Khum Gurung A4

Khum Gurung
MEMBRANE BIOREACTOR FOR THE REMOVAL

OF EMERGING CONTAMINANTS FROM
MUNICIPAL WASTEWATER AND ITS
VIABILITY OF INTEGRATING ADVANCED
OXIDATION PROCESSES
ACTA UNIVERSITATIS LAPPEENRANTAENSIS 866

Khum Gurung
MEMBRANE BIOREACTOR FOR THE REMOVAL

OF EMERGING CONTAMINANTS FROM
MUNICIPAL WASTEWATER AND ITS
VIABILITY OF INTEGRATING ADVANCED
OXIDATION PROCESSES
Dissertation for the degree of Doctor of Science (Technology) to be presented
with due permission for public examination and criticism in the Auditorium
of Mikkeli University Consortium (MUC), Mikkeli, Finland on the 1st of October
2019, at noon.
Acta Universitatis
Lappeenrantaensis 866
Supervisors Professor Mika Sillanpää
LUT School of Engineering Science
Lappeenranta-Lahti University of Technology LUT
Finland
Assistant Professor Mohamed Chaker Ncibi

LUT School of Engineering Science
Finland
Reviewers Professor Anastasios I. Zouboulis

Department of Chemistry
Aristotle University of Thessaloniki (AUTh)
Greece
Professor How Yong Ng

Department of Civil and Environmental Engineering
National University of Singapore (NUS)
Singapore
Opponent Professor of Practice Anna Mikola

Department of Built Environment
School of Engineering
Aalto University
Finland
ISBN 978-952-335-410-4
ISBN 978-952-335-411-1 (PDF)
ISSN-L 1456-4491
ISSN 1456-4491

LUT University Press 2019
Abstract
Khum Gurung
Membrane bioreactor for the removal of emerging contaminants from
municipal wastewater and its viability of integrating advanced
oxidation processes
Lappeenranta 2019
86 pages
Acta Universitatis Lappeenrantaensis 866
Diss. Lappeenranta-Lahti University of Technology LUT
ISBN 978-952-335-410-4, ISBN 978-952-335-411-1 (PDF), ISSN-L 1456-4491, ISSN
1456-4491
Water scarcity is undoubtedly a global concern due to exacerbated population growth,

economic development, climate change, and rapid urbanization. This has necessitated the
need for establishing more sustainable and well-functioning natural ecosystems by
transforming current water industries based on smart technologies and management
approaches. Municipal wastewater reclamation is one of the promising approaches to
relieve growing pressure on global water resources. However, adequate and efficient
treatment is always essential to achieve high quality reclaimed water, ensuring no
potential risks to human health and aquatic environment due to the presence of emerging
contaminants (ECs), for which legislations are being stricter with time. Since
conventional wastewater treatment plants are not designed to remove ECs, the capabilities
of advanced wastewater treatment technologies, for enhancing the removal of ECs before
releasing the treated effluent into the environment or reclamation, such as membrane
bioreactor (MBR) and its integration with advanced oxidation processes, are emerging
topics of research worldwide.
This dissertation presents a comprehensive practical study on the applicability of the

MBR system for treating municipal wastewater in terms of its performance on the
removal of diverse ECs, such as pharmaceutically active compounds, steroid hormones,
and endocrine disrupting compounds, under different operating conditions, including
Nordic cold environment and varying solid retention times. Additionally, a potential
viability of integrating MBR with two emerging advanced oxidation processes (AOP)
technologies, such as electrochemical oxidation (ECO) and photocatalytic oxidation
(PCO), were studied in batch modes to further enhancing the removal of ECs and to
produce high quality reclaimed water.
A significant membrane fouling, accompanying with about 75% permeability drop, was
observed when process temperature in MBR was below 10 °C, indicating high
deterioration of membrane performance. However, relatively consistent removal
efficiency of human enteric viruses, such as norovirus GI, norovirus GII, and adenovirus,
heavy metals, and selected ECs, was achieved Similarly, the removal and fate of 23
diverse ECs was studied at varying solid retention times (60 and 21 days). It was observed
that at long solid retention time, ECs removal majorly enhanced, while biopolymers
concentrations decreased. Moreover, diverse removal efficiencies of the selected ECs
were observed, which were explained based on their physicochemical properties and
other process operating parameters. The electrochemical oxidation of carbamazepine,
which is one of the highly recalcitrant EC, was studied in batch mode using Ti/Ta2O5-
SnO2 anode. The main operating parameters effecting the removal of carbamazepine,
including current density, initial substrate concentration, pH, and temperature, were
studied. A complete removal (>99.9 %) of carbamazepine, with concentration close to
environmental level (10 µg L-1), in real MBR effluent was observed when electrolyzed
under the optimized conditions (current density: 9 mA cm-2, pH: 6, and temperature: 30
°C) in synthetic solutions. Subsequently, the removal of ECs (carbamazepine and
diclofenac) was studied by heterogeneous photochemical oxidation using 5% Ag2O/P-25
photocatalyst under UV irradiation. The matrix effect, i.e., ECs in deionized water and
MBR effluent, was studied along with the catalyst dosage and initial ECs concentration.
The optimal removal of carbamazepine and diclofenac reached 89.1% and 93.5%,
respectively at catalyst dosage of 0.4 g L-1 in deionized water matrix, and further revealed
that optimal catalyst dosage for ECs removal in MBR effluent matrix increased by 1.5 to
2 fold to achieve the similar removal efficiency as deionized water matrix. High
reusability of 5% Ag2O/P-25 photocatalyst was observed for both ECs. Moreover, various
intermediates of carbamazepine generated during photochemical oxidation were analyzed
and identified, and a possible degradation pathway was proposed.
The current study has expanded the knowledge on the applicability and efficiency of
MBR operation in unique Nordic environment and integration possibilities of MBR with
selected innovative advanced oxidation processes. It has also revealed the need of many
other important related topics for further investigations.
Keywords: Membrane bioreactor, municipal wastewater treatment, emerging

contaminants, advanced oxidation processes, electrochemical oxidation, photochemical
oxidation
Acknowledgements
I have always been fascinated by science and technology. As the life goes on, my interests
and enthusiasm towards science have eventually led me to pursue PhD degree. Achieving
PhD degree has always been my academic dream, although it was not a piece of cake. I
should admit that it was full of challenges, hard work, and endurances. Now, I have
accomplished nearly four years of my PhD study and I have a great opportunity to finally
acknowledge many people who made this journey possible. In this occasion, I am feeling
overwhelmed and would like to convey my sincere gratitude to all of them.
First of all, I would like to express my deepest gratitude to my supervisor, Prof. Mika
Sillanpää for believing on me and providing me this opportunity to start and accomplish
my PhD study under his commendable guidance. I am also grateful to Assistant Prof.
Mohamed Chaker Ncibi for his valuable supports, suggestions and guidance throughout
my PhD journey.
I would like to thank Department of Green Chemistry (DGC), where most of my research
works for the doctoral dissertation was carried out, for facilitating incredible
infrastructures and resources.
I am thankful to Prof. Anastasios I. Zouboulis and Prof. How Yong Ng for reviewing this
dissertation and providing their valuable comments and suggestions that have enriched
the content of the dissertation further.
Many thanks to Professor of Practice Anna Mikola for agreeing to act as the opponent for
the public examination.
I would like to thank Business Finland Oy, which is the most prestigious public funding
agency for promoting high quality research in Finland, for financially supporting this
research work. The research project was granted with the title ‘‘Smart Effluents Project-
New generation wastewater treatment solutions for requirements of 2050’’ and project
decision number of 1043/31/2015. The project was started on October 2015 and ended
by September 2018.
My sincere thanks go to Kenkäveronniemi wastewater treatment plant, Mikkeli, Finland,
owned by Mikkeli water works (Mikkeli Vesilaitos), for being an active company partner
to this project, allow us to install and operate our pilot MBR plant within their premise,
and for providing technical supports during the project. I am grateful to Reijo Turkki
(Director of Mikkeli Vesilaitos), Anne Bergman, Risto Repo, Jani Koski, Päivi and all
other staffs of Kenkäveronniemi wastewater treatment plant, Mikkeli, for their
tremendous support on operating the MBR pilot plant and some laboratory works.
I would like to thank South-Eastern Finland University of Applied Sciences (XAMK) for
its incredible participation to Smart Effluents project. I would like to thank Dr. Heikki
Särkkä and Dr. Hanne Soininen for their continuous support on executing the Smart
Effluents Project steering group meetings. Moreover, especial thanks to all the Smart
Effluents Project consortium group members, Metsäsairila Oy, Suomen Ekolannoite Oy,
BioGTS Oy, and Mipro Oy, for their valuable comments and suggestions throughout the
project.
I would like to thank Dr. Marina Shestakova for helping me on establishing the
electrochemical study setup and guided me during the experimental period. I am thankful
to Dr. Bo Gao for her valuable suggestions on photocatalysis. I like to thank Dr. Deepika
Lakshmi Ramasamy for helping me on some analytical procedures.
I would like to thank all my senior and junior colleagues in the Department of Green
Chemistry for establishing a pleasant working environment with healthy discussions,
mutual support, and good scientific practices, especially Dr. Bhairavi, Dr. Varsha, Dr.
Sidra, Asst. Prof. Yuri, Mirka, Tam, Zhao, and Mahsa. All of you are amazing people.
Also, I like to thank Sanna Tomperi for helping me in many ways during the entire
journey.
Last but not the least, I am indebted to my wife and daughter, Jen and Nohyo, and all my
family members and relatives for being my strength, support and inspiration throughout
this PhD journey and sharing all the ups and downs in my life.
Khum Gurung
September 2019
Mikkeli, Finland
This doctoral dissertation is dedicated in memory of my late mother and
father.
Table of Contents
Abstract
Acknowledgements
Table of contents
List of publications ............................................................................................... 11
List of acronyms and symbols ............................................................................. 13
List of figures ........................................................................................................ 15
List of tables .......................................................................................................... 17
1 Introduction ....................................................................................................... 19
1.1 Problem statement and research motivations ............................................... 19
1.2 Research objectives ...................................................................................... 20
2 Municipal wastewater ....................................................................................... 21
2.1 Occurrence of emerging contaminants in municipal wastewater ................. 23
2.2 Environmental intimidations of ECs ............................................................ 26
2.3 Municipal wastewater treatment ................................................................... 26
2.4 EU legislation and removal of ECs from municipal wastewater.................. 27
3 State-of-the-art of MBRs: advanced municipal wastewater treatment........ 29
3.1 Introduction to MBRs ................................................................................... 29
3.2 Membrane materials, configuration and operating principles of MBRs ...... 30
3.3 Membrane fouling: a major shortcoming in MBRs...................................... 32
3.4 Fate of ECs in MBRs: application of mass balance ..................................... 34
3.5 Fate of human enteric viruses and heavy metals in MBRs........................... 37
4 Advanced oxidation processes for wastewater treatment .............................. 39
4.1 Electrochemical oxidation (ECO) process ................................................... 39
4.1.1 Active anodes ........................................................................................ 41
4.1.2 Non-active anodes ................................................................................. 42
4.2 Photocatalytic oxidation (PCO) .................................................................... 43
5 Integration potentials of MBR with AOPs ...................................................... 47
6 Materials and methods ...................................................................................... 51
6.2 Assessing the performance of MBR unit in different operating conditions . 51
6.1.1 The Pilot-scale MBR set-up .................................................................. 51
6.2 Assessing integration approaches to MBR ................................................... 52
6.2.1 Lab-scale set-up for batch ECO unit ..................................................... 52
6.2.2 Lab-scale set-up for PCO process ......................................................... 53
6.2.3 Analytical procedures ............................................................................ 54
7 Results and discussion ....................................................................................... 57
7.1 Assessing the performance of MBR at Nordic cold conditions (Paper I) .... 57
7.2 Assessing the performance of MBR under varying solid retention times- fate and
removal of ECs (Paper II) ................................................................................... 61
7.3 Assessing the AOPs as integration alternatives to MBR to enhance the removal of
ECs. .................................................................................................................... 63
7.3.1 Study of ECO process for carbamazepine removal using Ti/Ta2O5-SnO2
electrode (Paper III) ........................................................................................ 64
7.3.2 Study of PCO process for PhACs removal using Ag2O/P-25 photocatalyst
(Paper IV) ....................................................................................................... 66
8 Conclusions and further research .................................................................... 73
References.............................................................................................................. 75
11
List of publications
I. Gurung, K., Ncibi, M.C., Sillanpää, M., 2017. Assessing membrane fouling
and the performance of pilot-scale membrane bioreactor (MBR) to treat real
municipal wastewater during winter season in Nordic regions. Science of the
Total Environment. 579, 1289–1297.
II. Gurung, K., Ncibi, M.C., Sillanpää, M., 2019. Removal and fate of emerging
organic micropollutants (EOMs) in municipal wastewater by a pilot-scale
membrane bioreactor (MBR) treatment under varying solid retention times.
Science of the Total Environment. 667, 671–680.
III. Gurung, K., Ncibi, M.C., Shestakova, M., Sillanpää, M., 2018. Removal of
carbamazepine from MBR effluent by electrochemical oxidation (EO) using
a Ti/Ta2O5-SnO2 electrode. Applied Catalysis B Environment. 221, 329–338.
IV. Gurung, K., Ncibi, M.C., Thangaraj, S.K., Jänis, J., Seyedsalehi, M.,
Sillanpää, M., 2019. Removal of pharmaceutically active compounds
(PhACs) from real membrane bioreactor (MBR) effluents by photocatalytic
degradation using composite Ag2O/P-25 photocatalyst. Separation and
Purification Technology. 215, 317–328.
Author’s contribution
Khum Gurung is the principal investigator, and corresponding author for all the
publications I-IV.
12 List of publications
Related publication
Gurung, K., Ncibi, M.C., Fontmorin, J.M., 2016. Incorporating Submerged MBR in
Conventional Activated Sludge Process for Municipal Wastewater Treatment: A
Feasibility and Performance Assessment. Journal of Membrane Science and
Technology 6, 1-10.
Other publications
Gurung, K., Tang, W.Z., Sillanpää, M., 2018. Unit Energy Consumption as Benchmark
to Select Energy Positive Retrofitting Strategies for Finnish Wastewater Treatment Plants
(WWTPs): a Case Study of Mikkeli WWTP. Environmental Processes. 5, 667–681.
Habib, R., Asif, M.B., Iftekhar, S., Khan, Z., Gurung, K., Srivastava, V., Sillanpää, M.,
2017. Influence of relaxation modes on membrane fouling in submerged membrane
bioreactor for domestic wastewater treatment. Chemosphere 181, 19–25.
Seyedsalehi, M., Paladino, O., Hodaifa, G., Sillanpää, M., Gurung, K., Sahafnia, M.,
Barzanouni, H., 2018. Performance evaluation of several sequencing batch biofilm
reactors with movable bed in treatment of linear alkyl benzene sulfonate in urban
wastewater. International Journal of Environmental Science and Technology.
13
List of acronyms and symbols
Acronyms
AdVs Adenovirus
AFM Atomic force microscope
AOPs Advanced oxidation processes
BDD Boron-doped diamond
BET Brunauer-Emmett-Teller
BOD5 Biological oxygen demand of 5-days
CAS Conventional activated sludge
COD Chemical oxygen demand
CST Capillary suction time
CV Cyclic voltammetry
DOC Dissolved organic carbon
DSA Dimensionally stable anodes
DW Deionized water
ECs Emerging contaminants
EDCs Endocrine disrupting compounds
EDGs Electron donating groups
EDTA Ethylenediaminetetraacetic acid
EDX Energy-dispersive X-ray spectroscope
EPS Extracellular polymeric substance
EWGs Electron withdrawing groups
FOG Fat-oil-grease
FS Flat sheet
FT-ICR Fourier-transform ion cyclotron resonance mass spectrometry
FTIR Fourier-transform infrared spectroscopy
HER Hydrogen evolution reaction
HF Hollow fibre
LOD Limit of detection
LOQ Limit of quantification
MBR Membrane bioreactor
MBBR Moving-bed biofilm reactor
MLSS Mixed liquor suspended solid
MLVSS Mixed liquor volatile suspended solid
MMO Mixed metal oxide
MWCO Molecular weight cut-off
NoV GI Norovirus genome group I
NoV GII Norovirus genome group II
OERs Oxygen-evolution reactions
PCPs Personal care products
14 List of acronyms and symbols
PE Polyethylene
PES Polyethylsulphone
PhACs Pharmaceutically active compounds
PP Polypropylene
PVDF Polyvinylidene difluoride
P-25 Aeroxide® Titania nanoparticles
RME Real MBR effluent
SADm Specific aeration demand per membrane surface area
SCE Saturated Calomel electrode
SEM Scanning electron microscope
SHE Standard Hydrogen electrode
SMP Soluble microbial product
SRT Solid retention time
SVI Sludge volume index
TMP Trans-membrane pressure
TP Total phosphorus
TSS Total suspended solid
WWTP Wastewater treatment plant
Symbols
Ti Titanium
ℎ Planck constant (J.s)
𝜗 Frequency (Hz)
m/z Mass-to-charge ratio
Rt Total membrane filtration resistance (m-1)
𝜇 Dynamic viscosity of water (Pa.s)
𝐽 Membrane flux (Lm-2h-1)
log D Distribution coefficient
log Kd Sludge-water distribution coefficient
E° Standard redox potential (eV)
Rads Adsorbed organic pollutants at anode surface
Pads Oxidized adsorbed organic pollutants at anode surface
∆E Band gap energy (eV)
ecb- Excited electrons to conductance band
hvb+ Photogenerated holes at valence band
15
List of figures
Figure 1: MBR process configurations, (a) Submerged and (b) side-stream ......... 31
Figure 2: Different MBR membrane modules: (a) Tubular, (b) Hollow-fiber, (c) flat-
sheet (Reprinted with permission from (Bérubé, 2010)). ....................................... 31
Figure 3: Temporal and spatial development of membrane fouling in MBRs (Modified
from (Meng et al., 2017)). ...................................................................................... 34
Figure 4: Electrochemical oxidation scheme of organic pollutants on the metal oxide
anode surface (Modified from (Comninellis, 1994)). ............................................ 41
Figure 5: Decision tree for potential integration approaches of MBR and AOP
(Modified from (López et al., 2010)). .................................................................... 48
Figure 6: Schematic diagram of the pilot MBR unit at Kenkäveronniemi WWTP,
Mikkeli, Finland. .................................................................................................... 51
Figure 7: Experimental set-up for electrochemical oxidation. ............................... 53
Figure 8: Experimental set-up for photocatalytic oxidation ................................... 54
Figure 9: Membrane fouling behavior of MBR system under cold weather conditions
and adopted cleaning measures. ............................................................................. 58
Figure 10: Occurrences of viruses in wastewater stream under cold water conditions.
................................................................................................................................ 59
Figure 11: Removal efficiencies of selected ECs in MBR treatment under low-
temperature periods. ............................................................................................... 60
Figure 12: Fate of the selected ECs during MBR treatment at two different SRTs.63
Figure 13: Effect of operating conditions (a) applied current densities; (b) initial
carbamazepine concentrations; (c) initial solution pH; and (d) temperature on the
electrochemical oxidation of carbamazepine. ........................................................ 66
Figure 14: FTIR spectrum (a) and plots of (F (R∞) h𝑣)1/2 vs (h𝑣) for the approximation
of the optical band gap (b) of 5% Ag2O/P-25. ....................................................... 67
Figure 15: Effect of catalyst dosage in different water matrices on photocatalytic
removal of (a & d) carbamazepine in DW and RME; (b & e) diclofenac in DW and
RME; and effect of initial PhACs concentration (c) carbamazepine in RME; (f)
diclofenac in RME. ................................................................................................. 69
Figure 16. Cyclic reusability of 5% Ag2O/P-25 photocatalyst during the photocatalytic
oxidation of (a) carbamazepine and (b) diclofenac. ............................................... 70
Figure 17: Photogenerated intermediates and proposed photodegradation pathway of
carbamazepine by 5% Ag2O/P-25 photocatalyst under UV irradiation. ................ 72
16 List of figures
17
List of tables
Table 1. Characteristics of Municipal wastewater and their sources (Metcalf & Eddy
Inc., 2003; U.S. EPA, 1997) ................................................................................... 22
Table 2. Physicochemical properties of commonly identified ECs in municipal
wastewater. ............................................................................................................. 24
Table 3: Different operating conditions of MBR pilot plant depending on purpose of
studies. .................................................................................................................... 52
18 List of tables
19
1 Introduction
1.1 Problem statement and research motivations
The global water demand has been increasing constantly due to exacerbated population
growth, economic development, climatic change, intensive agricultural practices,
changing consumption patterns and urbanization. Undoubtedly, combating water scarcity
is a global issue (WHO, 2015a; UN GEMS, 2006). As a result, many of the natural water
systems that preserve balanced and thriving ecosystems, have become severely stressed.
The UN report in 2015 revealed that the global clean water scarcity will reach 40% by
2030 (WWAP, 2015). Indeed, water scarcity is not about having too little water, but it is
the scarcity of having good water management practices, which severely endangers
billions of people and the environment (World Water Council, 2019). Water scarcity
afflicted many countries globally, leading poor access to clean potable water and
sanitation (Cosgrove et al., 2014). Globally, about 1.1 billion people lack access to safe
drinking water and 2.7 billion experience water crisis at least a month per year (WHO,
2015a; WWF, 2019). It is obvious that aquatic environments can no longer be perceived
as constant water supply source, and rather these environments are complex matrices that
need careful practices to ensure sustainable and well-functioning ecosystems in the future
(UN GEMS, 2006). This enormous challenge requires a transformation of water
industries based on smart integrated innovative technologies and effective management
outlooks that protects and reuses water for many purposes (Cosgrove et al., 2014).
In these circumstances, municipal wastewater reclamation is one of the promising

approaches to relieve increasing pressure on natural water resources. Nevertheless,
adequate treatment is always required to achieve high quality reclaimed water for reuse
and to ensure no potential threats to human health and aquatic environments (Ma et al.,
2013). Especially, an active evaluation of possible risks of emerging contaminants (ECs)
in the treated effluent of wastewater treatment streams has gained a great attention among
the scientific community (Luo et al., 2014; Nassiri Koopaei and Abdollahi, 2017; Yang
et al., 2017). Moreover, according to EU framework, ECs must be adequately removed
before discharging wastewater or treated effluent to the aquatic environment due to their
persistency, risk of bioaccumulation, and acute or chronic toxicity. Over the decades, EU
water framework legislations are being more stringent day-by-day regarding the removal
of many ECs found in the aquatic environment.
Conventional municipal wastewater treatment processes are mainly aimed for treating
organic matters, suspended solids and nutrients, hence their poor efficiency to treat ECs.
To ensure compliance with the stringent discharge limits for quality water reclamation,
interests in the ability of membrane bioreactors (MBRs) in removing ECs from municipal
wastewater has increased in recent decades (Judd, 2008). MBRs have emerged in the
wastewater treatment sector as one of the most reliable alternatives to the conventional
activated sludge (CAS) processes (Hai et al., 2014). Nevertheless, many ECs are not
efficiently attenuated neither in CAS nor MBRs due to their refractory nature (Xiao et al.,
20 1 Introduction
2019). Therefore, it is essential to develop a treatment system, in which, MBRs can be
integrated with other advanced oxidation processes (AOPs), to enhance the removal of
ECs before discharging the treated effluent is into the environment or reusing it.
Nevertheless, the number of research studies on integrating MBR with AOPs are rather
limited.
1.2 Research objectives

The current thesis investigates the feasibility of advanced wastewater treatment
techniques to treat municipal wastewater containing traces of many ECs that are of
emerging concern. The main research objectives can be divided into two principal
aspects: (1) to assess the functional operation and performance of a conventional MBR
unit in a unique Nordic environment (esp. low water temperature conditions) for treating
real municipal wastewater, in which the pilot MBR unit was operated as an advanced
secondary treatment process and its treatment efficiency, mainly the removal of ECs
under varying operating conditions, such as varying temperatures and solid retention
times (SRT) was assessed; (2) to investigate the possible integration of alternative
advanced water treatment processes for enhancing the ECs removal efficiency of the
conventional MBR with selected AOPs, such as electrochemical oxidation (ECO) and
photocatalytic oxidation (PCO). The main objectives are summarized as follow:
- To study the performance of a conventional MBR to treat the real municipal

wastewater in Nordic winter environment and under varying SRTs (Paper I and II)
- To study the removal of PhACs from MBR treated effluent by utilizing
electrochemical oxidation on newly developed Ti/Ta2O5-SnO2 anode (Paper III)
- To study the removal of PhACs from MBR treated effluent by photocatalytic
oxidation with Ag2O/P-25 photocatalyst (Paper IV)
21
2 Municipal wastewater
Municipal wastewater comprises of either sanitary sewage from households and
commercial buildings or the mixture of sanitary sewage with industrial wastewater,
stormwater and agricultural run-off (Metcalf & Eddy Inc., 2003). Combined sewers are
designed to collect both sanitary sewage and stormwater run-off, while only sanitary
sewage are allowed under separate sewers (US EPA, 2015a). In general, it is reported that
about 70 - 130% of municipal fresh water consumption becomes wastewater (Qasim,
2017).
The characteristics of municipal wastewater varies from site to site depending on land
uses, water consumption patterns, discharges of commercial and industrial wastewater,
extent of separation between stormwater and sanitary sewage, and inclusive of both
diurnal and seasonal fluctuations (Hung et al., 2012). The major constituent of municipal
wastewater is approximately 99.9% of water and the rest fractions includes organics,
inorganics, total suspended solids, together with traces of pharmaceuticals and other
hazardous materials and microorganism (Sperling, 2007). It is difficult to achieve any
effective wastewater management without the detailed understanding of wastewater
characteristics. However, since water-consumption patterns of individuals can rise
haphazardly, the composition and strength of wastewater generated from diverse
standpoint sources fluctuates with time, thus impossible to quantify these fluctuations
accurately. According to US EPA, municipal wastewater can be characterized in terms of
physical, chemical and biological constituents (U.S. EPA, 1997). Table 1 shows the
characteristics of typical municipal wastewater and possible sources.
22 2 Municipal wastewater
Table 1. Characteristics of Municipal wastewater and their sources (Metcalf & Eddy
Inc., 2003; U.S. EPA, 1997)
Characteristics Sources
a) Physical characteristics
Color and Odor Domestic and industrial wastewater, natural
decay of organic matters
Solids Domestic and industrial wastewater, soil
erosion, inflow/infiltration
Turbidity Domestic and industrial wastewater
Temperature Domestic and industrial wastewater
Electrical conductivity Domestic and industrial wastewater
Particle size distribution Domestic and industrial wastewater, soil erosion
b) Chemical characteristics
Organics:
BOD5 /COD/TOC,
Domestic commercial and industrial wastewater
Carbohydrates and proteins,
Fats, oils and grease (FOG)
Refractory organics
Domestic, industrial and agricultural
(Surfactants, pesticides,
wastewater
phenols, pharmaceuticals,
PFCs etc.)
Inorganics:
Domestic water supply, domestic wastes,
Alkalinity, chlorides
groundwater infiltration, water softener
Domestic and agricultural wastewater
Nitrogen
Domestic, commercial and industrial
Phosphorus
wastewater, storm water run-off
Domestic, commercial and industrial
pH
wastewater
Industrial wastewater
Heavy metals
Gases:
Domestic water supply, decomposition of
Methane, hydrogen sulfides,
domestic wastewater, surface water infiltration
oxygen, nitrous oxides
c) Biological characteristics
Bacteria, protozoa, algae, Wastewater treatment plants
helminths
Pathogens (coliform, viruses) Domestic wastewater
23
2.1 Occurrence of emerging contaminants in municipal wastewater

ECs are new chemicals with no associated regulatory standards, and whose effects on
environment and human health are still largely unknown (Deblonde et al., 2011). Over
the past few years, the occurrence of ECs, such as pharmaceutically active compounds
(PhACs), personal care products (PCPs), stimulants, pesticides, steroid hormones,
endocrine disrupting compounds (EDCs) and UV filters, has attained a great concern due
to their ubiquitous presence in various compartments of the aquatic environment (e.g.,
water, sediment and biota) and their potential ecological threats (Barceló and Petrovic,
2008; Deblonde et al., 2011; Tran et al., 2018; Verlicchi and Zambello, 2015). ECs not
only pose threat to aquatic life, but also their constant accumulation in aquatic
environment could evolve antibiotic-resistant microbial strains (Ganiyu et al., 2015).
Currently, more than 700 ECs and their active metabolites are identified in the EU aquatic
environment (UN FAO, 2018). A large number of these ECs have been identified at
source, in treated effluent, and wasted biomass of sewage treatment plants (Clara et al.,
2005b; Trinh et al., 2016; Vieno et al., 2005). The significant spatial and temporal
variations in the concentrations of ECs have been reported by many researchers due to
number of factors, such as rate of production, specific sales and consumption practices,
metabolism, and specific water consumptions (Luo et al., 2014). The physicochemical
properties of some commonly identified ECs in municipal wastewater, most of which
have used as model ECs in this thesis, are summarized in Table 2.
Many ECs are present at extremely low concentrations (µg L-1 to ng L-1) in the
environment, which make their identification and assessment even more challenging
(Rodriguez-Narvaez et al., 2017; US EPA, 2015b). Most of the ECs found in wastewater
and aquatic environment are of anthropogenic origin, which are generally introduced
from various routes, including direct or treated discharge of wastewater from municipal,
hospital, and industrial wastewater treatment plants (WWTPs), surface run-off from
agricultural and veterinary sectors, landfill leachates, and sewer leakage or overflow (Luo
et al., 2014; Tran et al., 2018). Of the many sources, occurrence of ECs at WWTPs are of
main concern since they are frequently detected in both influent and treated effluent of
WWTPs, and subsequently discharge ECs into the environment (Behera et al., 2011). The
category and concentrations of ECs generally found in WWTPs depend on the
socioeconomic composition of the population generating the municipal wastewater (Tran
et al., 2018).
Table 2. Physicochemical properties of some commonly identified ECs in municipal

wastewater.
Molecu Henry’s Law

lar log D Dissociation constant
Category Compound weight (at pH constant (at 25oC)
(g.moL- 7) (pKa) (atm.m3.moL
1 -1
) )
1. PhACs
Analge Paracetamol 151.16 0.47 9.86;1.72 1.92 x 10-11
sics Diclofenac 296.15 1.77 4.18; -2.26 2.69x10-11
and Ibuprofen 206.28 0.94 4.41 5.54 x 10-10
anti- Ketoprofen 254.28 0.19 4.23 1.92 x 10-13
inflam Naproxen 230.26 0.73 4.84 6.08 x 10-12
matori
es
Antiepilepti Carbamazepine 236.27 1.89 13.94; -0.49 9.41 x 10-12
cs
Antibiotics Sulfamethoxazole 253.28 -0.22 5.81; 1.39 2.24 x 10-13
Trimethoprim 290.32 0.27 7.04 1.37 x 10-14
Tetracycline 444.44 -2.06 4.50; 11.02 7.35 x 10-29
Ciprofloxacin 331.35 -0.33 6.43; 8.68 2.10 x 10-17
Tylosin 916.11 0.15 13.06; 7.39 n.a.
Ofloxacin 361.37 -0.20 5.19; 7.37 1.76 x 10-16
Norfloxacin 319.33 -0.65 0.16; 8.68 4.13 x 10-16
Oxytetracycline 460.43 -2.25 4.50; 10.80 1.33 x 10-32
Metronidazole 171.15 -0.14 14.44;2.58 2.01 x 10-12
Doxycycline 444.44 -0.92 4.50; 10.84 3.20 x 10-26
Lipid Clofibric acid 214.65 -1.06 3.18 2.91 x 10-11
regulators Gemfibrozil 250.34 2.07 4.75 1.92 x 10-11
Bezafibrate 361.82 -0.93 3.29; -2.06 2.12 x 10-19
Simvastatin 418.57 4.72 13.49 4.93 x 10-16
β-Blocker Bisoprolol 325.45 -0.54 9.42 4.54 x 10-15
Atenolol 266.34 -2.09 13.88;9.43 1.34 x 10-17
Metoprolol 267.37 -0.81 13.89; 9.43 1.59 x 10-13
Propranolol 259.35 0.45 13.84; 9.50 6.27 x 10-14
Sotalol 272.36 -2.01 8.28; 9.31 2.63 x 10-14
Diuretics Furosemide 330.74 -0.79 3.04; -2.49 2.57 x 10-19
Hydrochlorothiazide 297.73 -0.03 8.95; -4.08 5.61 × 10-17
Enalapril 376.45 -0.14 3.15; 5.43 1.48 × 10-20
Psychostim Caffeine 194.19 -0.63 0.52 1.63 × 10-12
ulants
25
Molecu Henry’s Law
lar log D Dissociation constant
Category Compound weight (at pH constant (at 25oC)
(g.moL- 7) (pKa) (atm.m3.moL
1 -1
) )
Contrast Iopamidol 777.09 -2.54 10.87; -2.85 6.67 ×10-30
media Iopromide 791.12 -2.66 10.62; -2.60 3.29 × 10-35
Antidepress Fluoxetine 309.33 1.15 10.05 2.25 × 10-11
ants Paroxetine 329.37 1.16 9.68 3.53 × 10-12
Anticoagula Warfarin 308.33 0.67 4.50 1.42 × 10-15
nts
Immunosup Methylprednisolone 374.48 2.17 12.46 1.88 × 10-17
pressives Methotrexate 454.45 -4.98 3.47; 5.56 n.a.
Beclomethasone 408.92 2.44 12.17 2.22 × 10-18
Cyclophosphamide 261.08 0.73 2.84 9.46 × 10-10
Ifosfamide 261.08 0.76 1.44 9.46 × 10-10
Hydrocortisone 362.47 1.76 12.47 1.74 × 10-17
Bronchodila Terbutaline 225.28 -1.77 9.11; 9.65 3.59 × 10-14
tors Clenbuterol 277.19 0.24 13.29; 9.51 9.12 × 10-13
Salbutamol 239.31 -1.77 9.99; 9.62 8.75 × 10-15
Anti- Flubendazole 313.29 3.08 10.66; 4.45 n.a.
parasitics Fenbendazole 299.35 2.35 10.80 n.a.
Ketoconazole 531.43 3.80 6.88 2.18 × 10-24
Contracepti Norethindrone 298.43 2.86 13.09 9.5 × 10-12
c
2. EDCs
Plasticizer Bisphenol A 228.29 3.64 10.29 1.92 × 10-11
3. Steroid 17β-Estradiol 272.38 4.15 10.27 1.17 × 10-9
Hormon 17α-Ethynylestradiol 296.41 4.11 10.24 3.74 × 10-10
es Estriol 288.39 2.53 10.25 1.75 × 10-11
Estrone 270.37 3.62 10.25 9.61 × 10-10
Progesterone 314.46 3.83 n.a. 1.51 × 10-09
Testosterone 288.43 3.18 15.06 4.89 × 10-11
4. Industri Perfluoro-octanoic 414.07 2.69 0.50 1.13 × 10-05
al acid (PFOA)
Surfacta
nts
Perfluorooctane 212.28 4.49 <1.0 4.1 × 10-04
sulfonic acid (PFOS)
n.a.- data not available
(Data are extracted from SciFinder database
https://scifinder.cas.org/scifinder/view/scifinder/scifinderExplore.jsf)
2.2 Environmental intimidations of ECs
The major concern on environmental risk of ECs is evidenced only if they may adversely
pose impact on aquatic life and human health. There are three major risk factors related
to the ECs presence in the aquatic environment viz. persistency, bioaccumulation, and
toxicity (Ebele et al., 2017). Persistency is mainly defined by the physicochemical
properties of ECs, due to which many of them are difficult to remove by conventional
means of treatment. Indeed, their incomplete removal could generate many active
metabolites that could pose potential risks to aquatic life. Even though not all the detected
ECs in environmental matrix are persistent, their continuous replenishment and release to
the environment referred them as ‘‘pseudo-persistent’’ compounds, which pose higher
persistence even when acted by environmental processes, such as photodegradation,
biodegradation, and sorption into sediments (Bu et al., 2016; Ebele et al., 2017). Another
risk is the bioaccumulation of ECs in aquatic biota. Generally, ECs are designed and used
for specific target, but since many of ECs and their metabolites are biologically active,
they can adversely affect non-targeted aquatic organisms (Rodríguez-Mozaz et al., 2016).
Coogan et al reported an abundant bioaccumulation of ECs and their metabolites in algae
samples of WWTPs (Coogan et al., 2007). On the other hand, toxicity is one of the major
concern of ECs, which can emerge due to complex mixtures of ECs that leads to high
synergetic effects. This means that the presence of ECs even at low concentrations could
pose significant toxicity to aquatic microorganisms. Another risk related to the ECs in the
environment is the possible emergence and spread of antibiotic resistant strains in natural
bacterial populations, which is a very emerging topic in recent years. Antibiotic resistance
is aggravated by the misuse and overuse of antibiotic that has continuously undermined
many advances in health and medicine (WHO, 2015b).
2.3 Municipal wastewater treatment

The main objectives of wastewater treatment are to separate, convert, and eliminate
objectionable and hazardous contaminants and infectious pathogenic organisms via a
combination of physical, chemical and biological processes (EU, 1991). Wastewater
produced from municipalities and communities must ultimately be returned to receiving
waters or to land or reuse after the treatment. Wastewater treatment is a multi-step
process, which typically involves primary, secondary and tertiary treatments, according
to the treatment objectives, level of treatment efficiency, and environmental threats on
the receiving water bodies (Sperling, 2007). The municipal wastewater reclamation is an
important concept, in which the collected municipal wastewater is treated and reused for
recharging fresh water sources and recovering value-added resources and energy (Metcalf
& Eddy Inc., 2003).
The primary treatment mostly includes the physical treatment of wastewater in order to
reduce to treatment load to the subsequent steps by selectively removing settleable
organic and inorganic materials and floating (scums) materials via sedimentation. This
27
process can reduce about 25 – 50% of BOD5, 50 – 70% TSS and 65% FOG loads (Aziz
and Mojiri, 2014). However, advanced primary treatment or chemically enhanced
primary treatment could be attained by the addition of coagulants (aluminium sulphate,
ferric chloride or other polymers) to remove mainly phosphorus by precipitation
(Sperling, 2007).
The secondary treatment is intended for the removal of organic pollutants (either
suspended, particulate, or dissolved) and nutrients (nitrogen and phosphorus) (Punmia et
al., 1998). Secondary treatment typically includes biological reactor, where pollutants are
removed by bacteria metabolism, and followed by the secondary sedimentation tank to
separate settleable solids and treated water. The biological treatment process can be
suspended growth, attached growth, and their various combinations. The suspended
growth system includes CAS process (with or without biological nitrogen and phosphorus
removal), extended aeration, and sequencing batch reactors, in which microorganisms
attached to the biomass are maintained in suspension. The biomass and microorganism
assays develop in support medium in attached growth system, including trickling filter,
rotating biological contractor, and submerged aerated biofilters. On the other hand,
combined suspended attached growth system involves processes with internal and up-
flow suspended packing for attached growth such as moving-bed biofilm reactor (MBBR)
and fluidized-bed bioreactors (Metcalf & Eddy Inc., 2003).
2.4 EU legislation and removal of ECs from municipal wastewater

The main objective of EU framework is to protect natural water environment in Europe
and to replenish every river, lake, groundwater, wetland and other surrounding water
bodies (Buttiglieri and Knepper, 2008). Due to persistency, risk of bioaccumulation in
biota, and high synergetic toxicity, ECs must be adequately removed before releasing
with treated effluent to the environment. The EU Water Framework Directive
2000/06/CE announced a list of 33 priority substances or group of hazardous substances,
including heavy metals, pesticides, and polyaromatic hydrocarbon (EU 2000). In
December 2008, directive 2008/105/EC was introduced amending to directive
2000/06/CE for progressively reducing contamination of priority substances with
environmental quality standards (EQS) (EU 2008). In 2013, the directive 2008/105/EC
was further amended to 2013/60/EC by identifying additional 12 substances and a first
watch list of chemicals, including diclofenac, 17- 𝛽 -estradiol (E2) and 17- 𝛼 -
ethinylestradil (EE2) (EU 2013), was established. The watch list of these substances was
set out in Commission Implementing Decision (EU) 2015/495 according to Article 8b of
directive 2013/60/EC that additionally includes 2,6-Ditert-butyl-4-methylphenol, 2-
Ethylhexyl 4-methoxycinnamate, macrolide antibiotics, methiocarb, neonicotinoids,
oxadiazon and tri-allate (EU 2015). In 2018, the Commission Implementing Decision
(EU) 2015/495 was updated to EU 2018/840, in which the substances with adequately
available high-quality monitoring data, such as diclofenac, 2,6-Ditert-butyl-4-
methylphenol, 2-Ethylhexyl 4-methoxycinnamate, oxadiazon and tri-allate, were
removed from the watch list with inclusion of sensitive substances, including amoxicillin,
metaflumizone and ciprofloxacin (EU 2018).
Since municipal WWTPs receive diverse type and concentrations of ECs originated from
synthetic chemicals that have been used for many purposes, these WWTPs are the major
sources of many environmental pollution. More than 100,000 tons of ECs are produced
globally (Radjenović et al., 2008). As confirmed by many researchers, the conventional
wastewater treatment methods, such as CAS processes, are not efficient enough to remove
persistent ECs, and many of these ECs remain poorly degraded and subsequently
discharged into receiving waterbodies with treated effluent (Clara et al., 2005; Vieno et
al., 2007; Vieno and Sillanpää, 2014; Zorita et al., 2009). Polar ECs are mostly
hydrophilic and highly persistent under activated sludge process conditions, which lead
them easily detected even in the treated effluent (Buttiglieri and Knepper, 2008).
Therefore, upgrading of WWTPs and implementation of more sustainable technologies
could be the possible solutions to abate such ECs emissions and to achieve high-quality
treated effluent for the safe reclamation (Radjenović et al., 2009). On this note, MBR is
one of the advanced wastewater technologies that has gained increased interest over three
decades (Xiao et al., 2019).
29
3 State-of-the-art of MBRs: advanced municipal
wastewater treatment
3.1 Introduction to MBRs
MBRs couple biological treatment with membrane separation to produce clarified and
highly disinfected effluent (Judd, 2016). MBR technology is based on suspended growth
process that applies microporous membranes to allow solid/liquid separation (Radjenović
et al., 2008). The combination of a biological reaction to membrane separation
strengthens the biological process and replaces many other subsequent operations, such
as secondary clarifiers (Coutte et al., 2017). MBRs have become a state-of-the-art
technology in water management practices for wastewater treatment and reclamation due
to their remarkable benefits such as high quality treated effluent, relatively small plant
footprint, and less sludge production (Drews, 2010). The MBR enables to maintain a high
MLSS concentration, which is important for slow growing microorganisms (Drews,
2010; Judd, 2008). MBRs can perform biological treatment and disinfection of the
effluent simultaneously with the optimum control of biological degradation and greater
stability and flexibility in operation (Radjenović et al., 2009).
There have been several historical milestones leading to the development of today’s
MBRs. The combination of membrane solid/liquid separators in biological treatment
system has been studied since 1960s (Visvanathan et al., 2000). The era between 1960s
and 1980s is often known as a golden age on membrane science development (Hai et al.,
2014). The first historical coupling of an activated sludge bioreactor with a cross flow
membrane filtration looped side-stream MBRs, was developed by Dorr-Oliveir Inc. in
1969 (Smith et al., 1969). However, the first generation MBRs associated with poor
economics due to higher membrane cost and higher energy consumption. The first
breakthrough of the MBR application emerged in 1989 when Yamamoto et al. introduced
the submerging membranes inside the bioreactor (Yamamoto et al., 1988). Currently,
more than 800 commercial MBRs are implemented in Europe, with the compound annual
global market growth rate up to 12% (Krzeminski et al., 2017). The global MBR market
should grow from 1.9 billion USD in 2018 to reach 3.8 billion USD by 2023 (BCC
Research, 2019). More than 60 super-large scale (capacity ≥ 100, 000 m3 d-1) MBR plants
will be in use by 2026, with Tuas Water Reclamation Plant being largest MBR (1200,
000 m3 d-1) in the world (The MBR site, 2019).
Stricter legislations concerning the effluent discharge, demands for wastewater reuse, and
dramatic reduction of membrane capital costs (~1/10) are the major drivers of rapid MBR
growth worldwide (Park et al., 2015). For instance, EU Urban Wastewater Treatment
Directive (91/271/EEC), amended directive 2006/7/EC for management of bathing water
quality, and EPA Clean Water Protection Act (2009), are the important legal complies
30 3 State-of-the-art of MBRs: advanced municipal wastewater treatment
on treated effluent discharge or reuse limits that expanded the MBR market growth
(Krzeminski et al., 2017).
3.2 Membrane materials, configuration and operating principles of

MBRs
Generally, pressure-driven membrane processes are comprised of microfiltration,
ultrafiltration, nanofiltration, and reverse osmosis. Of these membranes, including
microfiltration (0.1 to 1 µm) and ultrafiltration membranes ( 0.05 to 0.1 µm), which are
widely used to separate particulate materials from water and wastewater, are referred as
low-pressure membranes (Yang et al., 2006). On the other hand, nanofiltration and
reverse osmosis are high-pressure membranes, which are commonly used to separate
more soluble matters (Bérubé, 2010). The commercial MBR membrane pore sizes are
generally between 0.03 and 0.4 µm (Judd, 2016). The most widely used membrane
materials are cellulose, polyamide, polysulphone, polymeric materials, including
polyvinylidene difluoride (PVDF), polyethylsulphone (PES), polyethylene (PE), and
polypropylene (PP). Among the commercial membranes, PVDF membranes are the most
widely used membranes that account almost half of all the products both in flat sheet (FS)
and hollow fibre (HF) configurations (Judd, 2016). Most of the polymeric membranes are
resilient to chemical and physical reactions, but severely prone to fouling due to
hydrophobic nature of these membranes and biopolymers. Therefore, often the surface of
commercially available membrane modules are amended by chemical oxidation, metal
organic framework, or plasma treatment to achieve more hydrophilic surface (Radjenović
et al., 2008).
In general, there are two basic MBR configurations for the membrane module, depending
on the membranes placed either inside or outside the bioreactor, such as immersed or
submerged and side-stream or external MBR system as shown in Figure 1 (Gupta et al.,
2008; Judd, 2008; Melin et al., 2006). Submerged or immersed MBRs are more often
applied in municipal wastewater treatment due to their high compatibility with the
activated sludge process, compactness in design, relatively low energy consumption, and
easy wasting of excel sludge (Gupta et al., 2008). On contrary, external or side-stream
MBRs are stand-alone systems, are generally used to treat industrial high strength
wastewater with poor filterability and require relatively less membrane area than in
submerged MBRs. As pumping and recirculation of activated sludge is the integrated part
in external MBRs, high cost is required due to more energy dissipation (Judd, 2008).
31
Figure 1: MBR process configurations, (a) Submerged and (b) side-stream
There are three different modules of membranes commercially used in MBRs, such as
tubular membranes, hollow fibre and flat sheet membranes (Figure 2). Tubular
membranes are typically used in external MBRs, where membranes having internal
diameter greater than 5 mm are grouped into modules containing multiple tubes. Hollow
fiber and flat sheet membranes are used typically in submerged MBRs. Hollow fiber
membranes generally have external diameter in the range of 1- 3 mm and confined into
modules of dozens to thousands of fibers (Bérubé, 2010).
a) b) c)
Figure 2: Different MBR membrane modules: (a) Tubular, (b) Hollow-fiber,

(c) flat-sheet (Reprinted with permission from (Bérubé, 2010)).
3.3 Membrane fouling: a major shortcoming in MBRs
Despite the promising treatment technology, membrane fouling is inevitable phenomenon
in pressure driven MBRs, which normally increases TMP and reduces permeate flux (Le-
Clech et al., 2006). Membrane fouling is the possible deposition of sludge flocs,
particulates, solutes, microorganisms and cell debris in the membrane pores, and on the
membrane surface (Burman and Sinha, 2018; Meng et al., 2009). However, due to its
complex nature, membrane fouling in MBRs is not yet entirely understood (Meng et al.,
2010).
The major fouling consequences in MBRs involve increase in TMP, rapid deterioration
of permeate flux, and increase in filtration resistance, leading to frequent cleanings
requirement, and high energy consumption for aeration (Drews, 2010). Moreover,
membrane fouling also exacerbates feed pressure, reduces effluent quality, weakens
membrane performance, and reduces membrane effective life (Burman and Sinha, 2018).
The specific energy consumption in MBRs operation generally ranges from 0.6 – 2.3 kWh
m-3 (Krzeminski et al., 2017), which is quite high in comparison to CAS processes (0.3 –
0.6 kWh m-3) (Dai et al., 2015). Even though, the membrane costs have dramatically
reduced these days, which ultimately reduce capital investment cost in MBRs, the overall
operating costs are still high mainly due to membrane aeration for fouling mitigation.
Moreover, the use of aggressive chemicals for repeated membrane cleaning events can
create additional environmental concerns (Drews, 2010).
Membrane fouling can be classified mainly into reversible and irreversible types.
Reversible fouling during filtration mainly occurs due to external deposition of loosely
bound materials on the membrane surface that causes cake layer formation. Reversible
fouling can simply be controlled by physical techniques, which can involve back flushing
or relaxation under crossflow conditions. Contrarily, irreversible fouling mainly occurs
due to strongly attached materials and pore blocking of the membrane, which can only be
removed by chemical cleaning (Burman and Sinha, 2018). Nevertheless, irrecoverable
fouling, which is a subsequently left to be treated by chemical cleaning, is difficult to
remove by any cleaning methods (Drews, 2010).
Furthermore, membrane fouling can be categorized as biofouling, inorganic fouling and

organic fouling depending on the fouling components (Meng et al., 2009). Biofouling
occurs because of biological growth, metabolism and its secretion, which ultimately
accumulate on the membrane and its pores and leading to the formation of biofouling.
Organic fouling is due to the organic substances contained in the biomass and organic by-
products produced by microorganisms, which can easily be deposited on membrane
surface. Colloidal initiate membrane pore clogging and biopolymers, such as protein,
polysaccharides and humics can further propagate organic fouling. Whereas, inorganic
fouling forms due to the chemical precipitation of inorganic ions (anions and cations),
including Ca2+, Mg2+, Al3+, Fe3+, PO43-,SO42- and OH-. Inorganic fouling mainly occurs
33
via two mechanisms, such as crystallization and particulate fouling. Of these three
fouling, biofouling and organic fouling are dominant in MBRs (Burman and Sinha, 2018;
Hai et al., 2014). Indeed, membrane fouling is a very complex mechanism that includes
complicated inter-relationships between various factors, such as reactor design and
operational conditions, wastewater composition, sludge rheology, ambient conditions
(Zhang et al., 2014). Membrane fouling can be quantitatively determined according to the
Darcy’s equation as follows (Meng and Yang, 2007):
TMP
Rt = (1)
μJ
Where, Rt is the total membrane filtration resistance (m-1); TMP is trans-membrane

pressure (Pa); µ dynamic permeate viscosity (Pa.s); and J is the membrane flux (L m-2 h-
1
).
The water temperature greatly influences the efficiency of biological treatment processes
(van den Brink et al., 2011). When temperature drops at about 15°C, methane-producing
bacteria becomes essentially inactive, autotrophic bacteria practically cease functioning
at about 5°C, and even chemoheterotrophic bacteria feeding on carbonaceous material
becomes dormant at 2°C (Metcalf & Eddy Inc., 2003). In MBRs, temperature not only
influences biological process but also effects on membrane performance. However, MBR
process operation in low-temperature zones are comparatively less studied than under
normal ambient conditions (Zhang et al., 2014).
At lower temperatures, MBR efficiency deteriorated mainly due to increased MLSS
viscosity, releasing of more EPS due to increased sludge deflocculation, and reduced back
transport velocity of particles from membrane surface (van den Brink et al., 2011).
Arévalo et al. reported the loss in permeability and increased membrane flux resistance
at temperatures < 15°C (Arévalo et al., 2014) in a full-scale MBR. Brink et al. showed
that at low temperatures, membrane fouling is intensified due to the release of
polysaccharides and submicron particles from sludge flocs (van den Brink et al., 2011).
The bulking of sludge has noticed with significant increase in the sludge volume index
(SVI) at 13°C, which was accompanied by rapid membrane fouling and thereby two-fold
increase in the membrane cleaning frequency (Zhang et al., 2014). The bulking sludge
can cause severe cake fouling due to the accumulation of more irregular shaped sludge
flocs, whereas deflocculated sludge can cause both cake fouling and pore blocking fouling
by the deposition of colloidal particles and dissolved matter onto/into the membrane
surface (Meng and Yang, 2007). Krzeminski et al. observed that the colloidal and soluble
fraction (<1 µm) plays a major role in increasing the membrane resistance during winter
period (Krzeminski et al., 2012). The possible fouling components in MBRs and their
spatial and temporal distributions are schematically shown in Figure 3 (Meng et al.,
2017).
Figure 3: Temporal and spatial development of membrane fouling in MBRs

