMBR-Digester system by incorporating a digester operated at 35 °C in order to
assess the performance of the AnMBR-Digester system on the same indicators. The
characteristics of the sludge that will be circulated to the digester determines the
viability of the AnMBR-Digester system. Results described in Chapter 6 showed
differences in sludge stability and a clear sludge stratification over the reactor height,
and indicated that the best location for sludge recirculation is at about 40-50% of the
sludge bed height. Additionally, the sampling location at about 75% of the sludge
bed height was comparatively evaluated for removing sludge. The overall
information obtained from this study will be valuable for the operation and
improvement of the AnMBR-Digester system treating municipal wastewater at low
temperatures.

138
7.2 Materials and Methods

7.2.1 Wastewater source

Synthetic municipal wastewater as described in Chapter 3 (Table 3.1 and Table 3.2)
was used as feed.

7.2.2 Seed sludge

The UASB reactor was seeded with flocculent anaerobic sludge obtained from a
pilot-scale UASB reactor treating black water (Sneek, the Netherlands). The
characteristics of the seed sludge are presented in Table 3.3 in Chapter 3. The excess
sludge withdrawn from the AnMBR at different intervals during the operation at 15
°C was used as inoculum for the digester.

7.2.3 Experimental setup

The experimental setup consisted of a laboratory-scale CSTR type digester with an

effective volume of 4 L and UASB type reactor with an effective volume of 7 L
coupled with an external membrane module. The experimental period was divided
into three stages: (i) before digester incorporation, when the system was operated as
an AnMBR, (ii) after digester incorporation, when the system was operated as an
AnMBR-Digester system with sludge recirculation from the sludge transfer point of
S3 (Figure 7.1) and (iii) after digester incorporation, when the system was operated
as an AnMBR-Digester system with sludge recirculation from the sludge transfer
point of S2 (Figure 7.1).

A schematic diagram of the experimental setup is depicted in Figure 7.1. The

AnMBR was equipped with feed, circulation pump and effluent pump (Watson
Marlow 120U/DV). The circulation pump (Watson Marlow 120U/DV) was used for
circulating the flow between the UASB reactor and the membrane module, and
obtaining permeate through the membrane. Furthermore, recirculation over the
membrane surface was applied by a peristaltic pump (Watson Marlow 620UN/R) in
order to maintain a cross-flow velocity of 1 m/s across the membrane surface
independently of the circulation flow between the UASB reactor and the membrane
module. This concentrate circulation flow rate was adjusted enabling a constant
upflow velocity in the UASB reactor.

139
 Figure 7.1 : Schematic diagram of the AnMBR-Digester system.

UASB effluent was directed through an external cross-flow tubular module

containing 28 UF (0.03 µm pore size) membrane fibers (Pentair X-Flow), and the
concentrate was returned to the UASB reactor. Each membrane fiber had an internal

140
diameter of 1.5 mm and a length of 80 cm, with an effective filtration area of 0.0038
m2. The characteristics of the membrane are presented in Table 5.1 in Chapter 5.
Pressure sensors (AE Sensors, ATM -800/+600 mbar, the Netherlands) were
installed in the membrane feed, concentrate and permeate lines in order to measure
the TMP.

Taps were installed over the whole UASB reactor height at about 8 cm apart for
sludge recirculation and analysis. The biogas was collected by means of a three-
phase separator installed at the top part of the UASB reactor. Hereafter, the AnMBR
setup was modified to the AnMBR-Digester system by incorporating a CSTR type
digester. Biogas flow rates of the AnMBR and digester were measured separately
with a gas meter (Ritter, Milligas Counter MGC-1 PMMA). Produced biogas was
recycled by a diaphragm pump (KNF, N86 KTDCB) via a diffuser at the bottom of
the reactor to provide mixing inside the digester. Temperature was controlled by
means of a water bath (Tamson Instruments, the Netherlands). Temperature and pH
inside the reactors were monitored on-line with a probe combined with a transmitter
(Elscolab, M300 ISM). A computer running LabView software (LabVIEW 10.0.1,
National Instruments) was used to control the pumps and collect the data.

7.2.4 Experimental procedure

The experimental work was carried out over three successive periods. In the first
period, the AnMBR was operated for 63 days at 15 °C without incorporating the
digester. Afterwards, the AnMBR was modified to the AnMBR-Digester system by
incorporating a CSTR digester operated at 35 °C. Two sampling points at different
heights of the UASB sludge bed, represented by S2 (located at 25 cm height about
40-50% of the sludge bed height) and S3 (located at 41 cm height about 75% of the
sludge bed height), were comparatively investigated for optimal sludge recirculation
location. In the second and third stage, the AnMBR-Digester system was operated
for 126 days, of which during the first 63 days S3 was used for sludge circulation and
S2 during the remaining days. The operational and design details of the AnMBR-
Digester system are given in Table 7.1. As can be seen in Figure 7.1, the sludge was
transferred from one reactor to another once a day by recirculating the sludge from
the sludge transfer points of S2 and S3 in UASB reactor to the sludge transfer point
of T3 in digester and then from the sludge transfer point of T2 in the Digester to the

141
sludge transfer point of S1 in UASB reactor, ca. 9 cm from the bottom representing
the lowest part of the AnMBR. Sludge transfer of S1 to digester was not investigated
in order to avoid short-circuiting of the UASB influent. This sludge recirculation rate
between digester and AnMBR was 200 mL/d, which corresponded to 0.8% of the
influent flow rate of 25 L/d. During each stage, the AnMBR was operated with
cycles consisting of 180 seconds of filtration, followed by 20 seconds of backwash.
Backwash was applied by reversing the flow direction, keeping the flux at the same
value as during filtration.

Table 7.1 : Operational conditions of the AMBR-Digester system.

Parameters Unit AnMBR Digester
HRT day 0.25 20
OLR kg/m3.day 2 -
Temperature °C 15 35
Upflow Velocity m/h 0.6 -
Cross-flow Velocity m/s 1 -
Flux L/m2.h 12.3 -

7.2.5 Experimental methods

7.2.5.1 Analysis

COD, NH4+-N, TN, TP and TSS were measured in triplicate following Standard
Methods (APHA, 2005). The samples for soluble COD measurement were filtered
through a 0.45 µm filter. The turbidity and pH were measured by HACH 2100N
Turbidimeter and by a WTW multi720 pH meter, respectively. The PSD of the
effluent was assessed by a Mastersizer 2000 (Malvern Instruments, Hydro 2000 MU)
that has a detection range of 0.02-2000 µm. Laser diffraction technique was used to
measure the size of the particles. The methane content in the biogas was measured by
a Varian 3800 gas chromatograph equipped with a flame ionization detector (Varian,
Palo Alto, CA). Statistical calculations were carried out by using Minitab 16
software (Minitab Inc., USA).

SMA of the sludge samples was assessed by an AMPTSII (Bioprocess Control,

Sweden) (Li et al., 2011a). Stability of sludge samples was assessed by obtaining the
specific ultimate methane production. Detailed explanations about SMA and stability
tests can be found in Chapter 4.

142
7.2.5.2 Filtration resistance

After long-term reactor operation, the fouled membranes were cleaned with a
sequential cleaning procedure to calculate the various fouling resistances and
evaluate the characteristics of membrane fouling. Details about the assessment of the
various resistances can be found in Chapter 5.

7.2.5.3 Microbial community analysis

Microbial community structure was investigated using 454-pyrosequencing. Details

about sample preparation, DNA extraction, pyrosequencing and post analysis of
pyrosequencing data can be found in Chapter 4.

7.3 Results and Discussion

7.3.1 Treatment performance

Figure 7.2 shows the total and soluble COD in the UASB effluent and final permeate
of the AnMBR system during all experimental stages. Stable operation performance
was considered to be achieved once stable effluent COD and biogas production were
obtained. During the single stage operation of the AnMBR, average total and soluble
UASB effluent COD were 1759±92 mg/L and 153±3 mg/L, respectively, under
stabilised conditions. Following this period, sludge was started to be recirculated
from the sludge transfer point S3 from the UASB reactor, starting at the beginning of
stage 2. During the second stage, a gradual increase was observed in average total
COD to the value of 1966±32 mg/L. However, the average soluble COD
concentration of 149±6 mg/L was found to be similar to the values obtained during
the first stage. Therefore, the observed increase in total COD concentration may be
attributed to the release of colloidal and suspended particles from the sludge bed
resulting from the recirculation. Following the second stage, the sludge to be
recirculated from UASB reactor to digester was taken from S2 where the sludge had
the lowest stability, as recommended based on the results of Chapter 6. The effect of
the utilization of a certain sludge transfer point on the system performance was
investigated. The lowest average total COD concentration of 1405±51 mg/L was
achieved at this stage also with an additional decrease in average soluble COD
concentration to 133±3 mg/L. Compared to the second stage results, the observed

143
decrease in total and soluble COD concentration during the third stage confirmed the
effect of sludge transfer point on the performance of AnMBR-Digester systems.

Figure 7.2 : Total and soluble COD concentrations in the UASB effluent and final
permeate during three experimental stages.

The lowest average total and soluble COD concentration in the UASB effluent was
achieved during the third stage. Possibly improved methanogenic and hydrolytic
conditions in the combined system were established in the third stage. The
improvement in soluble COD removal might be either attributed to the soluble COD
transfer to the digester which is subsequently converted to methane in the digester, or
to the increased methanogenic capacity coming from the digester to the UASB
reactor. Higher colloidal and suspended COD removal may result from a better solids
entrapment (physical process), which is caused by improved digestion conditions
based on the existence of the capturing-de-capturing mechanism for solids removal
in the sludge bed. Elmitwalli et al. (2001) found that the colloidal particles in the
sewage are highly biodegradable and attributed the poor colloidal removal during
anaerobic sewage treatment at low temperatures to poor physical removal. Similarly,
the improvement of suspended COD removal may also have resulted from the
improved digestion conditions, which on its turn enhanced colloidal COD removal,
since colloids may be generated from the SS (Elmitwalli et al., 2000). Mahmoud et
al. (2004) also suggested the existence of a capturing-de-capturing mechanism for
improved colloidal and suspended COD removal efficiency in the UASB-Digester

144
system. They hypothesized that a particle to be digested, should firstly be captured
and once captured, it would be degraded in case the digestion conditions are
appropriate. Otherwise, it might detach from the sludge bed. Therefore, the physical
removal of solids might not only be affected by the characteristics of the influent
particles and the sludge but also by the prevailing digestion conditions (Mahmoud et
al., 2004).

Average total COD removal efficiencies, based on influent and permeate COD, were
about 90% with average COD concentrations of 51±5 mg/L, 55±2 mg/L and 49±4
mg/L in the permeate. A slight difference in the overall treatment performance of the
AnMBR during each experimental stage was due to the physical retention capacity of
the membrane. Higher overall suspended and colloidal COD removal efficiencies in
spite of the higher total COD in the UASB effluent during the first and second stages
were mainly due to the accumulation in the sludge bed.

SMA and stability measurement results of the UASB and the digester sludge are
shown in Table 7.2. The highest SMA of the UASB sludge was obtained during the
third stage indicating the improved conversion of sewage solids to methane and
biomass, and an increased supply of methanogens to the UASB sludge. SMA of the
sludge in the UASB reactor was about 90% of that in the digester during the third
stage. The sludge digested in the digester had also substantially higher stability than
the excess sludge from the UASB reactor of the UASB-Digester system.

The stability of the UASB sludge remained constant during the first and second
stages, showing that the sludge is relatively unstable, i.e. it still contains considerable
amounts of biodegradable solids and accumulation of such solids in the sludge bed.
This may be explained by a limited hydrolysis of the entrapped solids, which is in
agreement with the results presented by de Man (1990). However, the stability of the
UASB sludge drastically improved from 0.65±0.02 to 0.54±0.02 g CH4-COD/g COD
during the third stage. The sludge accumulation during the first and second stages
was accompanied with UASB effluent quality deterioration (Figure 7.2). However,
average total COD removal efficiencies, based on influent and permeate COD, were
similar during all experimental stages. Thus, although the total COD removal
efficiencies were similar, lower stability in the first and second stages compared with
the third stage was actually attributed to the organic solids accumulation during the
first and second stages. The incorporation of the digester and selection of the sludge

145
transfer point of S2 for sludge transfer resulted in improved stabilization of the
accumulated solids. Higher stability of the anaerobic sludge through the
biodegradation and removal of e.g. fats and other inactive substances coincided with
an observed increased SMA.

Table 7.2 : SMA and stability of the AnMBR and the digester sludge.
SMA Stability
Experimental Period (g CH4-COD/g VS.d) (g CH4-COD/g COD)
UASB Digester UASB Digester
AnMBR 0.38±0.01 - 0.64±0.03 -
AnMBR-Digester
0.35±0.02 0.37±0.01 0.65±0.02 0.11±0.01
(Sampling Point: S3)
AnMBR-Digester
0.44±0.01 0.49±0.01 0.54±0.02 0.09±0.01
(Sampling Point: S2)

Table 7.3 shows the average biogas production rate of the UASB and the digester at
different experimental stages. Average methane contents of the biogas were 69±4%,
70±3% and 73±2% during the three experimental stages. The highest total methane
production was achieved during the third stage when particulate and colloidal COD
transfer to the digester occurred, which could be then converted to methane. The
biogas production in the digester significantly increased during the third stage
resulting from the increase in the captured suspended and colloidal COD. Besides
that, the increase can be explained by the large amounts of methanogens supplied
from the digester to the UASB reactor that enhanced the conversion of soluble COD
to methane in the UASB reactor. This explanation is consistent with the results in
Figure 7.2 showing a decrease in the soluble COD concentration in the UASB
effluent during the third stage.

Lower biogas production rate was obtained from the UASB reactor in comparison to
the biogas from the digester during all the experimental stages. The recovered
methane in the gas amounted to 0.12±0.01, 0.13±0.01 and 0.23±0.01 L CH4/g
CODremoved for the 3 stages, respectively. During the single stage operation of the
AnMBR, an average methane yield of 0.12±0.01 L CH4/g CODremoved was obtained,
which represented 33% of the maximum theoretical value, i.e. 0.369 L CH4/g
CODremoved at 15 °C. A slight increase in the methane yield to 0.13±0.01 L CH 4/g
CODremoved, which was 35% of the maximum theoretical value, was observed during

146
the second stage. COD difference between influent and permeate of the AnMBR was
represented as the removed COD in the calculation of methane yield but apparently
the actually degraded COD that was converted to methane was lower than the
removed COD. Besides, a large part of the methane might be dissolved in the
permeate. Smith et al. (2013) found that up to 40-50% of total methane generated in
an AnMBR can be dissolved in the permeate. Methane yields below theoretical
values are therefore commonly observed in AnMBR studies. Martinez-Sosa et al.
(2011a) and Huang et al. (2011) reported methane yields ranging from 0.124 to 0.27
L CH4/g CODremoved in AnMBR studies. The highest methane yield was achieved
during the third stage confirming the higher biodegradability of COD removed.

Table 7.3 : Average biogas production rate in the UASB and the digester.
Biogas Biogas Total biogas
production in production in production
Experimental Period the UASB the digester
reactor
(L/ d) (L/d) (L/d)
AnMBR 2.0±0.1 - 2.0±0.1
AnMBR-Digester
0.9±0.1 1.3±0.1
(UASB Sampling Point: S3) 2.2±0.2
AnMBR-Digester
1.1±0.1 2.4±0.2 3.5±0.3
(UASB Sampling Point: S2)

Table 7.4 presents the overall removal efficiencies in terms of COD, TSS, TN, NH4+
and TP during each experimental stage. As a general comparative evaluation of the
system performance during all stages, regardless of the increases in COD and TSS
concentrations in the UASB effluent, SS-free (non-detected) AnMBR permeate was
obtained with average COD values between 49 and 55 mg/L, standing for overall
COD removal efficiencies of about 90% in the AnMBR. As expected nitrogen and
phosphorus concentrations in the permeate were higher compared to the influent
values, underlining the fertilising value of the treated effluent. However, when
nutrient removal is required, a separate post-treatment step needs to be combined
with the proposed AnMBR system.

147
 Table 7.4 : Influent and permeate characteristics in terms of COD, TSS, TN, NH4+
and TP during anaerobic municipal wastewater treatment in the single stage AnMBR
and the AnMBR-Digester system.
AnMBR-Digester AnMBR-Digester
Single stage system system
Influent AnMBR (UASB Sampling (UASB Sampling
Parameter
mg/L Point: S3) Point: S2)
Permeate Removal Permeate Removal Permeate Removal
mg/L % mg/L % mg/L %
COD 530±30 51±5 90±0.9 55±2 90±0.4 49±4 91±0.7
TSS 230±25 < 10 >99 < 10 >99 < 10 >99
TN 54±5.2 65±7 ~0 75±11 ~0 63±3 ~0
NH4+ 36±5.5 64±9 ~0 73±9 ~0 59±3 ~0
TP 12±0.8 14±4 ~0 15±3 ~0 14±2 ~0

7.3.2 Filtration performance

7.3.2.1 Flux and TMP

Figure 7.3 presents the applied flux and resulting TMP during each experimental
stage. The membrane flux was maintained constant at 12.3 L/m2.h throughout the
experimental study. During each run, membrane cleaning including physical, base
and acid cleaning steps was carried out nearly every 30 days to recover the P,
followed by another operational period. After stable reactor conditions were
achieved, gradual increase in the TMP was observed over time during operation of
the single stage AnMBR showing evidence of gradual membrane fouling. The
highest point in the curve corresponded to the TMP of 437 mbar measured at the end
of the first stage. Operation of the AnMBR at low temperature resulted in abrupt
membrane fouling evidenced by a substantial increase in TMP over the course of 28
days. Similar TMP increase profiles were developed during the second stage. TMP
also increased rapidly within each filtration cycle, reaching values of 418 and 426
mbar, respectively. However, TMP increased slowly with operation time and low
TMP values were maintained during the third stage, which is in agreement with the
results of UASB effluent characteristics (Figure 7.2). Single stage operation of the
AnMBR resulted in the fastest membrane fouling, followed by second and third
stage. The differences between first, second and third stage suggested an increased
filtration performance after digester incorporation and sludge transfer from the
sludge transfer point S2. The slow increase in TMP in the third stage is most likely
related to the low solids content in the UASB effluent and presence of bigger
particles as evidenced by the PSD measurements.

148
 Figure 7.3 : Filtration performance of a single stage AnMBR and an AnMBR-
Digester system.

7.3.2.2 TSS and turbidity

Figure 7.4 presents TSS and turbidity in the effluent of the UASB reactor. Average
TSS concentrations were 505±34 mg/L, 741±73 mg/L and 476±30 mg/L during the
experimental stages. A considerable increase from the range of 440-552 mg/L to
611-820 mg/L was observed in the TSS concentration during the second stage
confirming the release of suspended particles from the sludge bed. This increase was
followed by a decrease during the third stage. Consequently, there was no significant
change in TSS concentrations in the effluent between the first and third stage.
However, different average turbidities in the UASB effluent were observed during
each stage. The average turbidity in the effluent was 732±46 NTU during the first
stage, whereas it increased to 889±122 NTU during the second stage. Subsequently,
considerable decrease to 363±36 NTU was observed during the third stage.
Variations observed in effluent turbidity between the first and third stage seem
contradictory to TSS results which do not show significant differences. Observed
results are attributed to the decrease in colloidal particles washout from the system
instead of suspended particles washout after digester incorporation as turbidity is an
indicator of both suspended and/or colloidal particles. Since TSS concentrations were
very similar at first and third stages, the higher fouling rate at the first stage was
probably not associated with TSS concentration. Smaller colloidal particles measured
during the single stage operation of AnMBR could be partly responsible for the
higher fouling rate, related with their fouling tendency due to the low back-transport.