(Modified from (Meng et al., 2017)).
The most common techniques of cleaning membranes in MBRs are physical and chemical
methods. Physical cleaning includes backflushing or relaxation which is applied for 1-2
min in every 10-15 min, whereas alkaline (pH ~ 12) sodium hypochlorite often followed
by acidic (pH ~3) citric (or oxalic) acid with concentration up to 1% are used as chemical
cleaning (Judd, 2008). However, some innovative membrane cleaning protocols have
been emerged over the traditional methods to emphasize improved efficiency and
minimum energy expenditure. The emerging novel strategies involved the application of
various additives (flux enhancers), biocarriers, sludge granulation, membrane surface
modifications, and quorum quenching techniques (Drews, 2010).
3.4 Fate of ECs in MBRs: application of mass balance

The removal mechanism of ECs in MBRs is a complex phenomenon, which is
characterized by following four major pathways: (i) biotransformation or biodegradation;
(ii) sorption onto the sludge; (iii) volatilization; and (iv) physical retention via membrane
size exclusion (Cirja et al., 2008a). However, the ECs removal via volatilization and
physical retention by membrane is almost insignificant since Henry’s law constant (KH)
values of most of the ECs are absolutely low (< 10-6) ( Table 2) and the molecular weight
cut-off (MWCO) of most of the pressure driven MF and UF membranes are above several
35
thousand Daltons compared to molecular weight of commonly found ECs i.e., 150 to 900
Da. Therefore, biotransformation and sorption onto the sludge are the major removal
mechanisms of ECs in MBR treatments (Li et al., 2015; Luo et al., 2014).
The mass balance approach is useful to assess the fate of ECs considering their percentage
shares of biotransformation/biodegradation, sorption and remained in the treated effluent.
The mass loads (µg d-1) of ECs in influent, effluent and sludge can be calculated by
multiplying their measured concentrations (CInf, CEff, CSludge) with the equivalent flow
data (QInf, QEff, QSludge). Then, the relative distributions of ECs to the influent mass load
can be estimated according to the following equations (2)-(5)(Joss et al., 2006):
( C Inf ∗ QInf − C Eff ∗ QEff )

Overall removal (%) = ∗ 100 (2)
C Inf ∗ QInf
Removal via sorption onto sludge(%)
C Sludge ∗ QSludge
= ∗ 100 (3)
C Inf ∗ QInf
Remain in effluent (%)
C Eff ∗ QEff
= ∗ 100 (4)
C Inf ∗ QInf
Removal via biotransformation(%)
= Overall removal (%)
− Removal via sorption onto sludge (%) (5)
where C Inf ∗ QInf ; C Eff ∗ QEff ; 𝑎𝑛𝑑 C Sludge ∗ Q Sludge are the mass load (µg d-1) of
ECs in influent, treated effluent and sludge, respectively.
The removal mechanism of ECs in MBRs is under the influence of mainly ‘internal
factors’ and ‘external factors’(Luo et al., 2014). The internal factors are related to the
physicochemical properties of ECs, including hydrophobicity, biodegradability, degree
of ionization, and volatility (Cirja et al., 2008a; Li et al., 2015). In general, ECs with High
log D value (log DpH 7 > 3.2) are hydrophobic, poorly soluble and with high sorption
affinity on organic constituents (Cirja et al., 2008a). The higher removal efficiency of
hydrophobic compounds may be ascribed to the sorption of ECs onto sludge that assists
for enhanced biotransformation. Contrarily, the removal of hydrophilic ECs (log DpH 7 <
3.2) can be described based on the qualitative framework reported by Tadkaew et al.: (i)
compounds possessing only EDGs, (ii) compounds having both EDGs and e-withdrawing
(EWDs), (iii) compounds having only strong EWGs (Tadkaew et al., 2011). Likewise,
degree of ionization or polarity nature of ECs can influence their removal in MBR. In
general, polar and non-volatile ECs are the main concerns in wastewater treatment
processes as they can easily escape with treated effluent. The solid-water distribution
coefficient (Kd), which indicates the partition of ECs in water phase and sludge, is a
relative indicator of sorption behavior (Joss et al., 2006). The sorption tendency is
insignificant for compounds having log Kd < 2.48, whereas those having log Kd > 3.2 can
have higher removal (Luo et al., 2014; Tadkaew et al., 2011). Similarly, biodegradability
of ECs depends on their bioavailability, which involves primary uptake of ECs by
microorganism cell, leading to their degradation due to the increased affinity with the
bacterial enzymes (Cirja et al., 2008a). The molecular structure of compound greatly
influences the complexity of compound to biodegradation. In short, ECs with linear short
chains, unsaturated aliphatics, and EDGs, are more biodegradable than ECs with long and
more branched side chains, saturated or polycyclics, and sulfate, halogen or EWGs (Luo
et al., 2014).
The external factors are mainly WWTP-specific that involves process operating
conditions, such as SRT, temperature and pH (Luo et al., 2014). SRT better controls the
size and diversity of microorganism community. In general, the MBRs operated with
extended SRTs can improve building up of slowly growing bacteria (e.g., nitrifying
bacteria) and longer retention of sludge inside the bioreactor, which favors higher removal
of ECs (Suárez et al., 2012). Nitrifying conditions can have positive effect on the removal
of ECs via co-metabolism using ammonium monooxygenase enzyme. The SRTs above
10 days that allow nitrifying conditions can enhance the attenuation of many
biodegradable compounds (Clara et al., 2005a). Nevertheless, there are many studies that
revealed no enhanced elimination of ECs with wide variation of SRTs, indicating that
high SRT does not necessarily mean enhanced removal of ECs (Luo et al., 2014).
Likewise, pH value is another critical parameter influencing the removal of ECs in MBR
treatment. Various protonation states of ECs can be expected depending on their pKa
values. It is expected that at low pH, hydrophobicity of most of the ionizable ECs would
increase that ultimately increases their adsorption onto sludge mass, which indeed favors
increased time for biotransformation (Urase et al., 2005). Modest pH variation can also
have significant effects on the removal of acidic ECs due to the activation of enzymes or
enhanced affinity between sludge and ECs due to protonation phenomenon (Kimura et
al., 2007). On contrary, at varying pH, no changes in the removal efficiency of non-
ionizable compounds is noticed (Taheran et al., 2016).
On the other hand, seasonal variation of wastewater temperature can affect the removal
of ECs in MBRs, which influence mostly biodegradation and partition mechanisms (Luo
et al., 2014). In general, with increasing temperatures, adsorption equilibria are reached
earlier, and microbial activities and biodegradation ECs are enhanced. However, Hai et
al. reported the decreased removal of most hydrophobic ECs when process temperature
reached 45 °C (Hai et al., 2011).
37
3.5 Fate of human enteric viruses and heavy metals in MBRs
The human enteric viruses, including noroviruses and adenoviruses, are most common
pathogenic group associated with several waterborne diseases, such as epidemic
gastroenteritis, conjunctivitis, and respiratory diseases (Kuo et al., 2010; Miura et al.,
2015). Human enteric viruses are largely found in municipal wastewater, which are highly
resistant to inactivation. Adequate reduction of viruses is of a prime regulatory focus from
water reuse due to their acute impact on public health and hygiene (Chaudhry et al.,
2015b). The inadequate WWTPs treatment efficiency often led to the insufficient removal
of viruses from source waters, causing their outbreaks to the environment (Lodder and
Husman, 2005).
MBRs offer absolute barrier to solids and microorganisms, despite MBRs are not
significantly efficient retaining viruses via membrane size exclusion (Herath et al., 1999).
Even though virus removal in MBRs are ascribed to their aggregation and adsorption to
sludge and retention by the gel and cake layer formed on the membrane surface. On the
other hand, formation of fouling layer on the membrane surface plays a vital role on
controlling virus removal, yet current understanding of interactions between viruses and
membrane surface, impact of membrane cleaning as well as membrane imperfections in
MBRs is limited (Hai et al., 2014).
Miura et al. indicated that adsorption to sludge plays a major role in the removal of
norovirus GII in the MBR, in which diverse adsorptive behavior of viruses to sludge
influenced the overall removal efficiency (Miura et al., 2015). Kuo et al. reported the
attachment of adenovirus to sludge and membrane-biofilm as a major mechanism for their
removal in the MBR (Kuo et al., 2010). However, the biomass attachment removal
mechanism is mostly dependent on the type of virus (Chaudhry et al., 2015a). In another
study, Chaudhary et al. suggested the major removal contribution of viruses is majorly by
backwashed membrane, inactivation by biomass due to predation or enzymes, the cake
layer formation and attachment to solids (Chaudhry et al., 2015b).
Heavy metals refers to any metals with relatively high density (>5 g cm-3) and atomic
weights, which are most serious environmental threats due to their recalcitrance and
persistence in the environment (Chipasa, 2003; Fu and Wang, 2011). The risks associated
with the bioaccumulation in the food chains are the major environmental and human
health concerns (Cantinho et al., 2016). The presence of heavy metals at certain
concentrations not only pose environmental concerns but also strongly reduces microbial
activity, leading to adverse effects on biological treatment processes (Chipasa, 2003).
Trace quantities of many heavy metals are identified in municipal wastewater from
various sources, mainly anthropogenic emissions. Of the many heavy metals, cadmium
(Cd), lead (Pb), mercury (Hg) and nickel (Ni) are defined by EU directive 2013 as priority
hazardous elements, which could pose acute to chronic health problems to human and
aquatic life depending on their potency and time of exposure (EU, 2013).
Some studies have confirmed the potential of MBRs for enhanced removal of heavy
metals from wastewater than conventional treatment methods (Di Fabio et al., 2013;
Fatone et al., 2008; Katsou et al., 2011; Santos and Judd, 2010). There are several
mechanisms involved in heavy metal removal by MBRs, such as surface adsorption and
diffusion of metal ions in sludge flocs, adsorption and complexation formation of soluble
metal ions in EPS matrix, and to some extent by chemical precipitation of metals due to
alkaline nature of wastewater (Katsou et al., 2011). Di Fabio et al., who studied the
removal and fate of heavy metals in a pilot-scale MBR revealed that the biosorption of
heavy metals to fouling layer (esp. EPS matrix) is more evident than to sludge (Di Fabio
et al., 2013). However, the pH of the medium greatly influences the number of EPS
binding sites for heavy metals (Comte et al., 2008). Fatone et al. reported that operating
MBRs at high SRTs is not effective strategy to enhance the removal of heavy metals at
low concentrations (Fatone et al., 2008). Addition of vermiculite minerals could enhance
the heavy metal removals due to increased sites for biofilm formation and microorganism
growth, but can increase sludge concentration and membrane fouling (Katsou et al.,
2011).
39
4 Advanced oxidation processes for wastewater treatment
Biological treatment is usually the most economic process to treat ‘readily degradable’
organic pollutants present in wastewater. However, if the wastewater is comprised of
refractory (non-biodegradable) organic pollutants, such as several ECs, then it requires
especial type of treatment to meet the desired discharge limits (Comninellis, 1994). In
this regard, advanced oxidation processes (AOPs), based on aqueous phase oxidation
method on the intermediacy of powerful reactive species, mainly hydroxyl radicals (•OH),
leading to the nonselective degradation of the organic pollutants to their mineralization
(Sirés et al., 2014).
AOP was first proposed by Glaze in 1980s for potable water treatment using ozone
(Glaze, 1987). Since then AOPs are widely applied to treat wastewater as strong oxidants
that can degrade recalcitrant ECs and simultaneously increase the biodegradability (Deng
and Zhao, 2015). AOPs are promising, efficient and environmental-friendly approaches
to treat wastewater contaminated with wide range of ECs, mainly by electrochemical,
photochemical, ozonation, Fenton, or sonochemical reactions (Nidheesh et al., 2018;
Oturan and Aaron, 2014). AOPs can be utilized either as pre- or post-biological treatment
depending on the concentration of ECs in the wastewater influent stream. Pre-treatments
using AOPs are more feasible when the concentrations of ECs are higher than the organic
matters, whereas post-treatments are realistic when the concentrations of ECs are
relatively lower than organic matters (Neoh et al., 2016).
The hydroxyl radical is the second most powerful oxidizing agent after fluorine with a
redox potential of E° (•OH/H2O) = 2.8 (V vs. SHE), and a life time of less than a second
(Brillas and Martínez-Huitle, 2015). These in situ produced •OH radicals then rapidly
reacts with organic pollutants at high reaction rates mainly by the electron transfer or
redox reactions, H atom abstraction (aliphatics), or the addition on an unsaturated bond
(aromatics) to initiate a radical chain oxidation reaction (Nidheesh et al., 2018; Sirés et
al., 2014).
4.1 Electrochemical oxidation (ECO) process

Electrochemical oxidation processes have gained great attention over the last decade for
the decontamination of wastewater containing persistent organic pollutants and for
disinfection purposes (Brillas et al., 2009; Särkkä et al., 2015). ECO processes are
emerging as most appealing environmentally friendly and highly efficient
electrochemical treatments (Martínez-Huitle and Brillas, 2009). The major benefit of
electrochemical treatments is its ability to regulate and generate in situ •OH radicals
without adding chemicals or large amount of catalyst (Oturan and Aaron, 2014). Instead,
ECO process can also benefits on the prevention and remediation of pollution issues since
the electron is a clean reagent (Panizza and Cerisola, 2009). In ECO treatments, the
organic pollutants can be removed by: (i) direct anodic oxidation (or direct electron
40 4 Advanced oxidation processes for wastewater treatment
transfer between anode surface and pollutants), which usually yields relatively poor
oxidation rates, and (ii) chemical reaction with electrogenerated reactive oxygen species
at anode surface such as physisorbed or chemisorbed •OH radicals, which leads to total or
partial oxidation of pollutants, respectively (Martínez-Huitle and Brillas, 2009; Panizza
and Cerisola, 2009).
In direct anodic oxidation, pollutants are oxidized at anode surface after adsorption solely
due to the electron, without the involvement of any other substances. In this process,
oxygen is transferred from water to the organic pollutant via electrical energy, which is
called electrochemical oxygen transfer reaction as described by equation (6) (Panizza and
Cerisola, 2009):
Rads – ne- → Pads (6)
where, Rads and Pads are adsorbed and oxidized adsorbed organic pollutants at anode
surface. Direct electro-oxidation can be possible at low potentials, but usually possesses
low kinetics, depending on the electrocatalytic activity of anode materials. The use of low
cell potentials avoiding oxygen evolution could frequently cause the loss of anodic
activity. This is called poisoning effect and occurs due to the formation of polymer layer
on the anode surface, which has limited their application in wastewater treatment
practices (Martínez-Huitle and Brillas, 2009). The anode surface deactivation depends
on, (i) adsorption properties of the anode surface ,and (ii) nature and concentration of
organic pollutants and their transformation products (Panizza and Cerisola, 2009).
On the other hand, electrochemical oxidation of organic pollutants can be obtained

without the poisoning effect by electrolyzing the aqueous solution at high anodic voltages
with the involvement of oxygen evolution reactions (OERs), which generates in situ •OH
radicals with no oxidation catalysts. A number of anodes favored partial and selective
oxidation of organic pollutants (i.e., conversion method), whereas many others favored
complete mineralization (i.e., combustion/incineration) to CO2 (Brillas and Martínez-
Huitle, 2015; Panizza and Cerisola, 2009). Comninellis found that the nature of anode
material strongly effects both the selectivity and efficiency of the ECO process, and
proposed a comprehensive model of organics oxidation in acidic medium including the
competition with O2 evolution reaction and heterogeneous •OH radicals (Figure 4)
(Comninellis, 1994). The model assumes that in the initial step of oxygen transfer
reaction, water molecules split to form adsorbed •OH radicals at anode oxide surface
(MOx) as equation (7):
MOx + H2O → MOx (•OH) + H++ e- (7)

41
Figure 4: Electrochemical oxidation scheme of organic pollutants on

the metal oxide anode surface (Modified from (Comninellis, 1994)).
Then, the different behaviors of anodes in ECO process was described by considering two
limiting cases of electrodes, defined as ‘’active’’ and ‘’non-active’’ anodes in the
schematic model.
4.1.1 Active anodes
At active anodes, with low O2-overpotentials (such as IrO2, RuO2, and Pt) (Martínez-
Huitle and Brillas, 2009), a strong interaction occurs between the anode surface (MOx)
and •OH radicals (i.e., chemisorbed active oxygen in the lattice of metal oxide). The
chemisorbed •OH radicals may interact with anode, forming so-called higher oxide
(equation (8)). The surface redox couple MOx+1/MOx can act as a mediator in the
conversion of selective oxidation of organics (R with m carbon atom and without
heteroatoms) into short-chain carboxylic acids (equation (9)). Moreover, O2 evolution as
side reaction due to chemical decomposition of higher oxides is also involved in this
reaction (equation (10))(Martínez-Huitle and Ferro, 2006; Panizza and Cerisola, 2009) :
MOx (•OH) → MOx+1 + H++ e- (8)
MOx+1 + R → MOx + RO (9)

MOx → MOx + (1/2) O2 (10)
4.1.2 Non-active anodes
On the other hand, at non-active anodes, with relatively high O2-overpotentials (such as
BDD, PbO2, and SnO2 (Brillas and Martínez-Huitle, 2015; Nidheesh et al., 2018)), weak
interaction exists between physisorbed •OH radicals and the electrode surface. These
anodes do not contribute to the direct anodic oxidation of organic pollutants and do not
promote any catalytic sites for pollutant adsorption (Nidheesh et al., 2018). Therefore, the
formation of higher oxide is excluded and allows the direct oxidation of organics with
adsorbed •OH radicals, which may result in complete combustion to CO2 as follows
(equation (11)):
𝛼MOx (•OH) + R → 𝛼MOx + mCO2 +nH2O + zH++ ze- (11)
where, organic compound R needs 𝛼 = (2m+n) oxygen atoms for complete mineralization
to CO2. The oxidative reaction (9) is much more selective than mineralization reaction
(6). Like in active anode behavior, reaction (11) also undergoes competitive side reaction,
i.e. O2 evolution reaction, resulting in decreased anodic oxidation performance, either
direct oxidation to O2 (equation (12)) or indirect consumption via dimerization to
hydrogen peroxide (equation (13)).
MOx (•OH) → MOx + (1/2) O2 + H++ e- (12)
2 MOx (•OH) → 2 MOx + H2O2 (13)
According to the mechanisms, anodes with low O2-overpotentials (active anodes) (good
catalysts for O2 evolution) allow only partial oxidation of organics, whereas anodes with
high O2-overpotentials (non-active) (poor catalysts for O2 evolution) favor complete
oxidation of organics to CO2 (Panizza and Cerisola, 2009). Therefore, non-active anodes
are ideal for wastewater purification. Nevertheless, since both organic oxidation and O2
evolution reactions compete parallel, many anodes exhibit diverse behavior in practices
(Nidheesh et al., 2018).
It is well confirmed that the type of anode material is the most important factor for
determining the extent of organic pollutants degradation in ECO processes (Comninellis
and Vercesi, 1991). There are wide variety of anode materials including doped and
undoped PbO2, mixed metal oxides of Ti, Ru, Ir, Sn and Sb, Ti/Pt, carbon based anodes ,
and boron-doped diamond (BDD) thin films anodes (Brillas and Martínez-Huitle, 2015).
The PbO2 and BDD anodes, with high O2-overpotential, are the most commonly used
anodes for the electrochemical degradation of organic pollutants. The PbO2 anodes are
inexpensive, easy to prepare, chemically stable, low electrical stability, and a large
surface area (Martínez-Huitle and Brillas, 2009). Likewise, BDD anodes are made by
depositing a thin diamond films on non-diamond materials, usually silicon, tungsten,
43
titanium, tantalum, or glassy carbon, by energy-assisted chemical vapor deposition. BDD
anodes are very promising anode material for the electrochemical treatment of wastewater
contaminated with organic pollutants. They exhibit a great chemical and electrochemical
stability, a wide electrochemical working range ( HER at -1.25 V vs SHE and OER at +
2.3 V vs SHE), high inert surface with low adsorption, and a great mineralization
efficiency compared to other anodes (Oturan and Aaron, 2014; Panizza and Cerisola,
2009). The behavior of PbO2 and BDD anodes during the electrooxidation of organic
pollutant is well confirmed by Panizza and Cerisola (Panizza and Cerisola, 2009).
Nevertheless, despite of the high O2-overpotential, PbO2 anodes show low durability by
surface corrosion and leaching of highly toxic Pb2+ ions into to treated water (Brillas and
Martínez-Huitle, 2015). Similarly, the major drawbacks of BDD anodes includes high
material cost and difficulties in finding an appropriate substrate for thin diamond layer
deposition (Panizza and Cerisola, 2009), which tend to compromise their applicability in
large-scale applications.
Alternatively, MMO electrodes include a broad range of materials able to adsorb oxygen
on their structure and more suitable for anodic oxidation of organic pollutants.
Dimensionally stable anodes (DSA), are the commercial terms for MMO electrodes, due
to their excellent ECO properties and structural integrity (Shestakova, 2016). DSAs
usually constitutes a Ti substrates deposited by a thin semiconducting layer of metal oxide
or mixed metal oxides, such as Ti/TiO2-RuO2, Ti/Ta2O5-IrO2, Ti/IrO2-RuO2, Ti/SnO2-
Sb2O5, and Ti/Ir0.45O2-Ta2O2 (Brillas and Martínez-Huitle, 2015; Shestakova, 2016).
4.2 Photocatalytic oxidation (PCO)

Among other AOPs, heterogeneous photocatalysis is also a well-proven, efficient, and
eco-friendly method for attenuating ambient concentrations of organic contaminants in
aqueous or gas (air) phase (Guillard et al., 1999). Heterogeneous photocatalysis ascribes
to the speeding of photoreaction in the presence of semiconductor catalyst and the light
source. The major advantage of heterogeneous catalysis is the partial photocatalytic
oxidation at beginning or complete mineralization of organic pollutants essentially to
benign substances (Gaya and Abdullah, 2008). In early 70’s, Fujishima and Honda first
unfolded the possibility of water splitting using titania (TiO2) photocatalyst in a photo-
electrochemical solar cell, referred to as the ‘‘Honda-Fujishima effect’’(Fujishima and
Honda, 1972). Subsequently, the heterogeneous photocatalysis by titania photocatalyst
became environmental frontiers (Gaya and Abdullah, 2008; Oturan and Aaron, 2014).
In heterogeneous photocatalysis process, organic pollutants are decontaminated in the

presence of semiconductor photocatalysts, a light energy source, and reactive oxidizing
species. When the irradiated photon energy is greater than the band gap energy (∆E) of
the semiconductor, electrons are excited to the conductance band (ecb-) and positive holes
is generated on the valence band ( hvb+) site vacated by electrons and electron-hole pairs
(ecb-/ hvb+) are formed (Joo et al., 2019). This chemistry of photo-generated electron-hole
pairs could form reactive oxidants such as •OH radicals, and superoxide (•O2-) radicals
and adsorbed onto the semiconductor surface, and consequently organic pollutants are
then non-selectively oxidized by both photo-generated holes and reactive oxidant species
(Dong et al., 2015). However, the absorption of photons with lower energy than ∆E or
shorter wavelengths causes energy loss in the forms of heat due to rapid recombination
of charge carriers i.e., ecb-/ hvb+(Ahmed et al., 2011; Cavalcante et al., 2015). On this note,
TiO2 (∆E = 3.2 eV) assisted heterogeneous photocatalysis has been widely applied in
recent years for photocatalytic oxidation of refractory organic pollutants present in
wastewater. TiO2 is easy to produce, inexpensive, chemically and biologically stable, and
has an energy gap comparable to that of solar photons (Oturan and Aaron, 2014). The
activation of TiO2 by UV light irradiation and the photocatalytic degradation mechanism
of organic pollutant can be represented by the following equations (14)-(17) (Ahmed et
al., 2011):
ℎ𝑣
TiO2 → e- + h+ (14)
e- + O2 → •O2- (15)
h+ + H2O → •OH + H+ (16)
h+ / •OH / •O2- + R → Intermediates→ CO2 + H2O+ inorganic ions (17)
In heterogeneous photocatalysis, TiO2 catalyst can be utilized either as aqueous

suspension or in thin film form (Murgolo et al., 2017). The dispersed TiO2 form is easy
to use, provide great surface area, and aerated suspensions prevents the recombining of
ecb-/ hvb+ to some extent, despite progressive formation of dark catalytic sludge could
diminishes the efficiency of UV intensity. Contrarily, TiO2 thin films do not need to
separate the catalytic particles for reusability, but the catalytic layer should be very stable
and active (Oturan and Aaron, 2014). There are a number of studies performed on
heterogeneous UV/TiO2 photocatalysis of many emerging contaminants (Benotti et al.,
2009; De la Cruz et al., 2013; Doll and Frimmel, 2004; Ghaly et al., 2011; Hapeshi et al.,
2010; Kanakaraju et al., 2014; Rizzo et al., 2009; Xue et al., 2011).
However, like other AOPs, major drawbacks associated with TiO2-based heterogeneous
photocatalysis, includes low quantum efficiency, low adsorption capacity for
hydrophobic pollutants, post- recovery TiO2 particles, and catalysts deactivation (Ahmed
et al., 2011). To overcome these shortcomings, various countermeasures have been
studied in order to enhance photocatalytic efficiency, complete destruction of organic
pollutants, improve light adsorption range, increase stability and reproducibility, and to
improve recycle and reuse capabilities of TiO2 catalyst (Dong et al., 2015). Over the past
few decades, transition metals (e.g. Fe, Ni, Cr, Ag, Au) or their oxides doped TiO2
photocatalysts have been studied for enhancing the removal efficiency of various organic
pollutants (Anpo, 2000; Ren et al., 2015; Khan et al., 2016; Chen et al., 2015; Zhou et al.,
45
2010). TiO2 nanoparticles are modified by simply substitution reaction or intrinsic doping
with different metal cations, which can then form metal oxides or a mixture of oxides.
The transition metals doping improves trapping of excited electrons that inhibits rapid
recombination of ecb-/ hvb+. Therefore, metal nanoparticles act as an electron traps in
storing and shuttling photogenerated electrons from the TiO2 surface (Dong et al., 2015).
In that context, Ag2O with narrow energy band gap of 1.2 eV (Xu and Schoonen, 2000),
which is a proven photocatalyst for organic pollutant degradation, can be doped on TiO2
nanoparticles for increasing surface area, well-defined crystal structure and homogeneity
in particles size for enhanced photocatalytic effect (Wang et al., 2011). Enhanced
efficiency of dye removal via the heterogeneous photocatalysis using Ag2O doped TiO2
photocatalysts has been confirmed in many previous studies (Ren and Yang, 2017; Zhou
et al., 2010).
47
5 Integration potentials of MBR with AOPs
MBRs are established as environmental-friendly approaches for municipal wastewater
treatment. However, the removal efficiency of many biologically persistent and
hydrophilic ECs with biological treatment alone is not effective (Fatta-Kassinos et al.,
2016). Moreover, although some MBRs can treat wastewater to meet the current
guidelines and producing water for basic industrial reuses, the treated effluent still needs
to be further decontaminated for high grade applications. Generally, single classical
treatment techniques have been inadequate for the removal of persistent ECs, therefore it
is necessary to integrate MBR with other green advanced treatment processes, such as
AOPs, high retention membranes, MBBR, and granulation technology. Since usually
MBR effluent is of high quality (i.e., low level of suspended solids and turbidity), MBRs
offer great technical flexibility to be integrated with other technologies, and significant
synergy can be achieved.
Of the many integration potentials, coupling MBR process with different AOPs have
achieved a great deal of attention over the last decade. The ‘multi barrier approach’
integration of MBR with AOPs aims to enhance the quality of effluent, mitigate the
fouling and to robust the process stability (Neoh et al., 2016). As a multi barrier approach,
it is evidenced that the integrated approach can offer the recycling of oxidation products
back to the biological system for further biodegradation (Laera et al., 2011). The general
integration strategy of wastewater treatment using biological treatment and AOP can be
illustrated by Figure 5. If the wastewater stream contains higher ECs load, then integration
of AOP is beneficial as ‘pre-treatment’ of ECs, which converts recalcitrant ECs to more
biodegradables and subsequently treated in MBR with high biodegradation efficiency.
Whereas, integration of AOP as polishing step to MBR process could further improve the
removal of residual fractions of ECs after MBR treatment. The integration approach not
only benefits the synergy in the treatment performance, but also trade-off the
shortcomings/ challenges of one another (Ganiyu et al., 2015).
48 5 Integration potentials of MBR with AOPs
Figure 5: Decision tree for potential integration approaches of MBR

and AOP (Modified from (López et al., 2010)).
García-Gómez et al. showed an enhanced removal of carbamazepine in synthetic

wastewater from about 6.5 % to 99.99% (complete removal) when MBR system was
coupled with ECO process using Ti/PbO2 anode (García-Gómez et al., 2016). While
MBR was integrated with ozonation in the recirculation stream of the MBR effluent,
residual concentration of anti-viral drug (acyclovir) in the MBR effluent was further
removed by 99%. The major benefit of this integrated system was evident for the removal
of specific ozonation transformation products of acyclovir, which was about 20-fold
higher than MBR alone (Mascolo et al., 2010). Likewise, Pollice et al. reported the
complete degradation of antibacterial (nalidixic acid) when treated with integrated MBR-
ozonation process (Pollice et al., 2012). López et al. demonstrated that the integration of
49
MBR with solar photo-Fenton oxidation process could effectively treat wastewater
polluted with pesticides, by enhancing biodegrability and lessening ecotoxicity of the
treated effluent (López et al., 2010). A laboratory scale MBR system combined with
nanofiltration or reverse osmosis membrane filtration by Alturki et al. reported more than
95% removal of 40 trace organic compounds (Alturki et al., 2010). More than 95% of
carbamazepine removal efficiency was observed by Laera et al. when MBR was
integrated with a lab-scale heterogeneous TiO2-based annular slurry photoreactor (Laera
et al., 2011).
The literature survey indicates that there have been very few studies acquired on the
possibility of integrating MBR with AOPs to enhance the removal of ECs. Other than
AOPs, physical treatment processes (e.g, nanofiltration and reverse osmosis) could
possibly enhance the removal of persistent ECs, but ultimately ends-up in the
concentrates or sludge, which may need to be treated further. Moreover, most of the NF
or RO processes attenuate ECs based on their respective MWCO. On the other hand,
AOPs such as ECO and PCO are eco-friendly processes that non-selectively
degrade/combust highly persistent ECs via hydroxylation or dihydroxylation till their
total mineralization or until they converted into non-toxic and biodegradable short-
chained molecules (Ganiyu et al., 2015).
50 5 Integration potentials of MBR with AOPs
51
6 Materials and methods
6.2 Assessing the performance of MBR unit in different operating
conditions
6.1.1 The Pilot-scale MBR set-up
A pilot scale submerged MBR unit (ARTAS Ltd. Turkey) was installed at full-scale
Kenkäveronniemi WWTP, Mikkeli, Finland as depicted in Figure 6. The full-scale plant
treats about 5 Million cubic meter of wastewater per year, majorly of municipal sewage
with the influent load of about 63000 population equivalent. The treated wastewater is
discharged into the Saimaa Lake, which is the largest in Finland and fourth largest
freshwater lake in Europe.
Figure 6: Schematic diagram of the pilot MBR unit at Kenkäveronniemi WWTP,

Mikkeli, Finland.
The submerged flat-sheet membrane units (0.4 µm pore size), made of chlorinated
polyethylene polymers, with total surface area of 16 m2 (KUBOTA corporation, Japan),
were used for the solid-liquid separation. The in-situ membrane fouling was controlled
physically by providing crossflow scouring air via compressed air and adopting
intermittent suction cycle for filtration. The clean-in-place (CIP) tank was connected with
sodium hypochlorite (NaOCl) and Citric acid dosing tanks for the cleaning of membranes
during fouling occasions. The pilot MBR was operated under different operating
conditions depending on the purpose of studies (Paper I, II, III and IV) as described in
Table 2.
52 6 Materials and methods
Table 3. Different operating conditions of the MBR pilot plant depending on
purpose of studies.
Values
Parameter Units Paper
Paper I Paper II Paper III
IV
HRT Hours 35 42.13 40-45 40-45
SRT Days 25 - 30 60 and 21 55-60 55-60
Avg. Flux L m-2 h-1 7.80 5.84 4-6 4-6
MLSS 5300 - 8550 and 5000-
mg L-1 5000-8000
concentration 9800 3748 8000
kg COD
(kg 0.02 – 0.027 and 0.02 – 0.02 –
F/M ratio
MLSS. 0.05 0.09 0.09 0.09
d)-1
Aeration
0.2 -
intensity m3 m-2 h-1 0.4 – 0.6 0.2 0.2 -0.23
0.23
(SADm)
Process
°C 6.5 - 21 19 ± 2 15-22 15-22
temperature
pH Unitless 6.6 – 7.3 6-7.4 6-8 6-8
9-On/1- 9-On/1-Off 9-On/1- 9-On/1-
Suction cycle Minutes
Off Off Off
6.2 Assessing integration approaches to MBR

For further polishing the treated effluent from the MBR treatment, two widely used AOPs,
i.e., ECO and PCO processes were assessed for the removal of highly recalcitrant ECs,
including carbamazepine and diclofenac, as model compounds. These AOP units were
performed as ‘post-treatment’ to the residual concentrations of ECs from the MBR treated
effluent.
6.2.1 Lab-scale set-up for batch ECO unit
The lab-scale set-up used for the electrochemical oxidation system is as shown in Figure
7. The experiment was carried out in a glass reactor with cooling jacket arrangement. The
external electrical power was applied by using DC power supply. The inter-electrode gap
of 10 mm was maintained between both anode and cathode electrodes. The aqueous
solutions containing carbamazepine were electrolyzed under different operating
conditions and its attenuation in the concentration was analyzed using UV-vis
53
spectrophotometer. The newly developed Ti /Ta2O5-SnO2 electrodes were synthesized by
thermal decomposition method using precursor solution of tantalum (Ta) and tin (Sn)
salts on titanium (Ti) plates. The as-prepared electrodes were then characterized for their
microstructure, chemical analysis, and cyclic voltammetry (CV).
Figure 7: Experimental set-up for electrochemical oxidation.
6.2.2 Lab-scale set-up for PCO process
The photocatalytic experimental set-up using a self-made photocatalytic reactor is shown

in Figure 8. The photocatalytic degradation of ECs (carbamazepine and diclofenac) in
aqueous solutions (both in deionized water and real MBR effluent) was performed using
5% (w/w) Ag2O/P-25 photocatalysts under UV irradiation. The sample aliquots of ECs
were analyzed by using UV-vis spectrophotometer. The 5% (w/w) Ag2O/P-25
photocatalysts were prepared by a simple pH-mediated chemical precipitation method
and characterized for microstructure, morphology, elemental analysis, surface area, and
energy band gap using different characterization instruments.
Figure 8: Experimental set-up for photocatalytic oxidation
6.2.3 Analytical procedures
The common pollutants in wastewater, such as TSS, COD, NH4-N and TP contained in
influent and effluent of the MBR process, as well as MLSS and MLVSS concentrations
were analyzed according to the Standard Methods (APHA, 1999) (Paper I-IV).
For human enteric viruses, including Adv and NoV (GI and GII) from liquid samples,
pre-treated 500 mL samples (influent and effluent) were concentrated by two-phase
separation method and viral nuclei acids were extracted using the High Pure Viral Nucleic
Acid and High Pure Viral RNA Kit. For sludge samples, 500 mL of sample was
centrifuged to 10 min at 1000g at 4°C. The supernatant was concentrated by two-phase
separation process. For the solid fractions, 50 g of sludge pellet was mixed with 50 mL
of 3% beef extract (pH 9.5) for 30 min at reverse transcriptase (RT) to elute viruses. After
centrifugation for 20 min at 5000g at 4°C, the supernatant was recovered and adjusted to
a pH 7.2. The supernatant samples were then mixed with polyethylene glycol 6000 (PEG
6000) and NaCl, and further rocked overnight at 4°C. After centrifugation for 30 min at
10000g, supernatant was discarded, and the PEG pellet was collected for nucleic acid
extraction (U.S. EPA, 2003). The detection of Advs was carried out by performing the
55
real-time RT-PCR assays in a Rotor-GeneTM 3000 real-time rotary analyzer, whereas
the real-time RT-PCR assays were carried out in one step for the detection of both NoVs
using the QuantiTect Probe RT-PCR Kit. The possible effect of inhibitors in wastewater
samples were assumed to be constant throughout the experimental period (Paper I). On
the other hand, total number of bacteria indicators, such as E.coli and Enterococcus in the
influent, sludge and permeate were enumerated using Colilert ® and Enterolert ®
methods (Paper I).
For heavy metals analysis, liquid influent and effluent samples were acidified (pH<2)
with HNO3 acid, centrifuged at 4000 rpm for 20 min, and then supernatant was filtered
with 0.20 µm cellulose syringe filters. The sludge samples were dried in oven at 105°C.
About 0.5 g of dried sludge in 10 mL HNO3: HCl (1:3) mixture was digested in a
microwave digester. Finally, heavy metals were analyzed using inductively coupled
plasma-optical emission spectrometer (ICP-OES, iCAP 6300, Thermo Electron
Corporation, USA) from the pretreated samples (Fontmorin and Sillanpää, 2015) (Paper
I & III).
The occurrence of diverse ECs from aqueous samples (influent and effluent) and solid
samples (sludge) of the MBR treatment was carried out by standard solid phase extraction
(SPE) followed by LC-MS/MS analysis (U.S. EPA, 2007) (Paper I & II). For SPE,
typically, about 100 mL of effluent or 50 mL of influent or sludge samples were spiked
with the mass-labelled surrogates, for example 13C-carbamazepine, 13C-bisphenol A,
2H-diclofenac, and so on. Additionally, mixing, ultrasonic extraction and centrifugation
were repeatedly employed for sludge samples. The spiked samples were extracted,
conditioned with the mixture of 20 mL methanol and 20 mL deionized water, washed
with 10 mL of 5% methanol and eluted with 15 mL of methanol. The mixture was then
evaporated to dryness under nitrogen flow and 0.5 mL of methanol and 0.5 mL of
deionized water were added. The aliquots were then analyzed using LC-MS/MS (Waters
Acquity UPLC and Xevo TQ, USA). The recovery correction during the calculation was
made with respect to the mass-labelled surrogates. The measurement uncertainty (MU)
was about 40%. The MU % was calculated according to the Nordest TR 537 that covered
every step in the laboratory analysis. Moreover, the residual concentrations of ECs, such
as carbamazepine and diclofenac were analyzed by UV-vis spectrophotometer (Lambda
45, USA) (Paper III & IV). The photocatalytically induced intermediates of
carbamazepine were identified and quantified by using Fourier-transform ion cyclotron
resonance (FT-ICR) mass spectrometry (Bruker Daltonics, Germany) (Paper IV).
The fouling precursors, such as EPS and SMP from sludge were determined by
centrifuging 25 mL of sludge samples at 4000 rpm at 4 °C for 30 min. The supernatant
was filtered through a 0.45 μm glass filter and measured for SMP. The obtained pallets
were resuspended by adding 0.05% NaCl, rapidly mixed with vortex mixture and the
solution was kept in a water bath (~100 °C) for heat treatment for 60 min. Finally, the
supernatant was filtered through a 0.45 μm glass filter and measured for total EPS. Both
SMP and EPS were analyzed using a TOC analyzer (TOC-V Series-CPN, Shimadzu,
Japan) and expressed as mg of DOC g-1 MLVSS (Paper II).
57
7 Results and discussion
7.1 Assessing the performance of MBR at Nordic cold conditions (Paper
I)
In order to investigate the performance of MBR at Nordic cold conditions, the MBR pilot
unit was operated for more than 120 days in winter season of Finland (January to April
2016). This is the first comprehensive study on the performance of pilot MBR on treating
the real municipal wastewater in the unique Nordic environment. The wastewater
temperature inside the bioreactor varied from 7 to 20 °C. On this occasion, the effect of
low sludge temperatures on membrane fouling was evaluated by monitoring TMP and
deterioration of sludge settleability was measured by monitoring SVI. Moreover, the
performance of MBR on removing organic, solids and nutrients loads in the incoming
wastewater, reduction of common enteric viruses, and most importantly the attenuation
of ECs, such as PhACs, EDCs and some toxic heavy metals, were assessed. The
membrane fouling tendency under cold water conditions and adopted cleaning methods
and frequencies are depicted in Figure 9.
While operating MBR in cold-water conditions, many consequences related to the process
operation were observed. During the experimental period, the ambient air temperature
was recorded, which was fluctuated between 0 and -33 °C for more than a week, which
is very typical in winter seasons of Nordic regions. With the decreasing sludge
temperature inside the bioreactor, increasing trend of TMP profile was observed (Figure
9). Two distinct stages of membrane fouling, which is well known tendency in MBRs
(Park et al., 2015), such as slow and steady fouling and rapid (jump) fouling, were clearly
observed. While the sludge temperature was recorded below 10 °C, two consecutive
chemical cleanings for fouled membranes were adapted to reduce very frequent jumping
of TMPs within a week. However, the cleaning cycle was extended to about two weeks
when sludge temperature raised more than 10 °C. It was also noticed that other than cold
water temperature, changing in the specific aeration demand per membrane surface area
(SADm) could effect on membrane fouling behavior to some extent. The permeability of
the membranes was dropped from about 345 L m-2 bar-1 to nearly 50 L m-2 bar-1 while
temperature decreased from 20.5 ± 0.6 °C to below 10°C. About 75% of permeability
drop was observed during the lowest sludge temperature period. After the final manual
cleaning, the permeability was restored gradually and reached nearly to 300 L m-2 bar-1.
The results indicated that the cold-water conditions can deteriorate the hydraulic
performance of MBR due to accelerated membrane fouling propensities. The sludge
settleability was also highly deteriorated, as indicated by the SVI values being increased
from 120 to 168 mL g-1, during the coldest temperature events observed inside the
bioreactor. On the other hand, despite of low temperature operation, relatively high
quality of treated effluent was observed with the corresponding removal efficiencies of
100%, 92.8%, >99.8%, and 90.6% for TSS, COD, NH4-N, and TP respectively.
58 7 Results and discussion
Figure 9: Membrane fouling behavior of MBR system under cold weather

conditions and adopted cleaning measures.
Regarding the fate of human enteric viruses, such as Noroviruses (NoV GI and NoV GII)
and adenovirus (AdVs), all three viruses were constantly detected in the influent samples
during the experimental period as shown in Figure 10. NoV GI was detected in 85% of
the influent samples with an average concentration of 95 GC mL−1, whereas NoV GII and
AdVs were positively detected in 100 % with corresponding average concentrations of
896 and 124 GC mL−1, respectively. The average log removal of 1.82, 3.02 and 1.94 log
units were achieved during MBR treatment for NoV GI, NoV GII and AdVs, respectively.
It was further confirmed that the virus removal efficiency was not correlated with the
degree of membrane fouling. Moreover, none of the concentration of Escherichia coli
and enterococci were detected in the MBR permeate over the experimental period,
showing excellent retaining of these bacteria.
59
Figure 10: Occurrences of viruses in wastewater stream under cold

water conditions.
Similarly, the occurrences and removal efficiencies of four ECs, including PhACs, such
as carbamazepine, bisoprolol, diclofenac, and EDC, i.e., bisphenol A, in the MBR
treatment under cold weather conditions were investigated. All the selected ECs were
frequently identified in the influent wastewater samples, but the relatively higher
concentrations were observed from January up to the end of March, ascribing the high
specific consumption of these ECs during cold season. The mean influent concentrations
of carbamazepine, bisoprolol, diclofenac, and bisphenol A were 0.63, 1.84, 0.43 and 0.95
μg L−1, respectively, where diclofenac was highly detected amongst other refractory
compounds in cold season. In this study, the removal efficiency of individual ECs varied
greatly under MBR treatment. The average removal efficiencies of bisoprolol and
diclofenac were 65 ± 6% and 38 ± 7%, respectively (Figure 11). A decreasing trend in
the removal rates of bisoprolol and diclofenac was observed from Jan-Feb to April, which
might be partially due to the very low sludge temperature inside the bioreactor that might
have affected the biodegradability of the ECs. Despite the low temperature MBR
operation, the bisphenol A removal efficiency was substantially stable (> 97%)
throughout the experimental period, which might be due to its high hydrophobicity that
enhanced its adsorption to biomass and consequently biotransformed. On the other hand,
very poor removal efficiency of carbamazepine ranging from -89 to 28% (only for one
occasion) was observed with highest negative removal efficiency during January to
February (Figure 11).
Figure 11: Removal efficiencies of selected ECs in MBR treatment under

low-temperature periods.
On the other hand, inorganic micropollutants i.e. heavy metals including Zn, Ni, Co, Cu,
As, Pb, Cd, and V, were also consistently detected in influent, permeate and sludge
samples over the experimental period. The average concentration of trace metals in the
incoming wastewater was found in the range of 0.1 to 80 µg L-1, with their relative
abundance in the order of Zn> Ni > Co ≥ Cu >V > Cr > As ≥ Pb > Cd. However, the
removal efficiencies of the trace metals greatly varied from element to element in the
MBR treatment. An average removal efficiency of ≥ 80% was achieved for Cd, Pb, and
V, while ≥ 60% was observed for Zn and Co. Whereas, comparatively lower removal
efficiencies (30 - 50%) were observed for Ni, Cu, As, and Cr under the identical operating
conditions. In this study, chemical precipitation and biosorption mechanisms were
expected to be dominant for the removal of trace metals in the MBR process. Moreover,
metal complexation via metal-ligand co-ordinate bonding and the biofilm layer developed
on the membrane surface could have contributed to the retention/adsorption of insoluble
metal species to some extent other than solubility and stable or unstable forms of trace
metals in the MBR treatment (Bolzonella et al., 2010).
61
Since trace metals are non-biodegradables, their incoming load in MBR process can either
ends up into the sludge or remains in the treated effluent. Therefore, it is crucial to assess
the concentration of trace metals in both treated effluent and produced sludge, especially
when they have some specific utilization. In this study, relatively abundant concentrations
of trace metals were observed in MBR effluent and sludge samples, although their
concentrations were observed below the allowable concentrations limit specified in the
EU and US regulations.
7.2 Assessing the performance of MBR under varying solid retention

times- fate and removal of ECs (Paper II)
Two different SRTs: 60 days and 21 days, were selected to mainly study the removal
efficiency of 23 ECs, including PhACs (antibiotics, β-blockers, analgesics, diuretics,
psychostimulants, antiepileptics, immunosuppressives, anticoagulants) and steroid
hormones (testosterone, progesterone, estrone, and estroil), commonly detected in
municipal wastewater. Overall, a consistently high removal efficiencies of conventional
wastewater pollutants such as TSS, COD, NH4-N, and TP were observed at both SRTs.
Majority of ECs detected in the influent wastewater during the testing SRT periods of 60
and 21 days revealed similar tendencies. Of the 23 ECs, compounds such as iopamidol,
furosemide, caffeine, hydrochlorothiazide, paracetamol, naproxen, ibuprofen and
diclofenac were detected in high levels, thus accounted for more than 50% of the total
ECs. On the other hand, steroid hormones were observed at relatively low levels as
compared to other PhACs.
The removal of selected ECs varied greatly during the MBR treatment process under the
maintained operating conditions. However, the concentrations of many ECs in the treated
effluent were below the limit of quantification (LOQ), which indicates their efficient
removal under MBR treatment regardless of SRT variations. In this study, 17 ECs,
including 13 PhACs (iopamidol, atenolol, tetracycline, caffeine, bisoprolol,
ciprofloxacin, enalapril, ketoprofen, trimethoprim, paracetamol, naproxen, ibuprofen,
and hydrocortisone) and four steroid hormones (estriol, testosterone, estrone,
progesterone) showed the median removal efficiency of more than 90% (90 to ≥ 99.7%).
Three PhACs including metoprolol, furosemide and propranolol showed median removal
efficiencies in the range of 50 – 87%, diclofenac showed poor removal of 39 – 46%, and
two PhACs (carbamazepine and hydrochlorothiazide) were very poorly removed (-8 to
10%). The fate of anti-coagulant (iopamidol) and immunosuppressive (hydrocortisone)
was studied for the first time in the MBR treatment. Out of four β-blockers, the removal
rates of metoprolol and propranolol were slightly higher at longer SRT of 60 days (82 -
84%) than in shorter SRT of 21 days (50 - 60%). Even though the attenuation of
hydrochlorothiazide was very low, its median removal efficiency improved from -1 % at
SRT of 21 days to 10% at SRT of 60 days. Similarly, the median removal efficiency of
carbamazepine enhanced to 2% from - 8% while SRT varied from 21 to 60 days. In most
of the occasions, concentrations of hydrochlorothiazide and carbamazepine in the treated
effluent were higher than in the influent, therefore showing negative removal efficiency.
The major reason for the phenomenon of ‘negative removal’ was ascribed to the
transformation of metabolites, such as glucuronide-conjugates back to the parent
compounds during the biological process, and eventually released with higher
concentrations in the treated effluent (Jelic et al., 2011).
Furthermore, to understand the pathway of ECs removal, mass-balance study was carried
out as described by equations (2)-(5). The shared proportions of ECs relative to the
influent load based on biotransformation, sorption onto sludge and remained in the treated
effluent are elucidated in Figure 12. The dominant mechanism for ECs removal was
presumably via biosorption and biotransformation. The ECs, including atenolol, caffeine,
paracetamol, naproxen, ibuprofen, esteriol and progesterone were removed significantly
with biotransformation rate of 83 to ≥ 99.9%. High biotransformation rate of 58 - 92%
was assigned for iopamidol, furosemide and trimethoprim. Slightly enhanced
biotransformation of metoprolol, hydrochlorothiazide and propranolol was observed at
SRT of 60 days than at SRT of 21 days. On the other hand, some of the ECs were found
to have been removed via sorption onto the biomass (sludge). More than 90% removal
rate was accounted for ciprofloxacin, which could pose serious risk if the excess sludge
has multiple utilization. Some other ECs, such as propranolol, carbamazepine and
testosterone, showed about 10 - 50% sorption efficiency, whereas most of the ECs posed
a very weak sorption affinity to the biomass with the sorption efficiency ranging from 0
- 21%. The physicochemical properties of individual ECs, such as hydrophobicity
(hydrophilic or hydrophobic, characterized by log D values), degree of ionization (pKa),
sorption coefficient (log Kd), could have additionally influenced their removal in MBR
treatment (Cirja et al., 2008b).
The calculated values of log Kd revealed that the selected ECs in MBR treatment were
removed either predominantly via sorption process or via simultaneous biodegradation
followed by partitioning onto sludge (Guerra et al., 2014). However, it was observed that
the calculated sorption coefficients were not correlated with the degree of
biotransformation for many ECs, which indicates that log Kd values of ECs may vary
plant-to-plant depending on different operating conditions (Kim et al., 2014).
Furthermore, noticeable changes in the concentrations of biopolymers such as SMP and
EPS were observed while changing the SRTs. The SMP concentration increased from 5.6
to 15.5 mg (g MLVSS)-1 (about 3 times) and EPS concentration increased from 51 to 103
mg (g MLVSS)-1 while SRT changed from 60 to 21 days, respectively. Similarly, it was
revealed that the dewaterability (measured in terms of CST) of the MBR sludge correlates
well with the SRT values. The sludge with high CST values possesses difficulty in
dewatering due to the presence of more cell bound water.
63
Figure 12: Fate of the selected ECs during MBR treatment at two different SRTs.
7.3 Assessing the AOPs as integration alternatives to MBR to enhance

the removal of ECs.
It was clear from the above studies that the removal of many ECs was not optimal
although they meet most of the recent limit standards. This indicates that the MBR
treatment itself is not the ultimate solution to remediate those highly recalcitrant ECs,
which advocates the need of highly efficient techniques to meet the future discharge
limits. In recent years, developing a more robust treatment processes to ensure high
quality reclaimed effluent from municipal wastewater, especially with the focus on the
removal of ECs has gained increasing interests. In this context, we studied the potential
integration options of two AOP based green technologies: ECO and PCO processes, to
further enhance the removal of ECs from MBR effluent.
7.3.1 Study of ECO process for carbamazepine removal using Ti/Ta2O5-SnO2
electrode (Paper III)
In this study, for the first time, the removal efficiency of one of the highly recalcitrant
PhACs, carbamazepine, in synthetic solutions and the real MBR effluent was assessed
via ECO process using newly developed Ti/Ta2O5-SnO2 electrode. The characterization
of Ti/Ta2O5-SnO2 electrode revealed that the electrocatalytic activity was enhanced due
to doped metallic impurities on the Ti substrate, roughness factor of 12.5 as compared to
smooth surface oxide, and high-overvoltage potentials (appeared between 1.7 and 2 V vs.
SCE) that endorsed the generation of powerful •OH radicals and subsequent oxidation of
carbamazepine. It was also observed that carbamazepine molecules oxidized at the
potential of 1.45 V vs. SCE with distinct anodic current peak. Moreover, the influence of
four major operating parameters, such as current density, initial concentration of
carbamazepine, initial pH of solution and reaction temperature, were also studied.
The current densities of 1, 5, 7, 9, 15 and 17 mA cm-2 were tested to electrolyze

carbamazepine (20 mg L-1) with supporting 0.1 M Na2SO4 electrolyte for 8 h (Figure
13a). With the increasing current density from 1 to 9 mA cm-2, carbamazepine removals
reached about 53 - 71.7%. However, beyond the current density of 9 mA cm-2, no
significant increase in either removal efficiency or kinetics was observed. The reasons
might be attributed to the parasitic reactions of •OH radicals to O2 gas evolution and mass
transfer rate of pollutants towards electrode surface (Song et al., 2010). Furthermore, the
electrical energy consumption to treat one cubic meter of water contaminated with
carbamazepine by electrochemical treatment for 8 h using Ti/Ta2O5-SnO2 was observed
in the range of 3.2 - 172.7 kWh m-3, when current densities varied from 1 - 17 mA cm-2.
The energy consumption was increased from 3.2 - 60.3 kWh m-3 to achieve the
carbamazepine removal efficiency of 53 - 71.7 % with corresponding current density
applied from 1 to 9 mA cm-2. Beyond the current density of 9 mA cm-2, it was noticed
that the energy consumption increased by a factor of more than 2 for achieving the
comparable degree of carbamazepine removal efficiency. Therefore, 9 mA cm-2 was
selected as an optimal current density due to comparably high degree of carbamazepine
removal efficiency and less energy consumption. The reaction kinetics fitted both pseudo-
first order and pseudo-second order model, but latter described better electro-degradation
of carbamazepine with slightly higher regression coefficients (R2). To study the effect of
initial concentration of carbamazepine on the electrochemical treatment, four different
batches of carbamazepine solutions, such as 2, 10, 20 and 30 mg L-1 were electrolyzed
under the optimized current density (9 mA cm-2) (Figure 13b). After the 8 h of
electrolysis, the removal efficiency reached about 80.3%, 73.3%, 71.8% and 70.2% at
initial concentrations of 2, 10, 20 and 30 mg L-1, respectively. The results indicated that
at low initial concentration of substrates, the electrochemical treatment is more efficient.
This might be due to the reason that at low initial concentrations of pollutants,
electrochemical oxidation of organic molecules can be faster than the diffusion of side-
products (Dai et al., 2014).
65
Similarly, to assess the effect of initial solution pH on the electrochemical treatment of
carbamazepine, acidic to alkaline conditions with the pH values of 2, 4, 8 and 10, were
studied (Figure 13c). The results showed that initial pH conditions do not have that
significant effect on the removal efficiency of carbamazepine under electrochemical
treatment. Nevertheless, optimal removal efficiency of 75.2% was achieved at pH 6,
which is very relevant since MBR effluent normally observed at the pH of 6 - 7. On the
other hand, effect of temperature was assessed by electrolyzing carbamazepine solution
at different temperature conditions ranging from 10 - 30°C (Figure 13d). The
carbamazepine removal efficiency of 71.7%, 74.1% and 75.5% were achieved at 10, 20
and 30°C, respectively under the electrochemical treatment for 8 h, indicating the slightly
enhanced trend of treatment efficiency with the increasing solution temperature.
Furthermore, under the optimal current density of 9 mA cm−2 and pH of 6.0, the TOC
removal of carbamazepine were 61.2%, 69.8% and 71.1% at solution temperature of 10
°C, 20 °C and 30 °C, respectively. Since the solution temperature during this work was
about 11 °C, the average degree of mineralization was expected about 61 %. In order to
study the effect of electrode type, the carbamazepine removal efficiency of as-prepared
Ti/Ta2O5-SnO2 and commercially available Ti/PbO2 (Magneto, Netherland) was
compared. After the 8 h of electrolysis, the carbamazepine removal efficiency at
Ti/Ta2O5-SnO2 and Ti/PbO2 anodes reached 71.7% and 77.9%, respectively, showing
relatively high removal efficiency of commercial electrode. However, a significant
leaching of PbO2 was observed when extending electrolysis to 16 h and 24 h, releasing
Pb concentrations of 37.3 and 25.8 μg L−1, respectively, which is higher than the
maximum allowable concentration of Pb (14 μg L−1) in inland and other surface water
(EU, 2013). On the other hand, leaching of Sn, Ti and Ta was observed in the range of
145.7 - 407.3 μg L−1, 53.6 - 92.9 μg L−1 and 2 - 14 μg L−1, respectively while electrolyzing
for 24 h. The high dissolution of Sn to bulk solution was followed respectively by Ti and
Ta, yet there are no specific regulations on these compounds in surface water sources.
Since the removal efficiencies of either electrodes are quite comparable, the as-prepared
Ti/Ta2O5-SnO2 is eco-friendly with no heavy metal leaching into the treated water.
Figure 13: Effect of operating conditions (a) applied current densities; (b) initial
carbamazepine concentrations; (c) initial solution pH; and (d) temperature on
the electrochemical oxidation of carbamazepine.
Furthermore, the optimal conditions (current density= 9 mA cm−2; pH= 6; T= 11 ± 1 °C;

at Ti/Ta2O5-SnO2 electrodes) for the removal of carbamazepine from synthetic solution
were applied to treat the real MBR effluent spiked with carbamazepine (10.75 μg L−1). A
complete carbamazepine removal efficiency of >99.9% was achieved after the 4h of
electrolysis, which required lowest energy consumption of about 57 kWh m-3. The results
revealed that the ECO process based on the use of newly developed Ti/Ta2O5-SnO2
semiconducting electrode is a reliable process to remediate carbamazepine from
contaminated water streams, with promising potential to be integrated with MBR
technology.
7.3.2 Study of PCO process for PhACs removal using Ag2O/P-25 photocatalyst
(Paper IV)
This study advocates the photocatalytic removal of commonly detected PhACs in

municipal wastewater, such as carbamazepine and diclofenac, using newly developed UV
light driven 5% Ag2O/P-25 photocatalyst for the first time ever. Two different water
67
matrices, including deionized water (DW) and real MBR effluent (RME) contaminated
with selected PhACs, were tested. For the preliminary screening experiments 0.4 g L-1 of
four different photocatalysts including Ag2O/P-25, Sn3O4/P-25, PdO/P-25 and NiO/P-25
were dispersed into 50 mg L-1 of carbamazepine and diclofenac and the extent of
adsorption under dark conditions, UV-photolysis and UV irradiated photocatalysis were
studied. In any case, the removal of carbamazepine and diclofenac via adsorption alone
was only about 2 – 9%. Similarly, UV-photolysis for 180 min showed the carbamazepine
and diclofenac degradation rates of nearly 13% and 25%, respectively. Whereas,
photocatalytic degradation under UV irradiation for 180 min using four as prepared
photocatalysts showed significant attenuation of carbamazepine and diclofenac. Since 5%
Ag2O/P-25 revealed a comparatively higher removal efficiency of the selected PhACs
than other photocatalysts, hence it was selected for the subsequent experiments.
The characterization of 5% Ag2O/P-25 by SEM and EDX showed the agglomeration and
homogenous dispersion of Ag+ nanoparticles on the surface of P-25. The FTIR analysis
showed the main absorption peaks at 3000, 1533.67, 1420, 1050 and 433.33 cm-1,
supporting the presence of chemical bonds (O-H stretching vibration, stretching of C=O
group or O-H bending and M-O stretching vibration, respectively) in the bulk
compositions (Figure 14a). The BET analysis indicated that the surface area of P-25
increased slightly by the deposition of Ag2O nanoparticles from 47.27 to 54.1 m2 g-1. The
energy band gap estimated based on the reflectance spectra of 5% Ag2O/P-25 was about
2.96 eV, which revealed the high possibility of harnessing absorption spectra in UV-vis
range (Figure 14b).
a) b)
Figure 14: FTIR spectrum (a) and plots of (F (R∞) h𝑣)1/2 vs (h𝑣) for the
approximation of the optical band gap (b) of 5% Ag2O/P-25.
In this study, operating parameters involved in the photocatalytic removal of selected
PhACs using 5% Ag2O/P-25, such as solution matrix, catalyst dosage, and initial PhACs
concentration were experimented as shown in Figure 15. The removal of carbamazepine
and diclofenac in two solution matrices (DW and RME) was assessed at the catalyst
dosage ranging from 0.2 to 1.0 g L-1. In DW matrix (Figure 15 a-d), carbamazepine and
diclofenac removal efficiencies reached 80.4 - 89.1% and 87.5 – 93.5% at catalyst dosage
ranging from 0.2 - 0.4 g L-1, respectively. Since the removal efficiency of both
carbamazepine and diclofenac did not improve significantly beyond the catalyst dose of
0.4 g L-1, so this value was considered as optimal dose for the photocatalytic removal of
both PhACs in DW matrix. For RME matrix, the removal efficiency of carbamazepine
reached 89.74% from 76.6% while increasing catalyst dosage from 0.4 - 0.8 g L-1, but no
significant improvement was observed beyond 0.8 g L-1. Whereas, optimal diclofenac
removal efficiency of 90.7% was achieved at catalyst dose of 0.6 g L-1. Therefore, the
optimal dosage of 5% Ag2O/P-25 in RME matrix to achieve comparable extent of
carbamazepine and diclofenac removal efficiency as in DW matrix, was increased nearly
by 2 and 1.5-fold, respectively. Furthermore, the effect of initial concentration on the
photocatalytic removal of PhACs by 5% Ag2O/P-25 was studied in RME matrix (Figure
15 e-f). The decreasing tendency in the removal efficiency of carbamazepine and
diclofenac was observed while gradually increasing their initial concentrations. The
removal efficiency of carbamazepine and diclofenac decreased about 38% and 23% while
initial concentrations were increased from 20 to 50 mg L-1. Moreover, the photocatalytic
degradation kinetics of PhACs at 20 min of initial reaction fitted the Langmuir-
Hinshelwood kinetics model very well with regression coefficients of the linear fitting ≥
98% for both carbamazepine and diclofenac. The mineralization efficiencies of about
68% and 65% were observed during the photocatalytic degradation of carbamazepine in
DW and RME matrices, respectively, whereas slightly lower mineralization efficiencies
of diclofenac i.e., 60% and 55% were accounted in DW and RME matrices, respectively.
69
Figure 15: Effect of catalyst dosage in different water matrices on photocatalytic

removal of (a & b) carbamazepine in DW and RME; (c & d) diclofenac in DW and
RME; and effect of initial PhACs concentration (e) carbamazepine in RME; (f)
diclofenac in RME.
The quenching experiments indicated that the addition of ethylenediaminetetraacetic acid
(EDTA) and methanol (for suppressing h+ and •OH), the photocatalytic degradation of
carbamazepine was substantially reduced nearly by 70% and 90% in 60 min, respectively,
whereas only about 18% reduction was observed with the O2-purging (suppressing •O2-).
On the other hand, diclofenac removal was substantially decreased to 62% in the presence
of EDTA, but no significant effect in removal efficiency (~ 7%) was observed due to
methanol and O2-purging. The results indicated that the induced h+ and •OH are the
dominant reactive oxidation species during the photocatalytic degradation of
carbamazepine, whereas h+ mainly attributed to diclofenac degradation.
Similarly, to assess the reusability of 5% Ag2O/P-25, cyclic photocatalytic experiments

were conducted for both the PhACs (Figure 16). The 5% Ag2O/P-25 did not show any
significant loss in the photocatalytic degradation efficiency of carbamazepine until five
repetitive cycles, exhibiting a remarkable reusability. Whereas, gradually decreased
photocatalytic removal efficiency (dropped by about 35% until 5th cycle) of diclofenac
was observed during cyclic reuse of 5% Ag2O/P-25. Nonetheless, since the inhibition rate
of the photocatalytic activity was almost stable after 2nd cycle, it can be concluded that
the as-prepared 5% Ag2O/P-25 is moderately stable for the attenuation of diclofenac and
could be useful for the long-term industrial application. Moreover, in either cases, the
recovery of photocatalyst was achieved by centrifuging the solution just for 15 min, which
is probably the more cost efficient than high filtration costs associated with TiO2
nanoparticles.
Figure 16. Cyclic reusability of 5% Ag2O/P-25 photocatalyst during

the photocatalytic oxidation of (a) carbamazepine and (b) diclofenac.
71
The intermediate identification study on the photocatalytic degradation of carbamazepine

(one of the selected PhACs) using high-resolution FT-ICR mass spectrometry detected
the common transformation products of carbamazepine, such as 10,11- dihydro-
carbamazepine -10,11-epoxide (m/z 253), acridine-9-carboxaldehyde (m/z 208), and
hydroxyacridine-9-carboxaldehyde (m/z 224). The identification of many intermediate
products indicated the complexities involved during the photocatalysis of carbamazepine.
Moreover, a possible multi-step carbamazepine photodegradation path was proposed
based on the unambiguously identified intermediate products in this study and suggested
in the other literatures as shown in Figure 17.
Figure 17: Photogenerated intermediates and proposed photodegradation pathway

of carbamazepine by 5% Ag2O/P-25 photocatalyst under UV irradiation.
73
8 Conclusions and further research
MBRs are very potential technology for municipal wastewater treatment and reclamation.
However, the depth of understandings on their performance and operational
consequences in the regions where low temperature conditions remain for longer period
of the year are very limited. For this aim, the pilot scale MBR plant was operated under
varying operating conditions to study the membrane fouling and efficiency of MBR on
removing ECs. It is noticed that severe membrane fouling is associated with MBR
operation during cold water conditions. However, MBR showed excellent removal of
ECs, heavy metals and human enteric viruses even under extreme cold water conditions
including other common pollutants. Diverse removal efficiencies of ECs were observed
during the MBR treatment and several influencing factors were evaluated.
Biotransformation and sorption to sludge were found to be the predominating reasons of
the removal of ECs, although metal complexation, retention by membrane fouling layer
and membrane material have also attributed on the attenuation of heavy metals and
viruses. Moreover, physicochemical properties of ECs and operational conditions had
also contributed on removal by MBR treatment. Overall, a very good quality of treated
effluent was achieved by MBR treatment despite of varying operating conditions, even
though the removal efficiencies of many ECs were not optimal.
This indicated that the MBR treatment is not the ultimate solution to remediate many of
highly recalcitrant ECs. Therefore, integration possibilities of MBR with some AOPs,
viz. ECO and PCO to ensure high quality reclaimed effluent with emphasizing the
removal efficiency of ECs from municipal wastewater treatment, were explored. In ECO,
a newly developed Ti/Ta2O5-SnO2 electrode was used for the electrochemical degradation
of carbamazepine. The optimal removal of carbamazepine in aqueous solution was
reached 75.5% when electrolyzed under the current density of 9 mA cm-2, pH of 6, 30 °C
for 8 hrs. A complete removal of carbamazepine in real MBR effluent was observed when
electrolyzed under the optimized conditions in aqueous solutions. On the other hand, In
PCO, heterogenous 5% Ag2O/P-25 photocatalyst was used for the photocatalytic
degradation of carbamazepine and diclofenac from aqueous solutions and real MBR
effluent matrices. About 89.1% and 93,5% removal of carbamazepine and diclofenac was
achieved using 5% Ag2O/P-25 at catalyst dosage of 0.4 g L-1 under UV irradiation for
180 min. However, to achieve the same degree of attenuation, the optimal catalyst
dosages for both carbamazepine and diclofenac were increased by the factor of 2 and 1.5,
respectively. The photo-induced holes and ●OH were the primary oxidation species
involved in the photocatalytic reactions. Moreover, some important degradation by-
products of carbamazepine were identified and quantified, and multi-step degradation
pathway was proposed. Therefore, the batch experiments on ECO and PCO systems
revealed the enhanced removal of selected ECs than in the MBR treatment alone, which
indicated their potentials of integrating with MBR technology to improve effluent quality
and the better stability of the treatment process.
74 8 Conclusions and further research
This dissertation has expanded the knowledge on the applicability and efficiency of MBR
operation at rather low water temperature conditions in Nordic environment and
integration possibilities of MBR with selected AOPs. However, it has also unfolded the
need of many other important topics for further investigations. Of the many research
topics related to the current work, some of them are briefly described below.
As substantial membrane fouling was observed in MBR during winter periods, more
research could be focused on finding fouling mitigation strategies. On this note, feasibility
of integrating MBR with electrocoagulation can be studied, which is expected to
simultaneously destroy major organic fouling fractions (e.g. EPS and SMP), enhance ECs
removal efficiency, and generate less sludge. Similarly, the ability of low-cost adsorbents,
such as biochars derived from wasted sludge, could be studied as subsequent fouling
mitigation approach. On the other side, scaling-up of integrated MBR-AOPs options
could be studied in a pilot-scale. Further, optimizing the operation of pilot-scale
integrated system, regeneration of electrodes and photocatalysts, and overall cost
evaluation of the system should be investigated. Similarly, toxicity associated with the
treated effluent after AOPs could be a challenge due to the generation of more toxic
transformation by-products. Therefore, further studies should be focused on detail
identification of the generated by-products of parent ECs and their relative toxicities in
the treated effluent.
75
References
Ahmed, S., Rasul, M.G., Martens, W.N., Brown, R., Hashib, M.A., 2011. Advances in
Heterogeneous Photocatalytic Degradation of Phenols and Dyes in Wastewater:
A Review. Water. Air. Soil Pollut. 215, 3–29. https://doi.org/10.1007/s11270-
010-0456-3
Alturki, A.A., Tadkaew, N., McDonald, J.A., Khan, S.J., Price, W.E., Nghiem, L.D.,
2010. Combining MBR and NF/RO membrane filtration for the removal of trace
organics in indirect potable water reuse applications. J. Membr. Sci. 365, 206–
215. https://doi.org/10.1016/j.memsci.2010.09.008
Anpo, M., 2000. Use of visible light. Second-generation titanium oxide photocatalysts
prepared by the application of an advanced metal ion-implantation method. Pure
Appl. Chem. 72, 1787–1792. https://doi.org/10.1351/pac200072091787
APHA, 1999. Standard Methods for the Examination of Water and Wastewater. Stand.
Methods 541–541.
Arévalo, J., Ruiz, L.M., Pérez, J., Gómez, M.A., 2014. Effect of temperature on
membrane bioreactor performance working with high hydraulic and sludge
retention time. Biochem. Eng. J. 88, 42–49.
https://doi.org/10.1016/j.bej.2014.03.006
Aziz, H.A., Mojiri, A., 2014. Wastewater Engineering: Advanced Wastewater
Treatment Systems. IJSR Publications.
Barceló, D., Petrovic, M. (Eds.), 2008. Emerging Contaminants from Industrial and
Municipal Waste: Occurrence, Analysis and Effects, Water Pollution. Springer-
Verlag, Berlin Heidelberg.
BCC Research, 2019. Membrane Bioreactors (MBR): Market Size, Trends.
Behera, S.K., Kim, H.W., Oh, J.-E., Park, H.-S., 2011. Occurrence and removal of
antibiotics, hormones and several other pharmaceuticals in wastewater treatment
plants of the largest industrial city of Korea. Sci. Total Environ. 409, 4351–
4360. https://doi.org/10.1016/j.scitotenv.2011.07.015
Benotti, M.J., Stanford, B.D., Wert, E.C., Snyder, S.A., 2009. Evaluation of a
photocatalytic reactor membrane pilot system for the removal of
pharmaceuticals and endocrine disrupting compounds from water. Water Res.
43, 1513–1522. https://doi.org/10.1016/j.watres.2008.12.049
Bérubé, P., 2010. Chapter 9 Membrane Bioreactors: Theory and Applications to
Wastewater Reuse, in: Escobar, I.C., Schäfer, A.I. (Eds.), Sustainability Science
and Engineering, Sustainable Water for the Future: Water Recycling versus
Desalination. Elsevier, pp. 255–292. https://doi.org/10.1016/S1871-
2711(09)00209-8
Bolzonella, D., Fatone, F., di Fabio, S., Cecchi, F., 2010. Application of membrane
bioreactor technology for wastewater treatment and reuse in the Mediterranean
region: Focusing on removal efficiency of non-conventional pollutants. J.
Environ. Manage. 91, 2424–2431.
https://doi.org/10.1016/j.jenvman.2010.07.024
Brillas, E., Martínez-Huitle, C.A., 2015. Decontamination of wastewaters containing
synthetic organic dyes by electrochemical methods. An updated review. Appl.
76 References
Catal. B Environ. 166–167, 603–643.
https://doi.org/10.1016/j.apcatb.2014.11.016
Brillas, E., Sirés, I., Oturan, M.A., 2009. Electro-Fenton Process and Related
Electrochemical Technologies Based on Fenton’s Reaction Chemistry. Chem.
Rev. 109, 6570–6631. https://doi.org/10.1021/cr900136g
Bu, Q., Shi, X., Yu, G., Huang, J., Wang, B., 2016. Assessing the persistence of
pharmaceuticals in the aquatic environment: Challenges and needs. Emerg.
Contam. 2, 145–147. https://doi.org/10.1016/j.emcon.2016.05.003
Burman, I., Sinha, A., 2018. A Review on Membrane Fouling in Membrane
Bioreactors: Control and Mitigation, in: Environmental Contaminants, Energy,
Environment, and Sustainability. Springer, Singapore, pp. 281–315.
https://doi.org/10.1007/978-981-10-7332-8_13
Buttiglieri, G., Knepper, T.P., 2008. Removal of Emerging Contaminants in Wastewater
Treatment: Conventional Activated Sludge Treatment, in: Barceló, D., Petrovic,
M. (Eds.), Emerging Contaminants from Industrial and Municipal Waste:
Removal Technologies, The Handbook of Environmental Chemistry. Springer
Berlin Heidelberg, Berlin, Heidelberg, pp. 1–35. https://doi.org/10.1007/978-3-
540-79210-9_1
Cantinho, P., Matos, M., Trancoso, M.A., dos Santos, M.M.C., 2016. Behaviour and
fate of metals in urban wastewater treatment plants: a review. Int. J. Environ.
Sci. Technol. 13, 359–386. https://doi.org/10.1007/s13762-015-0887-x
Cavalcante, R.P., Dantas, R.F., Wender, H., Bayarri, B., González, O., Giménez, J.,
Esplugas, S., Machulek, A., 2015. Photocatalytic treatment of metoprolol with
B-doped TiO2: Effect of water matrix, toxicological evaluation and
identification of intermediates. Appl. Catal. B Environ. 176–177, 173–182.
Chaudhry, R.M., Holloway, R.W., Cath, T.Y., Nelson, K.L., 2015a. Impact of virus
surface characteristics on removal mechanisms within membrane bioreactors.
Water Res. 84, 144–152. https://doi.org/10.1016/j.watres.2015.07.020
Chaudhry, R.M., Nelson, K.L., Drewes, J.E., 2015b. Mechanisms of Pathogenic Virus
Removal in a Full-Scale Membrane Bioreactor. Environ. Sci. Technol. 49,
2815–2822. https://doi.org/10.1021/es505332n
Chen, G., Ji, S., Sang, Y., Chang, S., Wang, Y., Hao, P., Claverie, J., Liu, H., Yu, G.,
2015. Synthesis of scaly Sn 3 O 4 /TiO 2 nanobelt heterostructures for enhanced
UV-visible light photocatalytic activity. Nanoscale 7, 3117–3125.
https://doi.org/10.1039/C4NR05749J
Chipasa, K.B., 2003. Accumulation and fate of selected heavy metals in a biological
wastewater treatment system. Waste Manag. 23, 135–143.
https://doi.org/10.1016/S0956-053X(02)00065-X
Cirja, M., Ivashechkin, P., Schäffer, A., Corvini, P.F.X., 2008a. . Rev. Environ. Sci.
Biotechnol. 7, 61–78. https://doi.org/10.1007/s11157-007-9121-8
Cirja, M., Ivashechkin, P., Schäffer, A., Corvini, P.F.X., 2008b. Factors affecting the
removal of organic micropollutants from wastewater in conventional treatment
plants (CTP) and membrane bioreactors (MBR). Rev. Environ. Sci. Biotechnol.
7, 61–78. https://doi.org/10.1007/s11157-007-9121-8
77
Clara, M., Kreuzinger, N., Strenn, B., Gans, O., Kroiss, H., 2005a. The solids retention
time—a suitable design parameter to evaluate the capacity of wastewater
treatment plants to remove micropollutants. Water Res. 39, 97–106.
https://doi.org/10.1016/j.watres.2004.08.036
Clara, M., Strenn, B., Gans, O., Martinez, E., Kreuzinger, N., Kroiss, H., 2005b.
Removal of selected pharmaceuticals, fragrances and endocrine disrupting
compounds in a membrane bioreactor and conventional wastewater treatment
plants. Water Res. 39, 4797–4807. https://doi.org/10.1016/j.watres.2005.09.015
Comninellis, C., 1994. Electrocatalysis in the electrochemical conversion/combustion of
organic pollutants for waste water treatment. Electrochimica Acta 39, 1857–
1862. https://doi.org/10.1016/0013-4686(94)85175-1
Comninellis, C., Vercesi, G.P., 1991. Problems in DSA® coating deposition by thermal
decomposition. J. Appl. Electrochem. 21, 136–142.
https://doi.org/10.1007/BF01464294
Comte, S., Guibaud, G., Baudu, M., 2008. Biosorption properties of extracellular
polymeric substances (EPS) towards Cd, Cu and Pb for different pH values. J.
Hazard. Mater. 151, 185–193. https://doi.org/10.1016/j.jhazmat.2007.05.070
Coogan, M.A., Edziyie, R.E., La Point, T.W., Venables, B.J., 2007. Algal
bioaccumulation of triclocarban, triclosan, and methyl-triclosan in a North Texas
wastewater treatment plant receiving stream. Chemosphere 67, 1911–1918.
https://doi.org/10.1016/j.chemosphere.2006.12.027
Cosgrove, W.J., Rijsberman, F.R., Rijsberman, F.R., 2014. World Water Vision :
Making Water Everybody’s Business. Routledge.
https://doi.org/10.4324/9781315071763
Coutte, F., Lecouturier, D., Firdaous, L., Kapel, R., Bazinet, L., Cabassud, C., Dhulster,
P., 2017. Recent Trends in Membrane Bioreactors, in: Larroche, C., Sanromán,
M.Á., Du, G., Pandey, A. (Eds.), Current Developments in Biotechnology and
Bioengineering. Elsevier, pp. 279–311. https://doi.org/10.1016/B978-0-444-
63663-8.00010-0
Dai, Q., Xia, Y., Sun, C., Weng, M., Chen, Jun, Wang, J., Chen, Jianmeng, 2014.
Electrochemical degradation of levodopa with modified PbO2 electrode:
Parameter optimization and degradation mechanism. Chem. Eng. J. 245, 359–
366. https://doi.org/10.1016/j.cej.2013.08.036
Dai, W., Xu, X., Liu, B., Yang, F., 2015. Toward energy-neutral wastewater treatment:
A membrane combined process of anaerobic digestion and nitritation–anammox
for biogas recovery and nitrogen removal. Chem. Eng. J. 279, 725–734.
https://doi.org/10.1016/j.cej.2015.05.036
De la Cruz, N., Dantas, R.F., Giménez, J., Esplugas, S., 2013. Photolysis and TiO2
photocatalysis of the pharmaceutical propranolol: Solar and artificial light. Appl.
Catal. B Environ. 130–131, 249–256.
Deblonde, T., Cossu-Leguille, C., Hartemann, P., 2011. Emerging pollutants in
wastewater: A review of the literature. Int. J. Hyg. Environ. Health, The second
European PhD students workshop: Water and health ? Cannes 2010 214, 442–
448. https://doi.org/10.1016/j.ijheh.2011.08.002
78 References
Deng, Y., Zhao, R., 2015. Advanced Oxidation Processes (AOPs) in Wastewater
Treatment. Curr. Pollut. Rep. 1, 167–176. https://doi.org/10.1007/s40726-015-
0015-z
Di Fabio, S., Lampis, S., Zanetti, L., Cecchi, F., Fatone, F., 2013. Role and
characteristics of problematic biofilms within the removal and mobility of trace
metals in a pilot-scale membrane bioreactor. Process Biochem. 48, 1757–1766.
https://doi.org/10.1016/j.procbio.2013.08.005
Doll, T.E., Frimmel, F.H., 2004. Kinetic study of photocatalytic degradation of
carbamazepine, clofibric acid, iomeprol and iopromide assisted by different
TiO2 materials—determination of intermediates and reaction pathways. Water
Res. 38, 955–964. https://doi.org/10.1016/j.watres.2003.11.009
Dong, H., Zeng, G., Tang, L., Fan, C., Zhang, C., He, X., He, Y., 2015. An overview on
limitations of TiO2-based particles for photocatalytic degradation of organic
pollutants and the corresponding countermeasures. Water Res. 79, 128–146.
Drews, A., 2010. Membrane fouling in membrane bioreactors—Characterisation,
contradictions, cause and cures. J. Membr. Sci. 363, 1–28.
https://doi.org/10.1016/j.memsci.2010.06.046
Ebele, A.J., Abou-Elwafa Abdallah, M., Harrad, S., 2017. Pharmaceuticals and personal
care products (PPCPs) in the freshwater aquatic environment. Emerg. Contam.
3, 1–16. https://doi.org/10.1016/j.emcon.2016.12.004
EU, 2013. Directive 2013/39/EU of the European Parliament and of the Council.
EU, 1991. Council Directive 91/271/EEC of 21 May 1991 concerning urban waste-
water treatment [WWW Document]. URL
http://data.europa.eu/eli/dir/1991/271/oj/eng (accessed 5.7.19).
Fatone, F., Eusebi, A.L., Battistoni, P., Pavan, P., 2008. Exploring the potential of
membrane bioreactors to enhance metals removal from wastewater: pilot
experiences. Water Sci. Technol. 57, 505–511.
https://doi.org/10.2166/wst.2008.115
Fatta-Kassinos, D., Dionysiou, D.D., Kümmerer, K., 2016. Advanced Treatment
Technologies for Urban Wastewater Reuse. Springer.
Fontmorin, J.-M., Sillanpää, M., 2015. Bioleaching and combined bioleaching/Fenton-
like processes for the treatment of urban anaerobically digested sludge: Removal
of heavy metals and improvement of the sludge dewaterability. Sep. Purif.
Technol. 156, Part 2, 655–664. https://doi.org/10.1016/j.seppur.2015.10.061
Fu, F., Wang, Q., 2011. Removal of heavy metal ions from wastewaters: A review. J.
Environ. Manage. 92, 407–418. https://doi.org/10.1016/j.jenvman.2010.11.011
Fujishima, A., Honda, K., 1972. Electrochemical Photolysis of Water at a
Semiconductor Electrode. Nature 238, 37. https://doi.org/10.1038/238037a0
Ganiyu, S.O., van Hullebusch, E.D., Cretin, M., Esposito, G., Oturan, M.A., 2015.
Coupling of membrane filtration and advanced oxidation processes for removal
of pharmaceutical residues: A critical review. Sep. Purif. Technol.,
Environmental Nanotechnology and Sustainability in Water Treatment in honor
of Professor Chin-Pao Huang, Ph.D, Harvard University, Donald C. Philips,
Professor Civil and Environmental Engineering, University of Delaware, USA
156, 891–914. https://doi.org/10.1016/j.seppur.2015.09.059
79
García-Gómez, C., Drogui, P., Seyhi, B., Gortáres-Moroyoqui, P., Buelna, G., Estrada-
Alvgarado, M.I., Álvarez, L.H., 2016. Combined membrane bioreactor and
electrochemical oxidation using Ti/PbO2 anode for the removal of
carbamazepine. J. Taiwan Inst. Chem. Eng. 64, 211–219.
https://doi.org/10.1016/j.jtice.2016.04.024
Gaya, U.I., Abdullah, A.H., 2008. Heterogeneous photocatalytic degradation of organic
contaminants over titanium dioxide: A review of fundamentals, progress and
problems. J. Photochem. Photobiol. C Photochem. Rev. 9, 1–12.
https://doi.org/10.1016/j.jphotochemrev.2007.12.003
Ghaly, M.Y., Jamil, T.S., El-Seesy, I.E., Souaya, E.R., Nasr, R.A., 2011. Treatment of
highly polluted paper mill wastewater by solar photocatalytic oxidation with
synthesized nano TiO2. Chem. Eng. J. 168, 446–454.
Glaze, W.H., 1987. Drinking-water treatment with ozone. Environ. Sci. Technol. 21,
224–230. https://doi.org/10.1021/es00157a001
Guerra, P., Kim, M., Shah, A., Alaee, M., Smyth, S.A., 2014. . Sci. Total Environ. 473–
474, 235–243. https://doi.org/10.1016/j.scitotenv.2013.12.008
Guillard, C., Disdier, J., Herrmann, J.-M., Lehaut, C., Chopin, T., Malato, S., Blanco, J.,
1999. Comparison of various titania samples of industrial origin in the solar
photocatalytic detoxification of water containing 4-chlorophenol. Catal. Today
54, 217–228. https://doi.org/10.1016/S0920-5861(99)00184-4
Gupta, N., Jana, N., Majumder, C.B., 2008. Submerged membrane bioreactor system
for municipal wastewater treatment process : An overview 15, 604–612.
Hai, F.I., Tessmer, K., Nguyen, L.N., Kang, J., Price, W.E., Nghiem, L.D., 2011.
Removal of micropollutants by membrane bioreactor under temperature
variation. J. Membr. Sci. 383, 144–151.
Hai, F.I., Yamamoto, K., Lee, C.-H., 2014. Membrane Biological Reactors- Theory,
Modeling, Design, Management and applications to Wastewater Reuse. IWA
Publishing.
Hapeshi, E., Achilleos, A., Vasquez, M.I., Michael, C., Xekoukoulotakis, N.P.,
Mantzavinos, D., Kassinos, D., 2010. Drugs degrading photocatalytically:
Kinetics and mechanisms of ofloxacin and atenolol removal on titania
suspensions. Water Res. 44, 1737–1746.
Herath, G., Yamamoto, K., Urase, T., 1999. Removal of viruses by microfiltration
membranes at different solution environments. Water Sci. Technol. 40, 331–
338. https://doi.org/10.1016/S0273-1223(99)00515-6
Hung, Y.-T., Wang, L.K., Shammas, N.K., 2012. Handbook of Environment and Waste
Management: Air and Water Pollution Control. World Scientific.
Jelic, A., Gros, M., Ginebreda, A., Cespedes-Sánchez, R., Ventura, F., Petrovic, M.,
Barcelo, D., 2011. Occurrence, partition and removal of pharmaceuticals in
sewage water and sludge during wastewater treatment. Water Res. 45, 1165–
1176. https://doi.org/10.1016/j.watres.2010.11.010
Joo, J.C., Kim, G.Y., Ahn, C.H., Lee, S., Park, J.-R., Kim, J.K., 2019. Development of
Titanium Dioxide (TiO2)-Embedded Buoyant Photocatalyst Balls Using
80 References
Expanded Polystyrene. J. Nanosci. Nanotechnol. 19, 1065–1068.
https://doi.org/10.1166/jnn.2019.15935
Joss, A., Zabczynski, S., Göbel, A., Hoffmann, B., Löffler, D., McArdell, C.S., Ternes,
T.A., Thomsen, A., Siegrist, H., 2006. Biological degradation of
pharmaceuticals in municipal wastewater treatment: Proposing a classification
scheme. Water Res. 40, 1686–1696.
Judd, S., 2008. The status of membrane bioreactor technology. Trends Biotechnol. 26,
109–116. https://doi.org/10.1016/j.tibtech.2007.11.005
Judd, S.J., 2016. The status of industrial and municipal effluent treatment with
membrane bioreactor technology. Chem. Eng. J., SI: New concepts in membrane
reactors: from membrane design to tissue engineering 305, 37–45.
Kanakaraju, D., Motti, C.A., Glass, B.D., Oelgemöller, M., 2014. Photolysis and TiO2-
catalysed degradation of diclofenac in surface and drinking water using
circulating batch photoreactors. Environ. Chem. 11, 51–62.
https://doi.org/10.1071/EN13098
Katsou, E., Malamis, S., Loizidou, M., 2011. Performance of a membrane bioreactor
used for the treatment of wastewater contaminated with heavy metals.
Bioresour. Technol. 102, 4325–4332.
https://doi.org/10.1016/j.biortech.2010.10.118
Khan, S.A., Ali, S., Sohail, M., Morsy, M.A., Yamani, Z.H., 2016. Fabrication of
TiO2/Ag/Ag2O Nanoparticles to Enhance the Photocatalytic Activity of
Degussa P25 Titania. Aust. J. Chem. 69, 41–46.
https://doi.org/10.1071/CH15250
Kim, M., Guerra, P., Shah, A., Parsa, M., Alaee, M., Smyth, S.A., 2014. Removal of
pharmaceuticals and personal care products in a membrane bioreactor
wastewater treatment plant. Water Sci. Technol. 69, 2221–2229.
https://doi.org/10.2166/wst.2014.145
Kimura, K., Hara, H., Watanabe, Y., 2007. Elimination of Selected Acidic
Pharmaceuticals from Municipal Wastewater by an Activated Sludge System
and Membrane Bioreactors. Environ. Sci. Technol. 41, 3708–3714.
https://doi.org/10.1021/es061684z
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. J. Membr. Sci. 423–424, 348–361.
Krzeminski, P., Leverette, L., Malamis, S., Katsou, E., 2017. Membrane bioreactors – A
review on recent developments in energy reduction, fouling control, novel
configurations, LCA and market prospects. J. Membr. Sci. 527, 207–227.
Kuo, D.H.W., Simmons, F.J., Blair, S., Hart, E., Rose, J.B., Xagoraraki, I., 2010.
Assessment of human adenovirus removal in a full-scale membrane bioreactor
treating municipal wastewater. Water Res. 44, 1520–1530.
81
Laera, G., Chong, M.N., Jin, B., Lopez, A., 2011. An integrated MBR–TiO2
photocatalysis process for the removal of Carbamazepine from simulated
pharmaceutical industrial effluent. Bioresour. Technol. 102, 7012–7015.
https://doi.org/10.1016/j.biortech.2011.04.056
Le-Clech, P., Chen, V., Fane, T.A.G., 2006. Fouling in membrane bioreactors used in
wastewater treatment. J. Membr. Sci. 284, 17–53.
Li, C., Cabassud, C., Guigui, C., 2015. Evaluation of membrane bioreactor on removal
of pharmaceutical micropollutants: a review. Desalination Water Treat. 55, 845–
858. https://doi.org/10.1080/19443994.2014.926839
Lodder, W.J., Husman, A.M. de R., 2005. Presence of Noroviruses and Other Enteric
Viruses in Sewage and Surface Waters in The Netherlands. Appl Env. Microbiol
71, 1453–1461. https://doi.org/10.1128/AEM.71.3.1453-1461.2005
López, J.L.C., Reina, A.C., Gómez, E.O., Martín, M.M.B., Rodríguez, S.M., Pérez,
J.A.S., 2010. Integration of Solar Photocatalysis and Membrane Bioreactor for
Pesticides Degradation. Sep. Sci. Technol. 45, 1571–1578.
https://doi.org/10.1080/01496395.2010.487465
Luo, Y., Guo, W., Ngo, H.H., Nghiem, L.D., Hai, F.I., Zhang, J., Liang, S., Wang, X.C.,
2014. A review on the occurrence of micropollutants in the aquatic environment
and their fate and removal during wastewater treatment. Sci. Total Environ.
473–474, 619–641. https://doi.org/10.1016/j.scitotenv.2013.12.065
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.
https://doi.org/10.1016/j.desal.2012.12.016
Martínez-Huitle, C.A., Brillas, E., 2009. Decontamination of wastewaters containing
synthetic organic dyes by electrochemical methods: A general review. Appl.
Catal. B Environ. 87, 105–145. https://doi.org/10.1016/j.apcatb.2008.09.017
Martínez-Huitle, C.A., Ferro, S., 2006. Electrochemical oxidation of organic pollutants
for the wastewater treatment: direct and indirect processes. Chem. Soc. Rev. 35,
1324–1340. https://doi.org/10.1039/B517632H
Mascolo, G., Laera, G., Pollice, A., Cassano, D., Pinto, A., Salerno, C., Lopez, A.,
2010. Effective organics degradation from pharmaceutical wastewater by an
integrated process including membrane bioreactor and ozonation. Chemosphere
78, 1100–1109. https://doi.org/10.1016/j.chemosphere.2009.12.042
Melin, T., Jefferson, B., Bixio, D., Thoeye, C., De Wilde, W., De Koning, J., van der
Graaf, J., Wintgens, T., 2006. Membrane bioreactor technology for wastewater
treatment and reuse. Desalination, Integrated Concepts in Water
RecyclingIntegrated Concepts in Water Recycling 187, 271–282.
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 Res. 43, 1489–1512.
Meng, F., Liao, B., Liang, S., Yang, F., Zhang, H., Song, L., 2010. Morphological
visualization, componential characterization and microbiological identification
82 References
of membrane fouling in membrane bioreactors (MBRs). J. Membr. Sci. 361, 1–
14. https://doi.org/10.1016/j.memsci.2010.06.006
Meng, F., Yang, F., 2007. Fouling mechanisms of deflocculated sludge, normal sludge,
and bulking sludge in membrane bioreactor. J. Membr. Sci. 305, 48–56.
Meng, F., Zhang, S., Oh, Y., Zhou, Z., Shin, H.-S., Chae, S.-R., 2017. Fouling in
membrane bioreactors: An updated review. Water Res. 114, 151–180.
Metcalf & Eddy Inc., 2003. Wastewater Engineering : Treatment and Reuse. McGraw-
Hill Series.
Miura, T., Okabe, S., Nakahara, Y., Sano, D., 2015. Removal properties of human
enteric viruses in a pilot-scale membrane bioreactor (MBR) process. Water Res.
Murgolo, S., Yargeau, V., Gerbasi, R., Visentin, F., El Habra, N., Ricco, G., Lacchetti,
I., Carere, M., Curri, M.L., Mascolo, G., 2017. A new supported TiO2 film
deposited on stainless steel for the photocatalytic degradation of contaminants of
emerging concern. Chem. Eng. J., Emerging Advanced Oxidation Processes for
the Elimination of Micro-Pollutants 318, 103–111.
Nassiri Koopaei, N., Abdollahi, M., 2017. Health risks associated with the
pharmaceuticals in wastewater. DARU J. Pharm. Sci. 25.
https://doi.org/10.1186/s40199-017-0176-y
Neoh, C.H., Noor, Z.Z., Mutamim, N.S.A., Lim, C.K., 2016. Green technology in
wastewater treatment technologies: Integration of membrane bioreactor with
various wastewater treatment systems. Chem. Eng. J. 283, 582–594.
Nidheesh, P.V., Zhou, M., Oturan, M.A., 2018. An overview on the removal of
synthetic dyes from water by electrochemical advanced oxidation processes.
Chemosphere 197, 210–227. https://doi.org/10.1016/j.chemosphere.2017.12.195
Oturan, M.A., Aaron, J.-J., 2014. Advanced Oxidation Processes in Water/Wastewater
Treatment: Principles and Applications. A Review. Crit. Rev. Environ. Sci.
Technol. 44, 2577–2641. https://doi.org/10.1080/10643389.2013.829765
Panizza, M., Cerisola, G., 2009. Direct And Mediated Anodic Oxidation of Organic
Pollutants. Chem. Rev. 109, 6541–6569. https://doi.org/10.1021/cr9001319
Park, H.-D., Chang, I.-S., Lee, K.-J., 2015. Principles of Membrane Bioreactors for
Wastewater Treatment. Taylor & Francis Group.
Pollice, A., Laera, G., Cassano, D., Diomede, S., Pinto, A., Lopez, A., Mascolo, G.,
2012. Removal of nalidixic acid and its degradation products by an integrated
MBR-ozonation system. J. Hazard. Mater. 203–204, 46–52.
https://doi.org/10.1016/j.jhazmat.2011.11.072
Punmia, D.B.C., Jain, Ashok Kr, Jain, Arun Kr, 1998. Waste Water Engineering,
Second edition. ed, Laxmi Publications. Firewall Media, New Delhi.
Qasim, S.R., 2017. Wastewater Treatment Plants : Planning, Design, and Operation,
Second Edition. Routledge. https://doi.org/10.1201/9780203734209
Radjenović, J., Matošić, M., Mijatović, I., Petrović, M., Barceló, D., 2008. Membrane
Bioreactor (MBR) as an Advanced Wastewater Treatment Technology, in:
83
Barceló, D., Petrovic, M. (Eds.), Emerging Contaminants from Industrial and
Municipal Waste: Removal Technologies, The Handbook of Environmental
Chemistry. Springer Berlin Heidelberg, Berlin, Heidelberg, pp. 37–101.
https://doi.org/10.1007/698_5_093
Radjenović, J., Petrović, M., Barceló, D., 2009. Fate and distribution of pharmaceuticals
in wastewater and sewage sludge of the conventional activated sludge (CAS)
and advanced membrane bioreactor (MBR) treatment. Water Res. 43, 831–841.
Ren, H.-T., Jia, S.-Y., Zou, J.-J., Wu, S.-H., Han, X., 2015. A facile preparation of
Ag2O/P25 photocatalyst for selective reduction of nitrate. Appl. Catal. B
Environ. 176–177, 53–61. https://doi.org/10.1016/j.apcatb.2015.03.038
Ren, H.-T., Yang, Q., 2017. . Appl. Surf. Sci. 396, 530–538.
https://doi.org/10.1016/j.apsusc.2016.10.191
Rizzo, L., Meric, S., Kassinos, D., Guida, M., Russo, F., Belgiorno, V., 2009.
Degradation of diclofenac by TiO2 photocatalysis: UV absorbance kinetics and
process evaluation through a set of toxicity bioassays. Water Res. 43, 979–988.
Rodríguez-Mozaz, S., Huerta, B., Barceló, D., 2016. Bioaccumulation of Emerging
Contaminants in Aquatic Biota: Patterns of Pharmaceuticals in Mediterranean
River Networks, in: Petrovic, M., Sabater, S., Elosegi, A., Barceló, D. (Eds.),
Emerging Contaminants in River Ecosystems: Occurrence and Effects Under
Multiple Stress Conditions, The Handbook of Environmental Chemistry.
Springer International Publishing, Cham, pp. 121–141.
https://doi.org/10.1007/698_2015_5015
Rodriguez-Narvaez, O.M., Peralta-Hernandez, J.M., Goonetilleke, A., Bandala, E.R.,
2017. Treatment technologies for emerging contaminants in water: A review.
Chem. Eng. J. 323, 361–380. https://doi.org/10.1016/j.cej.2017.04.106
Santos, A., Judd, S., 2010. The fate of metals in wastewater treated by the activated
sludge process and membrane bioreactors: A brief review. J. Environ. Monit. 12,
110–118. https://doi.org/10.1039/B918161J
Särkkä, H., Bhatnagar, A., Sillanpää, M., 2015. Recent developments of electro-
oxidation in water treatment — A review. J. Electroanal. Chem. 754, 46–56.
https://doi.org/10.1016/j.jelechem.2015.06.016
Shestakova, M., 2016. Ultrasound-assisted electrochemical treatment of wastewaters
containing organic pollutants by using novel Ti/Ta2O5-SnO2 electrodes.
Lappeenranta University of Technology.
Sirés, I., Brillas, E., Oturan, M.A., Rodrigo, M.A., Panizza, M., 2014. Electrochemical
advanced oxidation processes: today and tomorrow. A review. Environ. Sci.
Pollut. Res. 21, 8336–8367. https://doi.org/10.1007/s11356-014-2783-1
Smith, C.V., Di Gregorio, D., Talcott, R.M., 1969. Use of ultrafiltration membranes for
activated sludge separation, in: Dorr-Oliver Inc., Engineering Extension Series
No. 135. Presented at the 24th Industrial Waste Conference, pp. 1300–1310.
Song, S., Fan, J., He, Z., Zhan, L., Liu, Z., Chen, J., Xu, X., 2010. Electrochemical
degradation of azo dye C.I. Reactive Red 195 by anodic oxidation on Ti/SnO2–
Sb/PbO2 electrodes. Electrochimica Acta 55, 3606–3613.
https://doi.org/10.1016/j.electacta.2010.01.101
84 References
Sperling, M. von, 2007. Wastewater Characteristics, Treatment and Disposal. IWA
Publishing, UK.
Suárez, S., Reif, R., Lema, J.M., Omil, F., 2012. Mass balance of pharmaceutical and
personal care products in a pilot-scale single-sludge system: Influence of T, SRT
and recirculation ratio. Chemosphere 89, 164–171.
https://doi.org/10.1016/j.chemosphere.2012.05.094
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 Res.
Taheran, M., Brar, S.K., Verma, M., Surampalli, R.Y., Zhang, T.C., Valero, J.R., 2016.
Membrane processes for removal of pharmaceutically active compounds
(PhACs) from water and wastewaters. Sci. Total Environ. 547, 60–77.
https://doi.org/10.1016/j.scitotenv.2015.12.139
The MBR site, 2019. Largest MBR plants (over 100 MLD) – Worldwide [WWW
Document]. MBR Site. URL https://www.thembrsite.com/largest-membrane-
bioreactor-plants-worldwide/ (accessed 3.20.19).
Tran, N.H., Reinhard, M., Gin, K.Y.-H., 2018. Occurrence and fate of emerging
contaminants in municipal wastewater treatment plants from different
geographical regions-a review. Water Res. 133, 182–207.
Trinh, T., van den Akker, B., Coleman, H.M., Stuetz, R.M., Drewes, J.E., Le-Clech, P.,
Khan, S.J., 2016. Seasonal variations in fate and removal of trace organic
chemical contaminants while operating a full-scale membrane bioreactor. Sci.
Total Environ. 550, 176–183. https://doi.org/10.1016/j.scitotenv.2015.12.083
UN FAO, 2018. Pollutants from agriculture a serious threat to world’s water [WWW
Document]. Food Agric. Organ. U. N. URL
http://www.fao.org/news/story/en/item/1141534/icode/ (accessed 4.4.19).
UN GEMS, 2006. Water Quality for Ecosystem and Human Health [WWW Document].
URL https://wedocs.unep.org/handle/20.500.11822/12217 (accessed 4.10.19).
Urase, T., Kagawa, C., Kikuta, T., 2005. Factors affecting removal of pharmaceutical
substances and estrogens in membrane separation bioreactors. Desalination,
Membranes in Drinking and Industrial Water Production 178, 107–113.
US EPA, 2015a. Municipal Wastewater [WWW Document]. URL
https://www.epa.gov/npdes/municipal-wastewater (accessed 3.15.19).
US EPA, 2015b. Contaminants of Emerging Concern including Pharmaceuticals and
Personal Care Products [WWW Document]. URL
https://www.epa.gov/wqc/contaminants-emerging-concern-including-
pharmaceuticals-and-personal-care-products (accessed 4.4.19).
U.S. EPA, 2007. Method 1694 : Pharmaceuticals and Personal Care Products in Water ,
Soil , Sediment , and Biosolids by HPLC / MS / MS.
U.S. EPA, 2003. Environmental Regulations and Technology, Control of Pathogens and
Vector Attraction in Sewage Sludge. US Environ. Prot. Agency EPA/625/R-
92/013.
U.S. EPA, 1997. Wastewater Treatment Manuals - Primary, Secondary & Tertiary
Treatment. Environ. Prot. Agency EPA.
85
van den Brink, P., Satpradit, O.-A., van Bentem, A., Zwijnenburg, A., Temmink, H.,
van Loosdrecht, M., 2011. Effect of temperature shocks on membrane fouling in
membrane bioreactors. Water Res. 45, 4491–4500.
Verlicchi, P., Zambello, E., 2015. Pharmaceuticals and personal care products in
untreated and treated sewage sludge: Occurrence and environmental risk in the
case of application on soil — A critical review. Sci. Total Environ. 538, 750–
767. https://doi.org/10.1016/j.scitotenv.2015.08.108
Vieno, N., Sillanpää, M., 2014. Fate of diclofenac in municipal wastewater treatment
plant — A review. Environ. Int. 69, 28–39.
https://doi.org/10.1016/j.envint.2014.03.021
Vieno, N., Tuhkanen, T., Kronberg, L., 2007. Elimination of pharmaceuticals in sewage
treatment plants in Finland. Water Res. 41, 1001–1012.
Vieno, N.M., Tuhkanen, T., Kronberg, L., 2005. Seasonal Variation in the Occurrence
of Pharmaceuticals in Effluents from a Sewage Treatment Plant and in the
Recipient Water. Environ. Sci. Technol. 39, 8220–8226.
https://doi.org/10.1021/es051124k
Visvanathan, C., Aim, R.B., Parameshwaran, K., 2000. Membrane Separation
Bioreactors for Wastewater Treatment. Crit. Rev. Environ. Sci. Technol. 30, 1–
48. https://doi.org/10.1080/10643380091184165
Wang, X., Li, S., Yu, H., Yu, J., Liu, S., 2011. Ag2O as a New Visible-Light
Photocatalyst: Self-Stability and High Photocatalytic Activity. Chem. – Eur. J.
17, 7777–7780. https://doi.org/10.1002/chem.201101032
WHO, 2015a. Progress on drinking-water, sanitation and hygiene, 2017: Infographics
[WWW Document]. URL
http://www.who.int/water_sanitation_health/monitoring/coverage/jmp-update-
2017-graphics/en/ (accessed 4.9.19).
WHO, 2015b. Antibiotic resistance: Multi-country public awareness survey.
World Water Council, 2019. Water Crisis.
WWF, 2019. Water Scarcity [WWW Document]. URL
https://www.worldwildlife.org/threats/water-scarcity (accessed 4.9.19).
Xiao, K., Liang, S., Wang, X., Chen, C., Huang, X., 2019. Current state and challenges
of full-scale membrane bioreactor applications: A critical review. Bioresour.
Technol. 271, 473–481. https://doi.org/10.1016/j.biortech.2018.09.061
Xu, Y., Schoonen, M.A.A., 2000. The absolute energy positions of conduction and
valence bands of selected semiconducting minerals. Am. Mineral. 85, 543–556.
https://doi.org/10.2138/am-2000-0416
Xue, G., Liu, H., Chen, Q., Hills, C., Tyrer, M., Innocent, F., 2011. Synergy between
surface adsorption and photocatalysis during degradation of humic acid on
TiO2/activated carbon composites. J. Hazard. Mater. 186, 765–772.
https://doi.org/10.1016/j.jhazmat.2010.11.063
Yamamoto, K., Hiasa, M., Mahmood, T., Matsuo, T., 1988. Direct Solid-Liquid
Separation Using Hollow Fiber Membrane In Aa Activated Sludge Aeration
Tank, in: Lijklema, L., Imhoff, K.R., Ives, K.J., Jenkins, D., Ludwig, R.G.,
Suzuki, M., Toerien, D.F., Wheatland, A.B., Milburn, A., Izod, E.J. (Eds.),
86 References
Water Pollution Research and Control Brighton. Pergamon, pp. 43–54.
https://doi.org/10.1016/B978-1-4832-8439-2.50009-2
Yang, W., Cicek, N., Ilg, J., 2006. State-of-the-art of membrane bioreactors: Worldwide
research and commercial applications in North America. J. Membr. Sci. 270,
201–211. https://doi.org/10.1016/j.memsci.2005.07.010
Yang, Y., Ok, Y.S., Kim, K.-H., Kwon, E.E., Tsang, Y.F., 2017. Occurrences and
removal of pharmaceuticals and personal care products (PPCPs) in drinking
water and water/sewage treatment plants: A review. Sci. Total Environ. 596–
597, 303–320. https://doi.org/10.1016/j.scitotenv.2017.04.102
Zhang, C., Wang, G., Hu, Z., 2014. Changes in wastewater treatment performance and
activated sludge properties of a membrane bioreactor at low temperature
operation. Environ. Sci. Process. Impacts 16, 2199–207.
https://doi.org/10.1039/c4em00174e
Zhou, W., Liu, H., Wang, J., Liu, D., Du, G., Cui, J., 2010. Ag2O/TiO2 Nanobelts
Heterostructure with Enhanced Ultraviolet and Visible Photocatalytic Activity.
ACS Appl. Mater. Interfaces 2, 2385–2392. https://doi.org/10.1021/am100394x
Zorita, S., Mårtensson, L., Mathiasson, L., 2009. Occurrence and removal of
pharmaceuticals in a municipal sewage treatment system in the south of Sweden.
Sci. Total Environ. 407, 2760–2770.
Publication I
Khum Gurung, Mohamed Chaker Ncibi, Mika Sillanpää