149
 Figure 7.4 : TSS concentration and turbidity in the UASB effluent of a single stage
AnMBR and an AnMBR-Digester system.

7.3.2.3 Particle size distribution

Figure 7.5 presents PSD curves of the UASB effluent. During the single stage
operation of the AnMBR, there was a shift to smaller sizes of the distribution pattern
and an apparent decrease in the D50 was observed, resulting in an effluent having a
D50 of 25 μm. Likewise, the range of D50 that was measured in the UASB effluent
at the second stage, was between 27 µm to 37 µm, whereas an apparent increase in
the D50 was observed at the third stage, from 30 µm on Day 129 to 55 µm on Day
189, resulting in an extinction of a distribution curve at lower ranges. Finer particles
observed in the UASB effluent could explain the accelerated fouling development
(Figure 7.3) during the first and second stage since hydrodynamic back transport of
particles is positively related with particle size (Tardieu et al., 1998). From these
results, it was concluded that at the single stage operation of AnMBR, small sized
particles were washed out more vigorously from the system. Increase in the D50
after digester incorporation and sludge transfer from S2 is likely related to the
enhancement in the biodegradation of influent solids retained in the UASB reactor.

150
Figure 7.5 : PSD of the UASB effluent during anaerobic municipal wastewater treatment at different operational stages (a) AnMBR, (b)
AnMBR-Digester (UASB Sampling Point: S3), (c) AnMBR-Digester (UASB Sampling Point: S2).

151
7.3.2.4 Filtration resistances and cleaning tests

The filtration resistances are summarized in Table 7.5. The results clearly show that
the filtration resistances at the third stage were significantly lower than those at the
first and second stage. Higher contents of foulants in the UASB effluent and finer
flocs (Figure 7.5) tended to form a denser and nonporous cake layer, giving rise to
increased filtration resistance before digester incorporation. Massé et al. (2006) also
reported that a reduction in the particle size results in a more compact cake layer
structure and hence higher cake resistance. Membrane resistance after the cleaning
procedure was similar to the resistance of the new membrane at the third stage,
whereas membrane resistances after the cleaning procedure were higher in
comparison to the virgin ones during first and second stage.

Table 7.5 : Total filtration resistances.

AnMBR- AnMBR-
Digester system Digester system
Single stage
Procedure Parameter Unit (UASB (UASB
AnMBR
Sampling Point: Sampling Point:
S3) S2)
Day- Day- Day- Day- Day- Day-
31 61 96 126 159 189
Permeability L/(m2.h.bar) 779 679 438 392 377 344
Virgin membrane
Resistance 1012 m-1 0.48 0.56 0.86 0.96 1.07 1.1
Permeability L/(m2.h.bar) 119 100 95 83 139 136
Before cleaning
Resistance 1012 m-1 3.16 3.69 4.27 4.57 2.84 2.8
After physical Permeability L/(m2.h.bar) 194 184 169 164 274 270
cleaning Resistance 1012 m-1 1.9 2 2.18 2.46 1.47 1.4
After NaOCl-Divos Permeability L/(m2.h.bar) 339 264 254 256 345 313
solution cleaning Resistance 1012 m-1 1.11 1.4 1.5 1.58 1.17 1.2
After citric acid Permeability L/(m2.h.bar) 640 395 372 348 371 327
cleaning Resistance 1012 m-1 0.59 0.93 1.09 1.06 1.1 1.15

For all the runs, the Rintrinsic of the membrane and the resistance caused by pore-
clogging were found smaller compared to the cake layer resistance (Figure 7.6).
Cake layer resistances accounted for most of the total resistance, representing 43%,
47% and 52% of the RT for the 3 stages, respectively, which is consistent with the
results obtained by Chu et al. (2005). Normally, the sludge is retained in the sludge
bed of the UASB. However, some dispersed solids in the reactor reached the
membrane module with the upward liquid flow and adhered to the membrane surface
due to permeate extraction. It was clear that a foulant layer had built up on the
surface, while the fouling of the membrane pores was insignificant. Irreversible
fouling was marginal since it only contributed to about 35%, 28% and 11% of the

152
total resistance for the 3 stages, respectively, indicating the feasibility and
importance of the fouling control by minimizing the cake resistance in AnMBRs.
However, the importance of Rirreversible increased during the first and second stage. As
shown in Figure 7.6, a decrease was observed in the contribution of irreversible
fouling resistance to RT of the system at the third stage.

Figure 7.6 : Contribution of different resistances to RT at different operational stages

(a) AnMBR, (b) AnMBR-Digester (UASB Sampling Point: S3), (c) AnMBR-
Digester (UASB Sampling Point: S2).

7.3.3 Microbial community

The performance of the AnMBR was slightly negatively affected when the sludge
was recirculated from the sludge transfer point S3, but recovered and even enhanced
when the sludge transfer was done from the sampling point S2. It is expected that the
microbial community could have changed. In order to gain insights into the
relationship between microbial-ecological characteristics and changes in reactor
performance in response to the incorporation of the digester and the transfer from
different points, biomass samples were collected from both the digester and UASB at
different operational stages including the seeding sludge for the digester and UASB
for further analysis using 454-pyrosequencing.

From an ecological point of view, a repeated temperature shock caused by the

biomass exchange between the digester (under 35 °C) and UASB (15 °C) is an
environmental stress to both communities, and under such stress some species can
outgrow and some may be suppressed (Zhou et al., 2013; Li et al., 2014). A direct
result of such blending of biomass between the digester and UASB was that the
UASB community became more and more different to the initial or seed community,

153
but became more and more similar to the digester’s community. Such trend can be
clearly seen from the change of alpha-diversity (Figure 7.7), which is a
comprehensive assessment over the richness and evenness of each single biomass.
The alpha-diversity includes four indices: Shannon-Wiener index, Chao1 richness,
phylogenetic diversity and OSN. The alpha-diversity of the UASB significantly
decreased (p<0.001 for each index) from the first stage to the third stage.

Figure 7.7 : Alpha-diversity of the biomass samples collected from the UASB and
digester during AnMBR, AnMBR-Digester (UASB Sampling Point: S3), AnMBR-
Digester (UASB Sampling Point: S2) operation.

The transition from the second stage to the third was induced by lowering the
transfer point from S3 down to S2. S2 is closer to the bottom of the UASB, where
more and larger particles exist. Hence, the change in the microbial communities in
both the digester and UASB, along with the variation of reactor performance, can be
attributed to the difference in sludge samples that were transferred from S3 (smaller
size) or S2 (larger size). It is well known that large granules comprise abundant
methanogens (Lens et al., 1993), which were found at S2 more near the bottom of the
UASB reactor, close to the feed entrance point. Under the prevailing low temperature
conditions in the UASB, the in-reactor methanogenic activity of the granules is
relatively low, however, the activity will significantly increase when transfered to the

154
digester under mesophilic temperature. The observed results, such as an enhanced
methane production in the digester (Table 7.3), agree well with this assumption.

Clear variation in bacterial assemblies was observed in the UASB before and after
the incorporation of the digester (Figure 7.8, Table D.1, Appendix D). Some bacterial
species that were originally predominant in the first stage decreased by a great
percentage in the second stage, such as Lactococcus raffinolactis (from 12.8% to
2.0%) and Cytophaga sp. (from 7.1% to 2.4%). Lactococcus raffinolactis is prevalent
in many environments and like many species belonging to the genus Lactococcus,
due to their commercial values, they have been deeply investigated both
physiologically and genetically (Meslier et al., 2012). Lactococcus raffinolactis cells
are ovoid, always in pairs or short chains. They have peculiar property of
fermentation and are found in both industrial diary processing reactors (Boucher et
al., 2003) and AD reactors (Nelson et al., 2011; Espana-Gamboa et al., 2012).
Lactococcus raffinolactis were the most abundant species of the UASB in the first
stage. However, their relative abundance sharply decreased when the digester
biomass was introduced. Another previously dominant species, Cytophaga sp., are
typical fermenting bacteria that prefer mesophilic conditions in AD reactors
(Weissbrodt et al., 2014). Their relative abundance was high (7.1 %) in the first stage
but gradually decreased. The explanation of that is possibly a combination of the
unfavourable temperature (15 °C) and the interruption by the digester biomass.

There are also some bacterial species that were not dominant in the first stage but
became dominant in the second stage, including Flavobacterium sp. (from 3.2% to
18.5%), Acinetobacter sp. (from 4.8% to 18.9), Bacteroides sp. (from 2.9% to 9.8%),
Clostridium botulinum (from 3.2% to 11.8%). Flavobacterium sp., Acinetobacter sp.
and Bacteroides sp. are all common anaerobes that are able to ferment hydrocarbons
under stressed conditions (Álvarez et al., 2006b; Ueki et al., 2011; Mesle et al.,
2013). They are prevalent in soils, sewage and ADs (Madigan et al., 2011; Gao et al.,
2012; Ke et al., 2014; Li et al., 2013). It is intriguing to see an increased abundance
of Clostridium botulinum. Clostridium botulinum is a typical spore-forming species,
generally widespread in soils, sediments, feces and biogas systems (Bagge et al.,
2010; Vidal et al., 2013). A common characteristic shared among these four species
is that they are able to grow under stressed conditions, and such ability favours them

155
to possibly outcompete other bacteria during the exchange between the digester and
UASB.

During the third stage, most of the key species had lower abundance except for
Lactococcus raffinolactis and Cloacibacterium normanense, whose abundance
increased from 2.0% to 6.0% and from 1.6% to 15.8%, respectively. Lactococcus
raffinolactis apparently recovered in abundance at the third stage, following a
possible adaptation to the temperature shock. Cloacibacterium normanense is a
facultative anaerobic, rod-shaped microorganism that was originally isolated from
raw sewage (Allen et al., 2006). This species is able to degrade complex organic
matter and hence is frequently discovered in anaerobic wastewater treatment systems
(Suihko and Skytta, 2009; Maspolim et al., 2015). The significant increase in the
abundance of Cloacibacterium normanense might be related to the performance
enhancement during the third stage. It is reported that Cloacibacterium normanense
is able to grow in a broad range of temperatures with distinct growth between 18 and
36 °C (Allen et al., 2006).

There were also some substantial changes in the digester community. First, two very
dominant species, Trichococcus sp. and Nostocoida limicola actinobacteria were
gradually eliminated. Their abundance decreased from over 30% to less than 1%. In
addition, the relative abundance of Bellilinea sp. greatly increased when the sludge
transfer point of S3 changed to S2, from 24.6% to 31.2%. Members of Bellilinea are
obligate anaerobic fermenters that produce VFAs (Ma et al., 2013) The mesophilic
strains of Bellilinea sp. were originally isolated from rice paddy soil (Ma et al., 2013)
and have been found in UASB reactor (Yamada et al., 2007). It has been reported
that Bellilinea sp. are specialized in propionate-degradation and can cooperate with
hydrogenotrophic methanogens under mesophilic conditions (Yamada et al., 2007).
Similar to Bellilinea sp., another typical syntrophic bacteria, Carboxydibrachium sp.,
also showed an increased abundance from 1.2% at the second stage to 9.8% at the
third stage. Carboxydibrachium sp. are very versatile (Subbotina et al., 2003;
Fardeau et al., 2004). Such versatility can not only guarantee them a wide niche
under stressed conditions such as the temperature shock of the digester-UASB
exchange, but also enable cooperation with acetoclastic and hydrogenotrophic
methanogens. The presence of a stable methanogenic community agrees with the
improved biogas production in the third stage.

156
The archaeal community of the UASB was also affected by the inclusion of the
digester (Figure 7.9). The acetoclastic methanogen Methanosaeta sp. was the most
dominant archaea in the UASB before incorporation of the digester, but its relative
abundance decreased from 83.7% at the first stage to 64.3% at the second stage after
digester coupling (Figure 7.9, Table D.2, Appendix D). The UASB community was
again dominated by acetoclastic Methanosaeta sp. at the third stage, with an
abundance as high as 86.9%. A hydrogenotrophic methanogen, Methanospirillum
sp., had a high abundance (28.6%) at the second stage, whereas this species was not
detectable before the digester coupling. The increase in hydrogenotrophic
methanogens indicates the presence of syntrophic associations in the UASB (Stams
and Plugge, 2009). Methanosaeta prevailed in all the digester samples with high
abundance over 80%. It has been reported that Methanosaeta sp. have high substrate
affinity, which makes them dominant in the low residual acetate conditions like the
current system (Bialek et al., 2013).

Figure 7.8 : Relative abundance of major bacterial species in the digester and UASB
during AnMBR, AnMBR-Digester (UASB Sampling Point: S3), AnMBR-Digester
(UASB Sampling Point: S2) operation.

157
 Figure 7.9 : Relative abundance of all archaeal species in the digester and UASB
during AnMBR, AnMBR-Digester (UASB Sampling Point: S3), AnMBR-Digester
(UASB Sampling Point: S2) operation.

7.4 Conclusion

The results of these investigations reveal the high potential of the new configuration
AnMBR-Digester system for treating sewage at 15 °C as it combines wastewater
treatment and sludge stabilization. The filtration performance of the investigated
single stage AnMBR was limited by the low temperature. Single stage AnMBR
operation at 15 °C led to organic solids accumulation in the UASB reactor and thus,
greatly influenced the characteristics of UASB effluent, resulting in an increase in
COD concentration and turbidity and a decrease in D50, which led to a significant
increase in the RT. Similar TMP increase profiles were observed during the
operational period after digester incorporation with sludge transfer from S3, which
coincided with the slight changes in PSD, turbidity and COD in the UASB effluent.
However, change in sludge transfer location from about 75% of the sludge bed
height to about 40-50% of the sludge bed height after digester incorporation had
remarkably influenced the characteristics of UASB effluent, resulting in a decrease
in turbidity, soluble and total COD and an increase in D50, which led to a significant
decrease in the RT. Improved stability and SMA of the sludge were achieved in the
UASB reactor with an increase in total biogas production. The sludge recirculation

158
from S2 in the AnMBR-Digester system improved both the solids physical removal
and the conversion, confirming the importance of selecting the proper sludge transfer
point, especially for the filtration performance of the AnMBR-Digester systems.
Therefore, under optimized sludge recirculation conditions, the integrated AnMBR-
Digester system represents an efficient technology for anaerobic sewage treatment in
order to mitigate rapid membrane fouling for the low temperature applications of
membrane coupled UASB reactors.

159
160
8. CONCLUSIONS AND RECOMMENDATIONS

8.1 Conclusions from the Various Research Steps

Upgrading of existing anaerobic treatment processes with membranes can be of vital

importance, especially when high effluent quality and/or use of treated effluents are
considered. In recent years, the AnMBR has been gaining attention as a means to
treat low strength wastewater due to its inherent benefits including shorter start-up
periods, smaller footprint, superior effluent quality leading to reuse and recycling
possibility, complete biomass retention without any concern of sludge washout, and
energy recovery through methane production (Ho and Sung, 2010; Ozgun et al.,
2013a; Wei et al., 2014). AnMBR effluents are characterised by a very high
reduction in pathogen content, which makes them suitable for agricultural use
(Saddoud et al., 2006; Ellouze et al., 2009). Therefore, an AnMBR can be used not
only for on-site wastewater treatment, but also for the generation of nutrient-rich
irrigation water for agricultural applications (Prieto et al., 2013). However, the wide
application of the AnMBRs in domestic utilities, especially in full-scale systems, has
not yet become a reality, which is partly because of system novelty and reluctance
owing to membrane fouling problems (Chu et al., 2005). Therefore, AnMBR systems
will likely benefit from the development of efficient technologies to prevent fouling.

Treatment of municipal wastewater by different types of anaerobic bioreactors in

AnMBRs has drawn considerable attention of many researchers. Different types of
anaerobic bioreactors, including CSTR, UASB, EGSB, etc., have been investigated
in combination with various types of membranes in different configurations (Ozgun
et al., 2013a). In high-rate anaerobic reactors such as sludge bed systems and
anaerobic filters, biomass is retained either by the formation of granular and/or thick
flocculent sludge or by attachment to a support material. Because in membrane
coupled anaerobic high-rate reactors biomass is not directly subjected to membrane
filtration, dense cake layer formation and consolidation will be less apparent in
comparison with membrane coupled CSTRs. Therefore, UASB reactors may offer a
good opportunity for being combined with membranes. However, the impact of

161
membrane application on sludge immobilization and sludge bed stability is not yet
well understood.

Despite the expected more efficient membrane filtration step when applying sludge
bed reactors, for the treatment of municipal wastewater under sub-mesophilic
conditions, hydrolysis of the retained particulates likely becomes rate-limiting and
particulate matter accumulation in the sludge bed will occur, subsequently resulting
in activity loss (Lettinga et al., 2001). A significant reduction in the hydrolysis rate of
solids is typical at low temperatures due to the low activity of anaerobic
microorganisms (Lettinga et al, 2001; Zhang et al., 2010). Methane loss is another
important issue of concern especially at low temperature, since methane solubility in
the liquid phase increases with decreasing temperature. Gimenez et al. (2012)
proposed biogas-assisted mixing as a limitation for super-saturation and a guarantee
for a minimum concentration of dissolved methane in the effluent of AnMBR
systems. Furthermore, a thinner and less porous cake layer on the membrane surface
can be expected, possibly leading to more severe pore clogging in UASB coupled
AnMBRs, because the membrane would not be protected by a cake layer, but
exposed directly to only fine particles instead of a range of particle sizes. Therefore,
the degree of small particle retention by the sludge bed will be of prime importance
for the feasibility of coupled membrane filtration. Further studies should be done and
new configurations need to be evaluated in order to mitigate rapid membrane fouling
in membrane coupled UASB reactor systems under low temperature conditions.

An integrated AnMBR-Digester system may offer a solution for these challenges.

The combination of a UASB and a digester, the so-called UASB-Digester system,
has been shown to be successful for mutual sewage treatment with a high
concentration of suspended organic solids at low temperature and sludge stabilization
(Álvarez et al., 2004; Mahmoud et al., 2004; Mahmoud, 2008; Mahmoud et al.,
2008; Lew et al., 2011; Zhang et al, 2012).

This thesis investigated the development and optimization of AnMBR technology for
municipal wastewater treatment enabling energy recovery and effluent reuse in
irrigation. A membrane coupled UASB reactor was used to produce high quality and
nutrient-rich effluent for reuse purposes. Moreover, a digester was incorporated with
the AnMBR system to solve the bottlenecks of AnMBRs under moderate climate
conditions.