Assessing membrane fouling and the performance of pilot-scale membrane
bioreactor (MBR) to treat real municipal wastewater during winter season
in Nordic regions
Reprinted with permission from

Science of the Total Environment
Vol. 579, pp. 1289-1297, 2017
© 2017 Elsevier
Science of the Total Environment 579 (2017) 1289–1297
Contents lists available at ScienceDirect
journal homepage: www.elsevier.com/locate/scitotenv
Assessing membrane fouling and the performance of pilot-scale

membrane bioreactor (MBR) to treat real municipal wastewater during
winter season in Nordic regions
Khum Gurung a,⁎, Mohamed Chaker Ncibi a, Mika Sillanpää a,b
a
Laboratory of Green Chemistry, School of Engineering Science, Lappeenranta University of Technology, Sammonkatu 12, FI-50130 Mikkeli, Finland
b
Department of Civil and Environmental Engineering, Florida International University, Miami FL-33174, USA
H I G H L I G H T S G R A P H I C A L A B S T R A C T
• Performance of a pilot-scale MBR was

studied at low temperatures in Nordic
region.
• Substantial membrane fouling was ob-
served at low temperatures (below
10 °C).
• About 75% of the permeability drop was
observed due to fouling.
• High removal rates were achieved for
pathogens.
• High removal rates for the emerging
micropollutants were observed.
a r t i c l e i n f o a b s t r a c t
Article history: In this study, the performance of a pilot-scale membrane bioreactor (MBR) to treat real municipal wastewater
Received 23 October 2016 was assessed at low temperatures (7 to 20 °C) in Nordic regions. First, the effect of low temperatures on mem-
Received in revised form 17 November 2016 brane fouling was evaluated by monitoring trans-membrane pressure. A significant membrane fouling was ob-
Accepted 17 November 2016
served when the sludge temperature inside the MBR unit was below 10 °C with a 75% permeability drop, thus
Available online 29 November 2016
indicating high deterioration of the membrane performance at low temperatures. Moreover, increasing values
Editor: Jay Gan of sludge volume index (SVI) during low temperatures showed high deterioration of sludge settleability. As for
the pollution removal, MBR achieved high performances primarily for pathogens and emerging micropollutants.
Keywords: The average log reductions of 1.82, 3.02, and 1.94 log units were achieved for norovirus GI, norovirus GII, and ad-
Membrane bioreactor enoviruses, respectively. Among the four trace organic compounds (TrOCs), the average removal efficiencies of
Membrane fouling bisoprolol, diclofenac and bisphenol A were 65%, 38%, and N97%, respectively. However, carbamazepine was
Low temperature not efficiently removed (‐89% to 28%). Regarding trace metals, an average removal of N80% was achieved for
Municipal wastewater Cd, Pb, and V. For the rest of the metals, the removal capacities were between 30 and 60%.
Pathogens
© 2016 Elsevier B.V. All rights reserved.
Emerging micropollutants
⁎ Corresponding author.
E-mail addresses: khumgrg@yahoo.com (K. Gurung), mika.sillanpaa@lut.fi (M. Sillanpää).
http://dx.doi.org/10.1016/j.scitotenv.2016.11.122
0048-9697/© 2016 Elsevier B.V. All rights reserved.
1290 K. Gurung et al. / Science of the Total Environment 579 (2017) 1289–1297
1. Introduction arrangement of the MBR plant was explained elsewhere (Gurung et al.,
2016). The capacity of the MBR pilot plant was designed to treat 3 m3 of
Temperature influences the performance of biological treatment in con- wastewater per day. The MBR unit used flat sheet microfiltration mem-
ventional activated sludge processes (Arévalo et al., 2014; Metcalf and Eddy branes having pore size of 0.4 μm and effective surface area of 16 m2.
Inc., 2003). Moreover, in MBRs, temperature not only affects the bioconver- The MBR unit was fed with primary clarified wastewater of the full-
sion process but also influences microbial community, fouling rate and scale Kenkäveronniemi WWTP. The stabilized sludge was wasted from
sludge morphology (Zhang et al., 2014). The biological activity of the acti- membrane compartment daily to maintain designated sludge retention
vated sludge tend to decrease in winter periods. Indeed, when temperature time (SRT). The treated permeate water and wasted sludge were col-
drops to about 15 °C, methane-producing bacteria become essentially dor- lected in the permeate and sludge tanks, respectively and then
mant, and at about 5 °C, the autotrophic bacteria practically cease function- discharged to the WWTP sewer.
ing. Even the chemoheterotrophic bacteria performing on carbonaceous The MBR pilot unit was operated for N 120 days from January to April
material becomes inactive at 2 °C (Metcalf and Eddy Inc., 2003). 2016, coinciding with the winter season in Finland (temperature of in-
Membrane fouling, which is the potential deposition of the feed fluent wastewater around 7–10 °C). The water level in the membrane
stream on the membranes, is one of the major causes behind the dete- tank was maintained constant (b 5% change in water volume during
rioration of MBR performance (permeability). In this context, it was re- the study). No backwashing was performed, however, the chemical
ported that the low temperatures can cause rapid membrane fouling as washings of membranes were carried out whenever the TMP reached
compared to mesophilic temperature under which proteins are the a level of ≤ 20 kPa by using sodium hypochlorite followed by citric
dominant foulants (Gao et al., 2014). There has been many previous acid. The pH of the biological reactors was adjusted using NaOH solu-
studies on membrane fouling (Drews, 2010; van den Brink et al., tion. During the manual cleaning, membrane cartridges were taken
2011; Zhang et al., 2014). Membrane fouling at low temperatures are out of the bioreactor and then accumulated cake layer was cleaned by
also found to induce changes in the cake layer thickness and/or porosity using brush and water jet. The operating conditions of pilot operation
of the membranes (van den Brink et al., 2011). Moreover, other factors is summarized in Table 1.
attributed to the membrane fouling at low temperature include in-
creased viscosity, reduced sludge stabilization, reduced particle size of 2.2. Analytical methods
sludge and reduced mass transfer efficiency (Zhang et al., 2014).
The fate of human enteric viruses such as noroviruses and adenovi- The major pollutants in wastewaters such as total suspended solids
ruses are of major interest because they are common causes of water- (TSS), chemical oxygen demand (COD), ammonium nitrogen (NH4-N),
borne outbreaks. In MBR system, the removal mechanisms of and total phosphorus (TP) in influent and permeate were determined
pathogens can be either via attachment to biofilm developed on the according to the Standard Methods (American Public Health
membrane surface or adsorption in/onto the biomass flocs and to Association et al., 1999). Liquid samples were collected for 24 h using
some extent via retention by micro-membranes (Miura et al., 2015). automatic composite samplers in weekly basis. The SVI was determined
The active evaluation of possible risk of emerging organic weekly from mixed liquor grab samples according to the previously de-
micropollutants or TrOCs such as pharmaceutically active compounds scribed method (Fontmorin and Sillanpää, 2015). The mixed liquor
(PhACs) and endocrine disrupting chemicals (EDCs) in treated effluents of suspended solids (MLSS) concentrations of mixed liquor from both aer-
wastewater treatment systems has become a great concern (Trinh et al., obic and membrane bioreactors were determined once a week accord-
2012). In winter seasons in cold regions, higher concentrations of emerging ing to Standard Methods (American Public Health Association et al.,
micropollutants are detected due to the decrease in bacterial biodegradation 1999).
activities under low temperature (Kruglova et al., 2016), in addition to the Membrane fouling was quantified as a progressing phenomenon by
increasing consumption of pharmaceutical drugs, especially antibiotics, in monitoring trans-membrane pressure (TMP), while the permeate flux
winters (Suda et al., 2014). In Finland, carbamazepine (CBZ), diclofenac was maintained almost constant over the experimental period. Mem-
(DCF), and bisoprolol (BIS) are three commonly used PhACs ((FIMEA), brane performance was assessed according to the following equation
2015). The toxicity of these PhACs on fish, algae, and bacteria in concentra- (Park et al., 2015):
tions b1 mg L−1 have been recently reported (Kruglova et al., 2016).
On the other hand, trace metals are recalcitrance and persistence inor- J 1
Lp ¼ ¼ ð1Þ
ganic micropollutants (Fu and Wang, 2011) that could cause acute and TMP μ w Rt
chronic toxicity (Karvelas et al., 2003). Moreover, their presence at cer-
tain concentrations is not only of environmental concern but also strong- Lp is membrane permeability (L m−2 h−1 bar−1); J is membrane flux
ly reduces microbial activity, thus resulting adverse effects on biological (L m−2 h−1); TMP is trans-membrane pressure (Pa); μw is permeate
wastewater treatment processes (Chipasa, 2003). In MBRs, the removal viscosity (Pa·s); Rt is total filtration resistance (m−1).
of trace metals are highly enhanced primarily due to the reduction in The analysis of noroviruses (GI and GII) and adenoviruses were per-
suspended solids from effluents (Di Fabio et al., 2013). formed from influent, permeate and mixed liquor according to the
In this context, the present research aims at studying the perfor- methods described elsewhere (Miura et al., 2015; Simmons et al.,
mance of a submerged pilot MBR to treat real municipal wastewater
under the low temperature conditions during winter season in Finland.
The investigation mainly focused on assessing the membrane fouling, Table 1
Operational conditions of the MBR pilot plant.
pathogens removal and fate of emerging micropollutants (organic and
inorganic). To the best of our knowledge, this is the first detailed Parameter Units Values
study on MBR treating real municipal wastewater under low tempera- Operational period Days 120
ture periods in Nordic environment. Hydraulic retention time (HRT) Hours 35
Sludge retention time (SRT) Days 25–30
Flux (continuous) L m−2 h−1 7.80
2. Materials and methods
MLSS concentration mg L−1 5300–9800
F/M ratio kg COD/kg MLSS 0.02–0.05
2.1. The pilot scale MBR plant and operating conditions Aeration intensity m−3 m−2 h−1 0.4–0.6
Sludge temperature °C 6.5–21/7.0-20
In this research, a pilot-scale MBR was employed to treat a real mu- pH Unitless 6.6–7.3
Suction cycle Minutes 9-ON/1-OFF
nicipal wastewater in Mikkeli Finland. The detailed experimental set-up
K. Gurung et al. / Science of the Total Environment 579 (2017) 1289–1297 1291
2011). Moreover, the concentrations of E. coli and Enterococcus in the in- and the process resumed again. The average temperature of mixed li-
fluent, sludge and permeate were enumerated by using Colilert® and quor reached 20.5 ± 0.6 °C immediately after the process restarted
Enterolert® methods. until 8th February, which was expected due to the heating effects of
The occurrence of TrOCs both in liquid and solid phases were ana- air mass flow along with heaters and heating cables. Thus, the SADm
lyzed from influent, permeate and sludge, respectively using standard was reduced and maintained at 0.4 m3 h−1 m−2 from 9th February to
solid phase extraction (SPE) and LC-MS/MS methods (EPA, 2007). 6th April in order to ensure the lowest possible temperature inside the
The analysis of heavy metals in the influent, permeate and sludge bioreactors. As a consequence, the sludge temperature started decreas-
samples was conducted with an inductively coupled plasma-optical ing gradually from 10th February and reached nearly 10 °C by 18th
emission spectrometer (ICP-OES, iCAP 6300, Thermo Electron Corpora- March. In addition, the decreasing temperature of the sludge was also
tion, USA) from the pretreated samples (Fontmorin and Sillanpää, attributed to the incoming flow of melting ice to the municipal waste-
2015). water. Occasionally, heaters and heating cables were intermittently
used depending on outside air temperature in order to prevent serious
3. Results and discussion damages related to freezing.
The first TMP jump occurred between 23rd February and 18th March.
3.1. MBR operation and membrane fouling at low temperatures The temperature shock was evident when temperature dropped to
about 10 °C, which is consistent with the results of Gao et al. (Gao et
Fig. 1 shows the variations in sludge temperature, TMP profile and al., 2014). As the membrane fouling accelerated rapidly and the TMP
aeration intensity during MBR operation. At the beginning of this exper- reached nearly to 17 kPa (1st peak), first chemical washing of mem-
iment, the average sludge temperature was 11 ± 1.5 °C, the specific aer- branes was performed. After the washing TMP dropped to about
ation demand per membrane surface (SADm) was fixed at about 6 kPa, but increased rapidly and finally reached nearly to 12 kPa (2nd
0.6 m3 h−1 m−2 and the TMP was about 2 kPa until 9th February. How- peak) within a week from the 1st peak. Yet, the sludge temperature
ever, two weeks later, the atmospheric temperature began to drop was measured to be 7.7 ± 1.2 °C during the 2nd peak and the propensity
sharply. The outside air temperature was varying between 0 °C and of the membrane fouling was very rapid. Therefore, it was concluded
−33 °C for more than a week, which is the typical Nordic weather phe- that the rapid increase in the TMP was highly attributed to the low tem-
nomenon during winter seasons. Due to this sudden temperature drop, perature of sludge. Besides, due to the Easter holidays, the second in-
some pumps and connecting pipes were frozen and severely clogged. situ washing of membranes was performed during the 2nd fouling
Thus, the system malfunctioned and stopped for more than a week peak (a bit early when compared to the 1st peak). Again, the TMP in-
(18th to 25th January). Heating cables were installed to the stretch of creased gradually and reached nearly to 13 kPa within 12 days from
pipes from feed point to the pilot plant inlet. The broken pumps were the 2nd peak. The sludge temperature was increased to about 8–12 °C
changed with new ones and the clogged pipes were cleaned with hot and stayed almost stable after second chemical washing. The membrane
water. Additionally, three heaters (25 W capacity each): two inside fouling occurrence between the 1st peak and 2nd peak was within a
the pilot plant and one at the influent pumping station were installed week. However, it extended to nearly two weeks between 2nd and 3rd
Fig. 1. Membrane fouling behavior and cleaning measures. (Note: SADm is not in scale).
peak. The extended time between the fouling peaks can be attributed to membranes are recommended to be installed for the colder periods in
the increased sludge temperature. The results showed that membrane order to ensure constant membrane permeability, especially in coun-
fouling was important during low temperature periods, which is in tries where long low temperature periods exists throughout a year.
agreement with other studies (Gao et al., 2014; van den Brink et al., Membrane fluxes can be lowered using extra membranes, resulting in
2011; Zhang et al., 2014). As the sludge temperature stayed at about a slower propensity of membrane fouling due to the distribution of
12 °C (after 2nd fouling peak) for two consecutive weeks, maintaining trans-membrane pressures. This leads to maintain relatively high and
low temperatures in the pilot was no more possible as the weather constant permeability.
started to gradually warm up. Therefore, after 3rd fouling peak, the Fig. 2 shows the variations of the MLSS concentrations and SVI of
SADm was increased back to 0.6 m3 h−1 m−2. This increase helped re- mixed liquor. The average SVI was found to be increasing from
ducing the TMP by approximately 2.5 kPa within a week. Thus, the foul- 120 mL g−1 to 168 mL g−1 when the average sludge temperature varied
ing phenomena occurring during the MBR operation were not only between 15 and 10 °C. The higher SVI values at low operating tempera-
attributed to the low temperature conditions but also, to some extent, ture indicates that the compressibility and settleability of sludge are
related to the reduction in aeration intensity. During the chemical wash- highly deteriorated (Zhang et al., 2014). More fine particles, colloids or
ings the TMP dropped nearly to 6 kPa and then raised rapidly, illustrat- polymers are released when operating at low temperatures due to
ing that this cleaning procedure is not so effective. As for the manual loose morphology of mixed liquor, leading to high tendency of mem-
washing of membranes, it reduced the TMP to nearly 3 kPa. Moreover, brane fouling (Drews, 2010). More irregular shaped sludge flocs might
SADm was maintained at of 0.6 m3 h−1 m−2 and the TMP was almost be produced due to bulky sludge which facilitates the development of
stable (stayed around 3 kPa until the end of the experiment). In this highly dense cake layer on the membrane surface, thus resulting in a
context, it was reported that when wastewater temperature decreases, more significant fouling phenomenon (Zhang et al., 2014). On the
the mixed liquor viscosity increases, resulting in the rapid building of other hand, the MLSS concentration was found to be decreasing from
fouling resistance, leading to the rise of TMP (van den Brink et al., about 9800 mg L−1 at the beginning to 5600 mg L−1 when the sludge
2011; Zhang et al., 2014). Membrane fouling at low temperatures can temperature dropped by approximately 10 °C (from 20 °C to 10 °C).
also be accelerated due to the reduced shear stress proclaimed by course However, the MLSS concentration remained almost stable when the
air bubbling, low particle back transport velocity and high hydrophobic- sludge temperature was about 8 °C to 12 °C. The results are consistent
ity of the mixed liquor (Zhang et al., 2014). In addition, the pore size can with previous study where reduced biomass concentrations is reported
shrink in the low temperatures due to intrinsic characteristics of mem- during low temperature operations (Zhang et al., 2014). Under the low
branes. At low temperatures, a higher amount of submicron particles temperatures less energy is released from the oxidation of organic mat-
are found in the mixed liquor composition, which enhances the rapid ter, which lowers the rate of cell synthesis and thus deteriorate the
fouling by pore blocking or narrowing of the membrane pores, however, sludge yield (Zhang et al., 2014). Similar observations reported by
cake layer formation becomes dominant after the initial fouling (van Krzeminski et al. (2012) revealed that sludge concentrations are highly
den Brink et al., 2011). reduced in winter than in summer in a full-scale MBR plant.
At the beginning, the permeability was about 345 L m−2 h−1 bar−1, The treated water quality regarding the removal of TSS, organic mat-
and the correlation with temperature was not so clear. During the 1st ter and inorganic nutrients (NH4-N and TP) is shown in Table 2. Despite
peak of fouling, the permeability dropped to 50 L m− 2 h− 1 bar− 1. of low temperature operation, high quality treated water was achieved.
After the 1st membrane cleaning, the permeability increased back to 100% removal of TSS was accomplished, thus revealing an excellent re-
130 L m−2 h−1 bar−1, but rapidly declined to 75 L m− 2 h− 1 bar− 1 tention capability from inert suspended solids by the micro-filtration
until the 2nd peak of fouling. About 75% of permeability drop was ob- (MF) membranes. About 93% of COD removal was achieved with the ef-
served during the period with the lowest temperatures (from 9th fluent concentration of 18.7 ± 2.9 mg L−1. The present work is in good
March to 11th April), as compared to the initial permeability. About agreement with a previous study (Arévalo et al., 2014), in which no sig-
50% of permeability differences were reported between summer and nificant effect was observed on the removal of COD under a wide tem-
winter periods in the full scale MBRs (van den Brink et al., 2011). After perature variation (9–33 °C), when full-scale MBR was operated at
the manual cleaning of membranes, the permeability was restored SRT and HRT of 35 days and 35 h, respectively. High removal rate of
gradually and reached nearly to 300 L m−2 h−1 bar−1, though unable NH4-N (nearly 100%) during the experimental period indicated an effi-
to regain the initial value. The results showed that the accelerated mem- cient nitrification process in the MBR system, regardless of low temper-
brane fouling due to the low temperatures deteriorated the hydraulic ature conditions. Higher nitrification under low temperature operation
performance (permeability) of membranes. Therefore, extra was expected due to the relatively longer SRT maintained during the ex-
perimental period. Maximum growth and accumulation of nitrifying
bacteria (nitrifiers) are achieved when MBR is operated at relatively
longer SRT of 35 days which offsets slow growth of nitrifiers due to
low temperature, thus leading to excellent nitrification (Arévalo et al.,
2014). Furthermore, the average removal rate of TP reached nearly
91% with the concentration of 0.43 mg L−1 in permeate, even though
no chemical coagulants were added to the MBR system. Therefore, en-
hanced biological phosphorus removal process is expected at longer
SRT during MBR operation due to the enrichment of some phosphate-
accumulating organisms (PAOs).
3.2. Removal of human viruses: Noroviruses (NoV GI and NoV GII), and Ad-
enoviruses (AdVs)
NoVs GI and GII, and AdVs were consistently detected in all the influ-
ent samples, and their occurrences are elucidated in Fig. 3. The detection
rates varied depending on varieties of samples and behaviors of differ-
ent viruses types (Miura et al., 2015). In the present work, the detected
Fig. 2. Evolution of MLSS concentrations and SVI in MBR process. but not quantified virus concentrations were assigned as positive results
Table 2 2006; Simmons et al., 2011). In this study, NoV GI was detected under
Basic performance of MBR process regarding the removal of main pollution load. limit of detection in the mixed liquor, whereas NoV GII was observed
Parameter Influent concentration Effluent concentration Removal rate at very low concentrations. The log removal of NoV GI and NoV GII
[mg L−1] [mg L−1] [%] was about 0.5 log units and 1.7 log units, respectively in the mixed li-
TSS 126.9 ± 19.6 b1 100 quor. One reason might be attributed to the presence of inhibitors in
COD 265.6 ± 53.1 18.7 ± 2.9 92.8 ± 1.4 the samples, which may affect the sensitivity of the analytical method
NH4-N 45.9 ± 9.2 b0.02 N99.96 or even cause false-negative results (Schrader et al., 2012). In this
TP 4.7 ± 0.8 0.43 ± 0.1 90.6 ± 2.4
study, the influent to the MBR was fed from primary settling tank of
the main process (where returned sludge was recycled periodically),
which might have traces of calcium ions (inorganic inhibitors). The
for viruses, except those under the limit of detection (i.e. ND: not presence of organic inhibitors such as polysaccharides, humic acids, pro-
detected). teins (Schrader et al., 2012), which are highly available during cold tem-
NoV GI was detected in 85% of the influent samples with an average perature (Zhang et al., 2014), might have effects on PCR. On the other
concentration of 95 GC mL−1 and high positive results during February hand, AdVs were significantly higher in the mixed liquor samples than
and March. NoV GI in the mixed liquor samples were below the limit of in the influent, showing their high adsorption affinity towards mixed li-
quantification with an average concentration of 48 ± 28 GC mL−1. NoV quor (Kuo et al., 2010). However, as the surface characteristics of viruses
GI was detected only in two permeate samples, during February and varies greatly, therefore their adsorption to the complex biomass may
April. The log removal of NoV GI was ranging from N0.67 to N 2.73 vary too (Miura et al., 2015). It is likely that viruses which are adsorbed
(1.82 ± 0.6) log units. Similarly, NoV GII was constantly detected in to sludge more efficiently removed as waste sludge in the MBR pilot
100% of the influent samples with an average concentration of process (Miura et al., 2015). This indicates that even under the low tem-
896 GC mL− 1 having gradually increasing trend during January to perature conditions, appropriate control of MLSS concentration in MBR
March, and then declining tendency during April. NoV GII in the process can enhance efficient virus removal. Furthermore, differences in
mixed liquor samples were mostly below limit of quantification with the log removal efficiencies related to the morphology of viruses was
an average concentration of 25 ± 44 GC mL−1. In permeate, NoV GII not clearly observed. The removal of NoV GII (27–40 nm) was found
was detected in three samples with an average concentration of 28 ± to be higher than the AdVs (90–100 nm). This result is contradictory
37 GC mL−1. The log removal of NoV GII was achieved in the range of with the findings of Simmons et al. (2011), where higher removal of
N2.1 to N 3.62 ( 3.02± 0.5) log units. On the other hand, AdVs were de- human adenoviruses was achieved compared to NoV GII in MBR pro-
tected in 100% of the influent samples with an average concentration cess. This tendency could be explained by the variations in influent con-
124 GC mL−1. The concentrations of AdVs varied consistently through- centrations in the one hand, and that of the limits of detection on the
out the low temperature period. AdVs were highly detected in the other hand. The combination of those variations might lead to higher
mixed liquor with an average concentration of 196 GC mL−1. Interest- or lower removal efficiency of viruses during the calculation
ingly, only one positive sample was detected for AdVs (on 23rd March) (Simmons et al., 2011).
from the permeate sample. The log removal of AdVs ranged from In the present study, NoV GI, NoV GII and AdVs were detected in the
N0.76 to N2.76 (1.94 ± 0.7) log units. MBR permeate, though the occurrence was not consistent, which con-
The tendency of variations monitored for NoVs and AdVs in the in- firms that MBR is not an absolute barrier for retaining virus passage
fluent wastewater during winter period is consistent to the previous (Simmons et al., 2011). During this research work, variations in the
studies (Kuo et al., 2010; Miura et al., 2015; Simmons et al., 2011). In membrane fouling degree were observed with the TMP ranging from
this study, the overall removal efficiencies of NoV GI (N 0.67 to N2.73 2 to 18 kPa. Despite the cleaning of membranes, the overall removal ef-
log unit), NoV GII (N 2.1 to N 3.62 log units), and AdVs (N0.76 to N2.76 ficiency of viruses in the MBR process did not change. Moreover, pas-
log units) 'were achieved'. Miura et al. (2015) reported the log reduc- sage of all three viruses were observed in the permeate samples
tions of N 0.2 to N3.4 log units for NoV GII. Ottoson et al. (2006) reported before or after the membrane fouling peaks. These results suggested
mean log removal of 1.0 to 1.1 log units for NoVs in pilot-scale MBR with that the virus removal efficiency was not related to the degree of mem-
a 0.45 μm pore size membranes. Sima et al. (2011) reported the log re- brane fouling, which is in good agreement with the previous study
moval of 0.9 to ≥6.8 log units for NoVs in full-scale MBR treatment. Sim- (Miura et al., 2015). Further research would be needed for hourly or
ilarly, in the full-scale MBR with nominal pore size (0.04 μm) daily samples before and after the cleaning of membranes to determine
membranes, Simmons et al. (Simmons et al., 2011) reported the average the precise correlation between fouling and virus removal.
log removals of 3.0 and 4.7 log units for AdVs and NoV GII, respectively. Furthermore, the average concentrations of the Escherichia coli and
Virus adsorption to mixed liquor plays an important role in the re- enterococci were observed in the influent wastewater in the range of
moval of viruses by MBR process (Miura et al., 2015; Ottoson et al., N4.8 × 105 MPN 100 mL−1 and 1.9 × 104 to N 4.8 × 105 MPN
Fig. 3. Concentrations of NoV GI, NoV GII, and AdVs. Black triangles - influent; grey diamonds – mixed liquor; white squares - permeate; dashed line - Limit of Quantification; ND - below
Limit of Detection.
−1
100 mL , respectively. However, none of them were detected in the Table 3
MBR permeate over the experimental period. Nonetheless, the high Mean concentrations (±SD) and Mass balances of selected TrOCs in MBR process.
concentrations of Escherichia coli (1.6 × 103 to 8.05 × 105 MPN Selected TrOCs
100 mL−1) and enterococci (1.07 × 103 to 4.3 × 104 MPN 100 mL−1) CBZ BIS DCF BPA
were detected in the sludge supernatant. The average bacteria removal
Treated water flow (L d−1) 2520
efficiency of the MBR process during low temperature period was N 5.6
Wasted sludge (kg d−1) 0.80
log units, which is in the good agreement with other study (Ottoson LOQ aqueous (μg L−1) 0.05 0.02 0.1 0.05
et al., 2006). LOQ sludge (μg kg−1) 0.1 0.05 0.2 0.1
Influent concentration (μg L−1) 0.63 0.43 1.84 0.95
± 0.2 ± 0.2 ± 0.4 ± 0.2
Permeate concentration (μg L−1) 0.77 0.14 1.12 b0.05
3.3. Fate of organic micropollutants: TrOCs ± 0.1 ± 0.1 ± 0.4
Mixed liquor concentration 0.88 0.17 2.11 0.17
The occurrence and removal efficiencies of CBZ, BIS, DCF, and BPA in (μg kg−1) ± 0.2 ± 0.1 ± 0.6 ± 0.1
the MBR process are shown in Fig. 4a–b. All the selected TrOCs were de- Overall Elimination in MBR −353 730 1814 2268
(μg d−1)
tected in the influents throughout the experimental period, although
Elimination due to sorption 0.71 0.14 1.70 0.14
the relatively higher concentrations were measured from January up (μg d−1)
to the end of March as shown in Fig. 4a. The mean influent concentra-
tions of CBZ, BIS and BPA were 0.63, 0.43, and 0.95 μg L−1, respectively.
Similarly, for DCF, the concentration was 1.84 ± 0.4 μg L−1, which was
the highest among the other refractory compounds during cold season.
must be interpreted as rough estimates as they were calculated from
The mean concentrations of CBZ, BIS, DCF and BPA in the MBR per-
the averaged values considering a completely mixed and steady state.
meate were 0.77, 0.14, 1.12, and b0.05 μg L− 1, respectively. Most of
Table 3 clearly elucidates great variations between the overall elimina-
the TrOCs were found comparatively low in the permeate than in the in-
tion and the elimination due to sorption regardless of the compounds.
fluent (excluding CBZ), indicating the enhanced efficiency of the MBR
For all the selected TrOCs, a very weak removal via biomass sorption
process to remove many emerging TrOCs from the real wastewaters.
was observed. Thus, the removal mechanism for BIS, DCF and BPA was
TrOCs removal efficiency differ from plant to plant and even country
attributed primarily due to the biotransformation of TrOCs in the MBR
to country as it highly depends on wastewater characteristics and spe-
process. Similar studies on MBR demonstrated that sorption is not the
cific operating procedures implemented by particular treatment pro-
dominant mechanism, but biotransformation is crucial in the elimina-
cesses (Tambosi et al., 2010). The overall removal rates in this study
tion of TrOCs regardless of the treatment conditions (Clara et al., 2005;
varied largely among the individual TrOCs. The mean removal efficien-
Kimura et al., 2007; Trinh et al., 2016). The biological elimination of
cies of CBZ, BIS, DCF and BPA achieved during low temperatures
TrOCs could be due to the direct metabolization or by co-metabolization
(interpreted as monthly average) are shown in Fig. 4b. Negative remov-
where bacteria use TrOCs as direct carbon source, or just break them
al efficiencies of CBZ were mostly observed, except a relatively low effi-
down via specific enzymes which having potential to degrade or trans-
ciency of about 9% and 28% during last two weeks of March. CBZ
form trace organic molecules (Vieno and Sillanpää, 2014). It is observed
removal rates varied from − 89 to 28% (average of − 25%) with the
that membrane filtration alone does not allow any further retention of
highest negative efficiency during January–February. The average re-
TrOCs due to the size exclusion (Clara et al., 2005).
moval efficiencies of BIS and DCF were 65 ± 6% and 38 ± 7%, respective-
ly. Decreasing tendencies of the removal rates were observed both for
BIS and DCF from Jan–Feb to April. On the other hand, BPA concentra- 3.3.1. CBZ removal
tions in the permeate were significantly below b0.05 μg L−1 (b LOQ) re- In the present study, the MBR unit showed insufficient elimination
gardless of low temperature operation, thus substantially stable BPA of CBZ with high concentrations in permeate. In most of the permeate
removal efficiency of N 97% (N 95 to N 98%) was achieved. samples, the concentrations of CBZ were observed even higher than in
A mass balance of TrOCs in MBR was performed (Table 3) according the influent samples. Moreover, many researchers have reported CBZ
to the method described by Kimura et al. (2007), where quantity of the as ‘non-removal’ compounds in MBRs by documenting even negative
TrOCs adsorbed on mixed liquor solids are measured and compared removal cases (Clara et al., 2005). The increased amount of CBZ in per-
with the eliminated quantity estimated by considering the influent meate samples were most likely due to the conversion of CBZ glucuro-
minus the permeate concentrations. The presented mass balances nides and other conjugated metabolites to the parent compound via
Fig. 4. Concentrations of CBZ, BIS, DCF and BPA in influents during low-temperature periods [a]; removal efficiencies achieved for selected TrOCs (monthly average) [b].
enzymatic cleavage in the activated sludge process in MBRs (Vieno et trace metals were detected in the collected samples with an occurrence
al., 2007). The small percentage of CBZ removal during the MBR process frequency of 100%. Concentrations of trace metals in the influent were
is associated with retention by the sludge and elimination during sludge in the range of 0.1 to 80 μg L−1 (only dissolved fractions), which are
wasting (González-Pérez et al., 2016). In addition, large concentrations consistent with the range of trace metal concentrations documented
of CBZ found in mixed liquors and permeate suggested the recalcitrant by other researchers (Di Fabio et al., 2013). With an average concentra-
behavior of CBZ to the biological transformation or biodegradation tion of 75 μg L−1, Zn was the most dominant, whereas Cd was the lowest
(González-Pérez et al., 2016). The poor removal of CBZ is partly due to (0.1 μg L−1) detected among other trace metals. On the other hand, low
its hydrophilic characteristics (Log Kow b 2.5) and chemical stability concentrations of Cr, Pb and Cd in influent wastewater explained that
(González-Pérez et al., 2016). no industrial or mining wastewaters are channeled with the
municipal wastewater. The relative abundance of trace metals in
3.3.2. DCF removal influent and permeate samples followed the general order of
The average influent concentration of DCF (1.12 μg L−1) was higher Zn N Ni N Co ≥ Cu N V N Cr N As ≥ Pb N Cd. Similar order of abundances
than the values (0.23 to 1 μg L−1) reported by other researchers in mu- of trace metals have been reported by other researchers in common
nicipal sewage treatment plants. In this study, the partial removal effi- municipal wastewaters (Chipasa, 2003; Karvelas et al., 2003). The con-
ciency of diclofenac varied from 13% to 89% (average 38%), which is centrations of Zn, Ni, Cu, and Co were highly fluctuating in the permeate
similar to the results in the previous studies performed during the win- samples, whereas the concentrations of V, Cr, As, Pb, and Cd were almost
ter season (Clara et al., 2005; Trinh et al., 2016). In lab-scale MBR under stable.
SRTs of 30–90 days at varying temperatures (8 to 12 °C), Kruglova et al. The removal efficiencies of trace metals in MBR process varied with
(2016) reported substantially slow biological degradation of DCF. Like- respect to each element. An average removal of N80% was achieved for
wise, the elimination of DCF was higher in Jan–Feb than March–April. Cd, Pb, and V, whereas about 60% removal was registered for Zn and Co.
One reason might be the mixed liquor temperature being higher around On the other hand, between 30% and 50% of Ni, Cu, As, and Cr were re-
15 °C during Jan–Feb. The other reason could be the high MLSS concen- moved under the same operating conditions. Similar results are report-
trations (approximately 8300 mg L−1) in the MBR process. In March, ed by other researchers (Bolzonella et al., 2010). In fact, the removal
the process temperature was even below 10 °C for more than a week. efficiencies are largely dependent on the initial concentrations of
Longer SRTs allow high microbial population and biomass concentra- metals. In the present work, since the feeding source of wastewater
tions and thus would enhance the biological interactions with the struc- was the effluent from primary clariflocculators of WWTP, the removal
turally-complex pharmaceuticals (Kimura et al., 2007). Majority of the of some metals is expected to start before reaching MBR system.
studies reported in the literature revealed higher removal performance When a raw wastewater undergoes clariflocculation, a portion of trace
of DCF with longer SRTs in MBR processes. However, it is worth noticed metals is abated by adsorption onto solid particles (Chipasa, 2003; Di
that some contradictory results with no correlation between the elimi- Fabio et al., 2013). Moreover, the concentrations of trace metals in per-
nation of DCF and SRT were also reported (Vieno and Sillanpää, 2014). meate were found to be far lower than the maximum allowable concen-
trations reported in the water quality guidelines.
3.3.3. BIS removal Trace metals removal in MBRs can occur via various processes in-
The present work is probably the only study performed so far on BIS cluding precipitation of metals inside bioreactor, biosorption of metal
removal in MBR operated at low temperature conditions. The removal ions on sludge or bacterial flocs, and retention of the insoluble metal
efficiency of BIS varied over the investigation period with higher remov- species by membranes (Bolzonella et al., 2010). Trace metals in waste-
al rates observed during Jan–Feb (73%) than in March (64%) and April water are present in different phases (i.e. dissolved or particulate)
(55%), during which sludge temperature varied approximately from with different water solubility at pH of 7–9 (Karvelas et al., 2003). The
20 °C to 12 °C. The explanation for the higher removal of BIS in MBR pro- pH critically affects the metal solubility, the surface functional groups
cess might be attributed to the higher biodegradability at higher tem- of bacteria, and the competition of cations and protons (Hammaini et
perature and vice-versa. al., 2007; Malamis et al., 2011). In the present study, the pH inside the
biological reactors was maintained at 6.6–7, therefore, both precipita-
3.3.4. BPA removal tion and biosorption mechanisms were expected for the removal of
Relatively higher and stable removal of BPA (N97%) was achieved as trace metals. Katsou et al. (2011) reported an average removal of
compared to other compounds regardless of low temperature condi- 80%–98% for Cu and Pb and to a lower extent 50%–77% for Ni and Zn
tions. The average BPA removal efficiency of N 90% was documented by precipitation at pH 7–9. Malamis et al. (2011) has reported that be-
by (Clara et al., 2005; Trinh et al., 2016) in the MBR processes during yond pH 6 the sludge biosorption rate is decreasing and chemical pre-
cold periods of the year, which is consistent with our result. Due to its cipitation becomes dominant mechanism for metal removal. The
high hydrophobicity (Log D N 3), BPA might have a tendency to sorb gradual increase in pH generates more OH− ions in the mixture which
on suspended solids of mixed liquor matrix and subsequently be re- form complex precipitates with metals (Fu and Wang, 2011). A signifi-
moved with the withdrawal of excess sludge (Cirja et al., 2008). Howev- cant metals removal via biosorption can be suggested for the present
er, about 95% of BPA was found to be eliminated via biotransformation, work, since the precipitation mechanism should be dominant (and
relative to influent load in this study. Thus, biotransformation is the much faster than biosorption) in the pH of range 8–11 (Fu and Wang,
dominant pathway of BPA removal. One reason for the poor elimination 2011). Comte et al. (Comte et al., 2008) reported more favorable
of BPA via sorption process in MBR process might be due to its very biosorption of metals at pH range of 6–9, in which activated sludge is
polar nature (Cirja et al., 2008). Biotransformation was also reported negatively charged and provides potential sorption sites for metals.
as the main pathway for BPA removal by other researchers (Trinh et However, dominant metals removal mechanism via biosorption is ob-
al., 2016). served at pH 4 for Pb and 6 for Cd, Cu and Zn when sludge is employed
as sorbent (Hammaini et al., 2007; Malamis et al., 2011).
3.4. Fate of inorganic micropollutants: trace metals Higher concentrations of Zn and Ni spotted in permeate might be at-
tributed to the metal ions penetrated through the membranes due to
3.4.1. Concentrations of trace metals and removal efficiencies their high solubility at aforementioned pH, which is consistent with
Concentrations of the selected trace metals in the influent and per- the observation reported previously (Katsou et al., 2011). However, in
meate, corresponding removal efficiencies, and the maximum allowable addition to the solubilization, these observations can be partially attrib-
concentrations (MAC) according to Water Quality Guidelines or Stan- uted to the unstable forms of Zn and Ni. On the other hand, removal of
dards set by international organizations are shown in Table 4. All the Cr, Pb, As, Cd, As, and Co remained relatively high which was possibly
Table 4
Occurrence and removal of 9 trace metals in the MBR pilot plant (n = 42); comparison of permeate concentrations with Water Quality Guidelines or Standards set by international
organization.
Metal Influent concentration (average Permeate concentration (average Removal efficiency (average Maximum allowable Reference
± Sd) ± Sd) ± Sd) concentration
[μg L−1] [μg L−1] [%] [μg L−1]
Zn 75.86 ± 17.5 31.43 ± 14.3 59.6 ± 11.6 3000 (WHO, 2011)a

Ni 14.93 ± 3.9 9.75 ± 1.8 33.7 ± 6.1 34 (EU, 2013)b
Co 6.66 ± 2.3 2.6 ± 1.4 60.9 ± 17.7 Not mentioned
Cu 6.19 ± 3.5 3.87 ± 2.0 31.3 ± 32.6 2000 (WHO, 2011)a
V 2.73 ± 1.0 0.35 ± 0.1 85.7 ± 5.8 Not mentioned
Cr 1.64 ± 0.7 0.81 ± 0.2 47.5 ± 9.6 50 (WHO, 2011)a
As 0.68 ± 0.2 0.34 ± 0.1 45.7 ± 18.9 10 (WHO, 2011)a
Pb 0.65 ± 0.3 0.11 ± 0.1 80 ± 10.9 14 (EU, 2013)b
Cd 0.1 ± 0.1 0.01 ± 0.0 86.3 ± 7.0 1.5 (EU, 2013)b
a
Guidelines for drinking-water quality.
b
2013/39/EU directive of the European parliament and of the council.
due to the low pH required for solubilizing and their availability in sta- point-source of discharges (Karvelas et al., 2003). The frequency of oc-
ble form within the given pH. Enhanced removal of trace metals is re- currence of all trace metals in the sludge samples was 100% and same
ported with the increasing MLSS concentration as more sites are for the liquid samples. From the analytical results, Zn was the most
available for the sorption of soluble metal ions (Katsou et al., 2011; abundant metal in sludge, whereas Cd exhibited the lowest abundance.
Malamis et al., 2011). On the contrary, Hammaini et al. (2007) reported However, the concentrations of Zn, Ni, Cu, Pb, and Cd were found to be
relatively steady or reduced metal sorption capacity with the increase in under the allowable concentrations specified in the EU standard. As no
MLSS concentration due to the screen effect between cells. limits are set for Co, Cr, As by EU so far, their concentrations were com-
Promising removal efficiencies of trace metals by different types of pared with US regulations and also found under the MAC (the concen-
membranes such as ultra-filtration, reverse osmosis, nanofiltration tration threshold for V was not mentioned).
and electrodialysis have been reported (Fu and Wang, 2011; Katsou et
al., 2011). However, the higher removal efficiencies of most of the 4. Conclusion
metals in the present study was attributed to the complete retention
of TSS via MF membranes, which is in agreement with a previous In this study, the performance of pilot-scale MBR unit, operated at
study (Bolzonella et al., 2010). Furthermore, a substantial removal of low temperature conditions in Nordic regions, was investigated for
trace metals is reported by a research study (Malamis et al., 2011), in the treatment of real municipal wastewater. An accelerated membrane
which MF membrane was combined with activated sludge. On the fouling was observed during low temperature periods and while main-
other hand, an enhanced biosorption of trace metals on the problematic taining aeration intensity below the normal value. This resulted in a
biofilm developed on membrane surface (fouling layer) than to the substantial decrease in membrane permeability (about 75%), leading
sludge flocs is reported in pilot-scale MBR (Di Fabio et al., 2013). Even to an increased frequency of membrane cleaning. The average reduc-
though fouling has negative influence on membrane permeability, a tions of 1.82, 3.02, and 1.94 log units were achieved for norovirus GI,
fraction of metals removal could be originating from membrane fouling. norovirus GII, and adenoviruses, respectively. Removal of Escherichia
coli and enterococci was N5.6 log units. The average removal efficiencies
3.4.2. Trace metals in sludge (mixed liquor) of BIS, DCF and BPA were 65%, 38%, and N 97%, respectively. However,
Average trace metal concentrations in activated sludge of MBR pro- the MBR pilot did not show efficient removal of carbamazepine.
cess are summarized in Table 5. To assess the toxicity of excess sludge Overall, and based on European and American regulations, high
from MBR process, it is very important to monitor trace metal accumu- quality permeate was achieved, demonstrating the aptitude of MBR sys-
lation in the bioreactor for the long term functioning (Bolzonella et al., tem to perform satisfactory wastewater treatment during the winter
2010). Trace metals in sludge are present in the several forms including season in Nordic regions and reclaim treated water and sludge for the
metal precipitates in the sludge flocs, soluble metal ions, complexes of agricultural and industrial sectors.
soluble metals and accumulated soluble metal in the microbial cells
(Chipasa, 2003). Trace metals are non-biodegradable compounds, thus Acknowledgements
the total trace metal loading of feed wastewater will either end up in
the sludge or remain in the treated effluent. In fact, trace metals released This work was supported by TEKES-Finnish (1043/31/2015) funding
via excess sludge is more crucial than through treated effluent to the agency for innovation, Finland, the research fund for SMART EFFLUENTS
Table 5
Average trace metal concentrations in sludge with maximum allowable concentrations as per International Standards.
Metals Average ± Sd Maximum allowable concentration Reference

[mg kg−1 dry weight sludge] [mg kg−1 dry weight sludge]
Zn 600.7 ± 89.3 2500–4000 (EU, 1986)

Ni 29.5 ± 4.5 300–400 (EU, 1986)
Co 24.9 ± 3.7 77 (Harrison et al., 1999)
Cu 130.7 ± 13.1 1000–1750 (EU, 1986)
V 30.9 ± 4.8 Not mentioned
Cr 45.4 ± 6.9 75 (Harrison et al., 1999)
As 3.8 ± 0.8 41 (Harrison et al., 1999)
Pb 8.8 ± 1.3 750–1200 (EU, 1986)
Cd 0.6 ± 0.1 20–40 (EU, 1986)
project. The authors like to thank National Institute for health and wel- Katsou, E., Malamis, S., Loizidou, M., 2011. Performance of a membrane bioreactor used
for the treatment of wastewater contaminated with heavy metals. Bioresour.
fare (THL), Finnish environment services (SYKE) and Mikkeli WWTP for Technol. 102:4325–4332. http://dx.doi.org/10.1016/j.biortech.2010.10.118.
their support on laboratory works. Especial thanks go to all the staff of Kimura, K., Hara, H., Watanabe, Y., 2007. Elimination of selected acidic pharmaceuticals
Mikkeli WWTP, Finland. from municipal wastewater by an activated sludge system and membrane bioreac-
tors. Environ. Sci. Technol. 41:3708–3714. http://dx.doi.org/10.1021/es061684z.
Kruglova, A., Kråkström, M., Riska, M., Mikola, A., Rantanen, P., Vahala, R., Kronberg, L.,
References 2016. Comparative study of emerging micropollutants removal by aerobic activated
sludge of large laboratory-scale membrane bioreactors and sequencing batch reactors
(FIMEA), F.M.A, 2015. Kuluvan vuoden lääkekulutus (2015). pp. 1–57. under low-temperature conditions. Bioresour. Technol. 214:81–88. http://dx.doi.org/
American Public Health Association, American Water Works Association, Water 10.1016/j.biortech.2016.04.037.
Environment Federation, 1999. Standard methods for the examination of water and Krzeminski, P., Iglesias-Obelleiro, A., Madebo, G., Garrido, J.M., van der Graaf, J.H.J.M., van
wastewater. Stand. Methods 541. Lier, J.B., 2012. Impact of temperature on raw wastewater composition and activated
Arévalo, J., Ruiz, L.M., Pérez, J., Gómez, M.A., 2014. Effect of temperature on membrane sludge filterability in full-scale MBR systems for municipal sewage treatment.
bioreactor performance working with high hydraulic and sludge retention time. J. Membr. Sci. 423–424:348–361. http://dx.doi.org/10.1016/j.memsci.2012.08.032.
Biochem. Eng. J. 88:42–49. http://dx.doi.org/10.1016/j.bej.2014.03.006. Kuo, D.H.W., Simmons, F.J., Blair, S., Hart, E., Rose, J.B., Xagoraraki, I., 2010. Assessment of
Bolzonella, D., Fatone, F., di Fabio, S., Cecchi, F., 2010. Application of membrane bioreactor human adenovirus removal in a full-scale membrane bioreactor treating municipal
technology for wastewater treatment and reuse in the Mediterranean region: focus- wastewater. Water Res. 44:1520–1530. http://dx.doi.org/10.1016/j.watres.2009.10.
ing on removal efficiency of non-conventional pollutants. J. Environ. Manag. 91: 039.
2424–2431. http://dx.doi.org/10.1016/j.jenvman.2010.07.024. Malamis, S., Katsou, E., Haralambous, K.J., 2011. Study of Ni (II), Cu (II), Pb (II), and Zn (II)
Chipasa, K.B., 2003. Accumulation and fate of selected heavy metals in a biological waste- Removal Using Sludge and Minerals Followed by MF/UF 81–92. http://dx.doi.org/10.
water treatment system. Waste Manag. 23:135–143. http://dx.doi.org/10.1016/ 1007/s11270-010-0625-4.
S0956-053X(02)00065-X. Metcalf & Eddy Inc., 2003. Wastewater Engineering: Treatment and Reuse [WWW Docu-
Cirja, M., Ivashechkin, Æ.P., Corvini, Æ.P.F.X., 2008. Factors affecting the removal of organic ment]. URL https://www.bookdepository.com/Wastewater-Engineering-Metcalf-
micropollutants from wastewater in conventional treatment plants (CTP) and mem- Eddy-Inc/9780071122504 (accessed 2.23.16).
brane bioreactors (MBR). Rev. Environ. Sci. Biotechnol. 61–78. http://dx.doi.org/10. Miura, T., Okabe, S., Nakahara, Y., Sano, D., 2015. Removal properties of human enteric vi-
1007/s11157-007-9121-8. ruses in a pilot-scale membrane bioreactor (MBR) process. Water Res. 75:282–291.
Clara, M., Strenn, B., Gans, O., Martinez, E., Kreuzinger, N., Kroiss, H., 2005. Removal of se- http://dx.doi.org/10.1016/j.watres.2015.02.046.
lected pharmaceuticals, fragrances and endocrine disrupting compounds in a mem- Ottoson, J., Hansen, A., Björlenius, B., Norder, H., Stenström, T.A., 2006. Removal of viruses,
brane bioreactor and conventional wastewater treatment plants. Water Res. 39: parasitic protozoa and microbial indicators in conventional and membrane processes
4797–4807. http://dx.doi.org/10.1016/j.watres.2005.09.015. in a wastewater pilot plant. Water Res. 40:1449–1457. http://dx.doi.org/10.1016/j.
Comte, S., Guibaud, G., Baudu, M., 2008. Biosorption properties of extracellular polymeric watres.2006.01.039.
substances (EPS) towards Cd, Cu and Pb for different pH values. J. Hazard. Mater. 151: Park, H.-D., Chang, I.-S., Lee, K.-J., 2015. Principles of Membrane Bioreactors for Wastewa-
185–193. http://dx.doi.org/10.1016/j.jhazmat.2007.05.070. ter Treatment [WWW Document]. CRC Press. URL https://www.crcpress.com/
Di Fabio, S., Lampis, S., Zanetti, L., Cecchi, F., Fatone, F., 2013. Role and characteristics of Principles-of-Membrane-Bioreactors-for-Wastewater-Treatment/Park-Chang-Lee/
problematic biofilms within the removal and mobility of trace metals in a pilot- 9781466590373 (accessed 3.31.16).
scale membrane bioreactor. Process Biochem. 48:1757–1766. http://dx.doi.org/10. Schrader, C., Schielke, A., Ellerbroek, L., Johne, R., 2012. PCR inhibitors – occurrence, prop-
1016/j.procbio.2013.08.005. erties and removal. J. Appl. Microbiol. 113:1014–1026. http://dx.doi.org/10.1111/j.
Drews, A., 2010. Membrane fouling in membrane bioreactors—characterisation, contra- 1365-2672.2012.05384.x.
dictions, cause and cures. J. Membr. Sci. 363:1–28. http://dx.doi.org/10.1016/j. Sima, L.C., Schaeffer, J., Saux, J.-C.L., Parnaudeau, S., Elimelech, M., Guyader, F.S.L., 2011.
memsci.2010.06.046. Calicivirus removal in a membrane bioreactor wastewater treatment plant. Appl. En-
EPA, 2007. Method 1694: Pharmaceuticals and Personal Care Products in Water, Soil, Sed- viron. Microbiol. 77:5170–5177. http://dx.doi.org/10.1128/AEM.00583-11.
iment, and Biosolids by HPLC/MS/MS. Simmons, F.J., Kuo, D.H.-W., Xagoraraki, I., 2011. Removal of human enteric viruses by a
EU, 1986. Council Directive 86/278/EEC on the protection of the enviornment, and in par- full-scale membrane bioreactor during municipal wastewater processing. Water
ticular of the soil, when sewage sludge is used in agriculture. Off. J. Eur. Communities: Res. 45:2739–2750. http://dx.doi.org/10.1016/j.watres.2011.02.001.
6–12 [http://eur-lex.europa.eu/legal-content/EN/TXT/?uri=celex%3A31986L0278]. Suda, K.J., Hicks, L.A., Roberts, R.M., Hunkler, R.J., Taylor, H., 2014. Trends and Seasonal
EU, 2013. EU. Directive 2013/39/EU of the European Parliament and of the Council. 2013: Variation in Outpatient Antibiotic Prescription Rates in the United States, 2006 to
pp. 1–17 [http://eur-lex.europa.eu/legal-content/en/ALL/?uri=CELEX%3A32013L0039]. 2010. 58:pp. 2763–2766. http://dx.doi.org/10.1128/AAC.02239-13.
Fontmorin, J.-M., Sillanpää, M., 2015. Bioleaching and combined bioleaching/Fenton-like Tambosi, J.L., de Sena, R.F., Favier, M., Gebhardt, W., José, H.J., Schröder, H.F., de F.P.M.
processes for the treatment of urban anaerobically digested sludge: removal of Moreira, R., 2010. Removal of pharmaceutical compounds in membrane bioreactors
heavy metals and improvement of the sludge dewaterability. Sep. Purif. Technol. (MBR) applying submerged membranes. Desalination 261:148–156. http://dx.doi.
156 (Pt 2):655–664. http://dx.doi.org/10.1016/j.seppur.2015.10.061. org/10.1016/j.desal.2010.05.014.
Fu, F., Wang, Q., 2011. Removal of heavy metal ions from wastewaters: a review. Trinh, T., van den Akker, B., Stuetz, R.M., Coleman, H.M., Le Clech, P., Khan, S.J., 2012. Re-
J. Environ. Manag. 92:407–418. http://dx.doi.org/10.1016/j.jenvman.2010.11.011. moval of trace organic chemical contaminants by a membrane bioreactor. Water Sci.
Gao, D.-W., Hu, Q., Yao, C., Ren, N.-Q., 2014. Treatment of domestic wastewater by an in- Technol.:1856–1864 http://dx.doi.org/10.2166/wst.2012.374.
tegrated anaerobic fluidized-bed membrane bioreactor under moderate to low tem- Trinh, T., van den Akker, B., Coleman, H.M., Stuetz, R.M., Drewes, J.E., Le-Clech, P., Khan,
perature conditions. Bioresour. Technol. 159:193–198. http://dx.doi.org/10.1016/j. S.J., 2016. Seasonal variations in fate and removal of trace organic chemical contam-
biortech.2014.02.086. inants while operating a full-scale membrane bioreactor. Sci. Total Environ. 550:
González-Pérez, D.M., Pérez, J.I., Nieto, M.Á.G., 2016. Carbamazepine behaviour and effects 176–183. http://dx.doi.org/10.1016/j.scitotenv.2015.12.083.
in an urban wastewater MBR working with high sludge and hydraulic retention time. van den Brink, P., Satpradit, O.-A., van Bentem, A., Zwijnenburg, A., Temmink, H., van
J. Environ. Sci. Health A:1–6 http://dx.doi.org/10.1080/10934529.2016.1181462. Loosdrecht, M., 2011. Effect of temperature shocks on membrane fouling in mem-
Gurung, K., Ncibi, M.C., Fontmorin, J., Särkkä, H., Sillanpää, M., 2016. Incorporating sub- brane bioreactors. Water Res. 45:4491–4500. http://dx.doi.org/10.1016/j.watres.
merged MBR in conventional activated sludge process for municipal wastewater 2011.05.046.
treatment: a feasibility and performance assessment. J. Membr. Sci. Technol. 6. Vieno, N., Sillanpää, M., 2014. Fate of diclofenac in municipal wastewater treatment plant
http://dx.doi.org/10.4172/2155-9589.1000158. — a review. Environ. Int. 69:28–39. http://dx.doi.org/10.1016/j.envint.2014.03.021.
Hammaini, A., González, F., Ballester, A., Blázquez, M.L., Muñoz, J.A., 2007. Biosorption of Vieno, N., Tuhkanen, T., Kronberg, L., 2007. Elimination of pharmaceuticals in sewage
heavy metals by activated sludge and their desorption characteristics. J. Environ. treatment plants in Finland. Water Res. 41:1001–1012. http://dx.doi.org/10.1016/j.
Manag. 84:419–426. http://dx.doi.org/10.1016/j.jenvman.2006.06.015. watres.2006.12.017.
Harrison, E.Z., Mcbride, M.B., Bouldin, D.R., 1999. Land application of sewage sludges?: an WHO, 2011. Guidelines for Drinking-water Quality. World Health Organ. [http://www.
appraisal of the US regulations. Int. J. Environ. Pollut. 11, 1–36. who.int/water_sanitation_health/publications/2011/dwq_guidelines/en/].
Karvelas, M., Katsoyiannis, A., Samara, C., 2003. Occurrence and fate of heavy metals in the Zhang, C., Wang, G., Hu, Z., 2014. Changes in wastewater treatment performance and ac-
wastewater treatment process. Chemosphere 53:1201–1210. http://dx.doi.org/10. tivated sludge properties of a membrane bioreactor at low temperature operation.
1016/S0045-6535(03)00591-5. Environ. Sci. Process. Impacts 16:2199–2207. http://dx.doi.org/10.1039/c4em00174e.
Publication II
Khum Gurung, Mohamed Chaker Ncibi, Mika Sillanpää

Removal and fate of emerging organic micropollutants (EOMs) in
municipal wastewater by a pilot-scale membrane bioreactor (MBR)
treatment under varying solid retention times

Vol. 667, pp. 671-680, 2019
© 2019 Elsevier
Science of the Total Environment 667 (2019) 671–680
journal homepage: www.elsevier.com/locate/scitotenv
Removal and fate of emerging organic micropollutants (EOMs) in

municipal wastewater by a pilot-scale membrane bioreactor (MBR)
treatment under varying solid retention times
Khum Gurung ⁎, Mohamed Chaker Ncibi, Mika Sillanpää
Department of Green Chemistry, School of Engineering Science, Lappeenranta-Lahti University of Technology, Sammonkatu 12, FI-50130 Mikkeli, Finland
H I G H L I G H T S G R A P H I C A L A B S T R A C T
• Fate of 23 EOMs in pilot MBR treatment

at varying SRT was studied.
• At long SRT, EOMs removal majorly en-
hanced while SMP and EPS concentra-
tion lessened.
• Hormones and antibiotics were highly
removed than diuretics and anti-
epileptics.
• Iopamidol and hydrocortisone (N90% re-
moval) were studied for the first time in
MBR.
• MBR Effluent fulfilled the water quality
requirements for reuse.
a r t i c l e i n f o a b s t r a c t
Article history: This study investigates the removal and fate of 23 emerging organic micropollutants (EOMs) including a wide
Received 29 December 2018 range of pharmaceuticals (antibiotics, β-blockers, analgesics, diuretics, psychostimulants, antiepileptics, immu-
Received in revised form 13 February 2019 nosuppressives, anticoagulants), and steroid hormones detected in municipal wastewater by a pilot-scale mem-
Accepted 19 February 2019
brane bioreactor (MBR) plant at two different solid retention times (SRTs) of 60 and 21 days. Different removal
Available online 20 February 2019
efficiencies of the selected EOMs were observed and explained based on their physicochemical properties (such
Editor: Paola Verlicchi as distribution coefficient, log D; dissociation constant, pKa; solid-water distribution coefficients, and Kd) along
with process operating parameters. The dominant removal mechanisms of EOMs were biotransformation and
Keywords: sorption onto the sludge, which were confirmed by the mass balance study. Moreover, changes in the sludge
Municipal wastewater treatment properties, as a consequence of different SRTs, were evaluated based on variations in soluble microbial products
Membrane bioreactor (SMP), extracellular polymeric substances (EPS), and capillary suction time (CST). Finally, the quality of the MBR
Emerging organic micropollutants effluent was compared with some established guidelines, which confirmed the fulfilment of water quality re-
Biotransformation/biodegradation quirements for reuse purposes.
Solid retention times
© 2019 Elsevier B.V. All rights reserved.
1. Introduction
In recent decades, the occurrence of emerging organic micropollutants

⁎ Corresponding author.
(EOMs) in the aquatic environment has received a global attention be-
E-mail addresses: khumgrg@yahoo.com, Khum.Gurung@lut.fi (K. Gurung), cause of its possible environmental intimidations. Globally, a large num-
mika.sillanpaa@lut.fi (M. Sillanpää). ber of EOMs have been frequently detected at source, and in treated
0048-9697/© 2019 Elsevier B.V. All rights reserved.
effluent and wasted sludge at wastewater treatment plants (WWTPs) 2016) or are based on lab-scale batch experiments (Hai et al., 2018;
(Clara et al., 2005b; Trinh et al., 2012, 2016). Moreover, the occurrence Phan et al., 2014; Tadkaew et al., 2011; Wijekoon et al., 2013). The influ-
of EOMs could endanger the reuse of reclaimed wastewater in achieving ence of solid retention times (SRTs) on the removal mechanisms of
sustainable water management within the WWTPs (Behera et al., PPCPs by MBRs has been studied by a few researchers (Clara et al.,
2011). Most of the EOMs found in wastewater are of anthropogenic origin 2005b; Schröder et al., 2012; Tambosi et al., 2010), but these studies
from various sources such as households, hospitals, veterinary surgeries, counted either estimated values of Kd or negligible amounts of adsorbed
farms and pharmaceutical manufacturing facilities, which eventually fractions of PPCPs onto sludge, as well as a spiking of PPCPs in the MBR
end up at WWTPs (Clara et al., 2005b). Since conventional WWTPs are system.
not specifically designed to eliminate EOMs, many of these EOMs pass Based on the above identified research gaps, the current study inves-
through WWTPs and are thus eventually discharged into the environ- tigates the removal dynamics of 23 frequently detected EOMs in the
ment (Behera et al., 2011; Jelic et al., 2011; Miège et al., 2009; Sim et al., treatment of real municipal wastewater at an MBR pilot plant operated
2010; Vieno et al., 2005; Yang et al., 2011, 2017; Ying et al., 2009). Some at two different SRTs. To the best of our knowledge, the removal and fate
countries such as Australia and Switzerland have adopted discharge of EOMs such as iopamidol, hydrocortisone, furosemide, ciprofloxacin,
guidelines for some EOMs (NHMRC/EPHC/NRMMC, 2008; tetracycline, and enalapril have rarely (or not) been studied in the
Oekotoxzentrum Centre Ecotox, 2016). In the EU context, environmental past by MBR processes. A mass balance study was assessed to further
quality standards (EQS) for 45 priority EOMs including first watch list confirm the removal mechanisms associated with different EOMs. Be-
substances like diclofenac, 17-β-estradiol, and 17-α-ethinylestradiol, sides, the influence of physicochemical properties on the removal of
were stipulated in Directive 2013/39/EU (EU, 2013), but the majority of EOMs was discussed. Additionally, the influence of SRTs on sludge char-
other EOMs have not yet been addressed. However, recent amendment acteristics including dewaterability, major foulants such as soluble mi-
in directive 2013/39/EU has introduced more watch list substances, in crobial products (SMPs), and extracellular polymeric substances (EPS),
which diclofenac has removed while macrolide antibiotics, methiocarb, was discussed. Finally, the quality of MBR-treated effluent in terms of
neonicotinoids, metaflumizone, amoxicillin and ciprofloxacin have been EOMs was compared with some established guidelines.
added (EU, 2018).
To ensure compliance with future discharge requirements, interest 2. Materials and methods
in the ability of MBRs in the removal of EOMs from municipal wastewa-
ter has increased in recent decades. MBRs are becoming a mature tech- 2.1. MBR system, operational protocol, selected EOMs and sample collection
nology in water management practices for wastewater reclamation and
reuse due to their ability to produce superior-quality effluent, a rela- A pilot-scale MBR system was employed in this study. The schematic
tively small footprint and little sludge (Judd, 2008; Meng et al., 2009). diagram indicating major components of the MBR pilot plant is shown
The removal mechanisms of EOMs involved in MBR treatment are com- in Fig. 1. The key features of the MBR pilot plant is described in our pre-
plex and include biotransformation/biodegradation, adsorption to vious work (Gurung et al., 2016). Typically, the submerged flat-sheet
sludge, volatilisation, and physical retention by the membrane (Cirja membrane units (0.4 μm pore size) with membrane surface area of
et al., 2008). However, biotransformation and adsorption to biosolids 16 m2 (KUBOTA Co., Japan) were used for solid-liquid separation. A per-
are often reported to be the dominant mechanisms for the removal of meate pump (Thomas, Germany) was used to draw water in an inter-
EOMs in the MBR process (Clara et al., 2005b; Kim et al., 2014; mittent suction mode. For in-situ cleaning of membranes, the clean-
Radjenović et al., 2009). On that note, physicochemical properties in-place (CIP) tank was also installed and connected with Sodium hypo-
(such as distribution coefficient, log D; dissociation constant, pKa; chlorite and Citric acid dosing tanks. Scouring air was supplied to main-
solid-water distribution coefficients, and Kd) and functional groups as- tain biological activities and to mitigate membrane fouling. The feed to
sociated with EOMs showed notable effects on their removal (Li et al., the MBR pilot plant was taken continuously from the outlet of the pri-
2015; Tadkaew et al., 2011). At the pH above pKa, phenolic hydroxyl mary clarifier at the Kenkäveronniemi WWTP, Mikkeli, Finland. This
group of some hormones dissociates that leads to the charge repulsion plant treats about 15,000 m3 of wastewater per day.
between the negatively charged hormone and negatively charged In this study, the pilot plant was operated in two different SRTs of
membrane (Schäfer et al., 2011). Naproxen and diclofenac (pKa values 60 days and 21 days in order to investigate MBR performance in treating
of 4.2 and 4.15, respectively) deprotonated a at pH 7–8 and rejected the pollution load of nutrients, organics, solids and, preferably, the re-
by negatively charged membrane surface, whereas carbamazepine moval of EOMs. Table 1 summarises the in-depth operational conditions
(pKa = 13.9) showed no ionisation at same pH level, hence attained a of the MBR pilot plant for this study. The pilot plant was operated a pro-
poor removal (Röhricht et al., 2009). At pH 6–7, tetracyclines (pKa = tocol similar to that reported by other researchers (Van den Broeck
4.5) are not charged, hence removed only via adsorption to sludge et al., 2012). The SRT of 60 days was adapted N10 weeks before the ac-
(Kim et al., 2005). The sorption onto sludge can considered to be insig- tual sampling periods in order to stabilize the activated sludge. How-
nificant for EOMs with logKd b 2.48 (Joss et al., 2006). The overall re- ever, relatively less adaption period of sludge (about a week) was
moval efficiencies of most of hydrophilic chemicals (log D b 3.2) were maintained for SRT of 21 days by wasting required amount of sludge
above 92% in the MBR (Trinh et al., 2016). out of the system. Each of the two different SRTs were maintained for
Often previous studies on the removal of EOMs by MBRs focused on stable sludge period of 4 weeks. During the investigation, the perme-
the aqueous phase (Bo et al., 2009; Radjenovic et al., 2007; Schröder ation flow rate and the transmembrane pressure (TMP) were nearly
et al., 2012; Trinh et al., 2012) and their mass load adsorbed in the constant (4–5 kPa), so no in-situ cleaning of membranes was necessary.
solid phase was often neglected, which restricts the differentiation be- A set of 55 EOMs representing a wide range of pharmaceuticals, pes-
tween the removal mechanisms via adsorption onto sludge and bio- ticide and steroid hormones with diverse physicochemical properties
transformation. There is limited literature investigating the fate of were studied. The selection of these EOMs was made based on their in-
EOMs in MBRs when treating real municipal wastewater (Kim et al., creasing annual consumption, potential adverse impacts on health of
2014; Radjenović et al., 2009; Trinh et al., 2016). Some researchers human and aquatic animals, and stringent future regulations in the EU
have estimated the contributions of the adsorption and biotransforma- Water Framework Directive (EU, 2013) on pharmaceuticals and
tion of EOMs based on values of Kd from literature or by direct batch ex- hormones. The physicochemical properties of studied EOMs are
periments (Clara et al., 2005b; Kimura et al., 2007). Additionally, summarised in Table S1 (Supplementary Information). Weekly 24-h
previous studies on the fate of EOMs in MBR either mainly focused on composite aqueous samples of influent (1 L), MBR permeate (1 L) and
pharmaceuticals and personal care products (PPCPs) (Radjenović grab samples of sludge (0.5 L) were collected for 8 weeks (4 samples
et al., 2009; Suárez et al., 2012; Tambosi et al., 2010; Trinh et al., of each SRT). The sum total of samples included 8 influent, 8 effluent
Fig. 1. Schematic diagram of the pilot MBR unit with indicated sampling points.
and 8 sludge samples. All the collected samples including liquid and CEff Q Eff
Remain in effluent ð%Þ ¼ 100 ð3Þ
sludge phases were stored in dark and cold conditions (~ −18 °C) be- CInf Q Inf
fore the analysis.
2.2. Mass balance calculation Removal via biotransformationð%Þ

¼ Overall removal ð%Þ−Removal via sorption onto sludge ð%Þ ð4Þ
To assess quantitatively the fate of the EOMs, the percentage shares
of biotransformation/ degradation, sorption and remaining in effluent where CInf ∗ QInf; CEff ∗ QEff; and CSludge ∗ QExcess sludge are the mass
were estimated using a mass balance approach (Jelic et al., 2011). The flow rate (μg/d) of influent, treated effluent and sludge, respectively.
mass flow rates of EOMs in the influent, effluent and sludge were calcu- QExcess sludge is excess sludge production (g d−1).
lated by multiplying measured concentrations in given streams by the
corresponding flow data. Then, mass load lost per unit of time (μg 2.3. Analytical methods
d−1) due to all the processes that possibly occur during the MBR process
and their relative distributions to the influent mass load were calculated Chemical oxygen demand (COD), total suspended solids (TSS), am-
according to the following Eqs. (1)–(4). monium nitrogen (NH4-N), total nitrogen (TN), total phosphorus (TP),
mixed liquor suspended solids (MLSS) and mixed liquor volatile
ð CInf Q Inf −CEff Q Eff Þ suspended solids (MLVSS) were determined according to Standard
Overall removal ð%Þ ¼ 100 ð1Þ Methods for the Examination of Water and Wastewater (APHA, 1999).
CInf Q Inf
The CST of sludge was measured using a capillary suction timer (Type
304 M, Triton, UK). The concentrations of EPS and SMP of the sludge
were determined by previously described methods (Ramesh et al.,
CSludge Q Excess sludge 2006). At first, 25 mL of sludge sample was centrifuged at 4000 rpm at
Removal via sorption onto sludgeð%Þ ¼ 100 ð2Þ
CInf Q Inf 4 °C for 30 min. The supernatant was filtered through a 0.45 μm glass filter
and measured for SMP. The obtained pallet was resuspended by adding
Table 1
0.05% NaCl, mixed with vortex mixture and the solution was kept in a
Operational conditions of the MBR pilot plant. water bath (~100 °C) for heat treatment for 60 min. Finally, the superna-
tant was filtered through a 0.45 μm glass filter and measured for total EPS.
Parameter Units Values min – max (avg.)
Both SMP and EPS were analysed in term of dissolved organic carbon
SRT: 60 days SRT: 21 days (DOC) using a TOC analyser (TOC-V Series-CPN, Shimadzu, Japan).
Hydraulic retention time The concentrations of EOMs in influent, permeate and sludge sam-
Hours 40.1–45.5 (42.13)
(HRT) ples were analysed using standard solid phase extraction (SPE)
Flux (Continuous) L m−2 h−1 5.4–6.13 (5.84) followed by UPLC/MS/MS (modified EPA 1694 and modified EPA 539).
MLSS concentration mg L−1 7808–9688 2271–4865
(8550) (3748)
The measurement uncertainty (MU) was 40%. The MU % was calculated
MLSS/MLVSS _ 0.57–0.83 0.3–0.68 according to the Nordest TR 537 that covered every step in the labora-
(0.67) (0.56) tory analysis.
F/M ratio kg COD. (kg MLVSS. 0.022–0.034 0.06–0.11
d)−1 (0.027) (0.09)
3. Results and discussion
Specific aeration demand
m h−1 0.15–0.23 (0.20)
(SADm)
Dissolved oxygen (DO) mg O2 L−1 1.2–4.6 (2.9) 3.1. Basic performance of MBR
Reactor temperature °C 14–22 (19)
pH Unitless 6.0–7.4 (6.6) As real municipal wastewater was fed into the MBR, the influent
Intermittent filtration Minutes 9-ON/1-OFF
composition varies over time, and does the pollution removal efficiency.
Table 2 summarises the composition of influent wastewater during each were observed at a consistently low level (23 to 510 ng L−1) during
period of SRT and the treatment efficiency of the MBR on conventional both testing periods.
pollutions such as COD, NH4-N, TN, and TP. The fluctuations on influent
composition were not significant (Table 2). The pH of the effluent was 3.3. Removal of EOMs in MBR treatment
around 7. The removal efficiencies of TSS, COD and NH4-N were consis-
tently high (N90%,N99.96%, respectively) at both SRTs. For two SRTs, the The removal efficiencies of selected EOMs observed during the MBR
suspended solids in the effluent samples were below the limit of quan- treatment at two different SRTs (60 days and 21 days) are shown in
tification (LOQ), showing the excellent physical barrier of the MBR Fig. 3a–b. SRT is a very important operational parameter in MBR pro-
system. The pilot MBR plant was not designed to remove TN, but cesses influencing the removal of EOMs (Clara et al., 2005a; Schröder
about 15% of removal on both SRTs was achieved due to a simultaneous et al., 2012; Taheran et al., 2016; Tambosi et al., 2010). The concentra-
nitrification-denitrification environment created by intermittent aera- tions of the majority of the EOMs in the treated effluent were observed
tion inside the nitrification tank. A relatively consistent removal of TP to be lower (even b LOQs to many EOMs) than in the influent at both
was also observed (N87%) on both SRTs due to enhanced biological SRTs, indicating the efficient attenuation of EOMs by the MBR, regard-
phosphorus removal (EBPR) in the MBR process. Overall, a good effluent less of the changing of the SRTs. In general, increasing SRT can improve
quality was observed at both the SRTs with comparable treatment the biodiversity of slowly growing bacteria, sludge growth and the lon-
efficiencies. ger retention of sludge that favours the removal of EOMs (Kimura et al.,
2007; Maeng et al., 2013; Taheran et al., 2016; Tambosi et al., 2010).
3.2. Occurrence of EOMs in the MBR However, some previous researchers confirmed that the removal of ibu-
profen, diclofenac, naproxen, ketoprofen and carbamazepine is not sig-
The median concentrations of the EOMs detected in the influent of nificantly affected by changing SRT during MBR treatment (Li et al.,
the MBR pilot plant are presented in Fig. 2. The concentrations of 2015).
EOMs in WWTPs depend on several factors such as the consumption Of the selected EOMs in this study, 17 EOMs, including steroid hor-
pattern of pharmaceuticals, the wastewater influent load etc., which mones, showed consistent removal of N90%. Furthermore, three com-
might vary within a region or country and from country to country pounds showed moderate removal between 50% and 87% (metoprolol,
(Radjenović et al., 2009). Of the 55 EOMs, only 23 compounds were an- furosemide, and propranolol), one showed poor removal of 39–46%
alytically detected in the median range from 23 to 145,000 ng L−1 (diclofenac), and two (carbamazepine and hydrochlorothiazide) were
(Fig. 2) and the remaining 32 compounds were not detected above very poorly removed (almost not at all) (−8 to 10%). Psychostimulant
their LOQs during the testing of both SRTs (Table S2). Fig. 2 indicates (caffeine) was detected at highest concentration among other EOMs at
that the occurrences of the majority of EOMs including pharmaceuticals both SRTs, but efficiently attenuated with a removal efficiency of up to
and steroid hormones during the SRT of 60 days showed similar tenden- 99.7%. These results are in agreement with many previous works (Kim
cies to those observed during the SRT of 21 days, which is explained by et al., 2014; Trinh et al., 2012, 2016). An effective attenuation of caffeine
the fact that all the sampling sessions were performed during the spring is achievable even at a relatively short SRT (Maeng et al., 2013).
season. The median concentrations of the EOMs belonging to anti- Moreover, EOMs including anti-coagulant (iopamidol) and
coagulants, diuretics, psychostimulants and analgesics categories such immunosupressive (hydrocortisone), whose fate was studied for the
as iopamidol (2625 ± 2367 ng L−1), furosemide (1864 ± 91 ng L−1), first time in the MBR process, were efficiently removed in the range of
caffeine (140,437 ± 6453 ng L−1), hydrochlorothiazide (2035 ± 90–96%. Regarding β-blockers, atenolol and bisoprolol were attenuated
757 ng L−1), paracetamol (93,600 ± 3394 ng L−1), naproxen (3395 at consistently high removal efficiency (92–97%) at both SRTs, whereas
± 276 ng L−1), ibuprofen (14,400 ± 565 ng L−1) and diclofenac two other β-blockers (metoprolol and propranolol) were attenuated
(1620 ± 394 ng L−1), respectively were observed at the highest levels slightly higher at a longer SRT of 60 days (82–84%) than in shorter SRT
during the testing periods. N50% of the total EOMs in the influent was days of 21 (50–60%). Very high removal efficiencies of analgesics (para-
accounted for these eight compounds. The high concentrations of caf- cetamol, ibuprofen, ketoprofen, naproxen) were observed ranging from
feine, naproxen and ibuprofen are consistent with previous papers 96 to N99.9%. The higher removal of paracetamol is due to its structure
that studied the occurrence and treatment of EOMs in real municipal that allows unrestricted access of bacteria and enzymes to the sterically
wastewater (Clara et al., 2005b; Kim et al., 2014; Radjenović et al., unprotected molecule that is then subsequently modified (Tambosi
2009; Trinh et al., 2016; Vieno et al., 2005). The observed caffeine con- et al., 2010). Moreover, the removal of ibuprofen, ketoprofen and
centration is in agreement with previous literature (Kim et al., 2014), naproxen is partially accompanied by the heterotrophic bacteria com-
and is also about four times higher than the values reported in another munity, which is highly favourable in extended SRTs, and less so by
study (Trinh et al., 2012). Antibiotics such as tetracycline, ciprofloxacin, slow-growing ammonia-oxidising bacteria (Falås et al., 2012). The
and trimethoprim were detected at b1000 ng L−1, which is very com- higher removal results of analgesics are in agreement with those re-
mon in municipal wastewater (Guerra et al., 2014). The concentrations ported by many researchers (Kim et al., 2014; Radjenović et al., 2009;
of EOMs were arranged from lower to higher log D pH 7 (distribution co- Schröder et al., 2012; Tambosi et al., 2010). On the other hand,
efficient) values of compounds (Fig. 2). The log D pH 7 is the corrected diclofenac (analgesic) was attenuated at relatively low efficiency com-
value of log Kow (partition co-efficient) for ionisation at pH 7 (Trinh pared to other analgesics with median removal efficiencies of 38% and
et al., 2016). The concentrations of steroid hormones including testos- 45% at SRTs of 60 days and 21 days, respectively. A great discrepancy
terone, progesterone, estrone, and estroil having log D pH 7 values N3, in the removal efficiency of diclofenac via MBR was reported in some
Table 2
Basic performance of MBR system (mean ± standard deviation) at two different SRTs.
Parameter SRT:60 days SRT: 21 days
Influent (mg L−1) Effluent (mg L−1) Removal efficiency (%) Influent (mg L−1) Effluent (mg L−1) Removal efficiency (%)
TSS 128.0 ± 37.9 b1 100 133.33 ± 14.1 b1 100