162
Based on the various studies and results presented in this dissertation, the following
conclusions can be drawn.

In Chapter 3, an optimum upflow velocity was identified that will result in an

effluent with good filterability for the case of membrane coupled UASB systems.
Despite the fact that the functioning of UASB systems depends on both physical
parameters and biological processes, the physical parameters have been barely
reported in literature. The reason is that the underlying mechanisms are complex and
depend on various interrelated parameters. In addition, the lack of a serious attempt
to gather the entire physical theme in one picture, have resulted in just a superficial
understanding of this field of science. Better understanding of the interaction and role
of these parameters is essential for the development of anaerobic treatment
technologies (Mahmoud, 2002). One of the most important parameters that might
affect the solids-liquid separation process by filtration through the sludge bed of a
UASB reactor is upflow velocity. Thus, upflow velocity plays a major role in
controlling the fouling potential that is one of the most important challenges for
stable operational process performance of membrane coupled UASB reactors. A high
upflow velocity in a UASB coupled membrane system resulted in an increased load
of small particles to the membrane. PSD of the effluent varied significantly with
variation in upflow velocity. Effluent characterization results coincided with
filterability tests. The increased effluent particle size at the lower upflow velocity
was associated with a better filterability. Better filterability was obtained with an
upflow velocity of 0.6 m/h in comparison to 1.2 m/h. Results show that upflow
velocity is an important factor affecting the effluent filterability in UASB reactors.

In Chapter 4, the impact of membrane integration on the performance of the UASB

reactor was assessed based on both biological removal efficiency and physical
aspects. An upflow velocity of 0.6 m/h was selected based on the results of Chapter
3. Coupling a membrane to a UASB reactor resulted in significant changes in both
physical and biological aspects, mainly due to the elimination of hydraulic dilution as
bacterial selection pressure factor. EPS content decreased and a shift in PSD of the
sludge to smaller particles was observed in the UASB reactor after membrane
addition. Decrease in the sludge particle size deteriorated the sludge bed settleability,
which led to a higher sludge washout, resulting in an increase in COD, TSS and SMP
concentrations of the UASB effluent. A decrease in the sludge activity was observed

163
after membrane incorporation. Stability tests confirmed that the decrease in SMA
was related to the particulate matter accumulation. However, SS-free permeate with
an average COD of 42 mg/L was obtained and despite the sludge bed deterioration,
the average TMP value was 85 mbar during the entire operation period indicating
that no severe membrane fouling occurred in the AnMBR. Overall system
performance increased in terms of COD removal efficiency and methane production
after membrane incorporation. Considering all responses of the UASB reactor to
membrane incorporation, the overall results put UASB reactor forward as a suitable
alternative for coupling membranes in AnMBRs at 25 °C.

In Chapter 5, the effect of low temperature on the removal efficiency and filterability
of the AnMBR has been investigated. Temperature affected the filtration
performance of the membrane coupled UASB reactor. Decreasing temperature from
25 °C to 15 °C greatly influenced the characteristics of UASB effluent, which led to
a significant increase in the RT. However, overall treatment performance of the
AnMBR was not affected significantly because of the physical removal capacity of
the membrane. More severe fouling occurred at 15 °C in comparison to 25 °C.
Higher SMP production at lower temperature induced more membrane pore blocking
and enhanced cake layer development on the membrane surface. Based on the overall
results, it is concluded that membrane coupled UASB configuration is not found
technically feasible for the treatment of municipal wastewater at 15 °C considering
the deterioration of the filtration performance, which would be a bottleneck to the
practical engineering application of AnMBRs.

In Chapter 6, the sludge characteristics along the height of the UASB reactor were
investigated in terms of sludge morphology, activity and stability in order to identify
the most appropriate sludge transfer point between the UASB reactor and digester.
Differences were observed in sludge characteristics at different reactor heights along
the UASB reactor. Membrane incorporation led to a deterioration in the sludge bed
settleability due to the accumulation of particles (Chapter 4 and Chapter 5); however,
stratification was still pronounced in sludge characteristics along the UASB reactor
height and the reactor biomass became stratified into several zones. The lower zone
of the UASB reactor had dense granulated biomass, whereas the upper zone had
loose active and non-active biomass. The highest SMA was obtained in the sludge
taken from the bottom of the UASB reactor. The active biomass remained near the

164
inlet of the reactor; whereas, non-active biomass consisting of loose, suspended
particles and flocs accumulated at the top part of the sludge bed. Treatment was
achieved by a stratified microbial habitats responsible for removing organic matter in
the membrane coupled UASB reactor. Because of sludge stability and solid content
stratification, the sludge to be recirculated from UASB reactor to digester is
recommended to be taken from about 40-50% of the sludge bed height where the
sludge has the lowest stability.

In Chapter 7, the impact of digester coupling on treatment and filterability

performance of the AnMBR was evaluated by transferring the digested sludge to the
UASB reactor to improve its methanogenic capacity. Sludge recirculation from the
sludge transfer point S2, located at 40-50% of the sludge bed height in the AnMBR-
Digester system, improved both the solids physical removal and conversion,
confirming the importance of sludge transfer point location. The AnMBR-Digester
system with the proper sludge transfer point provided substantially better removal
efficiencies than the single stage AnMBR. Permeates with a very low or maybe
negligible pathogen content which is nutrient rich can be obtained with the AnMBR-
Digester system at stable TMP enabling compliance with the regulations for water
reuse. The results show that the sludge in the UASB of the AnMBR-Digester system
was substantially more stable compared to the sludge in the single stage AnMBR.
Also, the digester sludge has substantially higher stability than the excess sludge
from the UASB of the AnMBR-Digester system. Under optimized sludge
recirculation conditions, the integrated AnMBR-Digester system represents an
efficient technology for anaerobic sewage treatment in order to mitigate rapid
membrane fouling for low temperature applications of membrane coupled UASB
reactors.

8.2 Main Outcomes and Evaluation

SUR test proved to be capable of providing information about the filterability of the
UASB effluent since effluent characterization results coincided with the filterability
and reversibility tests. The reversibility and SUR values of the UASB effluent were
reciprocally correlated.

The colloidal and soluble fractions (<1 µm) play an important role in the filterability
of the UASB effluent. The effluent filterability goes hand in hand with the amount of

165
submicron particles, which in turn is influenced by the upflow velocity. Therefore,
the amount of submicron particles is very likely an important factor determining
effluent filterability.

The upflow velocity (Chapter 3) does not have a strong influence on the overall
AnMBR functioning since incorporation of the membrane to the UASB reactor
results in significant differences in the UASB effluent (Chapter 4).

The prevailing selection pressure proved to be capable of providing sludge with good
settling properties (Chapter 4). Under the applied experimental conditions when the
UASB reactor was operated as the sole process before membrane incorporation, light
and dispersed sludge was washed out while heavier particles were retained in the
reactor. After membrane incorporation, more dispersed biomass was clearly visible
in the UASB reactor, indicating the presence of a bulking type of sludge with very
poor settling characteristics. We therefore postulate that the addition of the
membrane barrier eliminates the hydraulic selection pressure required for a sludge
bed with good settling properties in the UASB reactor.

The metabolic capacity of the sludge sampled under the prevailing condititions were
assessed by using SMA tests. Addition of the membrane resulted in a significant
drop in the SMA, which can likely be attributed to the changes in both physical and
biological properties of the sludge. After membrane addition, influent particulate
matter accumulated within the system, due to the presence of a physical barrier, thus
lowering the SMA. Stability tests may be performed in order to test the extent of
non- or slowly biodegradable matter accumulation in the sludge bed.

Temperature is a dominating parameter influencing the filtration performance of the

AnMBR, which was shown to be linked to the UASB effluent characteristics.
Therefore, seasonal temperature changes which are apparent under full-scale
conditions, will likely impact both operation and performance of the AnMBR similar
to full-scale aerobic MBR reactors (Krzeminski et al., 2012). However, membrane
separation allows to achieve a relatively stable quality of permeate despite the
changes in temperature. The quality of permeate is most likely not affected by the
changes in filterability.

Flux-step determination of the critical flux should not be used to predict long-term
TMP behaviour in real AnMBR systems (Chapter 5), which supported the former

166
hypothesis of Le Clech et al. (2003). This method is based on a short-term run, but
nonetheless provides useful data on comparative fouling propensity.

Cake layer composition is directly related to the D50 and PSD. The presence of
particles <1 µm instead of TSS concentration, determined the cake layer resistance
and thereby the achievable operational fluxes. The cake layer resulting only from
small particles showed a compact structure, characterised by a higher resistance and
lower fluxes. In contrast, the cake built up with bigger particles was more porous and
therefore allowed a higher permeate flow (Chapter 5), which agrees with previous
findings (Jeison et al., 2009b). The foulant layer had been built up on the surface,
whereas the fouling of the membrane pores was insignificant. However, an increase
in the contribution of the irreversible fouling resistance to the R T of the system
indicated more pore blocking, which is a drawback of the extension of the membrane
filtration period, applying fluxes exceeding the critical flux.

Even though membrane incorporation led to a deterioration in sludge bed

settleability, a.o. due to the accumulation of particles, sludge bed stratification was
still pronounced, as evidenced by the sludge characteristics along the UASB reactor
height. After prolonged periods of operation, the reactor biomass became segregated
into several zones (Chapter 6).

The selection of the type of anaerobic reactor apparently has a strong influence on
the overall AnMBR functioning. Compared to AnMBRs consisting of a CSTR,
membrane fluxes may become less dependent on the reactor MLSS concentration,
possibly leading to high membrane fluxes in AnMBRs consisting of UASB reactors
at ambient temperatures. Depending on the operational temperature, the digester can
be by-passed. However, for low temperature applications, the incorporation of a
sludge digester in an AnMBR system is indispensible to improve the filtration
performance of the system.

Optimization of sludge recirculation should be done carefully in AnMBR-Digester

systems in order to maintain complete conversion of biodegradable dissolved COD.
For instance, selection of the optimum sludge transfer point resulted in significant
changes especially for the filtration performance of AnMBR-Digester systems
(Chapter 7).

167
8.3 Future Perspectives and Recommendations for Further Research

Based on the conducted research described in this dissertation several aspects that
deserve attention and require further focus are proposed.

 The excellent performance of the AnMBR-Digester system requires

confirmation by long-term bench-scale studies under defined conditions,
along with relevant research at pilot-scale. However, the technology requires
optimization in terms of digestion and compactness. When scaled-up, biogas
utilization for heating the digester should be considered.

 More research is needed into the interaction between the sludge

physicochemical characteristics, digestion conditions and sludge physical
behavior. The integrated AnMBR-Digester system requires further fine
tuning, especially with respect to the size of the digester and recirculation rate
between the digester and the AnMBR.

 The research presented in this thesis has been conducted in order to assess the
feasibility of anaerobic treatment of municipal wastewater under low
temperature conditions. Considering the wide variety of sewage influent
characteristics, the AnMBR-Digester system should be studied under field
conditions, which so far is only limitedly described in literature. Recently,
Zhang et al. (2012) reported the feasibility of the UASB-Digester system for
the treatment of domestic sewage at low temperature with a low influent
dissolved COD to total COD ratio.

 Inorganic fouling by the precipitation of struvite, potassium ammonium

phosphate and/or calcium carbonate may be one of the major concerns in
fouling due to the release of ammonia and PO43- from organic nitrogen and
phosphorus during AD and the pH increase as a result of changes in CO 2
partial pressure and alkalinity generation in AnMBRs (Liao et al., 2006).
Salazar-Pelaez et al. (2011b) estimated that the chemical equilibrium for
struvite precipitation is not reached in municipal wastewaters with lower
concentrations of NH4-N, PO43- and magnesium compared to industrial
wastewaters. Therefore, struvite precipitation would unlikely occur.
However, not only the concentration of these ions but also the membrane
properties can play a significant role in struvite precipitation (Kang et al.,

168
 2002). It is therefore recommended, to further investigate the possible
occurrence of struvite in AnMBRs treating municipal wastewaters and its
impact on irreversible fouling.

 One of the major concerns about using treated municipal wastewater in

agriculture is related with the removal of toxicity and endocrine disrupting
chemicals. Industrial discharges to municipal sewers can cause serious
problems such as toxicity and organic shock loads in the end-of-pipe
wastewater treatment plants. Various studies address the removal of
endocrine disrupting chemicals in municipal wastewater treatment by MBR
systems (Cases et al., 2011; Tadkaew et al., 2011; Boonyaroj et al., 2012). In
aerobic MBRs, phenolic compounds, phthalates and estrogens can be more
effectively removed in comparison to conventional activated sludge systems
through biodegradation, adsorption and membrane rejection mechanisms
(Cases et al., 2011). However, there exists quite limited information about
their fate and biodegradation in AnMBRs (Ho et al., 2007). Obviously, more
research should be conducted on the fate of these priority pollutants in
AnMBR systems.

 Salinity can be another problem for the treatment of municipal wastewater,

especially for coastal residential areas with improper infrastructure which
allows the infiltration of sea water into sewer systems. In this situation, more
attention should be given to fouling control in order to have a feasible
treatment technology.

 When agricultural use of the treated effluent is not considered, a nutrient

removal step is generally required to comply with discharge standards.
Innovative solutions as a nitrogen polishing step with Anammox processes
and autotrophic nitrogen removal through post-treatment are currently
investigated in the water line of wastewater treatment and can be integrated.
These processes are disturbed by presence of solids, organic compounds and
COD in the water. AnMBR-Digester systems will produce more clean
effluents and are therefore better compatible with these systems.

 A synthetic municipal wastewater recipe that aims to mimic a real municipal

wastewater was used in the study. However, real wastewater may contain

169
 antibiotics and other organic pollutants from domestic and industrial
discharges, which seldom appears in the synthetic wastewater. Therefore,
further experiments should be performed with real wastewater in order to
confirm the comparability of the synthetic wastewater composition to real
wastewater.

 Alternative membrane materials, reactor and membrane types, and membrane

integration possibilities for the AnMBR technology in municipal wastewater
treatment schemes are of interest for further development (Ozgun et al.,
2013a). More research should be conducted to understand the impact of
different membrane materials and membrane integration possibilities. Based
on our results we speculate that membrane coupled UASB reactor may result
in the first applications of nanofiltration coupled to bioreactors. Besides,
different membrane integration possibilities, including the one that employs
the membrane unit as a polishing step after the UASB reactor without the
concentrate stream return to the bioreactor, need to be investigated.

170
REFERENCES

Achilli, A., Cath, T. Y., Marchand, E. A. & Childress, A. E. (2009). The forward
osmosis membrane bioreactor: A low fouling alternative to MBR
processes, Desalination,239, 10-21.
Achilli, A., Marchand, E. A. & Childress, A. E. (2011). A performance evaluation
of three membrane bioreactor systems: Aerobic, anaerobic, and
attached-growth, Water Science and Technology,63, 2999-3005.
Agrafioti, E. & Diamadopoulos, E. (2012). A strategic plan for reuse of treated
municipal wastewater for crop irrigation on the Island of Crete,
Agricultural Water Management,105, 57-64.
Agrawal, L. K., Ohashi, Y., Mochida, E., Okui, H., Ueki, Y., Harada, H. &
Ohashi, A. (1997). Treatment of raw sewage in a temperate climate
using a UASB reactor and the hanging sponge cubes process, Water
Science and Technology,36 (6-7), 433-440.
Ahn, Y., Min, K. & Speece, R. E. (2001). Pre-acidification in anaerobic sludge bed
process treating brewery wastewater, Water Research,35 (18), 4267-
4276.
Aiyuk, S. & Verstraete, W. (2004). Sedimentological evolution in an UASB
treating SYNTHES, a new representative synthetic sewage, at low
loading rates, Bioresource Technology,93, 269-278.
Akram, A. & Stuckey, D.C. (2008). Flux and performance improvement in a
submerged anaerobic membrane bioreactor (SAMBR) using powdered
activated carbon (PAC), Process Biochemistry,43, 93-102.
Allen, T. D., Lawson, P. A., Collins, M. D., Falsen, E. & Tanner, R. S. (2006).
Cloacibacterium normanense gen. nov., sp. nov., a novel bacterium in
the family Flavobacteriaceae isolated from municipal wastewater,
International Journal of Systematic and Evolutionary
Microbiology,56 (6), 1311-1316.
Altmann, J. & Ripperger, S. (1997). Particle deposition and layer formation at the
crossflow microfiltration, Journal of Membrane Science,124 (1), 119-
128.
Álvarez, B., Secades, P., Prieto, M., McBride, M. J. & Guijarro, J. A. (2006b). A
mutation in Flavobacterium psychrophilum tlpB inhibits gliding
motility and induces biofilm formation, Applied and Environmental
Microbiology,72 (6), 4044-4053.
Álvarez, J. A., Armstrong, E., Presas, J., Gómez, M. & Soto, M. (2004).
Performance of a UASB-digester system treating domestic
wastewater, Environmental Technology,25 (10), 1189-1199.