COD 299.3 ± 55.2 22.7 ± 5.2 92.43 330.0 ± 19.6 23.0 ± 5.3 93.03
NH4-N 29.7 ± 6.3 b0.01 N99.96 38.85 ± 3.2 b0.01 N99.96
TP 3.94 ± 0.7 0.47 ± 0.2 88.12 3.86 ± 1.4 0.49 ± 0.1 87.40
TN 45.0 ± 7.5 38.33 ± 7.1 14.81 50.0 ± 0.74 42.5 ± 4.4 15.00
Fig. 2. Occurrences of EOMs in MBR process at two different SRTs (60 days and 21 days) arranged according to their Log D (at pH7) values. EOMs with concentration lower than LOQ are not
included.
Fig. 3. Median removal efficiencies of EOMs in MBR process at different SRTs. The EOMs with their detection in influent b LOQ are not presented. The LOQ values of EOMs in the effluent are
used to calculate % removal whenever measured concentrations are below LOQ. Error bars represents the standard deviation calculated from the duplicate samples taken once in a week.
literature (0 to 87%) (Clara et al., 2005a; Trinh et al., 2016). A slow or 10−6) (Table S1), the removal of these compounds by volatilisation is in-
non-biotransformation rate of diclofenac has been reported by other re- significant. Moreover, the typical molecular weight cut-off (MWCO) of
searchers (biotransformation rate constant, Kbiol b 0.1 g−1 ss d−1 in MF and UF membranes are well above several thousand daltons (Da)
MBR) (Trinh et al., 2016; Vieno and Sillanpää, 2014). (Taheran et al., 2016), so the retention of EOMs (MWCO between 200
The diuretics (furosemide and enalapril), which were rarely studied and 800 Da) in the MBR process cannot be expected due to size exclusion.
in MBRs, were highly attenuated at both the SRTs with median removal Therefore, the dominant removal mechanisms of EOMs were presumably
efficiencies of 87% and 96%, respectively. These results are consistent via biotransformation and sorption onto the sludge, which was con-
with the previous study (Kim et al., 2014). On the other hand, the atten- firmed by the mass balance calculations as shown in Fig. 4. The daily
uation of diuretic (hydrochlorothiazide) was very low or hardly re- mass loads of EOMs via biotransformation, sorption onto sludge and re-
moved at all in this study. During SRT of 21 days, the removal maining in effluent are interpreted as shared proportions relative to the
efficiency ranged from −52% to 13% (median − 1%), which was slightly influent load (Eqs. (1)–(4)).
enhanced at SRT of 60 days achieving removal efficiency between −19% Among the selected EOMs, the highest significant biotransformation
and 41% (median 10%). The comparably poor removal of hydrochloro- rates (83–99.9%, median 99%) were observed for atenolol, caffeine,
thiazide (~5%) is reported by other researchers when treating municipal paracetamol, naproxen, ibuprofen, estriol and progesterone. Similarly,
wastewater with full-scale MBR (Radjenović et al., 2009). The increased high biotransformation rate was accounted to iopamidol, furosemide
concentrations of hydrochlorothiazide in treated effluent, more than and trimethoprim (58–92%, median 86%). Moderate biotransformation
the influent levels (i.e. negative removal efficiency), might be due to was observed for diclofenac (median 33%). Enhanced biotransforma-
the transformation of metabolites such as unmeasured products of the tions of metoprolol, hydrochlorothiazide, and propranolol were
human metabolism and/or metabolites that could be converted back achieved at SRT of 60 days (78%, 27% and 51%) compared to SRT of
to their parent compounds during the biological treatment process. 21 days (37%, 0% and 0%, respectively). On the other hand, carbamaze-
This phenomenon of “negative removal” was also reported in other pine showed high resistance to biotransformation, so was highly de-
studies (Jelic et al., 2011). tected in the effluent. The biotransformation proportions observed for
The attenuation of an anti-epileptic drug (carbamazepine) was very naproxen, ibuprofen, trimethoprim, enalapril, atenolol, metoprolol and
low and frequently not removed, regardless of varying SRTs. At an SRT furosemide are consistent with previous studies (Jelic et al., 2011; Joss
of 60 days, the median removal efficiency of carbamazepine was close et al., 2006; Kim et al., 2014).
to 2% (not removed in two samples out of four), whereas non-removal On the other hand, some of the EOMs were observed to be removed
was observed (median – 8%) at an SRT of 21 days. This means that, in via sorption to sludge as indicated in Fig. 4. EOMs can be sorbed to
most of the events, effluent concentrations of MBR were greater than in- sludge and subsequently removed while wasting the excess sludge
fluent levels. The major metabolite of carbamazepine in humans is from the bioreactor. The mechanism of sorption occurs via absorption
10,11 epoxy-carbamazepine, which is excreted principally as glucuro- and adsorption (Li et al., 2015). Absorption occurs due to the hydropho-
nide and further to glucuronide-conjugates via the hydroxylation of bic interactions of the aliphatic and aromatic groups of EOMs with the
the aromatic ring. These glucuronide-conjugates can presumably be liphophilic cell membrane of some microorganisms and the fat fractions
cleaved in a biological process and so increase in the effluents (Ternes, of sludge, whereas adsorption involves the electrostatic interactions of
1998). The recalcitrant behaviour and non-removal of carbamazepine the positively charged (e.g., amino groups) to negatively charged sur-
in MBR treatment due to poor degradability have been reported in face of microorganisms (Cirja et al., 2008). Most of the EOMs were fre-
many other studies (Clara et al., 2005b; Radjenovic et al., 2007; quently detected in the sludge under LOQ (10 to 500 ng g−1, median
Wijekoon et al., 2013). 100 ng g−1) (Fig. S1). Other EOMs were detected in the range of
Concerning the removal of antibiotics, consistently high removal ef- 38–3750 ng g−1 (median 275 ng g−1) at which ciprofloxacin followed
ficiencies were achieved at 87 to 98%. Tetracycline, ciprofloxacin and tri- by hydrochloro-thiazide were the highest detected in the sludge. This
methoprim were attenuated at a comparable level at both SRTs with result is in agreement with a previous study, where ciprofloxacin was
removal efficiencies of 98%, 93%, and 94%, respectively. Comparable re- most frequently detected in MBR sludge, although the concentration
sults in the removal of tetracycline and ciprofloxacin were reported by level was considerably higher (N 103 ng g−1) (Kim et al., 2014). From
Kim et al., which was the only research studied to-date in the MBR pro- the results, ciprofloxacin posed a serious challenge due to very strong
cess (Kim et al., 2014), where the SRT was kept relatively slow at affinity to the sludge with a sorption fraction of N90%. Relatively higher
6–8 days. This indicates that the removals of tetracycline and ciproflox- sorption of ciprofloxacin is reported by many researchers (Kim et al.,
acin were not significantly affected by changes in SRTs. Nevertheless, a 2014). Biosorption as a dominant mechanism for the removal of cipro-
wide range of removal efficiencies were reported in literature from floxacin is reported by many authors (Githinji et al., 2011; Guerra
non-removal to nearly 67% (Kim et al., 2014; Radjenović et al., 2009). Al- et al., 2014). Similarly, propranolol, carbamazepine and testosterone
though trimethoprim is often presumed to be recalcitrant to activated showed moderate sorption efficiency to the sludge (10–50%, median
sludge bacteria, its poor biotransformation was reported to be achiev- 32%), whilst weak sorption affinities were observed for rest of the
able by slow-growing bacteria (Radjenović et al., 2009). EOMs (0–21%, median 6%).
The median removal efficiency of steroidal hormones (testosterone, In order to better understand the fate of EOMs due to sorption to the
progesterone, estriol, estrone) in the MBR process at both SRTs is pre- sludge, sludge-water distribution coefficients or sorption coefficients
sented in Fig. 3. Results indicated that steroid hormones were attenu- (log Kd) were experimentally assessed by dividing the concentration
ated effectively by the MBR, achieving steady removal efficiency of of EOMs in sludge (ng kg−1) by their concentrations in effluent water
over 98%. These results are in agreement with other studies performed (ng L−1) (Loke et al., 2002; Radjenović et al., 2009). The corresponding
on MBRs (Maeng et al., 2013; Trinh et al., 2012, 2016). log Kd for high sorption fractions of N90% and moderate fraction
(10–50%) were 3.0 to 4.9 (Fig. 4), demonstrating their removal by
3.4. Behaviour of EOMs in MBR treatment: application of mass balances strong sorption to sludge. These results are comparable with the find-
ings of other researchers (Kim et al., 2014). Nevertheless, many com-
The removal mechanisms of EOMs in MBR treatment form a complex pounds which showed a high degree of removal via biotransformation
process characterised by four major pathways: (i) biotransformation or (e.g. tetracycline, bisoprolol, enalapril, ketoprofen, ibuprofen, hydrocor-
degradation; (ii) sorption onto the sludge (via excess sludge removal); tisone, estriol, estrone and progresterone) accounted for the high log Kd
(iii) volatilisation or stripping by aeration; and (iv) physical retention values (3.9 to 4.3), indicating that these EOMs are attenuated simulta-
due to size exclusion by membranes (Cirja et al., 2008; Li et al., 2015). neously by biodegradation and partitioning onto biosolids (Guerra
Since the Henry's law constant (KH) values of the selected are low (b et al., 2014). On the other hand, the relatively low value of log Kd (2.6
Fig. 4. Fate of the selected EOMs during MBR process at two different SRTs. The log Kd values to the right showed the distribution of calculated sludge-water distribution coefficients/
sorption coefficients of EOMs.
to 3.0) was calculated for diclofenac, carbamazepine and hydrochloro- On the other hand, the removal of hydrophilic EOMs (log D pH 7 b 3.2)
thiazide, indicating their poor removal efficiencies via sorption. In gen- varied from ‘limited removal’ (carbamazepine and hydrochlorothiazide)
eral, the higher biotransformation of EOMs was followed by lower to almost complete removal (paracetamol, caffeine and ibuprofen). It is
sorption potential, which was not in accordance with the calculated very obvious that the removal of hydrophilic EOMs varies significantly
log Kd. These results suggest that log Kd values may vary plant-to- since they have diverse molecular structures and functional groups
plant depending on operating conditions and the extent of removal ef- (Table S1). Varying removal efficiencies accounted for in hydrophilic
ficiencies (Kim et al., 2014). Indeed, the biotransformation of EOMs is EOMs can be explained based on the qualitative framework by Tadkaew
enhanced by the sorption process due to the longer retention time of et al. (Tadkaew et al., 2011) as: (i) compounds possessing only EDGs
solids in the MBR process (Tadkaew et al., 2011). (slightly hydrophobic such as estriol) could achieve high removal
(~99%); (ii) compounds possessing both EDGs and e-withdrawing
3.5. Other factors influencing the removal of EOMs in MBR (EWGs) such as amide and chloride in their molecular structure
(Tadkaew et al., 2011; Wijekoon et al., 2013) including paracetamol, ibu-
Hyrophphobic interaction is also one of the key mechanisms con- profen, diclofenac, hydrocortisone, naproxen, propranolol, trimethoprim,
trolling biosorption of EOMs, hence the removal efficiencies in MBR pro- ciprofloxacin, bisoprolol, caffeine, furosemide, metoprolol, tetracycline,
cesses (Cirja et al., 2008). Most of the EOMs quantified in the influent atenolol and iopamidol with varying removal efficiencies ranging from
load at both testing periods were hydrophilic (log D pH 7 b 3.2) (Hai 37 to 99.9%; (iii) compounds possessing only strong EWGs such as hydro-
et al., 2018; Tadkaew et al., 2011; Wijekoon et al., 2013) ranging from chlorothiazide and carbamazepine (0–10%). Therefore, the dominating
pharmaceuticals to steroid hormones. Of the four hormones, only tes- removal mechanism of hydrophobic compounds is strongly influenced
tosterone, estrone and progresterone are hydrophobic (log D pH 7 by their intrinsic biodegradability as sorption to sludge is less significant
N 3.2) (Fig. 3). Hydrophobicity can be defined by log Kow for non-ionic (Tadkaew et al., 2011).
and log D for ionic compounds (Taheran et al., 2016). All three hydro- Another aspect may be the degree of ionisation (pKa) of the selected
phobic EOMs showed consistently high removal efficiency of 96–99% EOMs, which controls their ionisation state as a consequence of pH var-
(median 97%). Similar results have been reported in other studies iation in the solution. For EOMs including ibuprofen, diclofenac,
(Suárez et al., 2012). The higher removal efficiency of hydrophobic com- naproxen, ketoprofen, enalapril, caffeine, furosemide, tetracycline and
pounds might be due to: (i) the dominating mechanism of sorption to ciprofloxacin possessing a pKa of between 0.52 and 6.43 (Table S1), re-
the sludge that facilitates enhanced biological degradation, and (ii) moval efficiencies varied from 38 to 99%. The results are partially consis-
possessing only e-donating groups (EDGs) that enhance better oxida- tent with the previous work, where acidic EOMs showed low removal
tion (Tadkaew et al., 2011). High log D is characteristic of hydrophobic efficiency in neutral pH condition. This might be related to the fact
EOMs, poor hydrosolubility and high sorption tendency on organic con- that these EOMs are negatively charged at neutral pH due to the pres-
stituents of sludge matrix (Cirja et al., 2008). Moreover, the calculated ence of carboxyl functional group in their structures (Urase et al.,
log Kd values of these EOMs were above 4, indicating substantially 2005). Likewise, EOMs with pKa values of 7 to 15 (Table S1) showed var-
high sorption affinity to the sludge. ied removal efficiencies ranging from ‘non-removal’ to N97%, with the
lowest being carbamazepine and hydrochlorothiazide and the highest process, no clear correlation of SMP and EPS to membrane fouling was
being paracetamol. Carbamazepine and hydrochlorothiazide both observed in this study. The consistent rate of fouling might have been
have a basic amine functional group, which may be protonated to gain adequately mitigated by continuous physical membrane-cleaning pro-
a positive charge under neutral pH (Nghiem and Khan, 2007; Urase cesses such as intermittent membrane relaxation and the use of scour-
et al., 2005). This phenomenon may have increased their hydrophilicity ing air involved in the pilot operation. Many researchers reported that
which was confirmed by their noticeably higher concentrations in the the release of SMP and EPS concentration decreases as the sludge
MBR effluents (Fig. S1). stays longer in the bioreactor, and increases when lowering the SRT
The molecular weights (MWs) of the selected EOMs varied from (Grelier et al., 2006; Meng et al., 2009). The presence of EOMs in MBR
151 g moL−1 (paracetamol) to 777 g moL−1 (iopamidol) in this study may affect microbial activities and the structure community of microor-
(Table S1). However, the removal tendencies of EOMs in relation to ganisms in activated sludge, which increases the endogenous respira-
MWs were not clear. For example, paracetamol and caffeine were al- tion rates and subsequently releases EPS (Avella et al., 2010).
most completely removed although their MWs were the lowest on However, studying the extent of EOM influence on releasing SMP and
the list (151 and 194 g moL−1), whereas the removal efficiencies of EPS was out of the scope of this study.
other EOMs with MWs ranging from 206 to 777 g moL−1 varied from Furthermore, the dewaterability of MBR sludge was also assessed at
non-removal to N99.6%. EOMs with higher MW may provide more both SRTs in terms of their corresponding CST values (Fig. 5). The aver-
branches that allow microbes to selectively cleave certain target sites age CST of the sludge operated at an SRT of 60 days (31 s) was about
and subsequently initiate degradation (Tadkaew et al., 2011). twice as high as at an SRT of 21 days (15 s). An increasing trend of sludge
Furthermore, the average MLSS concentrations inside the reactor CST with increasing SRT is reported by other researchers, which indi-
were 8.5 ± 0.8 g L−1 and 3.7 ± 1.1 g L−1 at SRTs of 60 days and cates that the sludge with high CST is difficult to dewater due to the
21 days, respectively. Results showed that the removal of metoprolol, high content of bound water (Ng et al., 2006). Also, increasing SMP
hydrochlorothiazide, propranolol and carbamazepine was enhanced and EPS concentrations indicated the decreasing CST of sludge due to in-
while operating the MBR at high SRT (i.e. high MLSS concentration) creased fractions of fine and colloidal particles, and vice versa (Grelier
(Fig. 3). At increasing SRTs, the amount of sludge wasted from the bio- et al., 2006).
reactor decreases, which allows longer retention and subsequent degra-
dation of less polar EOMs in the system. The higher specific area 3.7. MBR effluent quality in compliance with Australian and Swiss
provided by increasing MLSS concentration enables more enzymatic ac- guidelines
tivities (Cirja et al., 2008).
To evaluate the quality of MBR-treated effluent in terms of EOMs
3.6. Influence of SRTs on sludge characteristics concentrations, the observed concentrations of EOMs were studied in
compliance with the Australian (NHMRC/EPHC/NRMMC, 2008) and
Fig. 5 shows the variations in SMP and EPS during the MBR treat- Swiss guideline values (Oekotoxzentrum Centre Ecotox, 2016)
ment at two different SRTs. Although the TMP profile of the MBR system (Table 3). Nevertheless, no guidelines or standards have yet been pro-
was quite stable during the testing periods of both the SRTs, noticeable posed for EOMs selected by the EU. A wider range of pharmaceuticals
changes in the concentrations of SMP and EPS were observed. SMP con- and steroid hormones was established in the Australian guidelines
centration increased from 5.6 to 15.5 mg gMLVSS−1 (about 3 times) than in the Swiss guidelines. Most of the studied EOMs were observed
while SRT was lowered from 60 days to 21 days. Similarly, more EPS well below the guideline values of both the standards (Table 3). No ste-
were released at an SRT of 21 days (103 mg gMLVSS−1) than the longer roid hormones were detected above their LOQs, which were lower than
SRT of 60 days (51 mg gMLVSS−1). A comparable trend of EPS variation both the guidelines. Among diuretics, enalapril was detected below the
was reported by other researchers (Massé et al., 2006). As SRT in- LOQ (10 ng L−1), whereas hydrochlorothiazide and furosemide were
creased, concentrations of SMP and EPS decreased as most of the sub- frequently detected in the range of 185–1700 ng L−1. Similarly, immu-
strates were consumed for the metabolism of microorganisms (Park nosuppressive (hydrocortisone) was not detected above its LOQ, but
et al., 2015). Nevertheless, extremely high SRTs can accumulate high no guidelines for diuretics and immunosuppressives are available or
soluble and particulate matters due to the increase in MLSS concentra- have been proposed so far. Caffeine was detected at a concentration of
tion, leading to an increase in viscosity that demands high scouring air 60–328 ng L−1, which was lower than the guideline value of
for membranes (Krzeminski et al., 2012). However, since effluent flow 350 ng L−1, but it showed excellent removal efficiency of N99% in MBR
was relatively stable, which substantiates the robustness of the MBR treatment since the influent concentration was very high (140,437 ±
6453 ng L−1). Regarding poorly removed carbamazepine, the observed
concentration of 480–665 ng L−1 was lower than both guidelines. Over-
all, the results indicated that the quality of effluent by MBR treatment
fulfilled the water quality requirements for reuse purposes, regardless
of varying SRTs.
4. Conclusion
The removal and fate of selected EOMs by the MBR pilot plant at
varying SRTs were studied. Only 23 EOMs were detected in the influent
wastewater during the testing periods above their LOQs. Diverse re-
moval dynamics for selected EOMs were observed. The application of
mass balance assessment proved that biotransformation and sorption
onto sludge were the dominant mechanisms associated with the re-
moval of EOMs. Even though the EOMs removal efficiencies were not
that significant at SRTs of 21 and 60 days, enhanced removal of the ma-
jority of EOMs were observed at longer SRT. Therefore, SRT of 60 days
would be preferred to be implemented, which would additionally help
in releasing less concentrations of SMP and EPS. This study has ex-
Fig. 5. Variation of SMP, EPS fractions, and CST as a function of varying SRTs. panded the understanding of the removal and fate of the EOMs (some
Table 3 and membrane bioreactors (MBR). Rev. Environ. Sci. Biotechnol. 7, 61–78. https://
Comparison of median concentrations of EOMs in MBR effluent with Australian and Swiss doi.org/10.1007/s11157-007-9121-8.
guidelines. Clara, M., Kreuzinger, N., Strenn, B., Gans, O., Kroiss, H., 2005a. The solids retention time—a
suitable design parameter to evaluate the capacity of wastewater treatment plants to
Compound Concentration (ng L−1) remove micropollutants. Water Res. 39, 97–106. https://doi.org/10.1016/j.
watres.2004.08.036.
MBR Australian guideline Swiss guideline Clara, M., Strenn, B., Gans, O., Martinez, E., Kreuzinger, N., Kroiss, H., 2005b. Removal of se-
Effluent - values values lected pharmaceuticals, fragrances and endocrine disrupting compounds in a mem-
SRT 60 (NHMRC/EPHC/NRMMC, (Oekotoxzentrum brane bioreactor and conventional wastewater treatment plants. Water Res. 39,
(SRT 21) 2008) Centre Ecotox, 4797–4807. https://doi.org/10.1016/j.watres.2005.09.015.
2016) EU, 2013. EU. Directive 2013/39/EU of the European Parliament and of the Council
[WWW Document].
Iopamidol 200 (350) 400 × 103 n.p. EU, 2018. Commission Implementing Decision (EU) 2018/840 [WWW Document]. URL.
Atenolol 4 (25) n.a. 330 × 103 http://data.europa.eu/eli/dec_impl/2018/840/oj/eng, Accessed date: 22 January 2019.
Tetracycline b10 (b10) 105 × 103 n.p. Falås, P., Andersen, H.R., Ledin, A., la Cour Jansen, J., 2012. Impact of solid retention time
Metoprolol 100 (350) 25 × 103 750 × 103 and nitrification capacity on the ability of activated sludge to remove pharmaceuti-
Furosemide 185 (230) n.a. n.p. cals. Environ. Technol. 33, 865–872. https://doi.org/10.1080/09593330.2011.601764.
Caffeine 328 (60.5) 350 n.p. Githinji, L.J.M., Musey, M.K., Ankumah, R.O., 2011. Evaluation of the fate of ciprofloxacin
Bisoprolol 11 (19.5) 630 × 103 n.p. and amoxicillin in domestic wastewater. Water Air Soil Pollut. 219, 191–201.
Ciprofloxacin 50 (51.5) 250 × 103 360 https://doi.org/10.1007/s11270-010-0697-1.
Enalapril b10 (b10) n.a. n.p. Grelier, P., Rosenberger, S., Tazi-Pain, A., 2006. Influence of sludge retention time on
membrane bioreactor hydraulic performance. Desalination, International Congress
1700
Hydrochlorothiazide n.a. n.p. on Membranes and Membrane Processes. 192, pp. 10–17. https://doi.org/10.1016/j.
(1350)
desal.2005.04.131.
Ketoprofen 5 (5) 3.5 × 103 n.p.
Guerra, P., Kim, M., Shah, A., Alaee, M., Smyth, S.A., 2014. Occurrence and fate of antibiotic,
Trimethoprim 28.5 (42) 70 × 103 210 × 103 analgesic/anti-inflammatory, and antifungal compounds in five wastewater treat-
Propranolol 33 (51.5) 40 × 103 12 × 103 ment processes. Sci. Total Environ. 473–474, 235–243. https://doi.org/10.1016/j.
Paracetamol 325 (50) 175 × 103 n.p. scitotenv.2013.12.008.
Naproxen 19.5 (84) 220 × 103 860 × 103 Gurung, K., Ncibi, M.C., Fontmorin, J., Särkkä, H., Sillanpää, M., 2016. Incorporating sub-
Ibuprofen 59 (50) 400 × 103 1700 × 103 merged MBR in conventional activated sludge process for municipal wastewater
Hydrocortisone b10 (b10) n.a. n.p. treatment : a feasibility and performance assessment. J. Membr. Sci. Technol. 6.
Diclofenac 760 (1050) 180 × 101 n.p. https://doi.org/10.4172/2155-9589.1000158.
Carbamazepine 480 (665) 100 × 103 2000 × 103 Hai, F.I., Yang, S., Asif, M.B., Sencadas, V., Shawkat, S., Sanderson-Smith, M., Gorman, J., Xu,
Estriol b5 (b5) 50 n.p. Z.-Q., Yamamoto, K., 2018. Carbamazepine as a possible anthropogenic marker in
Testosterone b1 (b1) 7 × 103 n.p. water: occurrences, toxicological effects, regulations and removal by wastewater
treatment technologies. Water 10, 107. https://doi.org/10.3390/w10020107.
Estrone b5 (b5) 30 n.p.
Jelic, A., Gros, M., Ginebreda, A., Cespedes-Sánchez, R., Ventura, F., Petrovic, M., Barcelo, D.,
Progesterone b1 (b1) 105 × 103 n.p.
2011. Occurrence, partition and removal of pharmaceuticals in sewage water and
n.a. values not available; n.p. not proposed; ‘b’ denotes below LOQ. sludge during wastewater treatment. Water Res. 45, 1165–1176. https://doi.org/
10.1016/j.watres.2010.11.010.
Joss, A., Zabczynski, S., Göbel, A., Hoffmann, B., Löffler, D., McArdell, C.S., Ternes, T.A.,
of which were never or rarely studied before in MBR) via MBR treat- Thomsen, A., Siegrist, H., 2006. Biological degradation of pharmaceuticals in munici-
ments under different operating conditions (i.e. SRTs), which certainly pal wastewater treatment: proposing a classification scheme. Water Res. 40,
1686–1696. https://doi.org/10.1016/j.watres.2006.02.014.
helps in achieving more sustainable water management. Judd, S., 2008. The status of membrane bioreactor technology. Trends Biotechnol. 26,
109–116. https://doi.org/10.1016/j.tibtech.2007.11.005.
Kim, S., Eichhorn, P., Jensen, J.N., Weber, A.S., Aga, D.S., 2005. Removal of antibiotics in
Acknowledgements wastewater: effect of hydraulic and solid retention times on the fate of tetracycline
in the activated sludge process. Environ. Sci. Technol. 39, 5816–5823. https://doi.
org/10.1021/es050006u.
The authors would like to thank Business Finland Oy Kim, M., Guerra, P., Shah, A., Parsa, M., Alaee, M., Smyth, S.A., 2014. Removal of pharmaceu-
(Innovaatiorahoituskeskus) (formerly Tekes- the Finnish Funding ticals and personal care products in a membrane bioreactor wastewater treatment
Agency for Technology and Innovation) for funding this work as part plant. Water Sci. Technol. 69, 2221–2229. https://doi.org/10.2166/wst.2014.145.
Kimura, K., Hara, H., Watanabe, Y., 2007. Elimination of selected acidic pharmaceuticals
of the Smart Effluents Project. Special thanks to Reijo Turkki, Risto Repo, from municipal wastewater by an activated sludge system and membrane bioreac-
Anne Bergman and Päivi Luukkonen for their continuous support in op- tors. Environ. Sci. Technol. 41, 3708–3714. https://doi.org/10.1021/es061684z.
erating and maintaining the MBR pilot plant during the sampling pe- 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
riods. We also thanks Eurofins Environment Testing Finland Oy for its sludge filterability in full-scale MBR systems for municipal sewage treatment.
laboratory analysis support. J. Membr. Sci. 423–424, 348–361. https://doi.org/10.1016/j.memsci.2012.08.032.
Li, C., Cabassud, C., Guigui, C., 2015. Evaluation of membrane bioreactor on removal of
pharmaceutical micropollutants: a review. Desalin. Water Treat. 55, 845–858.
Appendix A. Supplementary data https://doi.org/10.1080/19443994.2014.926839.
Loke, M.-L., Tjørnelund, J., Halling-Sørensen, B., 2002. Determination of the distribution
Supplementary data to this article can be found online at https://doi. coefficient (logKd) of oxytetracycline, tylosin A, olaquindox and metronidazole in
manure. Chemosphere 48, 351–361. https://doi.org/10.1016/S0045-6535(02)
org/10.1016/j.scitotenv.2019.02.308.
00078-4.
Maeng, S.K., Choi, B.G., Lee, K.T., Song, K.G., 2013. Influences of solid retention time, nitrifica-
tion and microbial activity on the attenuation of pharmaceuticals and estrogens in
References membrane bioreactors. Water Res. 47, 3151–3162. https://doi.org/10.1016/j.
watres.2013.03.014.
APHA, 1999. Standard methods for the examination of water and wastewater. Stand. Massé, A., Spérandio, M., Cabassud, C., 2006. Comparison of sludge characteristics and
Methods 541. performance of a submerged membrane bioreactor and an activated sludge process
Avella, A.C., Delgado, L.F., Görner, T., Albasi, C., Galmiche, M., de Donato, P., 2010. Effect of at high solids retention time. Water Res. 40, 2405–2415. https://doi.org/10.1016/j.
cytostatic drug presence on extracellular polymeric substances formation in munici- watres.2006.04.015.
pal wastewater treated by membrane bioreactor. Bioresour. Technol. 101, 518–526. Meng, F., Chae, S.-R., Drews, A., Kraume, M., Shin, H.-S., Yang, F., 2009. Recent advances in
https://doi.org/10.1016/j.biortech.2009.08.057. membrane bioreactors (MBRs): membrane fouling and membrane material. Water
Behera, S.K., Kim, H.W., Oh, J.-E., Park, H.-S., 2011. Occurrence and removal of antibiotics, Res. 43, 1489–1512. https://doi.org/10.1016/j.watres.2008.12.044.
hormones and several other pharmaceuticals in wastewater treatment plants of the Miège, C., Choubert, J.M., Ribeiro, L., Eusèbe, M., Coquery, M., 2009. Fate of pharmaceuti-
largest industrial city of Korea. Sci. Total Environ. 409, 4351–4360. https://doi.org/ cals and personal care products in wastewater treatment plants – conception of a da-
10.1016/j.scitotenv.2011.07.015. tabase and first results. Environ. Pollut., Special Issue Section: Ozone and
Bo, L., Urase, T., Wang, X., 2009. Biodegradation of trace pharmaceutical substances in Mediterranean Ecology: Plants, People, Problems. 157, pp. 1721–1726. https://doi.
wastewater by a membrane bioreactor. Front. Environ. Sci. Eng. China 3, 236–240. org/10.1016/j.envpol.2008.11.045.
https://doi.org/10.1007/s11783-009-0004-9. Ng, H.Y., Tan, T.W., Ong, S.L., 2006. Membrane fouling of submerged membrane bioreac-
Cirja, M., Ivashechkin, P., Schäffer, A., Corvini, P.F.X., 2008. Factors affecting the removal of tors: impact of mean cell residence time and the contributing factors. Environ. Sci.
organic micropollutants from wastewater in conventional treatment plants (CTP) Technol. 40, 2706–2713. https://doi.org/10.1021/es0516155.
Nghiem, L., Khan, S., 2007. Effects of operating conditions on the removal of trace organic water and wastewaters. Sci. Total Environ. 547, 60–77. https://doi.org/10.1016/j.
contaminants by a loose nanofiltration membrane. Fac. Eng. - Pap. Arch. 1–12. scitotenv.2015.12.139.
NHMRC/EPHC/NRMMC, 2008. National Water Quality Management Strategy [WWW Tambosi, J.L., de Sena, R.F., Favier, M., Gebhardt, W., José, H.J., Schröder, H.F., Moreira, R. de
Document]. URL. https://nhmrc.gov.au/health-advice/water-quality-and-health, F.P.M., 2010. Removal of pharmaceutical compounds in membrane bioreactors
Accessed date: 29 October 2018. (MBR) applying submerged membranes. Desalination 261, 148–156. https://doi.
Oekotoxzentrum Centre Ecotox, 2016. Swiss Proposals for Acute and Chronic Quality org/10.1016/j.desal.2010.05.014.
Standards [WWW Document]. Oekotoxzentrum Cent. Ecotox. URL. http://www. Ternes, T.A., 1998. Occurrence of drugs in German sewage treatment plants and
ecotoxcentre.ch/, Accessed date: 26 October 2018. rivers1Dedicated to Professor Dr. Klaus Haberer on the occasion of his 70th birth-
Park, H.-D., Chang, I.-S., Lee, K.-J., 2015. Principles of Membrane Bioreactors for Wastewa- day.1. Water Res. 32, 3245–3260. https://doi.org/10.1016/S0043-1354(98)00099-2.
ter Treatment [WWW Document]. URL. https://www.crcpress.com/Principles-of- Trinh, T., van den Akker, B., Stuetz, R.M., Coleman, H.M., Le-Clech, P., Khan, S.J., 2012. Re-
Membrane-Bioreactors-for-Wastewater-Treatment/Park-Chang-Lee/ moval of trace organic chemical contaminants by a membrane bioreactor. Water Sci.
9781466590373, Accessed date: 1 October 2018. Technol. 66, 1856–1863. https://doi.org/10.2166/wst.2012.374.
Phan, H.V., Hai, F.I., Kang, J., Dam, H.K., Zhang, R., Price, W.E., Broeckmann, A., Nghiem, L.D., Trinh, T., van den Akker, B., Coleman, H.M., Stuetz, R.M., Drewes, J.E., Le-Clech, P., Khan,
2014. Simultaneous nitrification/denitrification and trace organic contaminant S.J., 2016. Seasonal variations in fate and removal of trace organic chemical contam-
(TrOC) removal by an anoxic–aerobic membrane bioreactor (MBR). Bioresour. inants while operating a full-scale membrane bioreactor. Sci. Total Environ. 550,
Technol., Special Issue: CESE 2013 & Special Issue: ICABB 2013 165, pp. 96–104 176–183. https://doi.org/10.1016/j.scitotenv.2015.12.083.
https://doi.org/10.1016/j.biortech.2014.03.094. Urase, T., Kagawa, C., Kikuta, T., 2005. Factors affecting removal of pharmaceutical sub-
Radjenovic, J., Petrovic, M., Barceló, D., 2007. Analysis of pharmaceuticals in wastewater stances and estrogens in membrane separation bioreactors. Desalination, Membranes
and removal using a membrane bioreactor. Anal. Bioanal. Chem. 387, 1365–1377. in Drinking and Industrial Water Production. 178, pp. 107–113. https://doi.org/
https://doi.org/10.1007/s00216-006-0883-6. 10.1016/j.desal.2004.11.031.
Radjenović, J., Petrović, M., Barceló, D., 2009. Fate and distribution of pharmaceuticals in Van den Broeck, R., Van Dierdonck, J., Nijskens, P., Dotremont, C., Krzeminski, P., van der
wastewater and sewage sludge of the conventional activated sludge (CAS) and ad- Graaf, J.H.J.M., van Lier, J.B., Van Impe, J.F.M., Smets, I.Y., 2012. The influence of solids
vanced membrane bioreactor (MBR) treatment. Water Res. 43, 831–841. https:// retention time on activated sludge bioflocculation and membrane fouling in a mem-
doi.org/10.1016/j.watres.2008.11.043. brane bioreactor (MBR). J. Membr. Sci. 401–402, 48–55. https://doi.org/10.1016/j.
Ramesh, A., Lee, D.-J., Hong, S.G., 2006. Soluble microbial products (SMP) and soluble ex- memsci.2012.01.028.
tracellular polymeric substances (EPS) from wastewater sludge. Appl. Microbiol. Vieno, N., Sillanpää, M., 2014. Fate of diclofenac in municipal wastewater treatment plant —
Biotechnol. 73, 219–225. https://doi.org/10.1007/s00253-006-0446-y. a review. Environ. Int. 69, 28–39. https://doi.org/10.1016/j.envint.2014.03.021.
Röhricht, M., Krisam, J., Weise, U., Kraus, U.R., Düring, R.-A., 2009. Elimination of carba- Vieno, N.M., Tuhkanen, T., Kronberg, L., 2005. Seasonal variation in the occurrence of
mazepine, diclofenac and naproxen from treated wastewater by nanofiltration. pharmaceuticals in effluents from a sewage treatment plant and in the recipient
Clean: Soil, Air, Water 37, 638–641. https://doi.org/10.1002/clen.200900040. water. Environ. Sci. Technol. 39, 8220–8226. https://doi.org/10.1021/es051124k.
Schäfer, A.I., Akanyeti, I., Semião, A.J.C., 2011. Micropollutant sorption to membrane poly- Wijekoon, K.C., Hai, F.I., Kang, J., Price, W.E., Guo, W., Ngo, H.H., Nghiem, L.D., 2013. The
mers: a review of mechanisms for estrogens. Adv. Colloid Interface Sci. 164, 100–117. fate of pharmaceuticals, steroid hormones, phytoestrogens, UV-filters and pesticides
https://doi.org/10.1016/j.cis.2010.09.006. during MBR treatment. Bioresour. Technol. 144, 247–254. https://doi.org/10.1016/j.
Schröder, H.F., Tambosi, J.L., Sena, R.F., Moreira, R.F.P.M., José, H.J., Pinnekamp, J., 2012. The biortech.2013.06.097.
removal and degradation of pharmaceutical compounds during membrane bioreac- Yang, X., Flowers, R.C., Weinberg, H.S., Singer, P.C., 2011. Occurrence and removal of phar-
tor treatment. Water Sci. Technol. J. Int. Assoc. Water Pollut. Res. 65, 833–839. maceuticals and personal care products (PPCPs) in an advanced wastewater reclama-
https://doi.org/10.2166/wst.2012.828. tion plant. Water Res. 45, 5218–5228. https://doi.org/10.1016/j.watres.2011.07.026.
Sim, W.-J., Lee, J.-W., Oh, J.-E., 2010. Occurrence and fate of pharmaceuticals in wastewa- Yang, Y., Ok, Y.S., Kim, K.-H., Kwon, E.E., Tsang, Y.F., 2017. Occurrences and removal of
ter treatment plants and rivers in Korea. Environ. Pollut. 158, 1938–1947. https://doi. pharmaceuticals and personal care products (PPCPs) in drinking water and water/
org/10.1016/j.envpol.2009.10.036. sewage treatment plants: a review. Sci. Total Environ. 596–597, 303–320. https://
Suárez, S., Reif, R., Lema, J.M., Omil, F., 2012. Mass balance of pharmaceutical and personal doi.org/10.1016/j.scitotenv.2017.04.102.
care products in a pilot-scale single-sludge system: influence of T, SRT and recirculation Ying, G.-G., Kookana, R.S., Kolpin, D.W., 2009. Occurrence and removal of pharmaceuti-
ratio. Chemosphere 89, 164–171. https://doi.org/10.1016/j.chemosphere.2012.05.094. cally active compounds in sewage treatment plants with different technologies.
Tadkaew, N., Hai, F.I., McDonald, J.A., Khan, S.J., Nghiem, L.D., 2011. Removal of trace or- J. Environ. Monit. 11, 1498–1505. https://doi.org/10.1039/B904548A.
ganics by MBR treatment: the role of molecular properties. Water Res. 45,
2439–2451. https://doi.org/10.1016/j.watres.2011.01.023.
Taheran, M., Brar, S.K., Verma, M., Surampalli, R.Y., Zhang, T.C., Valero, J.R., 2016. Mem-
brane processes for removal of pharmaceutically active compounds (PhACs) from
Publication III
Khum Gurung, Mohamed Chaker Ncibi, Marina Shestakova, Mika Sillanpää

Removal of carbamazepine from MBR effluent by electrochemical
oxidation (EO) using a Ti/Ta2O5-SnO2 electrode

Applied Catalysis B: Environmental
Vol. 221, pp. 329-338, 2018
© 2018 Elsevier
Applied Catalysis B: Environmental 221 (2018) 329–338
Applied Catalysis B: Environmental

journal homepage: www.elsevier.com/locate/apcatb
Research Paper
Removal of carbamazepine from MBR effluent by electrochemical oxidation MARK

(EO) using a Ti/Ta2O5-SnO2 electrode
⁎
Khum Gurunga, , Mohamed Chaker Ncibia, Marina Shestakovaa, Mika Sillanpääa,b
a
Laboratory of Green Chemistry, School of Engineering Science, Lappeenranta University of Technology, Sammonkatu 12, FI-50130, Mikkeli, Finland
b
Department of Civil and Environmental Engineering, Florida International University, Miami, FL, 33174, USA
A R T I C L E I N F O A B S T R A C T
Keywords: This study aims at investigating the electrochemical oxidation (EO) of carbamazepine (CBZ) synthetic solutions
Electrochemical oxidation and real membrane bioreactor (MBR) effluent using newly developed Ti/Ta2O5-SnO2 electrodes to enhance CBZ
Ti/Ta2O5-SnO2 electrodes removal. The characterization of the prepared Ti/Ta2O5-SnO2 electrodes was performed by using scanning
Carbamazepine electron microscope, energy dispersive X-ray spectroscope, atomic force microscope, and cyclic voltammetry
Operating parameters
analyses. The main operating parameters influencing the CBZ removal efficiency in synthetic solutions using Ti/
Energy consumption
Ta2O5-SnO2 electrodes were evaluated including the applied current density, initial CBZ concentration, pH, and
MBR effluent
temperature. The optimum removal of CBZ (20 mg L−1) and TOC reached 75.5%, and 71.1%, respectively, after
8 h of electrolysis, under current density of 9 mA cm−2, pH 6, temperature of 30 °C, and using 0.1 M Na2SO4 as
supporting electrolyte. Increasing current density and temperature influenced the CBZ removal, unlike pH which
did not have significant influence on CBZ removal. The performance of Ti/Ta2O5-SnO2 electrode was compared
with conventional Ti/PbO2 electrode in terms of CBZ removal efficiencies and stability of the electrodes. The
results showed that under the same operating conditions, the CBZ removal efficiency of Ti/PbO2 electrode was
slightly higher than of Ti/Ta2O5-SnO2 (77.9 and 71.7%, respectively). Nonetheless, the use of this newly de-
veloped electrode is more energy-efficient as it required the lowest energy consumption of 60.3 kWh m−3 to
achieve optimum CBZ removal, in addition to fact that no heavy metals were leached (unlike the PbO2 elec-
trode). Furthermore, a complete degradative removal of real MBR effluents spiked with CBZ was achieved when
electrolyzed under the optimized conditions of CBZ synthetic solutions. Overall, the EO based on the use of Ti/
Ta2O5-SnO2 electrode was found to be a reliable approach to remove CBZ from contaminated waters, with
promising potential for integration with MBR technology to remediate CBZ.
1. Introduction [3,5]. In Finland, the annual consumption of CBZ is around 3.3 tons,
with comparatively more concentrations in winter seasons than in
Pharmaceutically active compounds (PhACs) have gained a great summer at WWTP effluents [8]. CBZ is highly recalcitrant to biode-
deal of attention among researchers in recent years as emerging mi- gradation in conventional activated sludge processes and photo-de-
cropollutants with long-term potential threats to aquatic environment gradation (more than 100 days) [2]. In fact, many researchers have
and human health [1–4]. These PhACs are introduced to aquatic eco- documented difficulties in removing CBZ, with biodegradation removal
systems primarily from the effluents of wastewater treatment plants of only less than 10% [3,4,6,8]. Thus, it is crucial to apply highly ef-
(WWTPs). In addition, direct and unsafe discharge of untreated was- ficient degradation technologies for the removal of CBZ in order to
tewaters from agriculture, industries and hospitals are also reported to achieve the highest mineralization or generate non (or less) toxic in-
contribute to the release of PhACs into the aquatic ecosystems termediates.
[2,3,5–7]. In this context, MBRs have been studied for the removal of various
Carbamazepine, which is used for the treatment of epilepsy, de- PhACs including CBZ by many researchers [9–11]. However, a rela-
pression, trigeminal neuralgia and wide variety of mental disorders, is tively low removal (0–28%) of CBZ was reported in MBR processes
one of the frequently identified PhAC in various aquatic environments when treating synthetic and real wastewaters [2,9,11]. Thus, biode-
[2,3,7]. The world-wide annual consumption of CBZ is more than 1000 gradation and membrane filtration alone are not efficient enough for
tons and about 28% of it ends up in WWTPs as unmetabolized forms. the removal of CBZ. Therefore, in order to meet the future legislations
⁎
Corresponding author.
E-mail addresses: khumgrg@yahoo.com (K. Gurung), mika.sillanpaa@lut.fi (M. Sillanpää).
http://dx.doi.org/10.1016/j.apcatb.2017.09.017
Received 1 July 2017; Received in revised form 29 August 2017; Accepted 5 September 2017
Available online 06 September 2017
0926-3373/ © 2017 Elsevier B.V. All rights reserved.
K. Gurung et al. Applied Catalysis B: Environmental 221 (2018) 329–338
regarding the stringent limits of PhACs from WWTP effluents, MBR wastewaters.
process integrated with various advanced oxidation processes (AOPs)
has gained increasing interest. 2. Materials and methods
AOPs are oxidation-based technologies applied in aqueous phase in
order to destruct organic pollutants via the in-situ generated highly 2.1. Reagents and materials
reactive oxygen species, mainly hydroxyl radicals (%OH) [12]. They are
promising and environmentally friendly methods for the treatment of All the reagents and materials used were of analytical grade. CBZ
wastewaters contaminated with a wide range of recalcitrant PhACs. (meets USP testing specifications), Titanium foil (99.7% trace metal
Among the various AOPs, EO processes are of growing interest for water basis, 2 mm thick), SnCl2·2H2O (≥99.99% trace metal basis), TaCl5
and wastewater decontamination due to their low cost and high effi- (≥99.99% trace metal basis), NaOH (≥98% anhydrous), H3PO4 (85%),
ciency [3]. In EO processes, dissolved organic pollutants are mainly Na2SO4 (≥99% anhydrous) were purchased from Sigma-Aldrich, USA.
oxidized by (i) direct anodic oxidation on the anode surface by charge Commercial Ti/PbO2 electrodes (5 cm × 10 cm × 2 mm) were sup-
transfer, and (ii) reaction with physio and/or chemisorbed hydroxyl plied by Magneto, Netherlands.
radical (%OH) generated from water oxidation (Eq. (1)) [13,14]. Hy- A synthetic stock solution containing CBZ (20 mg L−1) was pre-
droxyl radicals are extremely reactive oxidants (2.8 V (vs. SHE)) with pared using Milli-Q water (18.2 MΩ.cm), stirred at 500 rpm for 24 h at
the life time of less than a second, which makes their detection very room temperature [5]. Then, 0.1 M (14.2 mg L−1) of Na2SO4 (better
challenging [15]. The subsequent oxidation mechanism of organic supporting electrolyte than NaCl regarding economical, effectiveness
compounds depends on the behavior of anodes, which are divided as and environmental aspects) was added to increase conductivity of the
“non-active” and “active” [13–16]. In both cases, firstly, water mole- solution. In this regard, the possible formation of organic chlorine by-
cules splitted and leading to the formation of OH at the anode oxide products is also limiting the use of NaCl in electrochemical processes
surface (MOx) (Eq. (1))[17]. [5]. Finally, the synthetic stock solution was stored in the refrigerator at
4 °C for further use.
H2O + MOx = MOx (%OH) + H+ + e− (1)
At ‘active’ anodes, OH radicals are chemisorbed (strongly adsorbed), 2.2. Preparation of Ti/Ta2O5-SnO2 electrodes
thus allow only selective oxidation of organics (R), which involves in-
termediate step of higher oxide formation (Eqs. (2) and (3))[14,16]. Ti/Ta2O5-SnO2 electrodes were prepared by thermal decomposition
of precursor solutions on Ti substrates [25]. For the preparation of the
MOx (%OH) → MOx+1 + H+ + e− (2) electrodes, the methodology is detailed in a previous study conducted
MOx + 1 + R → MOx + RO (3) in our laboratory [26]. The optimized nominal composition of Ta (7.5
at.%) and Sn (92.5 at.%) were used to prepare precursor solutions by
On the other hand, at ‘non-active’ anodes, the formation of higher dissolving SnCl2·2H2O and TaCl5 salts in absolute alcohol (pro analysis,
oxide is excluded and OH radicals are physisorbed without interaction Panreac), where total concentration of metal ions was kept constant
on the anode surface [14,16]. Moreover, ‘non-active’ anodes with high (0.04 M). At first, Ti substrate plates (5 cm × 10 cm × 2 mm) were
O2-overpotential towards water oxidation allow good electrochemical pre-treated by mechanical polishing and followed by degreasing in
stability and can generate many strong (OH) radicals predominantly 10 wt.% NaOH for 10 min and etching in boiled 18 wt.% HCl (pro
leading to the complete combustion of organics (Eq. (4)) [15–17]. analysis, Fluka) for 30 min, and finally rinsed carefully using Milli-Q
water. Secondly, the precursor solution was applied over only one side
MOx (%OH) + R = MOx + CO2 + H2O + zH+ + ze− (4)
of the pre-treated Ti substrates by drop casting method. The precursor
The major problems of EO process are the polarization and passi- solution of 1 mL volume was dropped on to the Ti substrate, uniformly
vation of electrodes, related either to the poor mass transfer, corrosive distributed using the tip of a Pasteur glass pipette, dried in oven at 80 °C
environment, formation of polymer films or the inevitable formation of for 5 min, and then annealed in muffle furnace at 550 °C for 5 min. The
oxide interlayers due to the considerable porosity in supported catalyst process was repeated until 8 layers of ultrathin film deposition of the
films, [14,15,18]. In any cases, those phenomena result in the at- precursor solution, which attributed highest electro-catalytic properties
tenuation of the efficiency and service life of electrodes. of the electrodes [26]. After the final layer, electrodes were additionally
Various anodes such as PbO2, and more recently, boron-doped annealed at 550 °C for 10 h to enhance the formation of composite
diamond (BDD) have shown an interesting electro-degradation effi- oxides of Ta and Sn on the electrode surface via thermal decomposition.
ciency for CBZ [2,3,6]. Nevertheless, the higher cost of BDD electrodes The final composite oxides on Ti substrate gave the wide band gap n-
and the possible leaching of lead (Pb) from PbO2 coated electrodes tend type SnO2 semi-conductor extrinsically doped with Ta2O5. After 10 h,
to compromise their applicability in large-scale applications [19,20]. In electrodes were cooled down and kept in an ultrasonic bath for 5 min in
fact, PbO2 is one of the widely studied anode on the electrochemical order to remove the surface impurities and deformations, and then oven
degradation of PhACs [2,6,21,22]. PbO2 anodes are inexpensive, rela- dried at 105 °C. Thus, the precursor material deposited only one side of
tively easy to prepare, has low electrical resistivity, and has a large Ti substrates had surface area of 50 cm2 and the deposited material was
surface area [23]. In this context, Ti/Ta2O5-SnO2 (SnO2 semi-conductor equal to 1.10 mg cm−2 ± 5%.
doped with Ta2O5) are eco-friendly electrodes, which were proven to be
efficient in the oxidation of some organic compounds [18,24]. 2.3. Physicochemical and electrochemical characterization of the electrodes
The main aim of the current research is to evaluate the efficiency of
this novel electrode to remove persistent CBZ drug from synthetic so- Microstructure of the Ti/Ta2O5-SnO2 electrodes was analyzed by
lutions and the real MBR effluent. First, the EO of synthetic solution using scanning electron microscope (SEM) (Hitachi, SU3500, Japan)
loaded with CBZ was investigated by monitoring and optimizing the using 5 kV of acceleration voltage. The chemical analysis of selected
major influencing factors including current density, pH, and tempera- areas of the electrode surface was performed with energy-dispersive X-
ture, under various initial concentrations of CBZ. In addition, the de- ray spectroscope (UltraDry™ Compact EDS Detector, Thermo Fisher,
gradation efficiency and stability of Ti/Ta2O5-SnO2 were compared USA). The AFM images were taken by using atomic force microscope
with Ti/PbO2 electrodes. Furthermore, optimized operating conditions (AFM) (NX10, Park, Korea) using the non-contact mode.
in EO process were applied for the treatment of real MBR effluent The electrochemical characterization of the Ti/Ta2O5-SnO2 elec-
spiked with CBZ. This study is the first research work on the EO of CBZ trode was performed by conducting cyclic voltammetry (CV) mea-
using novel Ti/Ta2O5-SnO2 electrodes in synthetic solutions and real surements using an Autolab (PGSTAT 12 Potentiostat/Galvanostat,
330
Fig. 1. SEM and AFM images of Ti/Ta2O5-SnO2 electrodes before electrolysis (a–c) and after 8 h of electrolysis (d–f) under current density of 9 mA cm−2.
Metrohm, Switzerland) connected with GPES software (Eco Chemie, Supplementary data) in order to calculate the residual CBZ concentra-
Netherlands). All the CV measurements were performed in a three tion and thus the removal efficiency. As for the real MBR effluent, since
electrode cell (100 mL). A coiled platinum wire was used as counter the spiked samples have low concentration of CBZ (10 μg L−1), they
electrode and saturated calomel electrode (SCE) as reference electrode, were analyzed with LC–MS/MS (Acquity UPLC/Xevo TQMS, USA) using
whereas Ti/Ta2O5-SnO2 electrode (1 cm × 1 cm) was used as working 5 cm C18 column (Acquity UPLC BEH, USA).
electrode. First, the voltammograms were recorded at different scan
rates from 20 to 100 mV s−1 in 0.1 M Na2SO4 aqueous solution between 2.5.2. Other measurements
0.6 and 0.8 V (vs. SCE), where only double-layer currents (non-faradic) Non-Purgeable Organic Carbon (NPOC) was measured with TOC-
were observed for estimating the relative roughness factor. Second, the analyzer (TOC-V Series-CPN, Shimadzu, Japan) to evaluate the miner-
voltammograms were plotted between the anodic currents against alization efficiency. The pH was measured using pH meter (pH 110,
electrical voltage between 0.2–2.5 V (vs. SCE) at scan rate of VWR). The electrical conductivity was monitored by using conductivity
100 mV s−1 with or without CBZ in 0.1 M Na2SO4 aqueous solution. In meter (Euroscan Con 6, EUTECH, USA).
each of the assays, aqueous solution inside the CV cell was deox- The energy consumption (EC) needed for the electrochemical
ygenated by argon bubbling prior to the measurements. treatment per volume of working solution was calculated based on Eq.
(5) [15].
2.4. Electrochemical degradation assays
Ucell It
EC (kWh m−3) =
Vs (5)
The experimental set-up for the electrochemical degradation system
is shown in Fig. S1. The CBZ-containing solutions were electrolyzed where Ucell is the average potential difference of the cell (in V), I is the
using an open glass reactor of 500 mL capacity with cooling jacket for applied current (in A), t is the electrolysis time (in h), and Vs is the
the circulation of thermostated water, under a 500 rpm stirring. The working solution volume (in dm3).
reaction temperature was maintained at 11 ± 1 °C. The electro-
chemical unit consisted of Ti/Ta2O5-SnO2 electrode as anode and pure 3. Results and discussion
Ti plate (same area) as cathode, with an inter-electrode gap of 10 mm.
The electrical power was applied via DC power supply (EX752M- 3.1. Microstructural characterization
Multimode PSU, TTI, UK) with maximum current and potential rating
of 4A and 150 V, respectively. All the electrochemical degradation as- SEM and AFM micrographs of Ti/Ta2O5-SnO2 electrode before and
says were performed under galvanostatic conditions. Each experiment after the 8 h of electrolysis using the current density of 9 mA cm−2 are
was run for 8 h with sampling every hour. shown in Fig. 1a-f. In SEM images, the electrode surface before elec-
trolysis (Fig. 1a-b) showed foam-like porous microstructure with large
2.5. Analytical methods numbers of relatively homogenous agglomerates mainly related to the
density of Ta2O5 crystals, which can partially increase the surface
2.5.1. CBZ measurements roughness. In fact, the porosity and relative surface roughness depend
The removal of CBZ was analyzed by UV–vis spectrophotometer on the active oxide layers. When the electrode was subjected to elec-
(Lambda 45, Perkin Elmer, USA) by measuring the absorbance of UV trolysis for 8 h, the surface exhibited structural changes in the surface
light by CBZ at 210 nm. A calibration curve was created by scanning showing heterogeneously-distributed porous network and more rough
different concentrations of CBZ from 1 to 30 mg L−1 (Fig. S1, structure as shown in Fig. 1d–e.
331
Fig. 2. EDX mappings of Ti/Ta2O5-SnO2 electrodes before (a–c) and after 8 h of electrolysis (d–f) under current density of 9 mA cm−2.
Likewise, the AFM analysis before electrolysis shows a more even reasonable agreement with the nominal mixture used [26]. Moreover,
topography of the electrode surface as shown in Fig. 1c. Whereas, the the atomic ratio of Sn/Ta from EDX (10.9) was more closed to the
surface structure of the electrode after the 8 h of electrolysis depicts nominal ratio (12.33). This suggests that the Ti substrate is covered
more deformed surface with visible deep and extended valley (dark more uniformly with Sn, whereas Ta deposition is more dispersed.
spots in Fig. 1f). The depth of undulations was noted to be ± 1.5 μm Overall, the metal fractions did not change noticeably before and after
and ± 4 μm on the electrode surface before and after the electrolysis, the electrolysis, though morphological changes existed.
respectively.
3.3. Electrochemical characterization
3.2. Elemental characterization The electrocatalytic activity of any electrode depends mostly on
factors such as relative roughness factor of the electrode surface and the
The EDX elemental microanalysis was conducted to determine the voltammetric currents associated with the degree of electroactive sites
bulk composition of metal oxide thin films on Ti/Ta2O5-SnO2 electrode. [27]. For the roughness factor, the current values for different sweep
The EDX elemental mappings on the surface of Ti/Ta2O5-SnO2 elec- rates were measured at E = 0.7 V (vs. SCE), where an approximate
trode under the current density of 9 mA cm−2, before and after the 8 h symmetry for anodic and cathodic currents was recorded as shown in
of electrolysis, are as shown in Fig. 2a–c and d–f, respectively. The Fig. 3a. By plotting the anodic currents with respect to the scanning
atomic molar percentages of the precursor metals from EDX were found rates, a regression equation is deduced (shown in Fig. 3a, inset) as,
to be Sn- 91.7 at.% and Ta-8.3 at.% before and Sn-91.5 at.% and Ta-8.5
at.% after the electrolysis. The elemental composition of metals showed I (A) = 0.0001 v (Vs−1) + 2.6 × 10−6, R2 = 0.9977 (6)
Fig. 3. Cyclic voltammograms of Ti/Ta2O5-SnO2 in

(a) double-layer region (0.6–0.8 V) at scan rates from
20 to 100 mV s−1 in 0.1 M Na2SO4 solution. Inset:
Linear regression of I versus v measured at
E = 0.7 V; (b) CBZ (30 mg L−1, pH 5.7) with sup-
porting electrolyte 0.1 M Na2SO4, scanned between
0.2–2.5 V at scan rate of 100 mV s−1.
332
Fig. 4. Effect of applied current densities on the removal of CBZ

during electrolysis (CBZ (0): 20 mg L−1 + 0.1 M Na2SO4; pH-5.7;
T = 11 ± 1 °C) on Ti/Ta2O5-SnO2 anodes. Inset: fitting of the dif-
ferent current densities to a pseudo-second order kinetic model.
The linear relationship between electric current and sweep rate 17 mA cm−2, respectively. Thus, beyond current density of 9 mA cm−2,
shows that the Faradic current in this region is negligible, thus only the neither removal rate nor the kinetics increased noticeably. Two possible
electric double-layer current is measured [27]. The double-layer char- reasons might have associated with this phenomenon. First, the un-
ging current is much related to the double-layer capacitance of the desirable parasitic reactions of %OH radicals to O2 gas or dimerization
electrode/solution interface as conductive layer was made of metal to weak oxidants such as %OH2 radicals, which could compete the main
oxides [28]. So, the normalized capacitance of 0.0001 F cm−2 was oxidation reaction (via %OH radicals), and can reduce the oxidation
calculated by dividing the slope value of Eq. (6) with the geometric efficiency of the targeted pollutants [6,14,23,32]. Second, as anodic
surface area of the electrode. Further, the capacitance of the electrode oxidation takes place heterogeneously with physio-adsorbed %OH ra-
surface was compared with that of a smooth surface oxide layer, with dicals at electrode surface, electrochemical degradation rate after
the capacitance of 8 μF cm−2 [27], which gave a relative roughness reaching the limiting current density might be directly proportional to
factor of 12.5. Therefore, the roughness of electrode surface has slightly the mass transfer rate of pollutants towards the electrode surface [23].
improved by Ta doping over SnO2 layer, which in turn have attributed a The current density of 9 mA cm−2 was selected as optimum current
good electrocatalytic activity to the Ti/Ta2O5-SnO2 electrodes. How- density for further study to ensure both an improved removal efficiency
ever, roughness factors of metal oxide films might vary by factor of and a low energy consumption.
several thousands with high discrepancy in magnitude between them Furthermore, in order to evaluate the degradation kinetics of CBZ,
[28]. the kinetic data were analyzed by using different kinetic models:
Furthermore, CV tests were again performed to investigate the pseudo-first order kinetic model and pseudo-second order kinetic
electrochemical property of CBZ molecule at Ti/Ta2O5-SnO2 electrode, model. The following Eqs. (7) and (8) were used for the pseudo-first
as shown in Fig. 3b. The oxygen evolution reactions (OERs) appeared order and pseudo-second order kinetic models,
between 1.7 and 2 V (vs. SCE) in both the solutions. In the presence of
C0
CBZ, a small anodic current peak was noticed at potential of 1.45 V (vs. Ln = k1t
Ct (7)
SCE) in the voltammogram (Fig. 3b). The absent of more distinct peaks
for CBZ oxidation might probably be due to the CBZ oxidation at a
1 1
potential close to OER region. Furthermore, it can be noticed that there − = k2t
Ct C0 (8)
was no reduction peak recorded in the voltammogram during reverse
scanning, which might be due to the irreversible characteristics of CBZ where C 0 is the initial concentration of CBZ; C t is the concentration of
[29]. CBZ at time t; t is the reaction time; k1 is the pseudo-first order rate
constant (s−1); and k2 is the pseudo-second order rate constant
3.4. Effects of operating parameters on the removal of CBZ (L mol−1 s−1). The regression coefficients (R2) for both pseudo-first
order and pseudo-second order kinetics were compared corresponding
3.4.1. Effect of current density to the different current densities as indicated in Table 1. The calculated
In electrochemistry, applied current density is the most important values of R2 from the pseudo-first order kinetic model were slightly
parameter which controls electron transfer and the generation of re- lower than the value obtained from the pseudo-second order kinetic
active oxidants [30], thus directly influences the removal rates of pol- model. These results indicated that the pseudo-second order model
lutants. Such major influencing role of current densities during the better describes the degradation of CBZ by EO process. This is in
electrochemical removal of CBZ has been reported in many studies
[5,6,20,23,31]. In this study, different current densities ranging from 1 Table 1
to 17 mA cm−2 were applied in order to select an optimum current CBZ removal efficiency (%), regression coefficient (R2), pseudo-rate constant (kpseudo),
and half-life, as a function of current densities.
density with higher rate of CBZ removal as shown in Fig. 4.
It was observed that after the 8 h of electrolysis, the CBZ removal Current densities (mA cm−2) 1 5 7 9 15 17
reached 53.0%, 67.0%, 68.5%, and 71.7% with corresponding current
densities of 1, 5, 7, and 9 mA cm−2, respectively. This trend is in ac- % CBZ removal 53.0 67.0 68.5 71.7 73.9 72.1
Pseudo-First order model, R2 0.998 0.971 0.980 0.971 0.970 0.950
cordance with the fact that increasing current density increases the
Pseudo-Second order model, R2 0.997 0.996 0.993 0.995 0.991 0.983
organic pollution removal efficiency by the generation of more reactive k2 (L mol−1 s−1) 0.47 0.91 1.12 1.42 1.57 1.46
oxygen species (esp. %OH radicals) [3,7,23]. The generated powerful % Half-life (h) 6.9 3.6 2.9 2.3 2.1 2.3
OH radicals then attack CBZ molecules and destruct it. Likewise, the
CBZ removal rates were 73.8% and 72% at current densities of 15 and
333
agreement with the previous work by J. D. García-Espinoza et al. [6], more intermediate products, which in turn competes with CBZ in the
who also reported that the second-order was the best model to fit the reaction with %OH radicals and can reduce the CBZ removal efficiency
kinetics in the electrochemical removal of CBZ with Ti/PbO2 anodes. [21]. So, while increasing the initial concentration of CBZ under
However, the pseudo-first order kinetic model for CBZ removal using an identical current densities, the oxidation process can be limited by the
EO process have also been reported in other works [5,34,35]. In the formation of %OH radicals [20]. Further, the electrode surface might be
present work, since R2 values for both the models are comparable, the mediated with passivation (diffusion layer) due to the deposition of
pseudo-first order model is also valid to describe the reaction kinetics of more reaction products. On the other hand, at lower concentrations, the
CBZ degradation. In this regard, it is worth noting that within 8 h of the higher CBZ removal rate was achieved, which indicates that the greater
reaction time, the formation of intermediate compounds could slightly amount of CBZ molecules are attacked by the %OH radicals. This result
distort the kinetic analysis. can be explained by the phenomena that at low initial concentrations,
The tendency of the enhanced CBZ removal efficiencies corre- electrochemical degradation rates of organic molecules can be faster
sponding to the increased current densities are further supported with than the diffusion of side-products [21].
the results of pseudo-second order rate constant values, k2 (Table 1).
The increasing trend in the values of k2 as a function of applied current
densities indicates that amplified OERs are associated with increased 3.4.3. Effect of pH
current densities. With the increasing O2 gas evolution, the mass pH values can have a significant influence on the electrochemical
transport phenomena increases indicatively leading to the enhanced treatment processes [19]. To assess the effect of pH on the removal
diffusion flux of CBZ molecules towards anode surface, which subse- efficiency of CBZ at Ti/Ta2O5-SnO2 anode, different initial pH condi-
quently increases k values [20]. From Table 1, at optimum current tions ranging from acidic to basic media (2–10) were investigated.
density of 9 mA cm−2, the value of k2 was 1.42 L mol−1 s−1 and the Fig. 6 depicts the degradation kinetics of CBZ (20 mg L−1) when elec-
half-life was found to be 2.3 h. It indicates that 50% of the CBZ removal trolyzed for 8 h under the current density of 9 mA cm−2 with sup-
was reached within 2.3 h of electrolysis under the current density of porting electrolyte of 0.1 M Na2SO4. In Fig. 6 (inset), the residual
9 mA cm−2. concentrations of CBZ after the 8 h electrolysis are found to be de-
creasing from about 6.3, 5.7, and 5.0 mg L−1 at initial pH values of 2, 4,
3.4.2. Effect of initial CBZ concentration and 6, respectively. However, the residual concentrations are in in-
It is obvious that the concentrations of CBZ fluctuates periodically creasing trend in basic media ranging from 5.7 to 6.0 mg L−1 at cor-
or seasonally at wastewater treatment plants [89,36]. Therefore, it is responding pH values of 8 and 10, respectively. Under the identical
more significant to investigate the effect of the initial concentration of conditions, the optimum removal of 75.2% was achieved at pH value of
CBZ on the removal efficiency of electrochemical treatment process. In 6. Whereas, the removal efficiencies were 68.8%, 71.2%, 71.5%, and
this context, different concentrations of CBZ ranging from 2 to 70.1% at the pH values of 2, 4, 8, and 10, respectively. Thus, these
30 mg L−1 were electrolyzed under the constant current density of results indicate that the influence of pH was not significant for elec-
9 mA cm−2 for 8 h, in order to evaluate the influence of initial con- trochemical removal of CBZ, which is in agreement with other studies
centrations of CBZ on the removal efficiency, as shown in Fig. 5. [3,6,23]. The reason might be the high acid-dissociation constant (pKa)
The removal rates of about 80.3%, 73.3%, 71.8%, and 70.2% were of CBZ (13.94), for which no significant dissociation is generally re-
achieved after the 8 h of electrolysis when initial concentrations of CBZ ported, either in acidic or basic conditions [6]. In practice, this ten-
were 2, 10, 20, and 30 mg L−1, respectively (Fig. 5, inset). The removal dency will be useful in real treatment cases since MBR processes are
rate was much faster during the first 2 h of operation in all of the four normally operated at pH of 6–7 [9]. On the other hand, the decreasing
different assays, however, decreased from 80.3 to 70.2%, when the trend of the CBZ removal efficiency with increasing pH values might be
initial CBZ concentrations increased from 2 to 30 mg L−1. About 70% attributed to the significant shift of OER onset potential to less positive
of CBZ removal was achieved within 3.5 h of electrolysis for 2 mg L−1 values (low O2-overvoltage), which favors more oxygen formation than
of initial CBZ concentration, whereas 8 h were needed to achieve the pollutant oxidation [37]. It was also noticed that the solution pH did
same CBZ removal for initial CBZ concentration of 30 mg L−1. These not change significantly in acidic assays over the 8 h of time. Never-
results are consistent with other works in which electrochemical de- theless, the solution pH was found to be changing into a more acidic
gradation gave better results at low initial concentration [21,23,31]. character (5–6) when performed in alkaline media (pH 8–10 initially),
At higher initial concentrations, the mass transfer of organic sub- which might be due to the formation of weak acid products such as
stances increases, which ultimately inhibits the interaction between the carboxylic acids [32,38].
targeted pollutant and anode active sites [21,23]. This can generate
Fig. 5. Effect of the initial CBZ concentrations on the removal ten-

dency during electrolysis (CBZ (0): 1–30 mg L−1; 0.1 M Na2SO4; pH:
5.7; T = 11 ± 1 °C; current density: 9 mA cm−2) on Ti/Ta2O5-SnO2
anodes. Inset: CBZ removal efficiencies at different initial CBZ con-
centrations (2, 10, 20, and 30 mg L−1).
334
Fig. 6. Effect of initial pH on the removal tendency of CBZ during

electrolysis (CBZ (0): 20 mg L−1 + 0.1 M Na2SO4; current density:
9 mA cm−2; T = 11 ± 1 °C) on Ti/Ta2O5-SnO2 anodes. Inset: Residual
concentrations and corresponding CBZ removal% after 8 h of elec-
trolysis at different pH.
3.4.4. Effect of temperature Moreover, the degree of mineralization was also analyzed at this
The reaction rate constants are exponentially dependent on the stage under the optimum operating conditions. Using current density of
temperature of the medium following the Arrhenius law. The similar 9 mA cm−2, pH of 6.0 and initial CBZ concentration of 20 mg L−1,
principle can be expected in the electrochemical processes. Therefore, optimum removal of TOC were 61.2%, 69.8% and 71.1% under the
to investigate the effect of temperature on the electro-oxidation, solution temperatures of 10 °C, 20 °C, and 30 °C, respectively (Fig. 7,
20 mg L−1 of CBZ solution was electrolyzed on Ti/Ta2O5-SnO2 anode inset). Therefore, the average rate of mineralization during this work
under the current density of 9 mA cm−2 at three different temperatures was about 61%, as the operating temperature was maintained at about
ranging from 10 to 30 °C. Fig. 7 shows the effect of different tempera- 11 ± 1 °C. It can be concluded that increase in the temperature has
tures on the CBZ degradation kinetics during 8 h of electrolysis. The positive effect on the degree of mineralization too. Nevertheless, the
removal efficiencies of 71.7%, 74.1%, and 75.5% were achieved when mineralization efficiency was always lower than that of CBZ (71.1%
corresponding operating temperatures were maintained at 10 °C, 20 °C, and 75.5%, respectively). The most possible reason might be attributed
and 30 °C (Fig. 7, inset). Thus, increasing the temperature had a slight to the fact that carboxylic acids are highly recalcitrant, thus they de-
positive influence on the kinetic rates. One possible reason might be a grade slower than parent compounds [32].
better mass transfer of pollutants towards the anode surface at higher
temperature, due to high molecular agitation which could enhance the
oxidation efficiency [39]. Another reason might be the indirect oxida- 3.5. Comparison between Ti/Ta2O5-SnO2 and Ti/PbO2 electrodes
tion reaction of organics in the bulk solution by electrogenerated oxi-
dants peroxodisulphates (S2O82−), which is one of the powerful oxi- In order to evaluate the effect of different electrode materials on the
dizing species, from the oxidation of electrolyte (Na2SO4). Indeed, it CBZ removal efficiencies, Ti/Ta2O5-SnO2 and Ti/PbO2 electrodes were
was reported that the rate of oxidation of organic pollutants via per- compared. The stability of both electrodes were also compared by
oxodisulphates increases with positive increments in the solution tem- performing leaching tests. Fig. 8 shows the removal kinetics of
perature [31]. In terms of hydrodynamics of the bulk solution, viscosity 20 mg L−1 CBZ in aqueous 0.1 M Na2SO4 solution, pH 6.0, T = 10 °C,
decreases with increasing solution temperature, which enhances mass and using current density of 9 mA cm−2. The removal trends of both Ti/
transfer rate of organic species towards electrode surface thus con- Ta2O5-SnO2 and Ti/PbO2 electrodes were almost identical until 2 h of
tributes higher oxidation efficiency [20]. electrolysis. However, after 8 h of electrolysis, the CBZ removal effi-
ciency reached 71.7% and 77.9% on Ti/Ta2O5-SnO2 and Ti/PbO2
Fig. 7. Effect of temperature on the removal tendency of CBZ during

electrolysis (CBZ (0): 20 mg L−1 + 0.1 M Na2SO4; current density:
9 mA cm−2; pH: 6) on Ti/Ta2O5-SnO2 anodes. Inset shows% removal
of CBZ and TOC at different temperatures.
335
Fig. 8. The comparison on the removal tendencies of CBZ during

electro-oxidation (CBZ (0): 20 mg L−1 + 0.1 M Na2SO4; current den-
sity: 9 mA cm−2; pH: 6, T = 10 °C) on Ti/Ta2O5-SnO2 and Ti/PbO2
anodes.
anodes, respectively (Fig. 8). From the results, it can be concluded that electrolysis, respectively. In contrast, the leaching of Ta was measured
under the identical operating conditions, the Ti/PbO2 electrode is more only in the range of 2–14 μg L−1, indicating a low dissolution in to the
effective on CBZ removal. This tendency could be explained by the fact bulk solution. In this regard, currently there are no specific regulations
that the Ti/PbO2 anodes have greater electrocatalytic activity than Ti/ on the concentrations of Ti, Sn, or Ta in surface water sources. More-
Ta2O5-SnO2 anodes due to their higher relative roughness factor [40]. over, the CBZ removal efficiency of Ti/PbO2 and Ti/Ta2O5-SnO2 was
The increase in surface roughness can produce larger amount of phy- comparable, nonetheless the latter is more environmentally friendly
sisorbed %OH radicals under the identical current density [21], which since the leaching phenomenon (frequently associated with EO) will not
could react rapidly with all organic species at the vicinity of anode generate heavy metals into the treated waters.
surface [31].
Furthermore, stability of both the electrodes was examined in terms
of the leaching of Ta, Sn, Ti and Pb, by monitoring their ion con- 3.6. The energy consumption
centrations in the bulk solution. Each electrodes were electrolyzed for
24 h by reusing the same electrodes over three days (8 h per day with The electrical energy consumed to treat specific volumes polluted
fresh solutions each run), as mentioned in Fig. 9a,b. solutions is the prime concern to assess the viability of any electro-
In Fig. 9a, the dissolution profile of Ti/PbO2 anode exhibits no chemical degradation processes. The EC evolved during the electro-
leaching of Pb during the first 8 h of electrolysis. However, the dis- chemical degradation of CBZ at varying current densities (Cf. Section
solution of PbO2 was significant to bulk solution during 16 h and 24 h 3.4.1) are shown in Fig. 10. It indicates that the minimum electrical
of the electrolysis, releasing higher amounts of Pb (concentrations of energy consumption of 3.2 kWh m−3 was required to reach 53% of CBZ
37.3 and 25.8 μg L−1, respectively). Likewise, traces of Ti ions were removal while using the current density of 1 mA cm−2. The EC in-
also measured from the solution with maximum of 14.2 μg L−1 during creased from 26.3 to 60.3 kWh m−3 when current densities increased
the first 8 h of electrolysis. These results show that current density of from 5 to 9 mA cm−2. Beyond the current density of 9 mA cm−2 until
9 mA cm−2 is capable of dissolving PbO2 from the electrode surface. A 17 mA cm−2, the EC increased from 60.3 kWh m−3 and reached to
significant concentration of Pb leached from PbO2 oxide coated elec- 172.7 kWh m−3, even though the removal efficiencies were not im-
trodes has also been reported in another work [20]. It is worth noting proved significantly. For the optimum current density of 9 mA cm−2,
that the maximum allowable concentration of Pb in surface water re- the energy required to degrade 71.7% of CBZ was 60.3 kWh m−3. It has
sources, according to EU directive is 14 μg L−1 [41]. been observed that the current densities higher than the limiting one
On the other hand, leaching profile of Ti/Ta2O5-SnO2 shows the (9 mA cm−2 in the present study) can be associated with high energy
increasing trend for Sn and Ti, as compared to Ta (Fig. 9b). The consumption. The reason can be that a large mass of OH radicals might
leaching of Sn was the highest with concentrations of 145.7, 318.2, and be evolved to O2 due to parasitic reactions, then a large cell voltage is
407.3 μg L−1, respectively after 8, 16 and 24 h of electrolysis. Similarly, often used mostly for O2 generation, which does not help further in
Ti concentration increased from 53.6 to 92.9 μg L−1 after 8 and 24 h of oxidizing the organic pollutant [31].
Fig. 9. Leaching profile of Pb, Ta, Sn, and Ti from (a)

Ti/PbO2 and (b) Ti/Ta2O5-SnO2 anodes when elec-
trolyzed for 24 h by reusing the same electrodes but
using fresh working solutions (CBZ(0):
20 mg L−1 + 0.1 M Na2SO4) for each assays; under
current density of 9 mA cm−2.
336
Fig. 10. The electrical energy consumption trend against the CBZ
removal efficiency during the electrochemical oxidation (CBZ (0):
20 mg L−1; 0.1 M Na2SO4; pH: 5.7; T = 11 ± 1 °C) on Ti/Ta2O5-
SnO2 anode under different current densities.
3.7. Applicability of EO process in real MBR effluent for CBZ removal real MBR effluents using a newly developed Ti/Ta2O5-SnO2 electrodes
was investigated under galvanostatic conditions. The characterization
In practice, the optimum operating conditions during the removal of of the Ti/Ta2O5-SnO2 electrodes were performed by using SEM, EDX,
CBZ from synthetic solution were applied for the treatment of CBZ in AFM, and cyclic voltammetry analyses. The main operating parameters
the real MBR effluents, collected from a local MBR pilot. The char- influencing the CBZ removal efficiency in synthetic solutions on the
acteristics of the MBR effluent, detailed specifications and operating tested anodes were evaluated as a function of applied current density,
conditions of the MBR pilot plant are described in previous study [9]. initial CBZ concentration, pH, and temperature. The results showed that
The MBR effluent samples were then spiked with a precise amount of the maximum removal of CBZ (20 mg L−1) and TOC reached 75.5%,
CBZ to have an initial concentration of 10 μg L−1. Moreover, two dif- and 71.1%, respectively, after 8 h of electrolysis, current density of
ferent working solutions, with or without supporting electrolyte, were 9 mA cm−2, solution pH 6, temperature of 30 °C, and using 0.1 M
prepared and experimented. The details of experimental conditions and Na2SO4 as supporting electrolyte. The performance of Ti/Ta2O5-SnO2
results are presented in Table 2. electrode was compared with Ti/PbO2 electrode in terms of CBZ re-
From Table 2, it can be noticed that the EO method, which was moval efficiencies and stability. The results showed that, under the
operated under optimal conditions, has shown an excellent CBZ re- same conditions, the CBZ removal efficiency of Ti/PbO2 electrode was
moval efficiencies in the real MBR permeate, independent of the in- slightly higher when compared with the investigated electrode, but the
fluence of supporting electrolyte. About 90% removal of CBZ was at- latter was more environmentally friendly. Besides, achieving an optimal
tained after 2 h of electrolysis, which further reached 99.99% within CBZ removal efficiency of 71.7% required the lowest energy con-
4 h of electrolysis, for both assays (data not shown). These results are sumption of 60.3 kWh m−3.
consistent with the findings of García-Gómez et al. [2], where low Furthermore, the complete degradative removal of CBZ was
concentrations of CBZ spiked into the synthetic wastewater was elec- achieved in spiked real MBR effluents, under the optimal operating
trolyzed by using EO method. Nevertheless, the energy consumption conditions. Overall, the EO based on the use of novel Ti/Ta2O5-SnO2
was higher when no supporting electrolyte was used as compared to the electrode was found to be a reliable process to remove CBZ from con-
use of it (Table 2). High salinity of effluent wastewater could increase taminated waters, with promising potential for integration with MBR
the conductivity and thus can lower the energy consumption [1]. technology to remediate CBZ.
Nonetheless, the conductivity of the MBR effluent in this work was
reported to be 601 μS cm−1, which was not enough to maintain a
constant electrical pressure during the redox reactions. On the other Acknowledgements
aspect, the cost of using supporting electrolyte (approx. US$ 70 per
metric ton) is much cheaper than the consumption of electricity (ap- The authors are very grateful for the financial support provided by
prox. US$ 0.05 per kWh). TEKES-Finnish Funding Agency for Innovation 1043/31/2015
(Helsinki, Finland), a research fund for the Smart Effluents Project. Mr.
4. Conclusion Toni Väkiparta, for SEM/EDX analysis of electrodes, and Mr. Risto Repo
(Mikkeli Water works), for helping in operating the MBR unit and
The electrochemical oxidation of CBZ from synthetic solutions and providing the effluent samples.
Table 2
Electrochemical oxidation experiments on real MBR effluent for the validation of CBZ removal (Current density: 9 mA cm−2; pH: 6; T: 11 ± 1 °C; at Ti/Ta2O5-SnO2 anodes).
Parameter Initial concentration in MBR effluent Final concentration after 8 h of electrolysis Electrolyte Removal (%) EC (kWh m−3)
−1 a
CBZ (μg L ) 10.75 ± 0.35 < 0.07 (LOD) No electrolyte 99.99 109.4
< 0.07 (LOD)a 0.1 M Na2SO4 99.99 57.2
a
(LOD) Limit of detection.
337
Appendix A. Supplementary data J. Mol. Liq. 223 (November) (2016) 388–394.

[19] H. Särkkä, A. Bhatnagar, M. Sillanpää, Recent developments of electro-oxidation in
water treatment—a review, J. Electroanal. Chem. 754 (October) (2015) 46–56.
Supplementary data associated with this article can be found, in the [20] Y. Samet, L. Agengui, R. Abdelhédi, Anodic oxidation of chlorpyrifos in aqueous
online version, at http://dx.doi.org/10.1016/j.apcatb.2017.09.017. solution at lead dioxide electrodes, J. Electroanal. Chem. 650 (December (1))
(2010) 152–158.
[21] Q. Dai, et al., Electrochemical degradation of levodopa with modified PbO2 elec-
References trode: parameter optimization and degradation mechanism, Chem. Eng. J. 245
(June) (2014) 359–366.
[22] M.S. Morsi, A.A. Al-sarawy, W.A.S. El-dein, Electrochemical degradation of some
[1] S.O. Ganiyu, E.D. van Hullebusch, M. Cretin, G. Esposito, M.A. Oturan, Coupling of
organic dyes by electrochemical oxidation on a Pb/PbO 2 electrode, Desalin. Water
membrane filtration and advanced oxidation processes for removal of pharmaceu-
Treat. 26 (2011) 301–308.
tical residues: a critical review, Sep. Purif. Technol. 156 (December (Part 3)) (2015)
[23] S. Song, et al., Electrochemical degradation of azo dye C.I. Reactive Red 195 by
891––914.
anodic oxidation on Ti/SnO2-Sb/PbO2 electrodes, Electrochim. Acta 55 (Apr (11))
[2] C. García-Gómez, et al., Combined membrane bioreactor and electrochemical oxi-
(2010) 3606–3613.
dation using Ti/PbO2 anode for the removal of carbamazepine, J. Taiwan Inst.
[24] M. Shestakova, M. Vinatoru, T.J. Mason, M. Sillanpää, Sonoelectrocatalytic de-
Chem. Eng. 64 (July) (2016) 211–219.
composition of methylene blue using Ti/Ta2O5-SnO2 electrodes, Ultrason.
[3] T. Gonza, P. Palo, J. Sa, Electrochemical advanced oxidation of carbamazepine on
Sonochem. 23 (March) (2015) 135–141.
boron-Doped diamond anodes. influence of operating variables, Ind. Eng. Chem.
[25] C. Comninellis, G.P. Vercesi, Problems in DSA® coating deposition by thermal de-
Res. 49 (August) (2010) 8353–8359.
composition, J. Appl. Electrochem. 21 (February (2)) (1991) 136–142.
[4] N. Vieno, T. Tuhkanen, L. Kronberg, Elimination of pharmaceuticals in sewage
[26] M. Shestakova, P. Bonete, R. Gómez, M. Sillanpää, W.Z. Tang, Novel Ti/Ta2O5-
treatment plants in Finland, Water Res. 41 (March (5)) (2007) 1001–1012.
SnO2 electrodes for water electrolysis and electrocatalytic oxidation of organics,
[5] C. García-Gómez, et al., Experimental design methodology applied to electro-
Electrochim. Acta 120 (February) (2014) 302–307.
chemical oxidation of carbamazepine using Ti/PbO2 and Ti/BDD electrodes, J.
[27] T. Duan, Y. Chen, Q. Wen, Y. Duan, Different mechanisms and electrocatalytic
Electroanal. Chem. 732 (October) (2014) 1–10.
activities of Ce ion or CeO2 modified Ti/Sb-SnO2 electrodes fabricated by one-step
[6] J.D. García-Espinoza, Electrochemical removal of carbamazepine in water with Ti/
pulse electro-codeposition, RSC Adv. 5 (25) (2015) 19601–19612.
PbO 2 cylindrical mesh anode, Water Sci. Technol. 73 (March (5)) (2016)
[28] D. Santos, M.J. Pacheco, A. Gomes, A. Lopes, L. Ciríaco, Preparation of Ti/Pt/SnO2-
1155–1165.
Sb2O4 electrodes for anodic oxidation of pharmaceutical drugs, J. Appl.
[7] N. Tran, P. Drogui, S.K. Brar, A. De Coninck, Synergistic effects of ultrasounds in the
Electrochem. 43 (January (4)) (2013) 407–416.
sonoelectrochemical oxidation of pharmaceutical carbamazepine pollutant,
[29] N. Lavanya, et al., A novel disposable electrochemical sensor for determination of
Ultrason. Sonochem. 34 (January) (2017) 380–388.
carbamazepine based on Fe doped SnO2 nanoparticles modified screen-printed
[8] P.C. Lindholm-Lehto, H.S.J. Ahkola, J.S. Knuutinen, S.H. Herve, Widespread oc-
carbon electrode, Mater. Sci. Eng. C 62 (May) (2016) 53–60.
currence and seasonal variation of pharmaceuticals in surface waters and municipal
[30] M. Liu, H. Xia, W. Lu, T. Xu, Z. Zhu, W. Chen, Electrocatalytic degradation of or-
wastewater treatment plants in central Finland, Environ. Sci. Pollut. Res. 23 (April
ganic contaminants using carbon fiber coupled with cobalt phthalocyanine elec-
(8)) (2016) 7985–7997.
trode, J. Appl. Electrochem. 46 (February (5)) (2016) 583–592.
[9] K. Gurung, M.C. Ncibi, M. Sillanpää, Assessing membrane fouling and the perfor-
[31] E. Hmani, Y. Samet, R. Abdelhédi, Electrochemical degradation of auramine-O dye
mance of pilot-scale membrane bioreactor (MBR) to treat real municipal waste-
at boron-doped diamond and lead dioxide electrodes, Diam. Relat. Mater. 30
water during winter season in Nordic regions, Sci. Total Environ. 579 (February)
(2012) 1–8.
(2017) 1289–1297.
[32] C. Salazar, C. Ridruejo, E. Brillas, J. Yáñez, H.D. Mansilla, I. Sirés, Abatement of the
[10] F.I. Hai, K. Tessmer, L.N. Nguyen, J. Kang, W.E. Price, L.D. Nghiem, Removal of
fluorinated antidepressant fluoxetine (Prozac) and its reaction by-products by
micropollutants by membrane bioreactor under temperature variation, J. Membr.
electrochemical advanced methods, Appl. Catal. B: Environ. 203 (April) (2017)
Sci. 383 (November (1–2)) (2011) 144–151.
189–198.
[11] T. Trinh, B. van den Akker, Richard M. Stuetz, H.M. Coleman, P. Le Clech,
[34] N. Tran, P. Drogui, S.K. Brar, Sonoelectrochemical oxidation of carbamazepine in
S.J. Khan, Removal of trace organic chemical contaminants by a membrane bior-
waters: optimization using response surface methodology, J. Chem. Technol.
eactor, Water Sci. Technol. 66 (August (9)) (2012) 1856–1864.
Biotechnol. 90 (May (5)) (2015) 921–929.
[12] M.A. Oturan, J.-J. Aaron, Advanced oxidation processes in Water/Wastewater
[35] D.M. González-Pérez, J.I. Pérez, M.A. Nieto, Carbamazepine behaviour and effects
treatment: principles and applications. A review, Crit. Rev. Environ. Sci. Technol.
in an urban wastewater MBR working with high sludge and hydraulic retention
44 (December (23)) (2014) 2577–2641.
time, J. Environ. Sci. Health Part A 51 (May (10)) (2016) 1–6.
[13] I. Sirés, E. Brillas, M.A. Oturan, M.A. Rodrigo, M. Panizza, Electrochemical ad-
[36] M. Shestakova, J. Graves, M. Sitarz, M. Sillanpää, Optimization of Ti/Ta2O5-SnO2
vanced oxidation processes: today and tomorrow. A review, Environ. Sci. Pollut.
electrodes and reaction parameters for electrocatalytic oxidation of methylene blue,
Res. 21 (July (14)) (2014) 8336–8367.
J. Appl. Electrochem. 46 (March (3)) (2016) 349–358.
[14] M. Panizza, G. Cerisola, Direct and mediated anodic oxidation of organic pollutants,
[37] Y. Chu, W. Wang, M. Wang, Anodic oxidation process for the degradation of 2, 4-
Chem. Rev. 109 (December (12)) (2009) 6541–6569.
dichlorophenol in aqueous solution and the enhancement of biodegradability, J.
[15] E. Brillas, C.A. Martínez-Huitle, Decontamination of wastewaters containing syn-
Hazard. Mater. 180 (August (1–3)) (2010) 247–252.
thetic organic dyes by electrochemical methods. An updated review, Appl. Catal. B:
[38] F.C. Moreira, R.A.R. Boaventura, E. Brillas, V.J.P. Vilar, Electrochemical advanced
Environ. 166–167 (May) (2015) 603–643.
oxidation processes: a review on their application to synthetic and real wastewaters,
[16] C.A. Martínez-Huitle, S. Ferro, Electrochemical oxidation of organic pollutants for
Appl. Catal. B: Environ. 202 (March) (2017) 217–261.
the wastewater treatment: direct and indirect processes, Chem. Soc. Rev. 35 (12)
[39] X. Hao, G. Wuqi, W. Jia, F. Jiangtao, Y. Honghui, Y. Wei, Preparation and char-
(2006) 1324–1340.
acterization of titanium-based PbO2 electrodes modified by ethylene glycol, RSC
[17] C. Comninellis, Electrocatalysis in the electrochemical conversion/combustion of
Adv. 6 (January (9)) (2016) 7610–7617.
organic pollutants for waste water treatment, Electrochim. Acta 39 (August (11))
[40] EU, E.U. directive 2013/39/EU of the european parliament and of the council of-
(1994) 1857–1862.
ficial, Off. J. Eur. Union (August) (2013) pp.15.
[18] M. Shestakova, M. Vinatoru, T.J. Mason, E. Iakovleva, M. Sillanpää,
Sonoelectrochemical degradation of formic acid using Ti/Ta2O5-SnO2 electrodes,
338
Publication IV
Khum Gurung, Mohamed Chaker Ncibi, Senthil K. Thangaraj, Janne Jänis, Mahdi
Seyedsalehi, Mika Sillanpää
Removal of pharmaceutically active compounds (PhACs) from real
membrane bioreactor (MBR) effluents by photocatalytic degradation
using composite Ag2O/P-25 photocatalyst

Separation and Purification Technology
Vol. 215, pp. 317-328, 2019
© 2019 Elsevier
Separation and Purification Technology 215 (2019) 317–328
Separation and Purification Technology

journal homepage: www.elsevier.com/locate/seppur
Removal of pharmaceutically active compounds (PhACs) from real T

membrane bioreactor (MBR) effluents by photocatalytic degradation using
composite Ag2O/P-25 photocatalyst
⁎
Khum Gurunga, , Mohamed Chaker Ncibia, Senthil K. Thangarajb, Janne Jänisb,
Mahdi Seyedsalehic, Mika Sillanpääa
a
Department of Green Chemistry, School of Engineering Science, Lappeenranta University of Technology, Sammonkatu 12, FI-50130 Mikkeli, Finland
b
Department of Chemistry, University of Eastern Finland, Yliopistonkatu 7, FI-80101 Joensuu, Finland
c
School of Environment, Tsinghua University, Beijing 100084, China
A R T I C LE I N FO A B S T R A C T
Keywords: Pharmaceutically active compounds (PhACs) are emerging pollutants causing serious challenges to wastewater
PhACs treatment plants due to poor biodegradability. In this study, the enhanced removal of highly recalcitrant and
Photocatalytic degradation commonly monitored PhACs, carbamazepine (CBZ) and diclofenac (DCF) by heterogeneous photocatalysis was
Ag2O/P-25 photocatalysts investigated using 5% Ag2O/P-25 photocatalyst. The photocatalyst was characterized by scanning electron
Matrix effect
microscope (SEM-EDX), Brunauer-Emmett-Teller (BET), Fourier transfer infrared spectroscopy (FTIR) and
Transformation products
UV–vis diffuse reflectance spectra (UV-DRS). The effects of catalyst dose, initial pollutants concentration, and
mineralization during the photocatalytic degradation of PhACs were investigated. The matrix effect was assessed
in deionized water (DW) and real membrane bioreactor effluent (RME). Optimal CBZ and DCF removals of
89.10% and 93.5%, respectively for 180 min of UV irradiation were achieved at catalyst dosage of 0.4 g L−1 in
DW matrix. However, the optimal catalyst dosages for CBZ and DCF in RME matrix were increased by factor 2
and 1.5, respectively, to achieve the same degree of removal. Declining trends of removal rate were observed
when initial concentrations of both the PhACs were increased under optimal catalyst dosages, and kinetics seem
to fit the Langmuir-Hinshelwood model. Photo-induced holes and %OH were the dominant oxidation species
involved in the photocatalytic degradation of the PhACs. A plausible reusability of 5% Ag2O/P-25 photocatalyst
was observed for both the PhACs. Moreover, various aromatic/aliphatic intermediates generated during the
photodegradation CBZ were identified using fourier-transform ion cyclotron resonance (FT-ICR) mass spectro-
metry, and a possible multi-step degradation pathway was proposed. Overall, the removal of PhACs using 5%
Ag2O/P-25 photocatalyst showed promising results in real wastewater.
1. Introduction environment [3]. In recent years, the fate of the PhACs and their me-
tabolites has received a great deal of interest from the scientific com-
Pharmaceutically active compounds are used for the prevention, munity. Though PhACs are normally detected in the concentration
diagnosis, and treatment of several diseases in humans and animals. range of ng L−1 or µg L−1, they can pose serious issues due to their
The global annual average per capita consumption of pharmaceuticals potential adverse effects on human health and ecosystem [4,5]. How-
is estimated at about 15 g, whereas industrialized countries accounted ever, other factors such as increased persistence, time of exposure,
between 50 and 150 g [1]. The annual per capita expenditures on biotransformation, and degradation mechanisms can also influence the
pharmaceutical ranged from US$ 7.61 in low-income countries to US$ impacts of PhACs, other than their concentration alone [5]. Among the
431.6 in high-income countries [2]. As most of the PhACs are non-vo- large number of PhACs, CBZ and DCF are the most frequently identified
latile and often charged molecules composed of diverse moieties, these in aquatic environments [1].
emerging pollutants are highly recalcitrant and not entirely degraded CBZ is an antiepileptic drug often used to treat epilepsy and sei-
after the medical use, hence the excretion of unchanged (parent) zures, whereas DCF is a non-steroidal anti-inflammatory drug (NSAID)
compounds and metabolites and their channeling to the aquatic commonly used as analgesic, anti-rheumatic, and antiarthritic [6,7].
⁎
Corresponding author.
E-mail addresses: khumgrg@yahoo.com, Khum.Gurung@lut.fi (K. Gurung), mika.sillanpaa@lut.fi (M. Sillanpää).
https://doi.org/10.1016/j.seppur.2018.12.069
Received 22 October 2018; Received in revised form 24 December 2018; Accepted 25 December 2018
Available online 03 January 2019
1383-5866/ © 2018 Elsevier B.V. All rights reserved.
K. Gurung et al. Separation and Purification Technology 215 (2019) 317–328
CBZ has been proposed as an anthropogenic indicator in water bodies matrices, i.e. DW and RME, by heterogeneous photocatalytic degrada-
[8]. Moreover, DCF is included in the watch list of priority substances tion using Ag2O/P-25 photocatalyst. The composite heterojunction 5%
as emerging pollutants in EU directive that needs careful environmental Ag2O/P-25 photocatalyst was synthesized by a simple pH-mediated
monitoring in its member states [9]. chemical precipitation reaction method and its photocatalytic effi-
The wastewater treatment plants (WWTPs) are the primary receptor ciency for the removal of PhACs from both matrices were examined.
of PhACs via direct and unsafe discharge of untreated wastewaters from The emphasis was given on the influence of catalyst dose and miner-
domestic households, industries, agriculture lands, and hospitals alization in both DW and RME, the effect of initial pollutants con-
[10,11]. PhACs have been found in the effluents of municipal WWTPs centration, stability of catalyst, and determination of reactive oxygen
because of their incomplete removal [5]. It is well known that con- species and reaction mechanism. Moreover, various intermediate pro-
ventional WWTPs are not fully efficient in treating most of the PhACs ducts formed during the photodegradation of CBZ were identified, and
[1,12,13]. Even Membrane bioreactors (MBRs), which are regarded as a possible multi-step degradation pathway was proposed.
having superior capabilities in terms of performance and flexibility
[14], are not effective in the removal of certain PhACs [8,15]. In this 2. Materials and methods
regard, many researchers have recommended the development of
treatment systems where biological degradation is integrated with ad- 2.1. Chemicals
vanced oxidation processes (AOPs) as a pre-treatment of effluents, in
order to improve biodegradability and overall removal efficiency of All the chemicals used were analytical grade and purchased from
PhACs, before their discharge into water bodies [4,16]. Despite their Sigma-Aldrich Chemical Co. (St. Louis, USA), unless specified. The
cost, the use of AOPs is still justified as complete degradation of per- chemicals used for this experiment include CBZ (≥99%, meets USP
sistent pollutants is not always needed [4]. The prospective application testing specifications), DCF (2-[(2,6-dichlorophenyl) amino]benzenea-
of various methods such as Fenton oxidation [17,18], ozonation [19], cetic acid sodium salt), Aeroxide ® P-25 as commercial TiO2 (21 nm
chemical oxidation [20], and TiO2 photocatalysis [7] processes have primary size (TEM), ≥99.5% trace metal basis), AgNO3 (99.999% trace
been proposed to transform PhACs into non-toxic and biologically in- metal basis), NaOH (≥98%, anhydrous), Acetic acid (natural,
active compounds. ≥99.5%), N,N Dimethylformamide (anhydrous, 99.8%), Lithium
TiO2 is the most commonly used photocatalyst (EG = 3.2 eV) be- acetate dehydrate (99.999% trace metal basis), Titanium(IV) butoxide
cause of its cost, and remarkably active and stable features [7,21,22]. (reagent grade, 97%), Palladium(II) chloride (≥99.9%), Hydrochloric
Heterogeneous UV/TiO2 photocatalysis has become the most attractive acid (reagent grade, 37%), Tin (II) chloride dehydrate (≥99.99% trace
AOP to remove varieties of organic pollutants from water and waste- metal basis), Sodium citrate tribasic dehydrate (ACS reagent, ≥99.0%),
water [16]. In heterogeneous UV/TiO2 photocatalysis, upon irradiation Nickel nitrate (99.999% trace metal basis), Methanol (≥99.8%, anhy-
of TiO2 semiconductor by UV–vis light, an electron (e−) is promoted to drous), and Ethylenediaminetetraacetic acid sodium salt dehydrate (EP
the conduction band and a hole (h+) is generated on the valence band reference standard).
site vacated by the electron, thereby forming highly reactive oxygen
species (ROS) such as hydroxyl (%OH) and superoxide (%O2–) radicals
[21,23–25]. Therefore, heterogeneous photocatalysis is based on the 2.2. Synthesis of photocatalysts
chemistry of photo-generated h+ and/or the %OH, which promotes non-
selective oxidation of pollutants [3,26]. Yet, there are some drawbacks 5% (w/w) Ag2O/P-25 photocatalysts were synthesized via a simple
associated with conventional TiO2 such as rapid recombination of pH-mediated chemical precipitation reaction method [34]. The proce-
charge carriers (ecb−/hvb+), difficulty in separating the reacted TiO2 dure typically includes the dispersion of 1.0 g of P-25 in 117 mL of
particles, inefficient quantum yield, and wide energy band gap, hin- deionized water and the solution was stirred for 4 h. Subsequently,
dering its industrial applications [4,7,27]. Therefore, several R&D stu- 540 µL of 0.2 M AgNO3 solution was added dropwise into the suspen-
dies focused on developing highly efficient UV–vis light driven photo- sion. The mixture was then stirred in dark for 30 min in order to reach
catalysts by doping TiO2 with narrow band gap semiconductors, which Ag+ sorption equilibrium. Then, a measured amount of 0.1 M NaOH
is crucial having less energy use and highly efficient photocatalytic solution was slowly added into the Ag+ pre-equilibrated mixture and
oxidation of pollutants [27,28]. stirred for 1 h at room temperature. The mixture was filtered through
Many studies have suggested that fine particles of transition metals 0.45 µm nylon membranes (VWR, USA) and washed repeatedly with
or their oxides dispersed on the surface of photocatalyst matrix could deionized water and ethanol. The obtained pallets were then kept in
act as electron traps, which effectively reduces the combination of deep-freezer for 24 h and lyophilized. After grinding and sieving
charge carriers and improves photocatalytic activity [22,29,30]. Ag2O through 500 µ sieve, the dried powder was used for subsequent ex-
nanoparticles doped on TiO2 particles can enhance the optical absorp- periments.
tion spectrum and lattice structural stability of TiO2, improving pho- Furthermore, three other photocatalysts were also synthesized by
tocatalytic degradation of hazardous pollutants under UV–vis light ir- adopting the methods from literatures including Sn3O4/P-25 [33],
radiation [21,23,31]. Indeed, Ag2O is a proven photocatalyst for the PdO/P-25 [34], and NiO/P-25 [35] by doping different metallic sub-
degradation of organic pollutants under visible light irradiation having strates over pure P-25.
a narrow band gap of 1.2 eV [32]. The physicochemical changes in the
TiO2 particles after doping, such as increasing surface area, decrease of 2.3. Model solutions and matrix
TiO2 crystals, and homogeneity in particles surface size, are responsible
for the enhancement of photocatalytic efficacy [4,33]. Additionally, Model solutions containing target PhACs (CBZ and DCF) in two
recovery and reusability of TiO2 can be enhanced due to improved different matrices were adopted for this study. The properties of target
segregative specialty with impurity doping [30]. The photocatalytic pollutants in this study are shown in Table 1. The aqueous solution
efficiency of Ag2O doped TiO2 has been well confirmed by nitrate re- matrices were prepared by spiking CBZ or DCF in deionized (ultrapure
duction [34] and degradation of organic dyes [28,30,35]. Nevertheless, Milli-Q) water and in real effluent of the submerged MBR pilot plant
to the best of our knowledge, the application of Ag2O doped TiO2 (P-25) located in Kenkäveronniemi WWTP (Mikkeli, Finland). The detailed
nano-particles in photocatalytic degradation of CBZ and DCF from real descriptions of MBR pilot plant and operating conditions are described
MBR effluent has not been studied so far. in our previous studies [10,15]. The main physicochemical character-
The main objective of this study was to investigate the removal of istics of the MBR effluent are listed in Table S1 (Supplementary in-
commonly used PhACs including CBZ and DCF from two aqueous formation)
318
Table 1
The physico-chemical properties of target PhACs in this study.
Compound/formula Therapeutic Use Molecular weight Water solubility (mg L−1, Chemical structure
(g mol−1) 20 °C)
Carbamazepine (C15H12N2O) Antiepileptic, anticonvulsant drug 236.27 125 ± 2a
Diclofenac sodium (C14H10Cl2NNaO2) Non-steroidal anti-inflammatory and analgesics 318.13 50,000

drug
a
Ref. [52].
2.4. Characterization of photocatalysts aqueous solution (20 mg L−1) and reaction was started as in photo-
degradation experiment for 180 min. Samples were collected at the
Microstructure and morphology of the synthesized photocatalysts interval of 60 min, 120 min, and 180 min. After the first run, suspension
were examined by scanning electron microscope (SEM) (Hitachi, was centrifuged at 3500 rpm for 30 min, washed thoroughly with
SU3500, Japan) using acceleration voltage of 5 kV. The elemental deionized water, and again dispersed into the another fresh CBZ solu-
analysis of photocatlyst was performed using energy-dispersive X-ray tion to start the second run. Similarly, the experiment was repeated five
spectroscope (Thermo Fisher,USA). The Brunauer-Emmett-Teller (BET) times.
surface area (SBET) was determined by using Tristar II Plus. The UV–vis
diffuse reflectance spectra were measured in the range of 250–800 nm 2.8. Analytical procedures
using Agilent Cary 5000 spectrophotometer with DRA2500 integrating
sphere. The Fourier transfer infrared spectroscopy (FTIR) was per- The changes in the concentration of CBZ and DCF during photo-
formed using Bruker Platinum ATR VERTEX 70. catalytic reaction were analyzed using a UV–Vis spectrophotometer
(Lambda 45, Perkin Elmer, USA). The absorbance was measured at
2.5. Photodegradation experiments 285 nm and 276 nm for CBZ and DCF, respectively, since these wave-
lengths correspond to the characteristic peaks of these drugs in the UV
The photocatalytic degradation experiments were performed in a spectrum. Moreover, it should be noted that the absorbance measure-
self-made continuous stirred photochemical reactor as shown in Fig. S1 ments of the treated solutions might not indicate the absolute con-
(Supplementary information). The photocatalytic reaction chamber centration of parent CBZ or DCF as some of the treated intermediate
consists of a UV lamp (rated power is 14 W, rated voltage is 216–253 V products may also absorb at similar wavelengths, thus contributing to
at 50 Hz, λmax of 254 nm, Trojan, Canada) placed in a quartz socket increased absorbance. Therefore, the reported concentrations used to
tube with one end closed, magnetic stirrer, rotor, and a reaction vessel calculate percentage removal may account both residual parent pollu-
with double-jacketed cooling system. The model solution (300 mL) was tants and some of the related reaction by-products, which could un-
added inside the photochemical reactor and the temperature inside the derestimate to actual degradation efficiency of the parent compounds.
jacketed reactor was maintained at 21 ± 2 °C by continuous circula- Dissolved organic carbon (DOC) was also measured using TOC analyzer
tion of water. Prior to irradiation, measured quantity of photocatalyst (Shimazdu, Japan) to assess the extent of mineralization.
was dispersed into the solution and magnetically stirred in dark for The degradation products of CBZ were analyzed by 12-T Bruker
30 min to achieve adsorption–desorption equilibrium. Upon the irra- solariXTM XR hybrid Qh-Fourier Transform Ion cyclotron resonance
diation, an aliquot of 3 mL reaction solution was sampled at given time (FT-ICR) mass spectrometer (Bruker Daltonics, Bremen, Germany)
intervals. The reaction solution aliquots were then filtered through coupled with an electrospray ionization (ESI) source (Apollo-II), oper-
0.45 µm membranes prior to analysis. No buffer was used and the pH ated in the positive-ion mode. The photocatalytic degradation products
values ranged from 5 to 7 (before and after) during all set of the ex- of CBZ (10 µL) were prepared in methanol (diluted 10:200 v/v). Acetic
periments. acid (1 v-%) was added to the samples to aid protonation. The samples
were directly infused into the ion source at a flow rate of 2 µL min−1.
2.6. Trap experiments The ion accumulation time in the hexapole was 0.8 s and the time of
flight from the collision cell to the ICR cell was set at 0.260 ms. Dry
The radicals and holes trapping experiments were designed to elu- nitrogen was used as drying (4.0 L min−1) and nebulizing gas (80 °C,
cidate the types and roles of reactive species, and mechanisms involved 1.0 bar). For each spectrum, 200 time-domain transients (4 MWord
during the photocatalytic degradation of CBZ and DCF using 5% Ag2O/ each) were co-added and full-sine apodized. The mass spectra were
P-25. In practice, several chemical compounds were tested during the externally calibrated using an ES tuning mix (Agilent Technology).
photocatalytic processes, including ethylenediaminetetaacetic acid Bruker FTMS Control™ 2.1 software was used for the instrument control
(EDTA) as a hole (h+) scavenger [28], methanol (MeOH) as hydroxyl and data acquisition. The mass spectra were further processed and
radical (%OH) scavenger [36], and O2-purging (%O2– scavenger) via N2 analyzed by using Bruker DataAnalysis 4.4 software.
[37].
3. Results and discussion
2.7. Reusability experiments
3.1. Screening of photocatalysts: dark adsorption, UV- photolysis, and UV
Cyclic experiments were performed to check the reusability of the photocatalysis
investigated photocatalyst over the degradation of CBZ and DCF.
Typically, 0.24 g of 5% Ag2O/P-25 was dispersed into 0.3 L CBZ In a preliminary stage, photocatalytic degradation efficiencies of
319
photocatalyst such as surface chemical compositions and morphology

were characterized by X-ray diffraction (XRD), transmission electron
microscopy (TEM), and X-ray photoelectron spectroscopy (XPS) ana-
lysis in the previous work [34,35]. However, we further characterized
the as-prepared photocatalyst by other analytical techniques such as
SEM, EDX, BET, FTIR, and UV-reflectance spectra.
The morphology and elemental microanalysis of the selected 5%
Ag2O/P-25 photocatalyst were investigated by SEM and EDX. The SEM
micrographs with different magnifications are as shown in Fig. 2. The
surface of P-25 nanoparticles acts as underlying substrate to absorb
Ag2O nanoparticles and an average diameter of 4.4–20 nm could evenly
disperse on the surface of P-25 during pH-mediated precipitation pro-
cess [30,34]. Presence of Ag nanoparticles on the surface of P-25
photocatalyst was confirmed by the EDX elemental analysis, which
reveals the bulk compositions of Ti, O, C, and Ag elements (Fig. 2i). The
TiO2 nanoparticles provide numerous nucleation sites, which could be
responsible for the homogeneous dispersion of Ag2O nanoparticles on
the surface of TiO2 nanoparticles [30]. Furthermore, elemental map-
pings of 5% Ag2O/P-25 confirms the homogenous dispersion of Ag+
particles on P-25 as depicted by Fig. 2(d–h).
To further confirm the chemical composition of 5% Ag2O/P-25
photocatalyst, FTIR spectroscopy was performed as depicted in Fig. S2-
a. The main absorption peaks were observed at 3000, 1533.67, 1420,
1050 and 433.33 cm−1. The peak located at 433.33 cm−1 could be
assigned to MeO (i.e. either TieO or AgeO) stretching vibration
[41,42]. The broad absorption band at about 2976.75 cm−1 is char-
acteristic of OeH stretching vibration of the TieOH. The vibrational
peaks at 1558.30, 1419.43 and 1052 cm−1 are due to stretching of C]
O group, OeH bending and stretching of CeO bond, respectively
[42,43].
The BET surface area, pore size, and pore volume of 5% Ag2O/P-25
were estimated as reported in Table S2 (Supplementary information).
The SBET of pure TiO2 (P-25) increased slightly from 47.27 m2 g−1 to
Fig. 1. Dark adsorption, UV-photolysis, and photocatalytic degradation trends 54.1 m2 g−1 due to the deposition of Ag2O nanoparticles over P-25
of different photocatalyts for (a) CBZ, and (b) DCF. (Operating conditions: surface. Likewise, the average pore size and pore volume of P-25 were
[CBZ]0 = [DCF]0 = 50 mg L−1, [(5% w/w) Ag2O/P-25] = 0.4 g L−1). increased from 8.25 to 9.27 nm and 0.10 to 0.13 cm3 g−1, respectively.
However, this observation indicated that the contribution of the Ag2O
four different photocatalysts including Ag2O/P-25, Sn3O4/P-25, PdO/P- particles to the specific surface area of P-25 was relatively low. The
25, and NiO/P-25 were tested on the selected pollutants. First, in order increased surface area, pore size and pore volume showed that Ag2O
to evaluate the extent of adsorption and UV-photolysis of target pol- particles were well dispersed over P-25 surface with limited inter-
lutants, adsorption under dark conditions and UV-photolysis tests were ference with the existing pores. Similar trends in the characteristic
carried out (Fig. 1). Even though experiments were run for several modifications of P-25 after doping with Ag2O have been reported in
hours, adsorption-desorption equilibrium was reached in 30 min for all other studies [44].
the cases. At the catalyst dose of 0.4 g L−1, the percentage adsorption For the energy band gap (Eg) determination of semiconductors, the
removal of 50 mg L−1 of CBZ and DCF, was around 2–9%. Similarly, the reflectance spectra of photocatalysts were converted to the absorbance
conversion rates of CBZ and DCF during 180 min of UV-photolysis were spectra using the Kubelka-Munk functions [45] and then Eg evaluated
nearly 13% and 25%, respectively. The results concluded that both of by using Tau plot [44–46]. The undoped P-25 showed a band gap of
the target pollutants were slightly converted under UV-photolysis. 3.2 eV with the absorption peak at ∼ 395 nm. After doping with Ag2O,
Comparable extents of CBZ or DCF removal by adsorption on TiO2 the wavelength threshold of the absorption peak of P-25 was between
based photocatalysts and UV photolysis have been reported in many 400 and 500 nm, with the corresponding energy band gap at 2.96 eV as
studies [38–40]. On the other hand, the photocatalytic degradation shown in Fig. S2-b (supplementary information). The decreased band
kinetics of as prepared catalysts were tested by irradiating the model gap energy inferred the chemical interactions between P-25 and Ag2O
solutions under UV lamp for 180 min. Of the selected five photocatalyst, nanoparticles. Therefore, lower energy transitions are possible in Ag2O/
the highest photodegradation efficiency of CBZ (i.e. 55%) was observed P-25 semiconductors as Ag2O could create localized energy levels in the
with Ag2O/P-25 (Fig. 1). Similarly, the degradation of DCF by PdO/P- P-25 bandgap, which evidently increases the absorption spectra in
25 and Ag2O/P-25 was more than 80%. Based on the screening ex- UV–vis range [44].
periments, since 5% Ag2O/P-25 photocatalyst accounted the highest
photocatalytic activity either in case of CBZ or DCF, therefore selected 3.3. Effect of catalyst dose, solution matrix, and initial substrate
for the subsequent experiments involving both target micropollutants. concentration on the removal of selected PhACs
Moreover, the selection of Ag2O/P-25 over PdO/P-25 and NiO/P-25
was made because of the comparable efficiency and compatibility to In order to ensure the highest absorption of photons, without gen-
both the model pollutants. erating an substantial amount of waste, determining the optimal dose of
catalyst in an adopted photocatalytic system is of high importance
3.2. Characteristics of selected 5% Ag2O/P-25 photocatalyst [38,47]. In fact, several researchers have investigated the reaction ki-
netics as a function of catalyst dosing in photocatalytic processes [48].
The major characteristics of as-prepared 5% Ag2O/P-25 In this study, the optimal dose of 5% Ag2O/P-25 for the removal of CBZ
320
Fig. 2. SEM micrographs of different magnifications (a–c); Elemental mappings (d–h): (d) total mapping, (e–h) C, O, Ti, and Ag mappings; and (i) EDX microanalysis
of (5% w/w) Ag2O/P-25 selected and as-prepared photocatalyst.
and DCF was determined for two different solution matrices (i.e. DW to unfavorable phenomenon, such as light scattering due to enhanced
and RME) as shown in Fig. 3. The catalyst dose of 0.2–1.0 g L−1 was reflectance and increased opacity of the suspension [48,49]. Also, the
assessed for DW matrix, whereas 0.4–1.2 g L−1 was used in the case of agglomeration of solid particles due to excess dose may result in a re-
RME matrix. Since organic matter in RME could inhibit the degradation duction of active surface sites available for photon absorption, which
of organic pollutants, higher amount of catalyst dose is essential to eventually reduces photocatalytic efficiency [33]. Even though the
achieve enhanced removal efficiencies [4]. Therefore, a higher range of number of active sites increases with more catalyst dosing, the pene-
catalyst dosing was selected to assess the influence of organic matter tration of light is to be compromised because of the shielding effect
contained in the RME matrix. Table 2 shows the collective removal of [33,48]. The trade-off between these two divergent phenomena results
target pollutants and the extent of mineralization in two solution main an optimal catalyst dosing for photocatalytic processes [33]. The
trices under different operating conditions. shielding by TiO2 particles may occur due to deactivation phenomenon
The photocatalytic degradation kinetics of CBZ in DW and RME of activated TiO2 molecules by collision with ground state molecules
matrices are shown in Fig. 3a and b, respectively. In DW matrix, and [48,50]. Nevertheless, the optimal catalyst dosing is highly dependent
after an irradiation time of 180 min, the removal of CBZ reached from on the concentration and source/type of the pollutant and the operating
80.4% to 89.10% while increasing the catalyst dosing from 0.2 to conditions as well [33]. Therefore, the optimal catalyst dose of
0.4 g L−1. Likewise, for the case of DCF, the removal rate increased 0.4 g L−1 achieved for both CBZ and DCF in DW matrix might be dif-
from 87.5% to 93.5% in DW matrix for the similar range of catalyst. The ferent for other conditions.
higher removal rate of DCF compared to CBZ in identical operating On the other hand, relatively slow kinetics was observed in the
conditions exhibited greater affinity of DCF molecules towards photo- experiments performed with RME matrices of both CBZ and DCF
catalytic degradation by 5% Ag2O/P-25. From Fig. 3 and Table 2, it is (Fig. 3b and d). In Table 2, an increasing trend of CBZ removal rate was
clear that the photocatalytic degradation of both CBZ and DCF beyond clearly observed from 76.6 to 93.9% when the catalyst dose was in-
the 0.4 g L−1 of catalyst dose remained constant. Therefore, 0.4 g L−1 of creased from 0.4 to 0.8 g L−1. However, the removal rate decreased by
5% Ag2O/P-25 was concluded to be the optimum catalyst dose for the nearly 3% when catalyst dose was subsequently increased beyond
removal of both PhACs in DW matrix. Similar tendencies have been 0.8 g L−1. On the other hand, DCF removal rate was improved greatly
reported in earlier studies [4,47]. The limiting value 0.4 g L−1 might be when the catalyst dose was increased from 0.4 g L−1 to just 0.6 g L−1,
attributed to fact that photocatalytic rate increases with catalyst dosing though no further remarkable enhancement was observed beyond
at the beginning and then decreases (or saturates) at higher doses due 0.6 g L−1 of catalyst dose. Therefore, different optimal catalyst dose for
321
Fig. 3. Effect of catalyst dosing and water matrix on the photocatalytic degradation of (a) CBZ in DW, (b) CBZ in RME; and (c) DCF in DW, and (d) DCF in RME.
(Operating conditions: [CBZ]0 = [DCF]0 = 20 mg L−1; pH = 5–7; [5% Ag2O/P-25] = 0.2–1.0 g L−1 for DW matrix and 0.4–1.2 g L−1 for RME matrix;
Temperature = 21 ± 2 °C).
Table 2 between organic matter and the PhACs on the surface of catalyst was
Removal rates (% ± SD) of CBZ and DCF under the varying catalyst dosages and evidently observed in the RME matrix during the photocatalytic pro-
the extent of mineralization (%) in two different water matrices. cess. This may be attributed to either the organic carbon content of
Solution Target Catalyst Removal Mineralization (%) RME samples, or the presence of various species that may act as natural
matrices pollutants concentration efficiency (%) scavengers and other reactive moieties [12]. The organic matter in real
(g L−1) effluent not only competes for the active species generated in the pro-
cess but could also be adsorbed on the surface of photocatalysts, which
DW CBZ 0.2 80.40 ± 0.5 67.90
0.4 89.10 ± 1.5 inhibits the photocatalytic activity [4]. Additionally, a possible inter-
0.6 88.60 ± 1.4 ference of inorganic ions on catalyst surface deactivation, and a com-
0.8 91.70 ± 1.5 petition for UV absorption should also be taken into consideration [51].
1.0 89.02 ± 4.7 Subsequently, the optimized catalyst doses for CBZ and DCF re-
DCF 0.2 87.60 ± 2.2 64.80
0.4 93.50 ± 0.1
moval in RME matrix were used to study the influence of initial sub-
0.6 93.60 ± 0.1 strate concentration in photocatalytic degradation kinetics, as depicted
0.8 93.04 ± 0.1 in Fig. 4. The initial substrate concentration was varied in the range of
1.0 93.40 ± 0.3 20–50 mg L−1 at optimal catalyst doses of 0.8 g L−1 and 0.6 g L−1 for
RME CBZ 0.4 76.60 ± 5.2 60.30 CBZ and DCF, respectively. In Fig. 4a and b, decreasing trends in the
0.6 85.40 ± 6.5 removal of CBZ and DCF were observed when initial substrate con-
0.8 89.74 ± 0.4
centrations were gradually increased. The removal efficiency of CBZ
1.0 90.50 ± 0.4
1.2 90.95 ± 2.3
was decreased by 38%, when the concentration was increased from 20
DCF 0.4 86.60 ± 0.3 55.20 to 50 mg L−1 (Fig. 4a, inset). Whereas, only 23% reduction was ob-
0.6 90.70 ± 4.5 served in the DCF removal efficiency with the same range of con-
0.8 88.30 ± 3.4 centration increments (Fig. 4b, inset). These trends may be attributed to
1.0 90.40 ± 0.40
the increasing occupied sites of the catalyst due to more pollutant
1.2 92.10 ± 1.10
molecules adsorbed on the surface of catalyst, which inhibits the gen-
eration of reactive species [12]. On the other hand, high pollutants
CBZ (i.e. 0.8 g L−1) and DCF (i.e. 0.6 g L−1) were established in RME concentrations can lead to a decreasing penetration of photons into the
matrix. bulk solution and hence the solutions become more impermeable to UV
As clearly seen in Table 2, there is an obvious deviation in the op- irradiation due to an induced inner filtering effect. Thus, the photo-
timum catalyst dose of DW and RME matrices. The need for optimal excitation of catalyst particles by photon energy will be reduced leading
catalyst dose in RME matrix, in order to achieve approximately the to a diminished photodegradation efficacy [52].
same level of removal efficiency in DW matrix, is increased by factor of The photocatalytic oxidation of organic contaminants via hetero-
2 and 1.5 for CBZ and DCF, respectively. Therefore, the competition geneous catalytic process usually showed that the corresponding data
322
Fig. 5. Fitting of L-H model for photocatalytic degradation kinetics of CBZ and
DCF. (Operating conditions: RME solution matrix, [5% Ag2O/P-
25)]0 = 0.8 g L−1 for CBZ, and 0.6 g L−1 for DCF). Plot of Eq. (1).
Fig 4. Effect of initial substrate concentration in photocatalytic degradation of