171
Álvarez, J. A., Ruiz, I., Gómez, M., Presas, J. & Soto, M. (2006a). Start-up
alternatives and performance of an UASB pilot plant treating diluted
municipal wastewater at low temperature, Bioresource Technology,97,
1640-1649.
An, Y. Y., Yang, F. L., Bucciali, B. & Wong, F. S. (2009a). Municipal wastewater
treatment using a UASB coupled with cross-flow membrane filtration,
Journal of Environmental Engineering,135 (2), 86-91.
An, Y., Wang, Z., Wu, Z., Yang, D. & Zhou, Q. (2009b). Characterization of
membrane foulants in an anaerobic non-woven fabric membrane
bioreactor for municipal wastewater treatment, Chemical Engineering
Journal,155 (3), 709-715.
APHA. (2005). Standard Methods for Examination of Water and Wastewater, 21st
ed., Washington, USA: American Public Health Association.
Aquino, S. F., Hu, A. Y., Akram, A. & Stuckey, D. C. (2006). Characterization of
dissolved compounds in submerged anaerobic membrane bioreactors
(SAMBRs), Journal of Chemical Technology and Biotechnology,81
(12), 1894-1904.
Ayala, D. F., Ferre, V. & Judd, S. J. (2011). Membrane life estimation in full-scale
immersed membrane bioreactors, Journal of Membrane Science,378
(1-2), 95-100.
Bachmann, B. J. (1955). Studies on Cytophaga fermentans, n.sp., a facultatively
anaerobic lower myxobacterium, Journal of General Microbiology,13
(3), 541-551.
Bagge, E., Persson, M. & Johansson, K. E. (2010). Diversity of spore-forming
bacteria in cattle manure, slaughterhouse waste and samples from
biogas plants, Journal of Applied Microbiology,109 (5), 1549-1565.
Bandara, W. M. K. R. T. W., Kindaichi, T., Satoh, H., Sasakawa, M., Nakahara,
Y., Takahashi, M. & Okabe, S. (2012). Anaerobic treatment of
municipal wastewater at ambient temperature: Analysis of archaeal
community structure and recovery of dissolved methane, Water
Research,46, 5756-5764.
Berube, P. R., Hall, E. R. & Sutton, P.M. (2006). Parameters governing permeate
flux in an anaerobic membrane bioreactor treating low-strength
municipal wastewaters: A literature review, Water Environment
Research,78 (8), 887-896.
Bhunia, P. & Ghangrekar, M. M. (2007). Required minimum granule size in
UASB reactor and characteristics variation with size, Bioresource
Technology,98, 994-999.
Bialek, K., Cysneiros, D. & O'Flaherty, V. (2013). Low-temperature (10 °C)
anaerobic digestion of dilute dairy wastewater in an EGSB bioreactor:
Microbial community structure, population dynamics, and kinetics of
methanogenic populations, Archaea,2013, 346171.
Boonyaroj, V., Chiemchaisri, C., Chiemchaisri, W., Theepharaksapan, S. &
Yamamoto, K. (2012). Toxic organic micro-pollutants removal
mechanisms in long-term operated membrane bioreactor treating

172
 municipal solid waste leachate, Bioresource Technology,113, 174-
180.
Boucher, I., Vadeboncoeur, C. & Moineau, S. (2003). Characterization of genes
involved in the metabolism of α-galactosides by Lactococcus
raffinolactis, Applied and Environmental Microbiology,69 (7), 4049-
4056.
Bradford, M. M. (1976). A rapid and sensitive method for the quantitation of
microgram quantities of protein utilizing the principle of protein-dye
binding, Analytical Biochemistry,72, 248-254.
Brindle, K. & Stephenson, T. (1996). The application of membrane biological
reactor for the treatment of wastewaters, Biotechnology and
Bioengineering,49, 601-610.
Brockmann, M. & Seyfried, C. F. (1997). Sludge activity under the conditions of
crossflow microfiltration, Water Science and Technology,35, 173-181.
Bunani, S., Yörükoğlu, E., Sert, G., Yüksel, Ü., Yüksel, M. & Kabay, N. (2013).
Application of nanofiltration for reuse of municipal wastewater and
quality analysis of product water, Desalination,315, 33-36.
Calderon, K., Rodelas, B., Cabirol, N., Gonzales-Lopez, J. & Noyola, A. (2011).
Analysis of microbial communities developed on the fouling layers of
a membrane-coupled anaerobic bioreactor applied to wastewater
treatment, Bioresource Technology,102 (7), 4618-4627.
Caporaso, J. G., Kuczynski, J., Stombaugh, J., Bittinger, K., Bushman, F. D.,
Costello, E. K., …Knight, R. (2010). QIIME allows analysis of high-
throughput community sequencing data, Nature Methods,7 (5), 335-
336.
Cardinali-Rezende, J., Debarry, R. B., Colturato, L. F., Carneiro, E. V.,
Chartone-Souza, E. & Nascimento, A. M. (2009). Molecular
identification and dynamics of microbial communities in reactor
treating organic household waste, Applied Microbiology and
Biotechnology,84 (4), 777-789.
Cases, V., Alonso, V., Argandona, V., Rodriguez, M. & Prats, D. (2011).
Endocrine disrupting compounds: A comparison of removal between
conventional activated sludge and membrane bioreactors,
Desalination,272 (1-3), 240-245.
Cath, T. Y., Hancock, N. T., Lundin, C. D., Hoppe-Jones, C. & Drewes, J. E.
(2010). A multi-barrier osmotic dilution process for simultaneous
desalination and purification of impaired water, Journal of Membrane
Science,362, 417-426.
Chan, Y. J., Chong, M. F., Law, C. L. & Hassell, D. G. (2009). A review on
anaerobic–aerobic treatment of industrial and municipal wastewater,
Chemical Engineering Journal,155 (1-2), 1-18.
Chang, I. S., Le Clech, P., Jefferson, B. & Judd, S. (2002). Membrane fouling in
membrane bioreactors for wastewater treatment, Journal of
Environmental Engineering,128, 1018-1029.

173
Chen, S. & Dong, X. (2005). Proteiniphilum acetatigenes gen. nov., sp. nov., from a
UASB reactor treating brewery wastewater, Journal of Systematic and
Evolutionary Microbiology,55 (6), 2257-2261.
Chernicharo, C. A. L. (2007). Anaerobic Reactors, Biological Wastewater
Treatment Series. Volume 4. London, UK: IWA Publishing.
Chernicharo, C. A. L, van Lier, J. B., Noyola, A. & Ribeiro, T. B. (2015).
Anaerobic sewage treatment in Latin America. In H.H.P. Fang, T.
Zhang (Eds.), Anaerobic Biotechnology: Environmental Protection
and Resource Recovery (Chapter 12). London, UK : World Scientific,
Imperial College Press.
Cho, B. D. & Fane, A. G. (2002). Fouling transients in nominally sub-critical flux
operation of a membrane bioreactor, Journal of Membrane
Science,209, 391-403.
Choo, K. H., Kang, I. J., Yoon, S. H., Park, H., Kim, J. H., Adiya, S. & Lee, C.
H. (2000). Approaches to membrane fouling control in anaerobic
membrane bioreactors, Water Science and Technology,41(10-11),
363-371.
Choo, K. H. & Lee, C. H. (1998). Hydrodynamic behavior of anaerobic biosolids
during cross-flow filtration in the membrane anaerobic bioreactor,
Water Research,32, 3387-3397.
Chu, L. B., Yang, F. L. & Zhang, X. W. (2005). Anaerobic treatment of domestic
wastewater in a membrane-coupled expanded granular sludge bed
(EGSB) reactor under moderate to low temperature, Process
Biochemistry,40 (3-4), 1063-1070.
de Man, A. W. A. (1990). Anaerobic treatment of raw domestic sewage using
granular sludge UASB-reactors. Wageningen University,
Wageningen, the Netherlands : Final Report of a Pilot Plant Study.
de Man, A. W. A., van der Last, A. R. M. & Lettinga, G. (1988). The use of
EGSB and UASB anaerobic systems or low strength soluble and
complex wastewaters at temperatures ranging from 8 °C to 30 °C.
Proceedings of the 5th International Symposium on Anaerobic
Digestion. Bologna, Italy.
Dereli, R. K., Ersahin, M. E., Ozgun, H., Ozturk, I., Jeison, D., van der Zee, F.
& van Lier, J. B. (2012). Potentials of anaerobic membrane
bioreactors to overcome treatment limitations induced by industrial
wastewaters, Bioresource Technology,122, 160-170.
Dhaked, R. K., Singh, P. & Singh, L. (2010). Biomethanation under psychrophilic
conditions, Waste Management,30 (12), 2490-2496.
Drews, A. (2010). Membrane fouling in membrane bioreactors-Characterisation,
contradictions, cause and cures, Journal of Membrane Science,363, 1-
28.
Drews, A., Vocks, M., Bracklow, U., Iversen, V. & Kraume, M. (2008). Does
fouling in MBRs depend on SMP?, Desalination,231 (1-3), 141-149.

174
Drews, A., Vocks, M., Iversen, V., Lesjean, B. & Kraume, M. (2006). Influence of
unsteady membrane bioreactor operation on EPS formation and
filtration resistance, Desalination,192 (1-3), 1-9.
Dubois, M., Gilles, K. A., Hamilton, J. K., Rebers, P. A. & Smith, F. (1956).
Colorimetric method for determination of sugars and related
substances, Analytical Chemistry,28 (3), 350-356.
Dutta, K., Lee, M., Lai, W. W., Lee, C. H., Lin, A. Y., Lin, C. & Lin, J. (2014).
Removal of pharmaceuticals and organic matter from municipal
wastewater using two-stage anaerobic fluidized membrane bioreactor,
Bioresource Technology,165, 42-49.
Elimelech, M. & Phillip, W.A. (2011). The future of seawater desalination: Energy,
technology, and the environment, Science,333, 712-717.
Ellouze, M., Saddoud, A., Dhouib, A. & Sayadi, S. (2009). Assessment of the
impact of excessive chemical additions to municipal wastewaters and
comparison of three technologies in the removal performance of
pathogens and toxicity, Microbiological Research,164, 138-148.
El-Mamouni, R., Leduc, R. & Guiot, S. R. (1997). Influence of the starting
microbial nucleus type on the anaerobic granulation dynamics,
Applied Microbiology and Biotechnology,47 (2), 189-194.
Elmitwalli, T. & Otterpohl, R. (2007). Anaerobic biodegradability and treatment of
grey water in upflow anaerobic sludge blanket (UASB) reactor, Water
Research,41, 1379-1387.
Elmitwalli, T. A., Oahn, K. L. T., Zeeman, G. & Lettinga, G. (2002). Treatment
of domestic sewage in a two-step anaerobic filter/anaerobic hybrid
system at low temperature, Water Research,36 (9), 2225-2232.
Elmitwalli, T. A., Soellner, J., de Keizer, A., Bruning, H., Zeeman, G. &
Lettinga, G. (2001). Biodegradability and change of physical
characteristics of particles during anaerobic digestion of domestic
sewage, Water Research,35, 1311-1317.
Elmitwalli, T. A., van Dun, M., Bruning, H., Zeeman, G. & Lettinga, G. (2000).
The role of filter media in removing suspended and colloidal particles
in anaerobic reactor treating domestic sewage, Bioresource
Technology,72, 235-242.
Ersahin, M. E., Ozgun, H., Dereli, R. K., Ozturk, I., Roest, K. & van Lier, J.B.
(2012). A review on dynamic membrane filtration: Materials,
applications and future perspectives, Bioresource Technology,122,
196-206.
Ersahin, M. E., Ozgun, H. & van Lier, J. B. (2013). Effect of support material
properties on dynamic membrane filtration performance, Separation
Science and Technology,48(15), 2263-2269.
Espana-Gamboa, E. I., Mijangos-Cortes, J. O., Hernandez-Zarate, G.,
Maldonado, J. A. D. & Alzate-Gaviria, L. M. (2012). Methane
production by treating vinasses from hydrous ethanol using a modified
UASB reactor, Biotechnology for Biofuels,5(82).

175
Fakhru’l-Razi, A. & Noor, M. (1999). Treatment of palm oil mill effluent (POME)
with the membrane anaerobic system (MAS), Water Science and
Technology,39 (10-11), 159-163.
Fardeau, M. L., Salinas, M. B., L'Haridon, S., Jeanthon, C., Verhe, F., Cayol, J.
L., …Ollivier, B. (2004). Isolation from oil reservoirs of novel
thermophilic anaerobes phylogenetically related to
Thermoanaerobacter subterraneus: Reassignment of T-subterraneus,
Thermoanaerobacter yonseiensis, Thermoanaerobacter tengcongensis
and Carboxydibrachium pacificum to Caldanaerobacter subterraneus
gen. nov., sp nov., comb. nov. as four novel subspecies, International
Journal of Systematic and Evolutionary Microbiology,54, 467-474.
Field, R. W., Wu, D., Howell, J. A. & Gupta, B. B. (1995). Critical flux concept
for microfiltration fouling, Journal of Membrane Science,100, 259-
272.
Fox, R. A. & Stuckey, D. C. (2015). The effect of sparging rate on transmembrane
pressure and critical flux in an AnMBR, Journal of Environmental
Management,151, 280-285.
Fuchs, W., Binder, H., Mavrias, G. & Braun, R. (2003). Anaerobic treatment of
wastewater with high organic content using a stirred tank reactor
coupled with a membrane filtration unit, Water Research,37 (4), 902-
908.
Gao, D. W., Fu, Y., Tao, Y., Wu, W. M., An, R. & Li, X. X. (2009). Current
research and development of controlling membrane fouling of MBR,
African Journal of Biotechnology,8 (13), 2993-2998.
Gao, D. W., Tao, Y., An, R., Fu, Y. A. & Ren, N. Q. (2011). Fate of organic carbon
in UAFB treating raw sewage: Impact of moderate to low
temperature, Bioresource Technology,102 (3), 2248-2254.
Gao, D. W., Zhang, T., Tang, C. Y. Y., Wu, W. M., Wong, C. Y., Lee, Y. H.,
…Criddle, C. S. (2010). Membrane fouling in an anaerobic
membrane bioreactor: Differences in relative abundance of bacterial
species in the membrane foulant layer and in suspension, Journal of
Membrane Science,364 (1-2), 331-338.
Gao, R. F., Yuan, X. F., Li, J. J., Wang, X. F., Cheng, X., Zhu, W. B. & Cui, Z.
J. (2012). Performance and spatial succession of a full-scale anaerobic
plant treating high-concentration cassava bioethanol wastewater,
Journal of Microbiology and Biotechnology,22 (8), 1148-1154.
Gimenez, J. B., Marti, N., Ferrer, J. & Seco, A. (2012). Methane recovery
efficiency in a submerged anaerobic membrane bioreactor (SAnMBR)
treating sulphate-rich urban wastewater: Evaluation of methane losses
with the effluent, Bioresource Technology,118, 67-72.
Gimenez, J. B., Robles, A., Carretero, L., Duran, F., Ruano, M. V., Gatti, M. N.,
…Seco, A. (2011). Experimental study of the anaerobic urban
wastewater treatment in a submerged hollow-fibre membrane
bioreactor at pilot scale, Bioresource Technology,102, 8799-8806.

176
Gonçalves, R. F., Charlier, A. C. & Sammut, F. (1994). Primary fermentation of
soluble and particulate organic matter for wastewater treatment, Water
Science and Technology,30 (6), 53-62.
Grundestam, J. & Hellstrom, D. (2007). Wastewater treatment with anaerobic
membrane bioreactor and reverse osmosis, Water Science and
Technology,56, 211-217.
Hagi, T., Tanaka, D., Iwamura, Y. & Hoshino, T. (2004). Diversity and seasonal
changes in lactic acid bacteria in the intestinal tract of cultured
freshwater fish, Aquaculture,234 (1-4), 335-346.
Halalsheh, M. M. (2002). Anaerobic pre-treatment of strong sewage. (Ph.D.
Thesis). Wageningen University, Wageningen, the Netherlands.
He, Y., Xu, P., Li, C. & Zhang, B. (2005). High-concentration food wastewater
treatment by an anaerobic membrane bioreactor, Water Research,39
(17), 4110-4118.
Herrera-Robledo, M., Cid-Leon, D. M., Morgan-Sagastume, J. M. & Noyola, A.
(2011). Biofouling in an anaerobic membrane bioreactor treating
municipal sewage, Separation and Purification Technology,81, 49-55.
Herrera-Robledo, M., Morgan-Sagastume, J. M. & Noyola, A. (2010). Biofouling
and pollutant removal during long-term operation of an anaerobic
membrane bioreactor treating municipal wastewater, Biofouling,26
(1), 23-30.
Ho, J. & Sung, S. (2010). Methanogenic activities in anaerobic membrane
bioreactors (AnMBR) treating synthetic municipal wastewater,
Bioresource Technology,101, 2191-2196.
Ho, J. H., Khanal, S. K. & Sung, S. (2007). Anaerobic membrane bioreactor for
treatment of synthetic municipal wastewater at ambient temperature,
Water Science and Technology,55, 79-86.
Holloway, R. W., Childress, A. E., Dennett, K. E. & Cath, T. Y. (2007). Forward
osmosis for concentration of anaerobic digester centrate, Water
Research,41, 4005-4014.
Hu, A. Y. & Stuckey D. C. (2006). Treatment of dilute wastewaters using a novel
submerged anaerobic membrane bioreactor, Journal of Environmental
Engineering,132, 190-198.
Hu, A. Y. & Stuckey, D. C. (2007). Activated carbon addition to a submerged
anaerobic membrane bioreactor: Effect on performance,
transmembrane pressure, and flux, Journal of Environmental
Engineering,133 (1), 73-80.
Huang, Z., Ong, S. L. & Ng, H. Y. (2008). The effect of HRT and SRT on
performance of submerged anaerobic membrane bioreactor (SAMBR)
for low-strength wastewater treatment. Proceedings of the IWA North
American Membrane Research Conference. Amherst, USA.
Huang, Z., Ong, S. L. & Ng, H. Y. (2011). Submerged anaerobic membrane
bioreactor for low-strength wastewater treatment: Effect of HRT and
SRT on treatment performance and membrane fouling, Water
Research,45 (2), 705-713.

177
Hulshoff Pol, L. W., de Castro Lopes, S. I., Lettinga, G. & Lens, P. N. L. (2004).
Anaerobic sludge granulation, Water Research,38, 1376-1389.
Ince, B., Koksel, G., Cetecioglu, Z., Oz, N. A., Coban, H. & Ince, O. (2011).
Inhibition effect of isopropanol on acetyl-CoA synthetase expression
level of acetoclastic methanogen, Methanosaeta concilii, Journal of
Biotechnology,156 (2), 95-99.
Janssen, A. (2011). The applicability of the SUR measurement for ultrafiltration of
WWTP effluent. (Ph.D. thesis). Delft University of Technology, Delft,
the Netherlands.
Janssen, A. N., van Agtmaal, J., van den Broek, W. B. P., de Koning, J.,
Menkveld, H. W. H., Schrotter, J. -C., …Graaf, J. H. J. M. (2008).
Monitoring of SUR to control and enhance the performance of dead-
end ultrafiltration installations treating WWTP effluent,
Desalination,231 (1-3), 99–107.
Jarusutthirak, C. & Amy, G. (2006). Role of soluble microbial products (SMP) in
membrane fouling and flux decline, Environmental Science and
Technology,40 (3), 969-974.
Jeison, D., Plugge, C. M., Pereira, A. & van Lier, J. B. (2009a). Effects of the
acidogenic biomass on the performance of an anaerobic membrane
bioreactor for wastewater treatment, Bioresource Technology,100,
1951-1956.
Jeison, D., Telkamp, P. & van Lier, J. B. (2009b). Thermophilic side stream
anaerobic membrane bioreactors: The shear rate dilemma, Water
Environment Research,81 (11), 2372-2380.
Jeison, D., van Betuw, W. & van Lier, J. B. (2008). Feasibility of anaerobic
membrane bioreactors for the treatment of wastewaters with
particulate organic matter, Separation Science and Technology,43,
3417-3431.
Jeison, D. & van Lier J. B. (2006a). Cake layer formation in anaerobic submerged
membrane bioreactors (AnSMBR) for wastewater treatment, Journal
of Membrane Science,284 (1-2), 227-236.
Jeison, D. & van Lier, J. B. (2006b). On-line cake-layer management by trans-
membrane pressure steady state assessment in anaerobic membrane
bioreactors for wastewater treatment, Biochemical Engineering
Journal,29, 204-209.
Jeison, D. & van Lier, J. B. (2007). Thermophilic treatment of acidified and
partially acidified wastewater using an anaerobic submerged MBR:
Factors affecting long-term operational flux, Water Research,41,
3868-3879.
Judd, S. (2008). The status of membrane bioreactor technology, Trends in
Biotechnology,26, 109-116.
Kalogo, Y. & Verstraete, W. (1999). Development of anaerobic sludge bed (ASB)
reactor technologies for domestic wastewater treatment: Motives and
perspectives, World Journal of Microbiology and Biotechnology,15,
523-534.