(a) CBZ, and (b) DCF. (Operating conditions: All are RME matrix; [5% Ag2O/P-
25] = 0.8 g L−1 for CBZ; and 0.6 g L−1 for DCF). Removal percentages of CBZ
and DCF versus initial concentrations are shown in insets of (a), and (b), re-
spectively.
fit the Languir-Hinselwood (L-H) kinetic model as shown in Eq. (1)

[12,53]:
Ceq Ceq 1
= +
r0 kr Kkr (1)
where r0 is the initial reaction rate, Ceq is the equilibrium concentration
of substrate in solution after the completion of dark experiments, kr is
the reaction rate constant at maximum coverage, and K is equilibrium
constant for adsorption of the substrate onto the catalyst surface.
In order to check if the investigated photocatalytic degradation
follows the L-H kinetic model, the experimental data with initial con-
centrations of target pollutants (20–50 mg L−1) in RME matrix (Fig. 5)
were assessed and the initial rates were calculated at 20 min of initial
reaction. A graphical plot of Eq. (1) for runs at various initial substrate
concentrations is shown in Fig. 5, which indicates that the photo-
catalytic degradation well fitted the Langmuir-Hinshelwood kinetic
model. Moreover, the degradation kinetics beyond 20 min of irradiation
were observed to follow a curve rather than a straight line when Fig. 6. Mineralization during photocatalytic degradation of (a) CBZ, and (b)
modelled with pseudo-first order (PFO) kinetic for both CBZ and DCF DCF in two solution matrices. (Operating conditions: [5% Ag2O/P-
compounds. The possible formation of aromatic organic intermediates, 25] = 0.4 g L−1 in DW matrix for both CBZ and DCF; and 0.8 g L−1 and
which can be detected at the similar wavelengths of parent compounds, 0.6 g L−1 in RME matrix for CBZ and DCF, respectively).
could affect the corresponding final absorbance values and conse-
quently the reaction kinetics [54].
during photocatalytic treatment. The optimal catalyst doses were
adopted for different solution matrices and target pollutants. Fig. 6a
3.4. Mineralization of PhACs and b shows the normalized DOC removals during the photocatalytic
treatment of CBZ and DCF, respectively. About 68% of mineralization
In order to characterize the mineralization extent of CBZ and DCF in rate was observed when the CBZ removal was 90% in DW matrix,
different solution matrices, the DOC measurements were carried out
323
which suggests that the trapped holes were the main oxidative species
prompted due to UV irradiation of 5% Ag2O/P-25, and thus accoun-
table for the photocatalytic degradation of DCF. Interestingly, the
conversion of DCF was not influenced due to the addition of MeOH. To
confirm this result, other scavengers such as isopropyl alcohol and tert-
Butanol were also tested, however similar effects were observed (data
not shown). This might be due to the higher affinity of photo-generated
ROS with DCF molecules, compared to the added scavengers. Moreover,
the major role of photogenerated holes during the photocatalytic oxi-
dation of TiO2 based photocatalyts has been reported [55]. In addition,
the organic pollutants could also absorb actinic photons from UV irra-
diation and transfer energy or electrons to produce reactive oxidation
species, which in turn degrade the organic pollutants. This phenomenon
is referred to as self-sensitized photooxidation of pollutants [56].
On the basis of above results, a possible photocatalytic mechanism
scheme of the heterojunction 5% Ag2O/P-25 photocatalyst was pro-
posed as shown in Fig. S3 (Supplementary information). When com-
posite Ag2O/P-25 was irradiated by UV, both e− and h+ were excited in
the conductance band (CB) and valence band (VB), respectively ac-
cording to Eqs. (2) and (3). In normal case, most of e− – h+ pairs re-
combine rapidly to produce heat. As ECB (TiO2, −0.4 eV vs. SHE) is
lower than E (O2/%O2–) (−0.046 eV vs. SHE), %O2– could produce from
dissolved O2 [57,58]. However, the ECB (Ag2O, +0.2 eV vs. SHE) was
higher than E (O2/%O2–) , thus unable to produce %O2– from dissolved O2
[58,59]. The electron produced on CB of TiO2 can transferred easily to
Ag2O as ECB of TiO2 is positioned above that of Ag2O [57]. Subse-
quently, the partial lattice Ag+ in Ag2O nanoparticles on the surface of
TiO2 reacts with photoinduced electrons and reduced to metallic Ag
(Eq. (4)). Further, the metallic Ag acted as an electron pool and
transferred the produced electrons to combine with O2. The Ag2O as
electron absorbent inhibits the recombination of the photogenerated e−
– h+ pairs, which allows the holes to effectively oxidize organic pol-
Fig. 7. Trap experiments to assess the role of reactive species during the pho-
lutants, and thus the photocatalytic activity is highly improved [30].
tocatalytic degradation of (a) CBZ, and (b) DCF. (Operating conditions: [CBZ]
−1 Meanwhile, the generated electrons are trapped by dissolved O2 to
0 = [DCF] 0 = 20 mg L ; [5% Ag2O/P-25] = 0.4 g L−1).
produce %O2–, and transformed to reactive oxygen species %OH by
multistep reductions (Eq. (5)–(8)). Moreover, the EVB (TiO2, +2.8 eV vs.
whereas the mineralization rate was reduced to 65% when CBZ removal SHE) was higher than the E (%OH/H2O) (+2.68 eV vs. SHE) [57,59],
was 94% in RME matrix (Table 2). Similarly, the mineralization rates of indicating that the photogenerated holes at VB of TiO2 could oxidize
DCF for DW and RME matrices were 60% and 55%, whilst their re- H2O to %OH according to Eq. (9). Subsequently, photoinduced holes in
movals were 93% and 90%, respectively. The removal percentages of the VB of TiO2 can easily move and accumulate at the VB of Ag2O since
both CBZ and DCF molecules by photocatalysis were noticeably higher EVB (Ag2O, +1.0 eV vs. SHE) is less positive than that of TiO2. As a
than the corresponding removal percentage of DOC. This might be result, the recombination potential of e− - h+ pairs was largely re-
explained by the formation of aromatic organic intermediates from duced, thus greatly enhancing quantum efficiency [28]. Moreover,
parent molecules of target pollutants, which leads to the increase of photoinduced holes transferred and accumulated in the VB of Ag2O
organic molecules in the solution [12,40]. Moreover, results showed could directly oxidize the target pollutants to a great extent. The pho-
that the mineralization rates of both the target compounds were in- togenerated reactive species (h+, %OH, %O2–) could degrade CBZ or DCF
hibited to some extent in RME matrix as compared to DW matrix. This molecules into intermediates and finally to mineralization (Eq. (10)).
fact evidently confirms the influence of organic matter (cf. Table S1)
present in the RME matrix. hϑ
TiO2 → e− + h+ (2)
3.5. Role of reactive oxygen species and mechanism involved in the

hϑ
photocatalytic degradation Ag 2O → e− + h+ (3)
The influence of reactive species (h+, %OH, %O2–) involved during

the photocatalytic degradation of CBZ and DCF using 5% Ag2O/P-25 is
Ag2O + H2O + 2e− → 2Ag + 2OH– (4)
shown in Fig. 7. In Fig. 7a, the photocatalytic degradation of CBZ under
UV irradiation for 60 min was substantially suppressed with the addi- e−+O2 → %O2– (5)
tion of EDTA and MeOH by nearly 70% and 90%, respectively, as %
O2–+H+ → %OOH (6)
compared to the control test. However, the degradation percentage
%
decreased slightly (18%) when O2-purging test was performed. The OOH + H+ + e− → H2O2 (7)
results indicated that trapped holes and hydroxyl radicals adsorbed on
the surface of the 5% Ag2O/P-25 were the dominant oxidative species H2O2 + e− → %OH + OH– (8)
during the photocatalytic degradation of CBZ molecules. Similarly, the h+ + H2O → %OH + H+ (9)
degradation rate of DCF in the presence of EDTA was substantially
decreased to 62% as compared to the control test (Fig. 7b). However, h+, %O2–, %OH + CBZ or DCF → Intermediate products → CO2 + H2O
the effect of either MeOH or O2-purging was not significant (∼7%), (10)
324
Fig. 8. Cyclic experiments in the photocatalytic degradation of (a) CBZ, and (b) DCF over 5% Ag2O/P-25.
3.6. Reusability of 5% Ag2O/P-25 catalyst on PhACs removal under UV irradiation were analyzed using high-resolution FT-ICR mass
spectrometry and detected based on their accurate masses and deduced
The reusability of any catalyst during photocatalytic process is an elemental formulae (Table S3, supplementary information). The
important factor to evaluate its performance in practical applications. common transformation products of CBZ such as 10,11- dihydro-CBZ-
For this aim, cyclic photocatalytic experiments were performed for both 10,11-epoxide (m/z 253), acridine-9-carboxaldehyde (m/z 208), hy-
target pollutants on 5% Ag2O/P-25, as shown in Fig. 8. After, five droxyacridine-9-carboxaldehyde (m/z 224) [38,39,50,62] were de-
consecutive cycles of photocatalytic degradation of CBZ, the catalyst tected. The photocatalysis transforms CBZ to degradation products
did not exhibit any significant loss of its activity (Fig. 8a). The results showing other ecological properties [50]. The large number of com-
are indicative of the potential practical application of Ag2O/P-25 cat- pounds formed during the photocatalytic degradation evidences the
alyst based on its remarkable reusability for CBZ degradation. The complexity involved in the photocatalysis of CBZ, thus indicating the
plausible photocatalytic activity of the selected catalyst might be at- possible occurrence of multi-step degradation routes and inter-
tributed to the self-stability exhibited by Ag2O nanoparticles once the connected pathways [16].
complex structure of Ag-Ag2O is formed during the photocatalytic A photocatalytic degradation mechanism of CBZ shown in Fig. 9
process [60]. Moreover, the separation of catalyst from aqueous solu- was proposed based on the degradation products unambiguously
tion was achieved in all the cycles with just 15 min of centrifugation, identified photoproducts in this study and proposals of the previous
which leads to the conclusion that the complex catalyst synthesized can literatures [38,39,62,50]. The short-lived proposed intermediates are in
easily be recovered without high filtration costs, as normally associated the brackets, while stable intermediates are inside the parentheses.
with TiO2 nanoparticles [22]. Since substituents could appear in different positions in the aromatic
On the other hand, the photocatalytic removal efficiency of DCF rings, they have been represented as crossing the parenthesis marks. As
inhibited gradually during cyclic reuse of composite catalyst (Fig. 8b). verified by the quenching experiments, the major reactive species in-
The removal percentage of DCF dropped by 35% up to 5th cycle as volved in photocatalytic degradation of CBZ in aqueous Ag2O/P-25
compared to 1st cycle. Nevertheless, as the inhibition rate of the pho- suspensions were photogenerated h+ and %OH radicals.
tocatalytic activity was acceptable and remained almost stable after In photocatalytic process, hydroxylation of aromatic rings by %OH
2nd cycle, the synthesized 5% Ag2O/P-25 demonstrated a moderately radicals promotes the sequential ring cleavage reactions [7]. There
stable activity towards DCF degradation, and thus could be re- might be two different routes (a, and b) when photogenerated %OH
commended for the long-term application. However, the inhibitory radicals react with CBZ to form hydroxy-CBZ. First, %OH radical attacks
mechanism observed during the photocatalytic process of DCF under the aromatic ring of CBZ, which forms intermediate 1 via route b. The
the identical working conditions of CBZ was not clear. This might be intermediate 1 can further be transformed into intermediate 6 (oxcar-
due to the possible de-activation of catalyst by surface blockage, irre- bamazepine). Alternatively, intermediate 2 (epoxide-CBZ) and inter-
versible adsorption on active sites, and aggregation of catalyst particles mediate 5 ((2-carboxyphenyl)amino-5-hydroxybenzoic acid) form via
[47]. A clear transformation in a physical appearance of 5% Ag2O/P-25 route a due to short-lived reaction intermediate, hydroxy-CBZ. The
was observed before and after the reaction (Fig. S4). The interfering reaction of intermediate 2 with %OH radicals can form other degrada-
compounds formed during rapid initial photocatalytic process (organic tion products, such as intermediate 3 (acridine-9-carboxaldeyde). The
by-products etc.) may block the active surface sites, allowing neither ring cleavage in the intermediate 3 is followed by the loss of amide
target pollutants nor reactive species to adsorb and react with photo- group, which leads to an undetected intermediate, which further reacts
generated charges. Interfering compounds may act as a bridge or ad- with %OH to form intermediate 4 (hydroxyacridine-9-carboxaldehyde).
hesive between the particles, and also can scavenge the generated ROS Due to ring cleavage, intermediate 5 can further be transformed into
such as %OH, or %O2–, either discarding them or transforming into less other intermediates 6 (anthranilic/salicylic acid) and 7 (cathecol). Fi-
active ones [61]. The above hypothesis might be partially supported by nally, the complex hydroxylated intermediates and low molecular mass
the results from trapping experiment (cf. Fig. 7b), in which the influ- compounds can be oxidized through ring-rupturing reactions to ali-
ence of %OH, and %O2– was negligible on the photocatalytic degradation phatic compounds and further oxidation to mineralization (CO2 and
of DCF. H2O).
3.7. CBZ intermediates identification and possible degradation pathway 4. Conclusion
Of the selected PhACs, transformation products formed during the A heterojunction 5% Ag2O/P-25 was selected as an efficient pho-
photocatalytic degradation of CBZ (as an illustrative) using Ag2O/P-25 tocatalyst by a screening assessment of four different photocatalysts.
325
Fig. 9. Proposed photodegradation pathway of CBZ by Ag2O/P-25 under UV light irradiation. [CBZ]0 = 20 mg L−1, Catalyst dosage = 0.4 g L−1.
Dark adsorption, UV photolysis and photocatalytic kinetics of target initial pollutants concentration, and mineralization during the photo-
pollutants were investigated as preliminary tests. The selected 5% catalytic degradation of CBZ and DCF were investigated in both DW and
Ag2O/P-25 catalyst was prepared and characterized by SEM, EDX, BET, RME solution matrices. The optimal CBZ and DCF removals of 89.10%
and UV–vis diffuse reflectance spectra. The effects of catalyst dose, and 93.5%, respectively for 180 min of UV irradiation were observed at
326
catalyst dose of 0.4 g L−1 in DW matrix. However, the optimal catalyst mechanisms of ofloxacin and atenolol removal on titania suspensions, Water Res.
doses for CBZ and DCF in RME matrix were respectively increased by 44 (2010) 1737–1746, https://doi.org/10.1016/j.watres.2009.11.044.
[13] N. Vieno, T. Tuhkanen, L. Kronberg, Elimination of pharmaceuticals in sewage
factor 2 and 1.5, to achieve the same level of removal, which may be treatment plants in Finland, Water Res. 41 (2007) 1001–1012, https://doi.org/10.
due the influence of organic matters and natural scavengers in the real 1016/j.watres.2006.12.017.
effluents. A noticeable decreasing trends in removal rates were ob- [14] S. Judd, The status of membrane bioreactor technology, Trends Biotechnol. 26
(2008) 109–116, https://doi.org/10.1016/j.tibtech.2007.11.005.
served with the increased initial concentration of both PhAC, and ki- [15] K. Gurung, M.C. Ncibi, M. Sillanpää, Assessing membrane fouling and the perfor-
netics seem to fit the Langmuir-Hinshelwood model. A substantial im- mance of pilot-scale membrane bioreactor (MBR) to treat real municipal waste-
provement in mineralization was also achieved. The results showed that water during winter season in Nordic regions, Sci. Total Environ. 579 (2017)
1289–1297, https://doi.org/10.1016/j.scitotenv.2016.11.122.
photoinduced holes and %OH were the dominant oxidation species in [16] T.E. Doll, F.H. Frimmel, Kinetic study of photocatalytic degradation of carbama-
the photocatalytic degradation of PhACs. A possible reaction me- zepine, clofibric acid, iomeprol and iopromide assisted by different TiO2 materi-
chanism and energy band matching of complex 5% Ag2O/P-25 was also als—determination of intermediates and reaction pathways, Water Res. 38 (2004)
955–964, https://doi.org/10.1016/j.watres.2003.11.009.
demonstrated. Good reusability was exhibited by the selected photo-
[17] S.B. Hammouda, F. Zhao, Z. Safaei, I. Babu, D.L. Ramasamy, M. Sillanpää,
catalyst during photocatalytic degradation of target PhACs, ensuring its Reactivity of novel Ceria-Perovskite composites CeO2- LaMO3 (MCu, Fe) in the
aptitude for practical applications. Moreover, multiple routes of the catalytic wet peroxidative oxidation of the new emergent pollutant ‘Bisphenol F’:
CBZ transformation and various intermediates formed during photo- characterization, kinetic and mechanism studies, Appl. Catal. B Environ. 218 (2017)
119–136, https://doi.org/10.1016/j.apcatb.2017.06.047.
catalytic degradation were studied and a possible multi-step degrada- [18] F. Méndez-Arriaga, S. Esplugas, J. Giménez, Degradation of the emerging con-
tion path was proposed. taminant ibuprofen in water by photo-Fenton, Water Res. 44 (2010) 589–595,
https://doi.org/10.1016/j.watres.2009.07.009.
[19] M.M. Huber, S. Canonica, G.-Y. Park, U. von Gunten, Oxidation of pharmaceuticals
Acknowledgement during ozonation and advanced oxidation processes, Environ. Sci. Technol. 37
(2003) 1016–1024, https://doi.org/10.1021/es025896h.
This work was financially supported by Business Finland Oy [20] L. Hu, H.M. Martin, O. Arce-Bulted, M.N. Sugihara, K.A. Keating, T.J. Strathmann,
Oxidation of carbamazepine by Mn(VII) and Fe(VI): reaction kinetics and me-
(Innovaatiorahoituskeskus) (formerly Tekes – the Finnish Funding Agency chanism, Environ. Sci. Technol. 43 (2009) 509–515, https://doi.org/10.1021/
for Technology and Innovation) – A research funding for the Smart es8023513.
Effluents Project. Authors heartily thank Reijo Turkki (Director of [21] D. Chen, Z. Huang, H. Quan, S. Chen, J. Lin, X. Luo, L. Guo, Mesoporous single
crystal rutile TiO2 rods modified with Ag nanoparticles as a photocatalyst for de-
Mikkeli Vesilaitos), Risto Repo, Anne Bergman, and Päivi Luukkonen gradation of pollutants, Sci. Adv. Mater. 8 (2016) 1005–1009, https://doi.org/10.
for their consistent support to this work. The FT-ICR MS facility is 1166/sam.2016.2668.
supported by Biocenter Finland/Biocenter Kuopio and European [22] H. Dong, G. Zeng, L. Tang, C. Fan, C. Zhang, X. He, Y. He, An overview on lim-
itations of TiO2-based particles for photocatalytic degradation of organic pollutants
Regional Development Fund (Grant A70135).
and the corresponding countermeasures, Water Res. 79 (2015) 128–146, https://
doi.org/10.1016/j.watres.2015.04.038.
Appendix A. Supplementary material [23] J.C. Joo, G.Y. Kim, C.H. Ahn, S. Lee, J.-R. Park, J.K. Kim, Development of titanium
dioxide (TiO2)-embedded buoyant photocatalyst balls using expanded polystyrene,
J. Nanosci. Nanotechnol. 19 (2019) 1065–1068, https://doi.org/10.1166/jnn.2019.
Supplementary data to this article can be found online at https:// 15935.
doi.org/10.1016/j.seppur.2018.12.069. [24] Z. Li, F. Huang, H. Mo, A. Yan, H. Zhao, R. Xu, H. Dong, Q. Gao, Two-step sol-
vothermal method assisted by ultrasonic exfoliation to construct Nb3O7F/MoSe2
hybrids with enhanced photocatalytic activity, Sci. Adv. Mater. 9 (2018)
References 1626–1630, https://doi.org/10.1166/sam.2017.3182.
[25] P. Xiao, J.-H. Kim, S. Seo, Simple fabrication of highly sensitive photodetectors
[1] Y. Zhang, S.-U. Geißen, C. Gal, Carbamazepine and diclofenac: removal in waste- using MoS2 nanoparticles and Ag nanowires, Sci. Adv. Mater. 9 (2017) 1626–1630,
water treatment plants and occurrence in water bodies, Chemosphere 73 (2008) https://doi.org/10.1166/sam.2017.3182.
1151–1161, https://doi.org/10.1016/j.chemosphere.2008.07.086. [26] Q. Tong, T. Jiao, H. Guo, J. Zhou, Y. Wu, Q. Zhang, Q. Peng, Facile synthesis of
[2] Y. Lu, P. Hernandez, D. Abegunde, T. Edejer, The World Medicines Situation 2011 – highly crystalline α-Fe2O3 nanostructures with different shapes as photocatalysts
medicine expenditures, World Health Organ. 3 (2011) (2011) 1–34. for waste dye treatment, Sci. Adv. Mater. 8 (2016) 180–184, https://doi.org/10.
[3] O. Gimeno, J.F. García-Araya, F.J. Beltrán, F.J. Rivas, A. Espejo, Removal of 1166/sam.2016.2625.
emerging contaminants from a primary effluent of municipal wastewater by means [27] D. Wang, G. Xue, Y. Zhen, F. Fu, D. Li, Monodispersed Ag nanoparticles loaded on
of sequential biological degradation-solar photocatalytic oxidation processes, the surface of spherical Bi2WO6 nanoarchitectures with enhanced photocatalytic
Chem. Eng. J. 290 (2016) 12–20, https://doi.org/10.1016/j.cej.2016.01.022. activities, J. Mater. Chem. 22 (2012) 4751–4758, https://doi.org/10.1039/
[4] R.P. Cavalcante, R.F. Dantas, H. Wender, B. Bayarri, O. González, J. Giménez, C2JM14448D.
S. Esplugas, A. Machulek, Photocatalytic treatment of metoprolol with B-doped [28] X. Hu, Q. Zhu, X. Wang, N. Kawazoe, Y. Yang, Nonmetal–metal–semiconductor-
TiO2: effect of water matrix, toxicological evaluation and identification of inter- promoted P/Ag/Ag2O/Ag3PO4/TiO2 photocatalyst with superior photocatalytic
mediates, Appl. Catal. B Environ. 176–177 (2015) 173–182, https://doi.org/10. activity and stability, J. Mater. Chem. A 3 (2015) 17858–17865, https://doi.org/
1016/j.apcatb.2015.04.007. 10.1039/C5TA05153C.
[5] J. Zhang, B. Sun, X. Xiong, N. Gao, W. Song, E. Du, X. Guan, G. Zhou, Removal of [29] Z. Jiang, S. Le, Y. Xie, Q. Huang, B. Wang, T. Jiang, mpg-C3N4/Ag2O nano-
emerging pollutants by Ru/TiO2-catalyzed permanganate oxidation, Water Res. 63 composites photocatalysts with enhanced visible-light photocatalytic performance,
(2014) 262–270, https://doi.org/10.1016/j.watres.2014.06.028. J. Nanosci. Nanotechnol. 19 (2019) 721–728, https://doi.org/10.1166/jnn.2019.
[6] N. Vieno, M. Sillanpää, Fate of diclofenac in municipal wastewater treatment plant 15732.
– a review, Environ. Int. 69 (2014) 28–39, https://doi.org/10.1016/j.envint.2014. [30] W. Zhou, H. Liu, J. Wang, D. Liu, G. Du, J. Cui, Ag2O/TiO2 Nanobelts hetero-
03.021. structure with enhanced ultraviolet and visible photocatalytic activity, ACS Appl.
[7] J. Xu, L. Li, C. Guo, Y. Zhang, W. Meng, Photocatalytic degradation of carbama- Mater. Interfaces 2 (2010) 2385–2392, https://doi.org/10.1021/am100394x.
zepine by tailored BiPO4: efficiency, intermediates and pathway, Appl. Catal. B [31] J. Han, H. Jeong, H.-D. Kim, H. Song, K. Choi, J. Park, Erbium-doped TiO2 hollow
Environ. 130–131 (2013) 285–292, https://doi.org/10.1016/j.apcatb.2012.11.013. sphere as a visible-light-sensitive photocatalyst, Nanosci. Nanotechnol. Lett. 10
[8] M. Clara, B. Strenn, O. Gans, E. Martinez, N. Kreuzinger, H. Kroiss, Removal of (2018) 1152–1156, https://doi.org/10.1166/nnl.2018.2747.
selected pharmaceuticals, fragrances and endocrine disrupting compounds in a [32] Y. Xu, M.A.A. Schoonen, The absolute energy positions of conduction and valence
membrane bioreactor and conventional wastewater treatment plants, Water Res. 39 bands of selected semiconducting minerals, Am. Mineral. 85 (2000) 543–556,
(2005) 4797–4807, https://doi.org/10.1016/j.watres.2005.09.015. https://doi.org/10.2138/am-2000-0416.
[9] EU, EU. Directive 2013/39/EU of the European Parliament and of the Council, 2013 [33] S. Ahmed, M.G. Rasul, W.N. Martens, R. Brown, M.A. Hashib, Advances in het-
(2013) 1–17. erogeneous photocatalytic degradation of phenols and dyes in wastewater: a re-
[10] K. Gurung, M.C. Ncibi, M. Shestakova, M. Sillanpää, Removal of carbamazepine view, Water. Air. Soil Pollut. 215 (2011) 3–29, https://doi.org/10.1007/s11270-
from MBR effluent by electrochemical oxidation (EO) using a Ti/Ta2O5-SnO2 010-0456-3.
electrode, Appl. Catal. B Environ. 221 (2018) 329–338, https://doi.org/10.1016/j. [34] H.-T. Ren, S.-Y. Jia, J.-J. Zou, S.-H. Wu, X. Han, A facile preparation of Ag2O/P25
apcatb.2017.09.017. photocatalyst for selective reduction of nitrate, Appl. Catal. B Environ. 176–177
[11] L. Haroune, M. Salaun, A. Ménard, C.Y. Legault, J.-P. Bellenger, Photocatalytic (2015) 53–61, https://doi.org/10.1016/j.apcatb.2015.03.038.
degradation of carbamazepine and three derivatives using TiO2 and ZnO: effect of [35] H.-T. Ren, Q. Yang, Fabrication of Ag2O/TiO2 with enhanced photocatalytic per-
pH, ionic strength, and natural organic matter, Sci. Total Environ. 475 (2014) formances for dye pollutants degradation by a pH-induced method, Appl. Surf. Sci.
16–22, https://doi.org/10.1016/j.scitotenv.2013.12.104. 396 (2017) 530–538, https://doi.org/10.1016/j.apsusc.2016.10.191.
[12] E. Hapeshi, A. Achilleos, M.I. Vasquez, C. Michael, N.P. Xekoukoulotakis, [36] Y. Du, J. Rabani, The measure of TiO2 photocatalytic efficiency and the comparison
D. Mantzavinos, D. Kassinos, Drugs degrading photocatalytically: kinetics and of different photocatalytic titania, J. Phys. Chem. B 107 (2003) 11970–11978,
327
https://doi.org/10.1021/jp035491z. TiO2, Chemosphere 39 (1999) 781–786, https://doi.org/10.1016/S0045-6535(99)

[37] C. Pan, Y. Zhu, New type of BiPO4 oxy-acid salt photocatalyst with high photo- 00013-2.
catalytic activity on degradation of dye, Environ. Sci. Technol. 44 (2010) [50] T.E. Doll, F.H. Frimmel, Removal of selected persistent organic pollutants by het-
5570–5574, https://doi.org/10.1021/es101223n. erogeneous photocatalysis in water, Catal. Today 101 (2005) 195–202, https://doi.
[38] C. Martínez, M.L. Canle, M.I. Fernández, J.A. Santaballa, J. Faria, Kinetics and org/10.1016/j.cattod.2005.03.005.
mechanism of aqueous degradation of carbamazepine by heterogeneous photo- [51] N. De la Cruz, R.F. Dantas, J. Giménez, S. Esplugas, Photolysis and TiO2 photo-
catalysis using nanocrystalline TiO2, ZnO and multi-walled carbon nanotubes–a- catalysis of the pharmaceutical propranolol: solar and artificial light, Appl. Catal. B
natase composites, Appl. Catal. B Environ. 102 (2011) 563–571, https://doi.org/10. Environ. 130–131 (2013) 249–256, https://doi.org/10.1016/j.apcatb.2012.10.003.
1016/j.apcatb.2010.12.039. [52] J. Deng, Y. Shao, N. Gao, S. Xia, C. Tan, S. Zhou, X. Hu, Degradation of the anti-
[39] M. Nawaz, W. Miran, J. Jang, D.S. Lee, One-step hydrothermal synthesis of porous epileptic drug carbamazepine upon different UV-based advanced oxidation pro-
3D reduced graphene oxide/TiO2 aerogel for carbamazepine photodegradation in cesses in water, Chem. Eng. J. 222 (2013) 150–158, https://doi.org/10.1016/j.cej.
aqueous solution, Appl. Catal. B Environ. 203 (2017) 85–95, https://doi.org/10. 2013.02.045.
1016/j.apcatb.2016.10.007. [53] K.V. Kumar, K. Porkodi, F. Rocha, Langmuir-Hinshelwood kinetics – a theoretical
[40] L. Rizzo, S. Meric, D. Kassinos, M. Guida, F. Russo, V. Belgiorno, Degradation of study, Catal. Commun. 9 (2008) 82–84, https://doi.org/10.1016/j.catcom.2007.05.
diclofenac by TiO2 photocatalysis: UV absorbance kinetics and process evaluation 019.
through a set of toxicity bioassays, Water Res. 43 (2009) 979–988, https://doi.org/ [54] L. Rizzo, S. Meric, M. Guida, D. Kassinos, V. Belgiorno, Heterogenous photocatalytic
10.1016/j.watres.2008.11.040. degradation kinetics and detoxification of an urban wastewater treatment plant
[41] Ö. Kerkez, İ. Boz, Photodegradation of methylene blue with Ag2O/TiO2 under effluent contaminated with pharmaceuticals, Water Res. 43 (2009) 4070–4078,
visible light: operational parameters, Chem. Eng. Commun. 202 (2015) 534–541, https://doi.org/10.1016/j.watres.2009.06.046.
https://doi.org/10.1080/00986445.2013.853292. [55] Y. Tamaki, A. Furube, M. Murai, K. Hara, R. Katoh, M. Tachiya, Direct observation
[42] R. Lamba, A. Umar, S.K. Mehta, S.K. Kansal, Enhanced visible light driven photo- of reactive trapped holes in TiO2 undergoing photocatalytic oxidation of adsorbed
catalytic application of Ag2O decorated ZnO nanorods heterostructures, Sep. Purif. alcohols: evaluation of the reaction rates and yields, J. Am. Chem. Soc. 128 (2006)
Technol. 183 (2017) 341–349, https://doi.org/10.1016/j.seppur.2017.03.070. 416–417, https://doi.org/10.1021/ja055866p.
[43] R. Lamba, A. Umar, S.K. Mehta, S.K. Kansal, CeO2ZnO hexagonal nanodisks: effi- [56] N. Zhang, G. Liu, H. Liu, Y. Wang, Z. He, G. Wang, Diclofenac photodegradation
cient material for the degradation of direct blue 15 dye and its simulated dye bath under simulated sunlight: Effect of different forms of nitrogen and kinetics, J.
effluent under solar light, J. Alloys Compd. 620 (2015) 67–73, https://doi.org/10. Hazard. Mater. 192 (2011) 411–418, https://doi.org/10.1016/j.jhazmat.2011.05.
1016/j.jallcom.2014.09.101. 038.
[44] S.A. Khan, S. Ali, M. Sohail, M.A. Morsy, Z.H. Yamani, Fabrication of TiO2/Ag/ [57] J. Gao, C. Liu, F. Wang, L. Jia, K. Duan, T. Liu, Facile synthesis of heterostructured
Ag2O nanoparticles to enhance the photocatalytic activity of Degussa P25 titania, WS2/Bi2MoO6 as high-performance visible-light-driven photocatalysts, Nanoscale
Aust. J. Chem. 69 (2016) 41–46, https://doi.org/10.1071/CH15250. Res. Lett. 12 (2017) 377, https://doi.org/10.1186/s11671-017-2157-y.
[45] Y.C. Zhang, Z.N. Du, K.W. Li, M. Zhang, D.D. Dionysiou, High-performance visible- [58] X. Wang, S. Li, H. Yu, J. Yu, S. Liu, Ag2O as a new visible-light photocatalyst: self-
light-driven SnS2/SnO2 nanocomposite photocatalyst prepared via in situ hydro- stability and high photocatalytic activity, Chem. – Eur. J. 17 (2011) 7777–7780,
thermal oxidation of SnS2 nanoparticles, ACS Appl. Mater. Interfaces 3 (2011) https://doi.org/10.1002/chem.201101032.
1528–1537, https://doi.org/10.1021/am200102y. [59] L. Shi, L. Liang, J. Ma, F. Wang, J. Sun, Enhanced photocatalytic activity over the
[46] D. Wodka, R.P. Socha, E. Bielańska, M. Elżbieciak-Wodka, P. Nowak, P. Warszyński, Ag2O–g-C3N4 composite under visible light, Catal. Sci. Technol. 4 (2014) 758–765,
Photocatalytic activity of titanium dioxide modified by Fe2O3 nanoparticles, Appl. https://doi.org/10.1039/C3CY00871A.
Surf. Sci. 319 (2014) 173–180, https://doi.org/10.1016/j.apsusc.2014.08.010. [60] H. Yu, R. Liu, X. Wang, P. Wang, J. Yu, Enhanced visible-light photocatalytic ac-
[47] G. Laera, B. Jin, H. Zhu, A. Lopez, Photocatalytic activity of TiO2 nanofibers in tivity of Bi2WO6 nanoparticles by Ag2O cocatalyst, Appl. Catal. B Environ. 111–112
simulated and real municipal effluents, Catal. Today 161 (2011) 147–152, https:// (2012) 326–333, https://doi.org/10.1016/j.apcatb.2011.10.015.
doi.org/10.1016/j.cattod.2010.10.037. [61] A. Katz, A. McDonagh, L. Tijing, H.K. Shon, Fouling and inactivation of titanium
[48] M.Y. Ghaly, T.S. Jamil, I.E. El-Seesy, E.R. Souaya, R.A. Nasr, Treatment of highly dioxide-based photocatalytic systems, Crit. Rev. Environ. Sci. Technol. 45 (2015)
polluted paper mill wastewater by solar photocatalytic oxidation with synthesized 1880–1915, https://doi.org/10.1080/10643389.2014.1000763.
nano TiO2, Chem. Eng. J. 168 (2011) 446–454, https://doi.org/10.1016/j.cej.2011. [62] D. Vogna, R. Marotta, R. Andreozzi, A. Napolitano, M. d’Ischia, Kinetic and che-
01.028. mical assessment of the UV/H2O2 treatment of antiepileptic drug carbamazepine,
[49] M.S.T. Gonçalves, A.M.F. Oliveira-Campos, E.M.M.S. Pinto, P.M.S. Plasência, Chemosphere 54 (2004) 497–505, https://doi.org/10.1016/S0045-6535(03)
M.J.R.P. Queiroz, Photochemical treatment of solutions of azo dyes containing 00757-4.
328
ACTA UNIVERSITATIS LAPPEENRANTAENSIS
828. METSO, LASSE. Information-based industrial maintenance - an ecosystem perspective.

2018. Diss.
829. SAREN, ANDREY. Twin boundary dynamics in magnetic shape memory alloy Ni-Mn-
Ga five-layered modulated martensite. 2018. Diss.
830. BELONOGOVA, NADEZDA. Active residential customer in a flexible energy system - a

methodology to determine the customer behaviour in a multi-objective environment.
2018. Diss.
831. KALLIOLA, SIMO. Modified chitosan nanoparticles at liquid-liquid interface for

applications in oil-spill treatment. 2018. Diss.
832. GEYDT, PAVEL. Atomic Force Microscopy of electrical, mechanical and piezo
properties of nanowires. 2018. Diss.
833. KARELL, VILLE. Essays on stock market anomalies. 2018. Diss.
834. KURONEN, TONI. Moving object analysis and trajectory processing with applications in
human-computer interaction and chemical processes. 2018. Diss.
835. UNT, ANNA. Fiber laser and hybrid welding of T-joint in structural steels. 2018. Diss.
836. KHAKUREL, JAYDEN. Enhancing the adoption of quantified self-tracking wearable

devices. 2018. Diss.
837. SOININEN, HANNE. Improving the environmental safety of ash from bioenergy
production plants. 2018. Diss.
838. GOLMAEI, SEYEDMOHAMMAD. Novel treatment methods for green liquor dregs and
enhancing circular economy in kraft pulp mills. 2018. Diss.
839. GERAMI TEHRANI, MOHAMMAD. Mechanical design guidelines of an electric vehicle

powertrain. 2019. Diss.
840. MUSIIENKO, DENYS. Ni-Mn-Ga magnetic shape memory alloy for precise high-speed
actuation in micro-magneto-mechanical systems. 2019. Diss.
841. BELIAEVA, TATIANA. Complementarity and contextualization of firm-level strategic

orientations. 2019. Diss.
842. EFIMOV-SOINI, NIKOLAI. Ideation stage in computer-aided design. 2019. Diss.
843. BUZUKU, SHQIPE. Enhancement of decision-making in complex organizations: A

systems engineering approach. 2019. Diss.
844. SHCHERBACHEVA, ANNA. Agent-based modelling for epidemiological applications.

2019. Diss.
845. YLIJOKI, OSSI. Big data - towards data-driven business. 2019. Diss.
846. KOISTINEN, KATARIINA. Actors in sustainability transitions. 2019. Diss.
847. GRADOV, DMITRY. Experimentally validated numerical modelling of reacting

multiphase flows in stirred tank reactors. 2019. Diss.
848. ALMPANOPOULOU, ARGYRO. Knowledge ecosystem formation: an institutional and
organisational perspective. 2019. Diss.
849. AMELI, ALIREZA. Supercritical CO2 numerical modelling and turbomachinery design.
2019. Diss.
850. RENEV, IVAN. Automation of the conceptual design process in construction industry
using ideas generation techniques. 2019. Diss.
851. AVRAMENKO, ANNA. CFD-based optimization for wind turbine locations in a wind
park. 2019. Diss.
852. RISSANEN, TOMMI. Perspectives on business model experimentation in

internationalizing high-tech companies. 2019. Diss.
853. HASSANZADEH, AIDIN. Advanced techniques for unsupervised classification of remote

sensing hyperspectral images. 2019. Diss.
854. POPOVIC, TAMARA. Quantitative indicators of social sustainability applicable in

process systems engineering. 2019. Diss.
855. RAMASAMY, DEEPIKA. Selective recovery of rare earth elements from diluted
aqueous streams using N- and O –coordination ligand grafted organic-inorganic hybrid
composites. 2019. Diss.
856. IFTEKHAR, SIDRA. Synthesis of hybrid bio-nanocomposites and their application for
the removal of rare earth elements from synthetic wastewater. 2019. Diss.
857. HUIKURI, MARKO. Modelling and disturbance compensation of a permanent magnet

linear motor with a discontinuous track 2019. Diss.
858. AALTO, MIKA. Agent-based modeling as part of biomass supply system research.
2019. Diss.
859. IVANOVA, TATYANA. Atomic layer deposition of catalytic materials for environmental
protection. 2019. Diss.
860. SOKOLOV, ALEXANDER. Pulsed corona discharge for wastewater treatment and
modification of organic materials. 2019. Diss.
861. DOSHI, BHAIRAVI. Towards a sustainable valorisation of spilled oil by establishing a

green chemistry between a surface active moiety of chitosan and oils. 2019. Diss.
862. KHADIJEH, NEKOUEIAN. Modification of carbon-based electrodes using metal

nanostructures: Application to voltammetric determination of some pharmaceutical and
biological compounds. 2019. Diss.
863. HANSKI, JYRI. Supporting strategic asset management in complex and uncertain
decision contexts. 2019. Diss.
864. OTRA-AHO, VILLE. A project management office as a project organization’s

strategizing tool. 2019. Diss.
865. HILTUNEN, SALLA. Hydrothermal stability of microfibrillated cellulose. 2019. Diss.

ISBN 978-952-335-410-4
ISBN 978-952-335-411-1 (PDF)
ISSN-L 1456-4491
ISSN 1456-4491
Lappeenranta 2019

Khum Gurung A4

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Khum Gurung A4

Uploaded by

Copyright:

Available Formats

Khum Gurung

MEMBRANE BIOREACTOR FOR THE REMOVAL

ACTA UNIVERSITATIS LAPPEENRANTAENSIS 866

MEMBRANE BIOREACTOR FOR THE REMOVAL

Assistant Professor Mohamed Chaker Ncibi

Reviewers Professor Anastasios I. Zouboulis

Professor How Yong Ng

Opponent Professor of Practice Anna Mikola

Lappeenranta-Lahti University of Technology LUT

Water scarcity is undoubtedly a global concern due to exacerbated population growth,

This dissertation presents a comprehensive practical study on the applicability of the

Keywords: Membrane bioreactor, municipal wastewater treatment, emerging

In these circumstances, municipal wastewater reclamation is one of the promising

1.2 Research objectives

- To study the performance of a conventional MBR to treat the real municipal

2.1 Occurrence of emerging contaminants in municipal wastewater

Table 2. Physicochemical properties of some commonly identified ECs in municipal

Molecu Henry’s Law

2.3 Municipal wastewater treatment

2.4 EU legislation and removal of ECs from municipal wastewater

3.2 Membrane materials, configuration and operating principles of

Figure 1: MBR process configurations, (a) Submerged and (b) side-stream

Figure 2: Different MBR membrane modules: (a) Tubular, (b) Hollow-fiber,

Furthermore, membrane fouling can be categorized as biofouling, inorganic fouling and

Where, Rt is the total membrane filtration resistance (m-1); TMP is trans-membrane

Figure 3: Temporal and spatial development of membrane fouling in MBRs

3.4 Fate of ECs in MBRs: application of mass balance

( C Inf ∗ QInf − C Eff ∗ QEff )

Removal via sorption onto sludge(%)

Remain in effluent (%)

Removal via biotransformation(%)

= Overall removal (%)

− Removal via sorption onto sludge (%) (5)

4.1 Electrochemical oxidation (ECO) process

Rads – ne- → Pads (6)

On the other hand, electrochemical oxidation of organic pollutants can be obtained

MOx + H2O → MOx (•OH) + H++ e- (7)

Figure 4: Electrochemical oxidation scheme of organic pollutants on

4.1.1 Active anodes

MOx (•OH) → MOx+1 + H++ e- (8)

MOx+1 + R → MOx + RO (9)

4.1.2 Non-active anodes

𝛼MOx (•OH) + R → 𝛼MOx + mCO2 +nH2O + zH++ ze- (11)

MOx (•OH) → MOx + (1/2) O2 + H++ e- (12)

2 MOx (•OH) → 2 MOx + H2O2 (13)

4.2 Photocatalytic oxidation (PCO)

In heterogeneous photocatalysis process, organic pollutants are decontaminated in the

h+ + H2O → •OH + H+ (16)

h+ / •OH / •O2- + R → Intermediates→ CO2 + H2O+ inorganic ions (17)

In heterogeneous photocatalysis, TiO2 catalyst can be utilized either as aqueous

Figure 5: Decision tree for potential integration approaches of MBR

García-Gómez et al. showed an enhanced removal of carbamazepine in synthetic

Figure 6: Schematic diagram of the pilot MBR unit at Kenkäveronniemi WWTP,

6.2 Assessing integration approaches to MBR

6.2.1 Lab-scale set-up for batch ECO unit

Figure 7: Experimental set-up for electrochemical oxidation.

6.2.2 Lab-scale set-up for PCO process

The photocatalytic experimental set-up using a self-made photocatalytic reactor is shown

Figure 8: Experimental set-up for photocatalytic oxidation

6.2.3 Analytical procedures

Figure 9: Membrane fouling behavior of MBR system under cold weather

Figure 10: Occurrences of viruses in wastewater stream under cold

Figure 11: Removal efficiencies of selected ECs in MBR treatment under

7.2 Assessing the performance of MBR under varying solid retention