178
Kampmann, K., Ratering, S., Baumann, R., Schmidt, M., Zerr, W. & Schnell, S.
(2012). Hydrogenotrophic methanogens dominate in biogas reactors
fed with defined substrates, Systematic and Applied Microbiology,35
(6), 404-413.
Kang, I. J., Lee, C. H. & Kim, K. J. (2003). Characteristics of microfiltration
membranes in a membrane coupled sequencing batch reactor system,
Water Research,37 (5), 1192-1197.
Kang, I. J., Yoon, S. H. & Lee, C. H. (2002). Comparison of the filtration
characteristics of organic and inorganic membranes in a membrane-
coupled anaerobic bioreactor, Water Research,36, 1803-1813.
Kashyap, D. R., Dadhich, K. S. & Sharma, S. K. (2003). Biomethanation under
psychrophilic conditions: A review, Bioresource Technology,87 (2),
147-153.
Kataoka, N., Tokiwa, Y., Tanaka, Y., Fujiki, K., Taroda, H. & Takeda, K.
(1992). Examination of bacterial characteristics of anaerobic
membrane bioreactors in three pilot-scale plants for treating low-
strength wastewater by application of the colony-forming-curve
analysis method, Applied and Environmental Microbiology,58 (9),
2751-2757.
Kato, M. T., Field, J. A. & Lettinga, G. (1997). The anaerobic treatment of low
strength wastewaters in UASB and EGSB reactors, Water Science and
Technology,36 (6-7), 375-382.
Ke, X., Wang, C. Y., Li, R. D. & Zhang, Y. (2014). Effects of oxytetracycline on
methane production and the microbial communities during anaerobic
digestion of cow manure, Journal of Integrative Agriculture,13 (6),
1373-1381.
Khemkhao, M., Nuntakumjorn, B., Techkarnjanaruk, S. & Phalakornkule, C.
(2012). UASB performance and microbial adaptation during a
transition from mesophilic to thermophilic treatment of palm oil mill
effluent, Journal of Environmental Management,103, 74-82.
Kimura, S. (1991). Japan’s aqua renaissance ’90 project, Water Science and
Technology,23, 1573-1582.
Kiriyama, K., Tanaka, Y. & Mori, I. (1992). Field test of a composite methane gas
production system incorporating a membrane module for municipal
sewage, Water Science and Technology,25, 135-141.
Kiriyama, K., Tanaka, Y. & Mori, I. (1994). Field test on a methane fermentation
treatment system incorporating a membrane module for municipal
sewage, Desalination,98, 199-206.
Kleerebezem, R. & Macarie, H. (2003). Treating industrial wastewater: Anaerobic
digestion comes of age, Chemical Engineering,110 (4), 56–64.
Kobayashi, T., Yan, F., Takahashi, S. & Li, Y. Y. (2011). Effect of starch addition
on the biological conversion and microbial community in a methanol-
fed UASB reactor during long-term continuous operation, Bioresource
Technology,102 (17), 7713-7719.

179
Kocadagistan, E. & Topcu, N. (2007). Treatment investigation of the Erzurum City
municipal wastewaters with anaerobic membrane bioreactors,
Desalination,216, 367-376.
Kotsyurbenko, O. R., Glagolev, M. V., Nozhevnikova, A. N. & Conrad, R.
(2001). Competition between homoacetogenic bacteria and
methanogenic archaea for hydrogen at low temperature, FEMS
Microbiology Ecology,38, 153-159.
Kovacik, W. P., Scholten, J. J. C. M., Culley, D., Hickey, R., Zhang, W. &
Brockman, F. J. (2010). Microbial dynamics in upflow anaerobic
sludge blanket (UASB) bioreactor granules in response to short-term
changes in substrate feed, Microbiology,156 (8), 2418-2427.
Krzeminski, P., Iglesias-Obelleiro, A., Madebo, G., Garrido, J. M., van der
Graaf, J. H. J. M. & van Lier, J. B. (2012). Impact of temperature
on raw wastewater composition and activated sludge filterability in
full-scale MBR systems for municipal sewage treatment, Journal of
Membrane Science,423-424, 348-361.
Lawler, D. F., Chung, Y. J., Hwang, S. J. & Hull, B. A. (1986). Anaerobic
digestion: Effect on particle size and dewaterability, Journal (Water
Pollution Control Federation),58 (12), 1107-1117.
Le Clech, P., Jefferson, B., Chang, I. S. & Judd, S. J. (2003). Critical flux
determination by the flux-step method in a submerged membrane
bioreactor, Journal of Membrane Science,227 (1-2), 81–93.
Lee, S. M., Jung, J. Y. & Chung, Y. C. (2001). Novel method for enhancing
permeate flux of submerged membrane system in two phase anaerobic
reactor, Water Research,35 (2), 471-477.
Lens, P. N. L., de Beer, D., Cronenberg, C. C. H., Houwen, F. P., Ottengraf, S.
P. P. & Verstraete, W. H. (1993). Heterogeneous distribution of
microbial activity in methanogenic aggregates: pH and glucose
microprofiles, Applied and Environmental Microbiology,59 (11),
3803-3815.
Lettinga, G. & Hulshoff Pol, L. W. (1991). UASB process design for various types
of wastewater, Water Science and Technology,24, 87-107.
Lettinga, G., Rebac, S., Parshina, S., Nozhevnikova, A., van Lier J. B. & Stams,
A. J. M. (1999). High-rate anaerobic treatment of wastewater at low
temperatures, Applied and Environmental Microbiology,65 (4), 1696-
1702.
Lettinga, G., Rebac, S. & Zeeman G. (2001). Challenge of psychrophilic anaerobic
wastewater treatment, Trends in Biotechnology,19 (9), 363-370.
Lettinga, G., van Velsen, A. F. M., Hobma, S. W., de Zeeuw, W. & Klapwijk A.
(1980). Use of the upflow sludge blanket (USB) reactor concept for
biological wastewater treatment, especially for anaerobic treatment,
Biotechnology and Bioengineering,22 (4), 699-734.
Lew, B., Lustig, I., Beliavski, M., Tarre, S. & Green, M. (2011). An integrated
UASB-sludge digester system for raw domestic wastewater treatment
in temperate climates, Bioresource Technology,102, 4921-4924.

180
Lew, B., Tarre, S., Beliavski, M., Dosoretz, C. & Green, M. (2009). Anaerobic
membrane bioreactor (AnMBR) for domestic wastewater treatment,
Desalination,243, 251-257.
Li, A., Chu, Y., Wang, X., Ren, L., Yu, J., Liu, X., …Li, S. (2013). A
pyrosequencing-based metagenomic study of methane-producing
microbial community in solid-state biogas reactor, Biotechnology for
Biofuels,6 (3), 1-17.
Li, J., Ozgun, H., Ersahin, M. E., Spanjers, H. & van Lier, J. B. (2011a).
Comparative evaluation of the sludge characteristics along the height
of full-scale UASB and EGSB reactors treating paper-mill effluents.
Proceedings of the X Latin American Workshop and Symposium on
Anaerobic Digestion. Ouro Preto, Brazil.
Li, J., Wang, J., Luan, Z., Deng, Y. & Chen, L. (2011b). Evaluation of
performance and microbial community in a two-stage UASB reactor
pretreating acrylic fiber manufacturing wastewater, Bioresource
Technology,102 (10), 5709-5716.
Li, X. -K., Ma, K. -L., Meng, L. -W., Zhang, J. & Wang, K. (2014). Performance
and microbial community profiles in an anaerobic reactor treating
with simulated PTA wastewater: From mesophilic to thermophilic
temperature, Water Research,61, 57-66.
Liao, B., Kraemer, J. T. & Bagley, D. M. (2006). Anaerobic membrane
bioreactors: Applications and research directions, Critical Reviews in
Environmental Science and Technology,36 (6), 489-530.
Lin, H., Chen, J., Wang, F., Ding, L. & Hong, H. (2011). Feasibility evaluation of
submerged anaerobic membrane bioreactor for municipal secondary
wastewater treatment, Desalination,280 (1-3), 120-126.
Lin, H., Peng, W., Zhang, M., Chen, J., Hong, H. & Zhang, Y. (2013). A review
on anaerobic membrane bioreactors: Applications, membrane fouling
and future perspectives, Desalination,314, 169-188.
Lin, H. J., Xie, K., Mahendran, B., Bagley, D. M., Leung, K. T., …Liao, B. Q.
(2010). Factors affecting sludge cake formation in a submerged
anaerobic membrane bioreactor, Journal of Membrane Science,361,
126-134.
Liu, Y., Liu, H., Cui, L. & Zhang, K. (2012). The ratio of food-to-microorganism
(F/M) on membrane fouling of anaerobic membrane bioreactors
treating low-strength wastewater, Desalination,297 (3), 97-103.
Lu, L., Xing, D., Ren, N. & Logan, B. E. (2012). Syntrophic interactions drive the
hydrogen production from glucose at low temperature in microbial
electrolysis cells, Bioresource Technology,124, 68-76.
Ma, D., Gao, B., Hou, D., Wang, Y., Yue, Q & Li, Q. (2013). Evaluation of a
submerged membrane bioreactor (SMBR) coupled with chlorine
disinfection for municipal wastewater treatment and reuse,
Desalination,313, 134-139.
Ma, J., Nossa, C. W., Xiu, Z., Rixey, W. G. & Alvarez, P. J. (2013). Adaptive
microbial population shifts in response to a continuous ethanol blend

181
 release increases biodegradation potential, Environmental
Pollution,178, 419-425.
Madigan, M. T., Martinko, J. M., Stahl, D. A. & Clark, D. P. (2011). Brock
Biology of Microorganisms. 13th ed.. San Francisco: Benjamin
Cummings.
Mahadevaswamy, M., Murthy, B. M. S. & Girijamma, A. R. (2004). Performance
evaluation of up-flow anaerobic sludge blanket (UASB) reactor for
treatment of paper mill wastewater, Journal of Environmental
Sciences,16 (2), 194-198.
Mahmoud, N. (2002). Anaerobic pre-treatment of sewage under low temperature
(15 °C) conditions in an integrated UASB-digester system. (Ph.D.
Thesis). Wageningen University, Wageningen, the Netherlands.
Mahmoud, N. (2008). High strength sewage treatment in a UASB reactor and an
integrated UASB-digester system, Bioresource Technology,99, 7531-
7538.
Mahmoud, N., Zeeman, G., Gijzen, H. & Lettinga, G. (2003). Solids removal in
upflow anaerobic reactors, a review, Bioresource Technology,90 (1),
1-9.
Mahmoud, N., Zeeman, G., Gijzen, H. & Lettinga, G. (2004). Anaerobic sewage
treatment in a one-stage UASB reactor and a combined UASB-
Digester system, Water Research,38 (9), 2348-2358.
Mahmoud, N., Zeeman, G. & van Lier, J. B. (2008). Adapting UASB technology
for sewage treatment in Palestine and Jordan, Water Science and
Technology,57 (3), 361-366.
Maidak, B. L., Olsen, G. J., Larsen, N., Overbeek, R., McCaughey, M. J. &
Woese, C. R. (1997). The RDP (Ribosomal Database Project),
Nucleic Acids Research,25 (1), 109-110.
Marti, N., Bouzas, A., Seco, A. & Ferrer, J. (2008). Struvite precipitation
assessment in anaerobic digestion processes, Chemical Engineering
Journal,141, 67-74.
Martin Garcia, I., Mokosch, M., Soares, A., Pidou, M. & Jefferson, B. (2013).
Impact on reactor configuration on the performance of anaerobic
MBRs: Treatment of settled sewage in temperate climates, Water
Research,47, 4853-4860.
Martin, I., Pidou, M., Soares, S., Judd, S. & Jefferson, B. (2011). Modelling the
energy demands of aerobic and anaerobic membrane bioreactors for
wastewater treatment, Environmental Technology,32, 921-932.
Martin, M. A., De la Rubia, M. A., Martin, A., Borja, R., Montalvo, S. &
Sanchez, E. (2010). Kinetic evaluation of the psychrophylic anaerobic
digestion of synthetic domestic sewage using an upflow filter,
Bioresource Technology,101 (1), 131-137.
Martinez-Sosa, D., Helmreich, B. & Horn, H. (2012). Anaerobic submerged
membrane bioreactor (AnSMBR) treating low-strength wastewater
under psychrophilic temperature conditions, Process Biochemistry,47,
792-798.

182
Martinez-Sosa, D., Helmreich, B., Netter, T., Paris, S., Bischof, F. & Horn, H.
(2011a). Anaerobic submerged membrane bioreactor (AnSMBR) for
municipal wastewater treatment under mesophilic and psychrophilic
temperature conditions, Bioresource Technology,102 (22), 10377-
10385.
Martinez-Sosa, D., Helmreich, B., Netter, T., Paris, S., Bischof, F. & Horn, H.
(2011b). Pilot-scale anaerobic submerged membrane bioreactor
(AnSMBR) treating municipal wastewater: The fouling phenomenon
and long-term operation, Water Science and Technology,64, 1804-
1811.
Martin-Garcia, I., Monsalvo, V., Pidou, M., Le-Clech, P., Judd, S. J., McAdam,
E. J. & Jefferson, B. (2011). Impact of membrane configuration on
fouling in anaerobic membrane bioreactors, Journal of Membrane
Science,382 (1-2), 41-49.
Maspolim, Y., Zhou, Y., Guo, C., Xiao, K. & Ng, W. J. (2015). Comparison of
single-stage and two-phase anaerobic sludge digestion systems -
Performance and microbial community dynamics, Chemosphere,140,
54-62.
Massé, A., Spérandio, M. & Cabassud, C. (2006). Comparison of sludge
characteristics and performance of a submerged membrane bioreactor
and an activated sludge process at high solids retention time, Water
Research,40, 2405-2415.
McCarty, P. L., Bae, J. & Kim, J. (2011). Domestic wastewater treatment as a net
energy producer - can this be achieved?, Environmental Science and
Technology,45 (17), 7100-7106.
McGinnis, R. L. & Elimelech, M. (2008). Global challenges in energy and water
supply: The promise of engineered osmosis, Environmental Science
and Technology,42, 8625-8629.
McKeown, R. M., Scully, C., Enright, A., Chinalia, F. A., Lee, C., Mahony, T.,
…O'Flaherty, V. (2009). Psychrophilic methanogenic community
development during long-term cultivation of anaerobic granular
biofilms, The ISME Journal,3(11), 1231-1242.
Meabe, E., Deleris, S., Soroa, S. & Sancho, L. (2013). Performance of anaerobic
membrane bioreactor for sewage sludge treatment: Mesophilic and
thermophilic processes, Journal of Membrane Science,446, 26-33.
Meng, F., Chae, S. R., Drews, A., Kraume, M., Shin, H. S. & Yang, F. (2009).
Recent advances in membrane bioreactors (MBRs): Membrane
fouling and membrane material, Water Research,43 (6), 1489-1512.
Meng, F., Zhang, H., Yang, F., Li, Y., Xiao, J. & Zhang, X. (2006). Effect of
filamentous bacteria on membrane fouling in submerged membrane
bioreactor, Journal of Membrane Science,272 (1-2), 161-168.
Meng, F. G., Drews, A., Mehrez, R., Iversen, V., Ernst, M., Yang, F. L.,
…Kraume, M. (2009). Occurrence, source, and fate of dissolved
organic matter (DOM) in a pilot-scale membrane bioreactor,
Environmental Science and Technology,43 (23), 8821-8826.

183
Mesle, M., Dromart, G. & Oger, P. (2013). Microbial methanogenesis in
subsurface oil and coal, Research in Microbiology,164 (9), 959-972.
Meslier, V., Loux, V. & Renault, P. (2012). Genome sequence of Lactococcus
raffinolactis strain 4877, isolated from natural dairy starter culture,
Journal of Bacteriology,194 (22), 6364.
Mo, W. & Zhang, Q. (2012). Can municipal wastewater treatment systems be
carbon neutral?, Journal of Environmental Management,112, 360-367.
Montgomery, D., Peck, E. & Vining, G. (2001). Introduction to Linear Regression
Analysis. New York: Wiley.
Mori, K. & Harayama, S. (2011). Methanobacterium petrolearium sp. nov. and
Methanobacterium ferruginis sp. nov., mesophilic methanogens
isolated from salty environments, International Journal of Systematic
and Evolutionary Microbiology,61, 138-143.
Nagaoka, H., Ueda, S. & Miya, A. (1996). Influence of bacterial extracellular
polymers on the membrane separation activated sludge process, Water
Science and Technology,34 (9), 165-172.
Narnoli, S. K. & Mehrotra, I. (1997). Sludge blanket of UASB reactor:
Mathematical simulation, Water Research,31 (4), 715-726.
Nelson, M. C., Morrison, M. & Yu, Z. T. (2011). A meta-analysis of the microbial
diversity observed in anaerobic digesters, Bioresource
Technology,102 (4), 3730-3739.
Neyens, E., Baeyens, J., Dewil, R. & de Heyder, B. (2004). Advanced sludge
treatment affects extracellular polymeric substances to improve
activated sludge dewatering, Journal of Hazardous Materials,106 (2-
3), 83-92.
Nguyen, T. T., Ngo, H. H. & Guo, W. (2013). Pilot scale study on a new membrane
bioreactor hybrid system in municipal wastewater treatment,
Bioresource Technology,141, 8-12.
Norton-Brandao, D., Scherrenberg, S. M. & van Lier, J. B. (2013). Reclamation
of used urban waters for irrigation purposes-A review of treatment
technologies, Journal of Environmental Management,122, 85-98.
Noyola, A., Padilla-Rivera, A., Morgan-Sagastume, J. M., Guereca, L. P. &
Hernández-Padilla, F. (2012). Typology of municipal wastewater
treatment technologies in Latin America, Clean-Soil, Air, Water,40
(9), 926-932.
Nozhevnikova, A. N., Simankova, M. V., Parshina, S. N. & Kotsyurbenko, O. R.
(2001). Temperature characteristics of methanogenic archaea and
acetogenic bacteria isolated from cold environments, Water Science
and Technology,44 (8), 41-48.
Oh, S. E., Iyer, P., Bruns, M. A. & Logan, B. E. (2004). Biological hydrogen
production using a membrane bioreactor, Biotechnology and
Bioengineering,87 (1), 119-127.
Ozgun, H., Dereli, R. K., Ersahin, M. E., Kinaci, C., Spanjers, H. & van Lier, J.
B. (2013a). A review of anaerobic membrane bioreactors for

184
 municipal wastewater treatment: Integration options, limitations and
expectations, Separation and Purification Technology,118, 89-104.
Ozgun, H., Ersahin, M. E, Tao, Y., Spanjers, H. & van Lier, J. B. (2013b). Effect
of upflow velocity on the effluent membrane fouling potential in
membrane coupled upflow anaerobic sludge blanket reactors,
Bioresource Technology,147, 285-292.
Ozgun, H., Gimenez, J. B., Ersahin, M. E., Tao, Y., Spanjers, H. & van Lier, J.
B. (2015a). Impact of membrane addition for effluent extraction on
the performance and sludge characteristics of upflow anaerobic sludge
blanket reactors treating municipal wastewater, Journal of Membrane
Science,479, 95-104.
Ozgun, H., Tao, Y., Ersahin, M. E., Zhou, Z., Gimenez, J. B., Spanjers, H. &
van Lier, J. B. (2015b). Impact of temperature on feed-flow
characteristics and filtration performance of an upflow anaerobic
sludge blanket coupled ultrafiltration membrane treating municipal
wastewater, Water Research,83, 71-83.
Padmasiri, S. I., Zhang, J., Fitch, M., Norddahl, B., Morgenroth, E. & Raskin,
L. (2007). Methanogenic population dynamics and performance of an
anaerobic membrane bioreactor (AnMBR) treating swine manure
under high shear conditions, Water Research,41, 134-144.
Panichnumsin, P., Ahring, B., Nopharatana, A. & Chaiprasert, P. (2012).
Microbial community structure and performance of an anaerobic
reactor digesting cassava pulp and pig manure, Water Science and
Technology,66 (7), 1590-1600.
Pankratov, T. A., Ivanova, A. O., Dedysh, S. N. & Liesack, W. (2011). Bacterial
populations and environmental factors controlling cellulose
degradation in an acidic Sphagnum peat, Environmental
Microbiology,13 (7), 1800-1814.
Pastor, L., Mangin, D., Ferrer, J. & Seco, A. (2010). Struvite formation from the
supernatants of an anaerobic digestion pilot plant, Bioresource
Technology,101, 118-125.
Penning, H., Claus, P., Casper, P. & Conrad, R. (2006). Carbon isotope
fractionation during acetoclastic methanogenesis by Methanosaeta
concilii in culture and a lake sediment, Applied and Environmental
Microbiology,72 (8), 5648-5652.
Porcelli, N. & Judd, S. (2010). Chemical cleaning of potable water membranes: A
review, Separation and Purification Technology,71 (2), 137-143.
Prieto, A. L., Futselaar, H., Lens, P. N. L., Bair, R. & Yeh, D. H. (2013).
Development and start up of a gas-lift anaerobic membrane bioreactor
(GI-AnMBR) for conversion of sewage to energy, water and nutrients,
Journal of Membrane Science,441, 158-167.
Qui, G., Song, Y., Zeng, P., Duan, L. & Xiao, S. (2013). Characterization of
bacterial communities in hybrid upflow anaerobic sludge blanket
(UASB)-membrane bioreactor (MBR) process for berberine antibiotic
wastewater treatment, Bioresource Technology,142, 52-62.

185
Rademacher, A., Zakrzewski, M., Schluter, A., Schonberg, M., Szczepanowski,
R., Goesmann, A., …Klocke, M. (2012). Characterization of
microbial biofilms in a thermophilic biogas system by high-
throughput metagenome sequencing, FEMS Microbiology Ecology,79
(3), 785-799.
Rebac, S., van Lier, J. B., Lens, P. N. L., Stams, A. J. M., Dekkers, F., Swinkels,
K. T. M. & Lettinga, G. (1999). Psychrophilic anaerobic treatment of
low strength wastewaters, Water Science and Technology,39 (5), 203-
210.
Rebac, S., van Lier, J. B., Lens, P. N. L., van Capellen, J., Vermeulen, M.,
Stams, A. J. M., …Lettinga, G. (1998). Psychrophilic (6-15°C) high-
rate anaerobic treatment of malting wastewater in a two-module
EGSB system, Biotechnology Progress,14, 856-864.
Robles, A., Ruano, M. V., Ribes, J. & Ferrer, J. (2013). Performance of industrial
scale hollow-fibre membranes in a submerged anaerobic MBR (HF-
SAnMBR) system at mesophilic and psycrophilic conditions,
Separation and Purification Technology,104, 290-296.
Roest, K., Lutchmiah, K., Harmsen, D. J. H., Kappelhof, J. & Cornelissen, E. R.
(2010). Innovative water and energy recovery from sewage.
Proceedings of IWA Water & Energy. Amsterdam, the Netherlands.
Roorda. J. H. (2004). Filtration characteristics in dead-end ultrafiltration of
WWTP-effluent. (Ph.D. thesis). Delft University of Technology, Delft,
the Netherlands.
Ruiz, I., Soto, M., Veiga, M.C., Ligero, P., Vega, A. & Blazquez, R. (1998).
Performance of and biomass characterization in a UASB reactor
treating domestic wastewater at ambient temperature, Water SA,24,
215-222.
Saddoud, A., Ellouze, M., Dhouib, A. & Sayadi, S. (2006). A comparative study
on the anaerobic membrane bioreactor performance during the
treatment of domestic wastewaters of various origins, Environmental
Technology,27 (9), 991-999.
Saddoud, A., Ellouze, M., Dhouib, A. & Sayadi, S. (2007). Anaerobic membrane
bioreactor treatment of domestic wastewater in Tunisia,
Desalination,207, 205-215.
Salazar-Pelaez, M. L., Morgan-Sagastume, J. M. & Noyola, A. (2011a). Influence
of hydraulic retention time on UASB pot-treatment with UF
membranes, Water Science and Technology,64 (11), 2299-2305.
Salazar-Pelaez, M. L., Morgan-Sagastume, J. M. & Noyola, A. (2011b). Influence
of hydraulic retention time on fouling in a UASB coupled with an
external ultrafiltration membrane treating synthetic municipal
wastewater, Desalination,277 (1-3), 164-170.
Santos, A. & Judd, S. (2010). The commercial status of membrane bioreactors for
municipal wastewater, Separation and Science Technology,45, 850-
857.

186
Santos, A., Ma, W. & Judd, S. J. (2011). Membrane bioreactors: Two decades of
research and implementation, Desalination,273, 148-154.
Seghezzo L. (2004). Anaerobic treatment of domestic wastewater in subtropical
regions. (Ph.D. Thesis). Wageningen University, Wageningen, the
Netherlends.
Seghezzo, L., Gutierrez, M. A., Trupiano, A. P., Figueroa, M. E., Cuevas, C. M.,
Zeeman, G. & Lettinga, G. (2002). The effect of sludge discharges
and upflow velocity on the removal of suspended solids in a UASB
reactor treating settled sewage at moderate temperatures. Proceedings
of the VII Latin-American Work-Shop and Seminar on Anaerobic
Digestion. Merida, Mexico.
Seghezzo, L., Zeeman, G., van Lier, J. B., Hamelers, H. V. M. & Lettinga, G.
(1998). A review: The anaerobic treatment of sewage in UASB and
EGSB reactors, Bioresource Technology,65 (3), 175-190.
Shin, S. G., Lee, S., Lee, C., Hwang, K. & Hwang, S. (2010). Qualitative and
quantitative assessment of microbial community in batch anaerobic
digestion of secondary sludge, Bioresource Technology,101 (24),
9461-9470.
Shizas, I. & Bagley, D. M. (2004). Experimental determination of energy content of
unknown organics in municipal wastewater streams, Journal of
Energy Engineering,130 (2), 45-53.
Siggins, A., Enright, A. M. & O’Flaherty, V. (2011). Low-temperature (7 °C)
anaerobic treatment of a trichloroethylene-contaminated wastewater:
Microbial community development, Water Research,45 (13), 4035-
4046.
Singh, K. S. & Viraraghavan, T. (1998). Start-up and operation of UASB reactors
at 20 °C for municipal wastewater treatment, Journal of Fermentation
and Bioengineering,85 (6), 609-614.
Skouteris, G., Hermosilla, D., López, P., Negro, C. & Blanco, Á. (2012).
Anaerobic membrane bioreactors for wastewater treatment: A review,
Chemical Engineering Journal,198-199, 138-148.
Smith, A. L., Skerlos, S. J. & Raskin, L. (2013). Psychrophilic anaerobic
membrane bioreactor treatment of domestic wastewater, Water
Research,47 (4), 1655-1665.
Smith, A. L., Stadler, L. B., Love, N. G., Skerlos, S. J. & Raskin, L. (2012).
Perspectives on anaerobic membrane bioreactor treatment of domestic
wastewater: A critical review, Bioresource Technology,122, 149-159.
Sousa, D. Z., Salvador, A. F., Ramos, J., Guedes, A. P., Barbosa, S., Stams, A. J.
M., …Pereira, M. A. (2013). Activity and viability of methanogens
in anaerobic digestion of unsaturated and saturated long-chain fatty
acids, Applied and Environmental Microbiology,79 (14), 4239-4245.
Souza, C. L., Silva, S. Q., Aquino, S. F. & Chernicharo, C.A. (2006). Production
and characterization of scum and its role in odour control in UASB
reactors treating domestic wastewater, Water Science and
Technology,54 (9), 201-208.

187
Spagni, A., Casu, S., Crispino, N. A., Farina, R. & Mattioli, D. (2010).
Filterability in a submerged anaerobic membrane bioreactor,
Desalination,250 (2), 787-792.
Sponza, D. T. (2001). Anaerobic granule formation and tetrachloroethylene (TCE)
removal in an upflow anaerobic sludge blanket (UASB) reactor,
Enzyme and Microbial Technology,29, 417-427.
Stams, A. J. & Plugge, C. M. (2009). Electron transfer in syntrophic communities
of anaerobic bacteria and archaea, Nature Reviews Microbiology,7 (8),
568-577.
Stuckey, D. C. (2012). Recent developments in anaerobic membrane reactors,
Bioresource Technology,122, 137-148.
Subbotina, I. V., Chernyh, N. A., Sokolova, T. G., Kublanov, I. V., Bonch-
Osmolovskaya, E. A. & Lebedinsky, A. V. (2003). Oligonucleotide
probes for the detection of representatives of the genus
Thermoanaerobacter, Microbiology,72 (3), 331-339.
Suihko, M. L. & Skytta, E. (2009). Characterisation of aerobically grown non-
spore-forming bacteria from paper mill pulps containing recycled
fibres, Journal of Industrial Microbiology and Biotechnology,36 (1),
53-64.
Sutton, P. M., Rittman, B. E., Schraa, O. J., Banaszak, J. E. & Togna, A. P.
(2011). Wastewater as a resource: a unique approach to achieving
energy sustainability, Water Science and Technology,63 (9), 2004-
2009.
Tadkaew, N., Hai, F. I., McDonald, J. A., Khan, S. J. & Nghiem, L. D. (2011).
Removal of trace organics by MBR treatment: The role of molecular
properties, Water Research,45 (8), 2439-2451.
Talbot, G., Topp, E., Palin, M. F. & Masse, D. I. (2008). Evaluation of molecular
methods used for establishing the interactions and functions of
microorganisms in anaerobic bioreactors, Water Research,42, 513-
537.
Tang, F., Hu, H., Wu, Q., Tang, X., Sun, Y., Shi, X. & Huang, J. (2013). Effects
of chemical agent injections on genotoxicity of wastewater in a
microfiltration-reverse osmosis membrane process for wastewater
reuse, Journal of Hazardous Materials,260, 231-237.
Tao, Y., Gao, D. W., Fu, Y., Wu, W. M. & Ren, N. Q. (2012). Impact of reactor
configuration on anammox process start-up: MBR versus SBR,
Bioresource Technology,104, 73-80.
Tardieu, E., Grasmick, A., Geaugey, V. & Manem, J. (1998). Hydrodynamic
control of bioparticle deposition in a MBR applied to wastewater
treatment, Journal of Membrane Science,147 (1), 1-12.
Tchobanoglous, G. & Burton, F. L. (2003). Wastewater Engineering, Treatment
and Reuse. fourth ed.. Metcalf-Eddy. New York: McGraw-Hill.
Ueki, A., Abe, K., Ohtaki, Y., Kaku, N., Watanabe, K. & Ueki, K. (2011).
Bacteroides paurosaccharolyticus sp. nov., isolated from a

188
 methanogenic reactor treating waste from cattle farms, International
Journal of Systematic and Evolutionary Microbiology,61 (2), 448-453.
Uemura, S. & Harada, H. (2000). Treatment of sewage by a UASB reactor under
moderate to low tempareture conditions, Bioresource Technology,72,
275-282.
Vadlani, P. V. & Ramachandran, K. B. (2008). Evaluation of UASB reactor
performance during start-up operation using synthetic mixed-acid
waste, Bioresource Technology,99, 8231-8236.
van der Last, A. R. M. & Lettinga, G. (1992). Anaerobic treatment of domestic
sewage under moderate climatic (Dutch) conditions using upflow
reactors at increased superficial velocities, Water Science and
Technology,25 (7), 167-178.
van Dongen, U., Jetten, M. S. M. & van Loosdrecht, M. C. M. (2001). The
SHARON-Anammox process for treatment of ammonium rich
wastewater, Water Science and Technology,44, 153-160.
van Haandel, A. C. & Lettinga, G. (1994). Anaerobic Sewage Treatment: A
Practical Guide for Regions with a Hot Climate. United Kingdom:
John Wiley & Sons Ltd..
van Kempen, R., Mulder, J. W., Uijterlinde, C. A. & van Loosdrecht, M. C. M.
(2001). Overview: Full scale experience of the SHARON process for
treatment of rejection water of digested sludge dewatering, Water
Science and Technology,44 (1), 145-152.
van Lier, J. B. (2008). High-rate anaerobic wastewater treatment: diversifying from
end-of-the-pipe treatment to resource-oriented conversion techniques,
Water Science and Technology,57 (8), 1137-1148.
van Lier, J. B. & Jeison, D. (2010). Low permeate flux: intrinsic constraint of
anaerobic membrane bioreactors for wastewater treatment?
Proceedings of IWA Regional Conference on Membrane Technology
and Water Reuse. Istanbul, Turkey.
van Lier, J. B., Mahmoud, N. & Zeeman, G. (2008). Anaerobic wastewater
treatment. In M. Henze, M.C.M. van Loosdrecht, G.A. Ekama, D.
Brdjanovic (Eds.), Biological wastewater treatment, principles,
modelling and design (Chapter 16). London, UK : IWA Publishing.
van Lier, J. B., Rebac, S. & Lettinga, G. (1997). High-rate anaerobic wastewater
treatment under psychrophilic and thermophilic conditions, Water
Science and Technology,35 (10), 199-206.
van Lier, J. B., Vashi, A., van der Lubbe, J. & Heffernan, B. (2010). Anaerobic
Sewage Treatment using UASB Reactors: Engineering and
Operational Aspects. In H.H.P. Fang (Ed.), Environmental Anaerobic
Technology: Applications and New Developments (Chapter 4).
London, UK : World Scientific, Imperial College Press.
van Voorthuizen, E., Zwijnenburg, A., van der Meer, W. & Temmink, H.
(2008). Biological black water treatment combined with membrane
separation, Water Research,42 (16), 4334–4340.

189
Veiga, M. C., Jain, M. K., Wu, W., Hollingsworth, R. I. & Zeikus, J. G. (1997).
Composition and role of extracellular polymers in methanogenic
granules, Applied and Environmental Microbiology,63 (2), 403-407.
Vidal, D., Anza, I., Taggart, M. A., Perez-Ramirez, E., Crespo, E., Hofle, U. &
Mateo, R. (2013). Environmental factors influencing the prevalence
of Clostridium Botulinum type C/D in non-permanent Mediterranean
wetlands, Applied and Environmental Microbiology,79 (14), 4264-
4271.
Visvanathan, C. & Abeynayaka, A. (2012). Developments and future potentials of
anaerobic membrane bioreactors (AnMBRs), Membrane Water
Treatment,3, 1-23.
Visvanathan, C., Ben Aim, R. & Parameshwaran, K. (2000). Membrane
separation bioreactors for wastewater treatment, Critical Reviews in
Environmental Science and Technology,30 (1), 1-48.
Vyrides, I. & Stuckey, D. C. (2009). Saline sewage treatment using a submerged
anaerobic membrane reactor (SAMBR): Effects of activated carbon
addition and biogas-sparging time, Water Research,43 (4), 933-942.
Wang, W., Yang, Q., Zheng, S. & Wu, D. (2013). Anaerobic membrane bioreactor
(AnMBR) for bamboo industry wastewater treatment, Bioresource
Technology,149, 292-300.
Wang, Y. & Qian, P. Y. (2009). Conservative fragments in bacterial 16S rRNA
genes and primer design for 16S ribosomal DNA amplicons in
metagenomic studies, PLoS One,4 (10), e7401.
Wei, C., Huang, X., Aim, R. B., Yamamoto, K. & Amy, G. (2011). Critical flux
and chemical cleaning-in-place during the long-term operation of a
pilot-scale submerged membrane bioreactor for municipal wastewater
treatment, Water Research,45, 863-871.
Wei, C. H., Harb, M., Amy, G., Hong, P. Y. & Leiknes, T. (2014). Sustainable
organic loading rate and energy recovery potential of mesophilic
anaerobic membrane bioreactor for municipal wastewater treatment,
Bioresource Technology,166, 326-334.
Weissbrodt, D. G., Shani, N. & Holliger, C. (2014). Linking bacterial population
dynamics and nutrient removal in the granular sludge biofilm
ecosystem engineered for wastewater treatment, FEMS Microbiology
Ecology,88 (3), 579-595.
Wen, C., Huang, X. & Qian, Y. (1999). Domestic wastewater treatment using an
anaerobic bioreactor coupled with membrane filtration, Process
Biochemistry,35 (3-4), 335-340.
Westerholm, M., Leven, L. & Schnürer, A. (2012). Bioaugmentation of syntrophic
acetate-oxidizing culture in biogas reactors exposed to increasing
levels of ammonia, Applied and Environmental Microbiology,78 (21),
7619-7625.
Wilsey, B. & Stirling, G. (2007). Species richness and evenness respond in a
different manner to propagule density in developing prairie
microcosm communities, Plant Ecology,190, 259-273.

190
Wu, B., An, Y., Li, Y. & Wong, F.S. (2009). Effect of adsorption/coagulation on
membrane fouling in microfiltration process post-treating anaerobic
digestion effluent, Desalination,242 (1-3), 183-192.
Wunderlin, T., Junier, T., Roussel-Delif, L., Jeanneret, N. & Junier, P. (2014).
Endospore-enriched sequencing approach reveals unprecedented
diversity of Firmicutes in sediments, Environmental Microbiology
Reports,6 (6), 631-639.
Xiao, Y., Liu, X. D., Wang, D. X., Lin, Y. K., Han, Y. P. & Wang, X. L. (2013).
Feasibility of using an innovative PVDF MF membrane prior to RO
for reuse of a secondary municipal effluent, Desalination,311, 16-23.
Xie, K., Lin, H. J., Mahendran, B., Bagley, D. M., Leung, K. T., Liss, S. N. &
Liao, B. Q. (2010). Performance and fouling characteristics of a
submerged anaerobic membrane bioreactor for evaporator condensate
treatment, Environmental Technology,31 (5), 511-521.
Yamada, T., Imachi, H., Ohashi, A., Harada, H., Hanada, S., Kamagata, Y. &
Sekiguchi, Y. (2007). Bellilinea caldifistulae gen. nov., sp nov and
Longilinea arvoryzae gen. nov., sp nov., strictly anaerobic,
filamentous bacteria of the phylum Chloroflexi isolated from
methanogenic propionate-degrading consortia, International Journal
of Systematic and Evolutionary Microbiology,57, 2299-2306.
Yoo, R., Kim, J., McCarty, P. L. & Bae, J. (2012). Anaerobic treatment of
municipal wastewater with a staged anaerobic fluidized membrane
bioreactor (SAF-MBR) system, Bioresource Technology,120, 133-
139.
Yu, Z., Wen, X., Xu, M. & Huang, X. (2012). Characteristics of extracellular
polymeric substances and bacterial communities in an anaerobic
membrane bioreactor coupled with online ultrasound equipment,
Bioresource Technology,117, 333-340.
Yun, M. A., Yeon, K. M., Park, J. S., Lee, C. H., Chun, J. & Lim, D. J. (2006).
Characterization of biofilm structure and its effect on membrane
permeability in MBR for dye wastewater treatment, Water
Research,40 (1), 45-52.
Zanetti, F., De Luca, G. & Sacchetti, R. (2010). Performance of a full-scale
membrane bioreactor system in treating municipal wastewater for
reuse purposes, Bioresource Technology,101, 3768-3771.
Zeeman, G., Sanders, W. T. M., Wang, K. Y. & Lettinga, G. (1997). Anaerobic
treatment of complex wastewater and waste activated sludge-
Application of an upflow anaerobic solid removal (UASR) reactor for
the removal and pre-hydrolysis of suspended COD, Water Science and
Technology,35 (10), 121-128.
Zhang, D., Zhu, W., Tang, C., Suo, Y., Gao, L., Yuan, X., …Cui, Z. (2012).
Bioreactor performance and methanogenic population dynamics in a
low-temperature (5-18 °C) anaerobic fixed-bed reactor, Bioresource
Technology,104, 136-143.

191
Zhang, J., Padmasiri, S., Fitch, M., Norddahl, B., Raskin, L. & Morgenroth, E.
(2007). Influence of cleaning frequency and membrane history on
fouling in an anaerobic membrane bioreactor, Desalination,207 (1-3),
153-166.
Zhang, L., Hendrickx, T. L. G., Kampman, C., Temmink, H. & Zeeman, G.
(2013). Co-digestion to support low temperature anaerobic
pretreatment of municipal sewage in a UASB-digester, Bioresource
Technology,148, 560-566.
Zhang, L., Hendrickx, T. L. G., Kampman, C., Zeeman, G., Temmink, H., Li,
W., & Buisman, C. J. N. (2012). The effect of sludge recirculation
rate on a UASB-digester treating domestic sewage at 15 °C, Water
Science and Technology,66, 2597-2603.
Zhang, X., Wang, Z., Wu, Z., Lu, F., Tong, J. & Zang, L. (2010). Formation of
dynamic membrane in an anaerobic membrane bioreactor for
municipal wastewater treatment, Chemical Engineering Journal,165
(1), 175-183.
Zhang, X., Wang, Z., Wu, Z., Wei, T., Lu, F., Tong, J. & Mai, S. (2011).
Membrane fouling in an anaerobic dynamic membrane bioreactor
(AnDMBR) for municipal wastewater treatment: Characteristics of
membrane foulants and bulk sludge, Process Biochemistry,46 (8),
1538-1544.
Zhao, S., Zou, L., Tang, C. Y. & Mulcahy, D. (2012). Recent developments in
forward osmosis: Opportunities and challenges, Journal of Membrane
Science,396, 1-21.
Zheng, X., Su, Y., Li, X., Xiao, N., Wang, D. & Chen, Y. (2013). Pyrosequencing
reveals the key microorganisms involved in sludge alkaline
fermentation for efficient short-chain fatty acids production,
Environmental Science and Technology,47 (9), 4262-4268.
Zhou, A. F., He, Z. L., Qin, Y. J., Lu, Z. M., Deng, Y., Tu, Q. C., …Zhou, J. Z.
(2013). StressChip as a high-throughput tool for assessing microbial
community responses to environmental stresses, Environmental
Science and Technology,47 (17), 9841-9849.
Ziganshin, A. M., Liebetrau, J., Pröter, J. & Kleinsteuber, S. (2013). Microbial
community structure and dynamics during anaerobic digestion of
various agricultural waste materials, Applied Microbiology and
Biotechnology,97 (11), 5161-5174.

192
APPENDICES

APPENDIX A: Table A.1 and Table A.2

APPENDIX B: Table B.1 and Table B.2

APPENDIX C: Table C.1 and Table C.2

APPENDIX D: Table D.1 and Table D.2

193
APPENDIX A

Table A.1 : The dominant bacterial species detected by pyrosequencing in the

inoculum and sludge samples before and after membrane addition.
Relative Abundance (%)a
Before After
Seed
membrane membrane
Species Sludge
addition addition
Cytophaga sp 16.5 18.6 28.6
Bacteroides sp 13.1 8.2 1.8
Anaerophaga sp 6.0 2.6 3.1
Flavobacterium sp 0.0 4.1 10.1
Mucilaginibacter rigui 0.0 1.7 0.0
Desulfomicrobium baculatum 0.0 11.8 1.2
Syntrophus sp 2.6 0.6 0.6
Desulfovibrio carbinolicus 0.0 0.0 0.6
Geobacter sp 0.6 0.2 0.7
Aeromonas sharmana 0.0 4.5 4.0
Aeromonas sp 0.0 6.1 1.3
Aeromonas schubertii 0.0 3.0 0.1
Pseudoalteromonas sp 3.0 0.3 0.1
Acinetobacter sp 0.0 1.6 2.7
Dechloromonas sp 0.0 1.4 0.1
Ralstonia sp 0.3 0.0 2.4
Acidovorax sp 0.0 1.4 2.0
Hydrogenophaga sp 0.0 0.2 0.8
Imtechium assamiensis 0.0 0.2 0.8
Comamonas sp 0.0 0.6 0.8
Variovorax sp 0.0 0.0 0.7
Bellilinea sp 5.1 5.0 2.1
Levilinea saccharolytica 0.4 3.4 2.0
Chloroflexus sp 0.5 0.2 1.7
Clostridium sp 6.2 3.1 6.1
Thermoanaerobacter sp 8.2 0.3 0.1
Dethiobacter sp 8.5 0.1 0.2
Acetoanaerobium noterae 0.0 1.3 1.0
Acidaminococcus sp 2.3 0.1 0.0
Fusibacter sp 0.0 0.0 1.8
Syntrophomonas sp 0.4 0.3 0.6
Trichococcus sp 0.0 1.4 1.1
Bacillus sp 0.5 1.1 0.5
Carnobacterium maltaromaticum 0.0 1.1 0.4
Enterococcus sp 0.0 1.0 0.5
Leadbetterella sp 0.0 0.0 0.7
Acidobacterium sp 5.3 0.1 0.1
Synergistes sp 1.3 0.8 0.4
Spirochaeta sp 0.3 0.7 1.5
Leptospira sp 0.9 0.4 1.1
a
The relative abundance is defined as the number of sequences affiliated with that
taxon divided by the total number of sequences per sample (%).

194
Table A.2 : The all archaea detected by pyrosequencing in the inoculum and sludge
samples before and after membrane addition.
Relative Abundance (%)a
Before After
Seed
membrane membrane
Species Sludge
addition addition
Methanobacterium beijingense 2.1 0.0 2.7
Methanobacterium ferruginis 0.0 29.3 0.0
Methanobacterium oryzae 0.0 2.4 1.4
Methanobacterium petrolearium 14.6 24.4 15.1
Methanobacterium sp 0.0 24.4 16.4
Methanobrevibacter sp 43.8 4.9 8.2
Methanolinea mesophila 14.6 0.0 0.0
Methanomicrobium sp 2.1 0.0 0.0
Methanosaeta sp 14.6 14.6 56.2
Candidatus Nitrosocaldus
2.1 0.0 0.0
yellowstonii
Thermofilum sp 6.3 0.0 0.0
a
The relative abundance is defined as the number of sequences affiliated with that
taxon divided by the total number of sequences per sample (%).

195
APPENDIX B

Table B.1 : The dominant bacterial species detected by pyrosequencing in the

inoculum and sludge samples at different operational temperatures.
Relative Abundance (%)a
Seed
Species T= 25 °C T= 15 °C
Sludge
Cytophaga sp 16.5 28.6 7.1
Bacteroides sp 13.1 1.8 2.9
Desulfomicrobium baculatum 0.0 1.2 3.1
Bellilinea sp 5.1 2.1 1.2
Aeromonas sp 0.0 1.3 0.6
Clostridium sp 6.2 6.1 4.7
Anaerophaga sp 6.0 3.1 4.5
Aeromonas sharmana 0.0 4.0 0.2
Thermoanaerobacter sp 8.2 0.1 0.1
Dethiobacter sp 8.5 0.2 0.1
Flavobacterium sp 0.0 10.1 3.2
Levilinea saccharolytica 0.4 2.0 0.7
Syntrophus sp 2.6 0.6 0.6
Pseudoalteromonas sp 3.0 0.1 0.0
Acinetobacter sp 0.0 2.7 4.8
Trichococcus sp 0.0 1.1 4.1
Synergistes sp 1.3 0.4 0.5
Dechloromonas sp 0.0 0.1 2.0
Acetoanaerobium noterae 0.0 1.0 1.6
Carnobacterium maltaromaticum 0.0 0.4 1.1
Ralstonia sp 0.3 2.4 0.7
Fusibacter sp 0.0 1.8 0.6
Chloroflexus sp 0.5 1.7 0.4
Spirochaeta sp 0.3 1.5 1.1
Desulfovibrio carbinolicus 0.0 0.6 1.2
Lactococcus raffinolactis 0.0 0.0 12.8
Acidovorax sp 0.0 2.0 0.3
Acidovorax facilis 0.0 0.0 1.7
Leptospira sp 0.9 1.1 0.4
Comamonas sp 0.0 0.8 0.8
Nostocoida limicola actinobacteria 0.0 0.0 3.9
Ferrimonas marina 0.0 0.0 4.8
Clostridium botulinum 0.0 0.0 3.2
Cloacibacterium normanense 0.0 0.0 1.7
Acidobacterium sp 5.3 0.1 0.1
Syntrophorhabdus aromaticivorans 1.3 0.0 0.4
Acidaminococcus sp 2.3 0.0 0.0
Fervidobacterium sp 1.5 0.0 0.0
Eubacterium sp 1.6 0.0 0.0
Hydrogenophaga sp 0.0 0.8 0.6
Imtechium assamiensis 0.0 0.8 0.3
Leadbetterella sp 0.0 0.7 0.0
Variovorax sp 0.0 0.7 0.0

196
Table B.1 (continued) : The dominant bacterial species detected by pyrosequencing
in the inoculum and sludge samples at different operational temperatures.
Relative Abundance (%)a
Seed
Species T= 25 °C T= 15 °C
Sludge
Geobacter sp 0.6 0.7 0.2
Syntrophomonas sp 0.4 0.6 0.6
Enterococcus sp 0.0 0.5 0.0
Aeromonas schubertii 0.0 0.1 0.0
a
The relative abundance is defined as the number of sequences affiliated with that
taxon divided by the total number of sequences per sample (%).

197
Table B.2 : The all archaea detected by pyrosequencing in the inoculum and sludge
samples at different operational temperatures.
Relative Abundance (%)a
Seed
T= 25 °C T= 15 °C
Species Sludge
Methanobacterium beijingense 2.1 2.7 0.5
Methanobacterium ferruginis 0.0 0.0 1.0
Methanobacterium oryzae 0.0 1.4 0.0
Methanobacterium petrolearium 14.6 15.1 3.6
Methanobacterium sp 0.0 16.4 0.0
Methanobrevibacter sp 43.8 8.2 0.0
Methanobacterium congolense 0.0 0.0 5.1
Methanolinea mesophila 14.6 0.0 1.5
Methanomicrobium sp 2.1 0.0 0.5
Methanosaeta sp 14.6 56.2 83.7
Methanosaeta concilii 0.0 0.0 1.0
Candidatus Nitrosocaldus
2.1 0.0 0.0
yellowstonii
Thermofilum sp 6.3 0.0 0.0
Methanosarcina sp 0.0 0.0 1.0
Methanoculleus sp 0.0 0.0 0.5
Methanospirillum hungatei 0.0 0.0 1.5
a
The relative abundance is defined as the number of sequences affiliated with that
taxon divided by the total number of sequences per sample (%).

198
APPENDIX C

Table C.1 : The dominant bacterial species detected by pyrosequencing in the sludge
samples along the UASB reactor.
Relative Abundance (%)a

Species S1 S2 S3 S4
Clostridium botulinum 4.7 7.5 12.7 3.0
Cytophaga sp 20.0 13.1 4.2 2.0
Trichococcus sp 2.7 6.3 8.9 0.2
Nostocoida limicola
3.3 6.7 7.5 0.1
actinobacteria
Flavobacterium sp 4.7 4.6 5.7 0.8
Anaerophaga sp 2.8 2.1 0.5 6.5
Lactococcus raffinolactis 2.7 1.7 5.7 1.3
Bacteroides sp 2.0 2.2 1.4 5.1
Zoogloea oryzae 1.7 2.9 1.4 1.1
Desulfomicrobium baculatum 3.7 2.4 1.1 0.2
Acidovorax facilis 1.8 2.1 2.1 0.4
Carnobacterium maltaromaticum 1.7 1.5 4.3 0.6
Comamonas sp 1.2 1.3 1.3 0.3
Dechloromonas sp 1.7 1.9 1.1 2.0
Acetoanaerobium noterae 1.8 1.9 2.4 0.4
Bellilinea sp 2.4 2.0 1.7 0.0
Geobacter sp 0.8 1.0 0.2 0.2
Syntrophomonas sp 1.7 0.7 0.8 0.1
Novispirillum itersonii 0.0 0.2 0.1 14.3
Ferrimonas marina 0.0 0.0 0.0 28.0
Cloacibacterium normanense 0.7 0.7 0.7 6.4
Acinetobacter sp 1.0 0.7 1.1 1.5
Aeromonas sp 0.6 0.8 0.5 0.7
Spirochaeta sp 0.9 1.0 0.7 1.7
Desulfovibrio carbinolicus 1.1 0.4 1.4 0.2
Fusibacter sp 1.6 1.1 0.5 0.5
Ralstonia sp 1.7 1.0 1.1 0.1
Synergistes sp 0.4 0.7 0.2 0.1
Dehalobacter sp 0.6 0.6 0.2 0.0
Rhodocyclus sp 0.8 0.7 1.1 0.4
Clostridium sp 5.9 6.8 6.4 2.7
Obesumbacterium proteus 0.1 0.1 0.6 0.2
Levilinea saccharolytica 1.1 0.8 0.5 0.0
Clostridium sporogenes 0.6 0.7 0.8 0.1
Zoogloea caeni 0.8 0.7 0.4 0.2
Syntrophus 1.3 0.5 0.3 0.0
Fusibacter sp 1.6 1.1 0.5 0.5
Rahnella aquatilis 0.0 0.8 0.7 0.7
Leucobacter sp 0.2 0.7 0.6 0.4
Comamonas terrigena 0.8 1.1 0.4 0.0
a
The relative abundance is defined as the number of sequences affiliated with that
taxon divided by the total number of sequences per sample (%).

199
 Table C.2 : The all archaea detected by pyrosequencing in the sludge samples along
the UASB reactor.
Relative Abundance (%)a

Species S1 S2 S3 S4
Methanosaeta sp 94.5 84.7 48.0 31.2
Methanosaeta concilii 0.0 4.2 16.0 22.9
Methanomicrobium sp 0.0 0.0 0.0 22.9
Methanospirillum sp 0.0 0.0 0.0 22.9
Methanoculleus sp 0.0 2.8 12.0 0.0
Methanosarcina sp 0.0 4.2 8.0 0.0
Methanobacterium
0.0 0.0 12.0 0.0
petrolearium
Methanobacterium formicicum 0.0 0.0 4.0 0.0
Methanobacterium sp 1.8 1.4 0.0 0.0
Methanobacterium oryzae 0.0 1.4 0.0 0.0
Methanobacterium congolense 0.0 1.4 0.0 0.0
Methanobrevibacter sp 3.6 0.0 0.0 0.0
a
The relative abundance is defined as the number of sequences affiliated with that
taxon divided by the total number of sequences per sample (%).

200
APPENDIX D

Table D.1 : The dominant bacterial species detected by pyrosequencing in the sludge
samples from the UASB and digester during AnMBR, AnMBR-Digester (UASB
Sampling Point: S3), AnMBR-Digester (UASB Sampling Point: S2) operation.
Relative Abundance (%)a
AnMBR-Digester AnMBR-Digester
AnMBR UASB Sampling UASB Sampling
Species Point: S3 Point: S2
UASB UASB Digester UASB Digester
Flavobacterium sp 3.2 18.5 1.2 18.1 0.5
Cytophaga sp 7.1 2.4 12.3 1.9 11.6
Acinetobacter sp 4.8 18.9 1.4 11.4 1.3
Bellilinea sp 1.2 2.4 24.6 1.8 31.2
Lactococcus raffinolactis 12.8 2.0 0.0 6.0 0.8
Bacteroides sp 2.9 9.8 1.8 9.3 2.1
Anaerophaga sp 4.5 2.7 18.2 2.0 7.8
Cloacibacterium
1.7 1.6 0.0 15.8 0.2
normanense
Trichococcus sp 4.1 0.7 0.4 0.5 0.4
Nostocoida limicola
3.9 0.2 0.1 0.8 1.2
actinobacteria
Clostridium sp 4.7 3.2 3.9 3.8 3.5
Clostridium botulinum 3.2 11.8 5.5 4.0 4.7
Ferrimonas marina 4.8 0.0 0.0 0.0 0.2
Dechloromonas sp 2.0 2.8 3.9 0.8 0.5
Acetoanaerobium
1.6 1.5 0.3 1.9 1.2
noterae
Desulfomicrobium
3.1 0.3 0.4 0.2 0.4
baculatum
Carboxydibrachium sp 0.0 0.8 1.2 1.1 9.8
Dethiobacter sp 0.1 0.5 0.1 0.0 0.5
Thermoanaerobacter sp 0.1 0.1 0.4 0.0 0.6
Acidobacterium sp 0.1 0.0 0.0 0.0 0.0
a
The relative abundance is defined as the number of sequences affiliated with that
taxon divided by the total number of sequences per sample (%).

201
Table D.2 : The all archaea detected by pyrosequencing in the sludge samples from
the UASB and digester during AnMBR, AnMBR-Digester (UASB Sampling Point:
S3), AnMBR-Digester (UASB Sampling Point: S2) operation.
Relative Abundance (%)a
AnMBR-Digester AnMBR-Digester
AnMBR UASB Sampling UASB Sampling
Species Point: S3 Point: S2
UASB UASB Digester UASB Digester
Methanosaeta sp 83.7 64.3 83.3 86.9 85.8
Methanobrevibacter sp 0.0 0.0 0.0 0.0 5.3
Methanospirillum sp 0.0 28.6 5.6 0.0 1.2
Methanobacterium
3.6 0.0 5.6 0.0 0.0
petrolearium
Methanobacterium
5.1 7.1 0.0 0.0 0.0
congolense
Methanoculleus sp 0.5 0.0 0.0 5.6 0.0
Methanolinea mesophila 1.5 0.0 0.0 0.0 0.0
Methanospirillum
1.5 0.0 0.0 2.8 3.0
hungatei
Methanobacterium
0.0 0.0 0.0 3.3 0.0
formicicum
Methanosarcina sp 1.0 0.0 5.6 0.0 0.0
Methanosaeta concilii 1.0 0.0 0.0 0.0 3.0
Thermofilum sp 0.0 0.0 0.0 0.0 0.0
Methanobrevibacter
0.0 0.0 0.0 1.4 0.6
arboriphilus
Methanobacterium
1.0 0.0 0.0 0.0 0.0
ferruginis
Methanomicrobium sp 0.5 0.0 0.0 0.0 0.6
Methanobacterium
0.5 0.0 0.0 0.0 0.0
beijingense
a
The relative abundance is defined as the number of sequences affiliated with that
taxon divided by the total number of sequences per sample (%).

202
CURRICULUM VITAE

Name Surname : Hale ÖZGÜN

Date and Place of Birth : 14.06.1981 - İstanbul
E-Mail : [email protected]

EDUCATION:
 B.Sc. : 2004, İstanbul Technical University, Faculty of Civil
Engineering, Environmental Engineering Department
 B.Sc. : 2005, İstanbul Technical University, Faculty of
Management, Industrial Engineering Department (double major programme)
 M.Sc. : 2007, İstanbul Technical University, Environmental
Engineering Department, Environmental Sciences and Engineering Programme

PROFESSIONAL EXPERIENCE AND REWARDS:

 Hale Özgün graduated as an environmental engineer from the Environmental
Engineering Department of İstanbul Technical University Civil Engineering
Faculty in 2004 holding the first rank both in the department and faculty.
 In 2005, she graduated as an industrial engineer from the Industrial Engineering
Department of İstanbul Technical University. She received her MSc degree from
İstanbul Technical University, Environmental Sciences and Engineering
Program.
 She got the National Scholarship Award for MSc Students provided by The
Scientific and Technological Research Council of Turkey (TÜBİTAK).
 She has been working as a research assistant in Environmental Engineering
Department of Istanbul Technical University since 2005.
 She took several roles in research projects and assisted courses including
wastewater treatment, water treatment and environmental economics.

203
 Her main research areas are membrane processes, anaerobic biotechnology,
wastewater and water treatment technologies, industrial pollution control and
environmental economics.
 She was selected as a participant for EC funded Marie Curie Training Course
“EF2: Nanostructured materials and membrane modelling and simulation”
Actions in 2008.
 In November 2010, she also has joined the TU Delft, Sanitary Engineering
Section as a PhD researcher to conduct her PhD studies on anaerobic membrane
bioreactors under the supervision of Prof. Cumali Kınacı, Prof. Jules B. van Lier
and Assoc. Prof. Henri Spanjers.
 She got the PhD Fellowship Award provided by Turkish Academy of Sciences
(TUBA) in 2010.
 She received German Water Partnership-Award Turkey 2015 for young water
professionals, specialising in the issue of water and wastewater from her PhD
topic.

PUBLICATIONS, PRESENTATIONS AND PATENTS ON THE THESIS:

 Ozgun, H., Dereli, R. K., Ersahin, M. E., Kinaci, C., Spanjers, H., van Lier, J. B.,
2013. A Review of Anaerobic Membrane Bioreactors for Municipal Wastewater:
Integration Options, Limitations and Expectations, Separation and Purification
Technology, 118, 89-104.
 Ozgun, H., Ersahin, M. E., Tao, Y., Spanjers, H., van Lier, J. B., 2013. Effect of
Upflow Velocity on the Effluent Membrane Fouling Potential in Membrane
Coupled Upflow Anaerobic Sludge Blanket Reactors, Bioresource Technology,
147, 285-292.
 Ozgun, H., Gimenez, J.B., Ersahin, M.E., Tao, Y., Spanjers H., van Lier J.B.,
2015. Impact of Membrane Addition for Effluent Extraction on the Performance
and Sludge Characteristics of Upflow Anaerobic Sludge Blanket Reactors
Treating Municipal Wastewater. Journal of Membrane Science, 479, 95–104.
 Ozgun, H., Tao, Y., Ersahin, M.E., Zhou, Z., Gimenez, J.B., Spanjers, H., van
Lier, J.B., 2015. Impact of Temperature on Feed-Flow Characteristics and
Filtration Performance of an Upflow Anaerobic Sludge Blanket Coupled
Ultrafiltration Membrane Treating Municipal Wastewater. Water Research, 83,
71-83.
 Ozgun, H., Ersahin, M.E., Zhou, Z, Tao, Y., Spanjers, H., van Lier, J.B.
Comparative Evaluation of the Sludge Characteristics Along the Height of
Upflow Anaerobic Sludge Blanket Coupled Ultrafiltration Systems. under
review.
 Ozgun, H., Zhou, Z, Ersahin, M.E., Tao, Y., Spanjers, H., van Lier, J.B. An
Integrated Anaerobic Membrane Bioreactor-Digester System to Support Low
Temperature Anaerobic Treatment of Municipal Wastewater. under review.
 Ozgun H., Ersahin M.E., Tao Y., Spanjers H., van Lier J.B., 2013. Effect of
Upflow Velocity on the Ultrafiltration Resistance of UASB Effluents. 13th IWA
World Congress of Anaerobic Digestion, June 25-28, 2013 Santiago de
Compostela, Spain.

204
 Tao Y., Gao D.W., Wang H.Y., Zhang X.D., Ghasimi D., Ozgun H., Ersahin
M.E., Zhou Z.B., Liu G., Temudo M.F., Kloek J., Spanjers H., de Kreuk M., Ren
N.Q, van Lier J.B., 2007. Metagenomic Insights into the Bio-Functionality of 21
Anaerobic Biogas Reactors. IWA World Water Congress & Exhibition,
September 21-26, 2014 Lizbon, Portugal.

OTHER PUBLICATIONS, PRESENTATIONS AND PATENTS:

International Journal Papers

 Kayaalp, N., Ersahin, M. E. , Ozgun, H., Koyuncu, I., Kinaci, C., 2010. A New
Approach for Chemical Oxygen Demand (COD) Measurement at High Salinity
and Low Organic Matter Samples, Environmental Science and Pollution
Research, 17(9), 1547-1552.
 Dereli, R. K., Ersahin, M. E., Ozgun, H., Ozturk, I., Aydin, A. F., 2010.
Applicability of Anaerobic Digestion Model No. 1 (ADM1) for a Specific
Industrial Wastewater: Opium Alkaloid Effluents, Chemical Engineering
Journal, 165(1), 89-94.
 Ari, P. H., Ozgun, H., Ersahin, M. E., Koyuncu, I., 2011. Cost Analysis of Large
Scale Membrane Treatment Systems for Potable Water Treatment, Desalination
and Water Treatment, 26(1-3), 172-177.
 Ersahin, M. E., Dereli, R. K., Ozgun, H., Donmez, B. G., Koyuncu, I., Altinbas,
M., Ozturk, I., 2011. Source Based Characterization and Pollution Profile of a
Baker’s Yeast Industry, Clean- Soil, Air, Water, 39(6), 543-548.
 Kose, B., Ozgun, H., Ersahin, M. E., Dizge, N., Koseoglu-Imer, D. Y., Atay, B.,
Kaya, R., Altinbas, M., Sayili, S., Hoshan, P., Atay, D., Eren, E., Kinaci, C.,
Koyuncu, I., 2012. Performance Evaluation of a Submerged Membrane
Bioreactor for the Treatment of Brackish Oil and Natural Gas Field Produced
Water, Desalination, 285, 295-300.
 Ozgun, H., Karagul, N., Dereli, R. K., Ersahin, M. E., Coskuner T., Ciftci, D. I.,
Ozturk, I., Altinbas, M., 2012. Confectionery Industry: A Case Study on
Treatability-Based Effluent Characterization and Treatment System Performance,
Water Science and Technology, 66(1), 15-20.
 Ersahin, M. E., Ozgun, H., Dereli, R. K., Ozturk, I., Roest, K., van Lier, J. B.,
2012. A Review on Dynamic Membrane Filtration: Materials, Applications and
Future Perspectives, Bioresource Technology, 122, 196-206.
 Dereli, R. K., Ersahin, M. E., Ozgun, H., Ozturk, I., Jeison, D., van der Zee, F.,
van Lier, J. B., 2012. Potentials of Anaerobic Membrane Bioreactors to
Overcome Treatment Limitations Induced by Industrial Wastewaters,
Bioresource Technology, 122, 160-170.
 Ozgun, H., Ersahin, M. E., Erdem, S., Atay, B., Kose, B., Kaya, R., Altınbas,
M., Sayili, S., Hoshan, P., Atay, D., Eren, E., Kinaci, C., Koyuncu, I., 2013.
Effects of the Pre-Treatment Alternatives on the Treatment of Oil-Gas Field
Produced Water by Nanofiltration and Reverse Osmosis Membranes, Journal of
Chemical Technology and Biotechnology, 88(8), 1576-1583.

205
 Ersahin, M. E., Ozgun, H., van Lier, J. B., 2013. Effect of Support Material
Properties on Dynamic Membrane Filtration Performance, Separation Science
and Technology, 48(15), 2263-2269.
 Ozgun, H., Ersahin, M. E., Erdem, S., Atay, B., Sayılı, S., Eren, E., Atay, P.,
Hoshan, D., Altınbas, M., Kınacı, C., Koyuncu, I., 2013. Comparative Evaluation
for Characterization of Produced Water Generated from Oil, Gas and Oil-Gas
Production Fields, Clean-Soil, Air, Water, 41(12), 1175-1182.
 Ersahin, M. E., Ozgun, H., Tao, Y., van Lier, J. B., 2014. Applicability of
Dynamic Membrane Technology in Anaerobic Membrane Bioreactors, Water
Research, 48, 420-429.
 Yetilmezsoy, K., Ozgun H., Dereli, R.K., Ersahin, M.E., Ozturk, I., 2015.
Adaptive Neuro-Fuzzy Inference-Based Modeling of a Full-Scale Expanded
Granular Sludge Bed Reactor Treating Corn Processing Wastewater, Journal of
Intelligent & Fuzzy Systems, 28, 1601–1616.
 Ersahin, M.E., Tao, Y., Ozgun, H., Spanjers, H., van Lier, J.B. Characteristics
and role of dynamic membrane layer in anaerobic membrane bioreactors,
Biotechnology and Bioengineering, in press.
International Conference Papers
 Mantas E., Aydın E., Insel G., Ozgun H., Cakmakci M., Ozturk D., Akça L.,
Çokgor E., Ozturk I., 2007. Conceptual Approach for the Determination of Point
Sources in Melen Watershed. International Congress - River Basin Management,
March 22-24, 2007 Antalya, Turkey.
 Erturk A., Gurel M., Ekdal A., Tavsan C., Seker D. Z., Çokgor E., Insel G.,
Mantas E.P., Aydin E., Ozgun H., Cakmakci M., Tanik A., Ozturk I., 2007.
Estimating the Impact of Nutrient Emissions Via Water Quality Modelling in the
Melen Watershed. 11th International Conference on Diffuse Pollution, 1st Joint
Meeting of the IWA Diffuse Pollution and Urban Drainage Specialist Groups,
August, 26-31, 2007 Belo Horizonte, MG, Brazil.
 Ozgun H., Kayaalp N., Dogan N.B., Ersahin M.E., Cakmakci M., Unal A.,
Kinaci C., 2008. The Effects of Manganese Presence and Filter Media on Iron
Removal from Potable Water. 1st IWA Mexico National Young Water
Professional Conference, April 9-11, 2008 Mexico City, Mexico.
 Ozgun H., Kayaalp N., Dogan N.B., Ersahin M.E., Çakmakci M., Unal A.,
Kinaci C., 2008. Performance Evaluation of One-Layer and Double-Layer Filters
Used for Iron Removal in Drinking Water: A Pilot Scale Study. International
Symposium on Sanitary and Environmental Engineering, June 24-27, 2008
Florence, Italy.
 Ari P.H., Ozgun H., Ersahin M.E., Koyuncu I., 2009. Cost Analysis of Large
Scale Membrane Treatment Systems for Potable Water Treatment. International
Workshop on Urbanisation, Land Use, Land Degradation, and Environment
(ULE 2009), September 28-October 1, 2009 Denizli, Turkey.
 Ersahin M.E., Ozgun H., Kayaalp N., Atay B., Kiratli T., Erdem S., Koyuncu I.,
Kinaci C., 2009. Determination of Chemical Oxygen Demand in High-Salinity
Produced Water. First International Conference on “Advances in Wastewater
Treatment and Reuse”, November 10-12, 2009 Tehran, Iran.

206
 Dereli R.K., Ersahin M.E., Ozgun H., Koyuncu I., Ozturk I., Yildiz S., 2010.
Performance Evaluation of a Full Scale Membrane Bioreactor and Nanofiltration
System Treating Landfill Leachate. IWA Regional Conference and Exhibition on
Membrane Technology & Water Reuse (IWA MTWR 2010), October 18-22, 2010
Istanbul, Turkey.
 Erdem S., Ozgun H., Ersahin M.E., Atay B., Kose B., Eliduzgun S., Yilmaz F.,
Altinbas M., Sayili S., Hoshan P., Atay S.D., Eren E., Kinaci C., Koyuncu I.,
2010. Treatment of Produced Water Generated from Natural Gas Production
Fields by Microfiltration and Nanofiltration/Reverse Osmosis Membranes. IWA
Regional Conference and Exhibition on Membrane Technology & Water Reuse
(IWA MTWR 2010), October 18-22, 2010 Istanbul, Turkey.
 Atay B., Erdem S., Kose B., Ozgun H., Ersahin M.E., Kayaalp N., Altinbas M.,
Gencsoy E.B., Eliduzgun S., Yilmaz F., Sayili S., Hoshan P., Atay S.D., Eren E.,
Kinaci C., Koyuncu I., 2010. Effect of Pre-Ozonation on the Produced Water
Treatment Performance by Membrane Bioreactor Systems. IWA Regional
Conference and Exhibition on Membrane Technology and Water Reuse (IWA
MTWR 2010), October 18-22, 2010 Istanbul, Turkey.
 Arslan H., Ozgun H., Cakmakci M., Koyuncu I., 2010. Reuse of High and Low
Contaminated Wastewater Streams in Textile Industry by Membrane
Technologies. IWA Regional Conference and Exhibition on Membrane
Technology and Water Reuse (IWA MTWR 2010), October 18-22, 2010 Istanbul,
Turkey.
 Guney K., Arslan H., Eisele I., Ozgun H., Minke R., Koyuncu I., Steinmetz H.,
2010. Water Reuse Potential of Dye-Wash Process: In Turkey and In Germany.
IWA Regional Conference and Exhibition on Membrane Technology and Water
Reuse (IWA MTWR 2010), October 18-22, 2010, Istanbul, Turkey.
 Ozgun H., Karagul N., Dereli R.K., Ersahin M.E., Coskuner T., Ciftci D.İ.,
Altinbas M., Ozturk I., 2011. Confectionery Industry: A Case Study Of
Characterization and Treatment System Performance. 8th IWA International
Symposium on Waste Management Problems in Agro-Industries, June 22-24,
2011, Cesme, Izmir, Turkey.
 Altinbas M., Atay B., Erdem S., Kose B., Eliduzgun S., Yilmaz F., Ozgun H.,
Ersahin M.E., Sayili S., Hoshan P., Atay S.D., Eren E., Kinaci C., Koyuncu I.,
2011. Bacterial Communities Induced Biofilm Fouling in Aerobic Membrane
Bioreactor (MBR) Treating Produced Water from Oil Extraction Wells. 6th IWA
Specialist Conference on Membrane Technology for Water&Wastewater
Treatment, October 4-7, 2011 Aachen, Germany.
 Li J., Ozgun H., Ersahin M.E., Spanjers H., van Lier J.B., 2011. Comparative
Evaluation of the Sludge Characteristics Along the Height of Full-Scale UASB
and EGSB Reactors Treating Paper-Mill Effluents. X Latin American Workshop
and Symposium on Anaerobic Digestion, October 23-27, 2011 Ouro Preto, Brasil.
 Guney K., Arslan H., Ozgun H., Minke R., Koyuncu I, Steinmetz H., 2011.
Cleaning of Membranes Fouled by Bleaching-Washing Wastewater. 8th IWA
International Conference on Water Reclamation and Reuse, September 26-29,
2011 Barcelona, Spain.

207
 Guney K., Arslan H., Ozgun H., Minke R., Koyuncu I., Steinmetz H., 2011.
Investigating Pre-treatment Methods for Water Reclamation from Bleaching-
Washing Wastewater. 6th IWA Specialist Conference on Membrane Technology
for Water &Wastewater Treatment, October 4-7, 2011 Eurogress, Aachen,
Germany.
 Ersahin M.E., Gimenez J.B., Ozgun H., Tao Y., van Lier J.B., 2013. Anaerobic
Dynamic Membrane Bioreactors for High Strength Wastewater Treatment. 13th
IWA World Congress of Anaerobic Digestion, June 25-28, 2013 Santiago de
Compostela, Spain.
 Aydin A.F., Sezer B., Dereli R.K., Ozgun H., Ersahin M.E., 2014. Two Stage
Co-Digestion of Fruit and Vegetable Wastes with Waste Activated Sludge. 2nd
EurAsia Waste Management Symposium, April 28-30, 2014 Istanbul, Turkey.
 Ersahin M.E., Ozgun H., Tao Y., van Lier J.B., 2014. Application of Dynamic
Membrane Technology in Anaerobic Membrane Bioreactors. MBR for the Next
Generation VI Workshop, September 11-12, 2014 Antalya, Turkey.
Books, Book Chapters and Editorships
 Koyuncu, I., Yildiz, S., Altinbas, M., Ozgun, H., Ersahin, E., Dereli, K., Ozcan,
O., 2010. Proceedings Book of IWA Regional Conference and Exhibition on
Membrane Technology and Water Reuse (IWA MTWR 2010), October 18-22,
2010, Istanbul, Turkey.
 Ersahin, M.E., Ozgun, H., Dereli, R.K., Ozturk, I., 2011. “Anaerobic Treatment
of Industrial Effluents: An Overview of Applications”, Waste Water - Treatment
and Reutilization, ed. F.S.G. Einschlag, Chapter 1, InTech, India, 2011. ISBN
978-978-953-307-249-4.
 Ozturk, I., Kinaci, C., Genceli, E.A., Ozgun, H., Cuceloglu, G., Sabuncugil,
A.O., 2015. Environmental Economics and Determination of Environmental
Tariffs Based on Exact Costs, Union of Municipalities of Turkey, Ankara,
Turkey.
 Sarıkaya H. Z., Ozgun H., 2015. General Water Supply Design Considerations,
In: Water and Wastewater Engineering Design Principles and Practice, ed.
Toroz I (Edited Davis M. L.), Chapter 2, ISBN: 978-605-320-058-1, Ankara,
Turkey.

208