forward osmosis for resource recovery from anaerobically
TRANSCRIPT
University of WollongongResearch Online
University of Wollongong Thesis Collection 2017+ University of Wollongong Thesis Collections
2019
Forward osmosis for resource recovery fromanaerobically digested sludge centrateTruong Minh VuUniversity of Wollongong
Unless otherwise indicated, the views expressed in this thesis are those of the author and donot necessarily represent the views of the University of Wollongong.
Research Online is the open access institutional repository for the University of Wollongong. For further information contact the UOW Library:[email protected]
Recommended CitationVu, Truong Minh, Forward osmosis for resource recovery from anaerobically digested sludge centrate, Master of Philosophy thesis,School of Civil, Mining and Environmental Engineering, University of Wollongong, 2019. https://ro.uow.edu.au/theses1/494
School of Civil, Mining and Environmental Engineering
Faculty of Engineering and Information Sciences
University of Wollongong Australia
FORWARD OSMOSIS FOR RESOURCE RECOVERY FROM
ANAEROBICALLY DIGESTED SLUDGE CENTRATE
A thesis submitted in fulfilment of the requirements for the award of the degree of
Master of Philosophy
from
UNIVERSITY OF WOLLONGONG
By
Truong Minh Vu
February 2019
i
CERTIFICATION
I, Truong Minh Vu, hereby declare that this thesis, submitted in fulfilment of the requirements
for the award of Master of Philosophy to the School of Civil, Mining and Environmental
Engineering, Faculty of Engineering and Information Sciences, University of Wollongong is
wholly my own work unless otherwise referenced or acknowledged. The document has not
been submitted for qualification at any other academic institution
Truong Minh Vu
February 2019
Thesis supervision:
Assoc. Prof. Faisal I. Hai, School of Civil, Mining and Environmental Engineering
UOW, Wollongong, NSW 2522. Tel: +612 42213054. Email: [email protected]
Prof. Long Nghiem, School of Civil and Environmental Engineering
UTS, Sydney, NSW 2522. Tel: +612 95142625. Email: [email protected]
Dr. Ashley J. Ansari, School of Civil, Mining and Environmental Engineering
UOW, Wollongong, NSW 2522 Tel: +612 42214590. Email. [email protected]
ii
THESIS RELATED PUBLICATIONS
Peer-reviewed Journal Paper
1. M.T. Vu, A.J. Ansari, F.I. Hai, L.D. Nghiem, Performance of a seawater-driven forward
osmosis process for pre-concentrating sludge centrate: organic enrichment and membrane
fouling, Environmental Science: Water Research & Technology, 4 (2018) 1047-1056.
Conference Presentation
1. M.T. Vu, A.J. Ansari, F.I. Hai, L.D. Nghiem, Fouling behaviour and performance of
seawater-driven forward osmosis for pre-concentrating sludge centrate. Oral presentation at
the 11st conference of the Aseanian Membrane Society, AMS 11, Brisbane, Queensland,
Australia, 3 – 6 July 2018.
iii
ABSTRACT
Digested sludge centrate also known as sludge centrate is a by-product of the anaerobic
digestion process obtained from sludge dewater. Sludge centrate contains most of the
nutrients and some organics from the initial sewage sludge. Nutrients, namely ammonia and
phosphorus can be extracted from sludge centrate in the forms of calcium phosphate or
struvite precipitate. However, the concentration of organic matter and nutrients in raw sludge
centrate is too low for for direct anaerobic digestion and precipitation. Thus, a novel approach
of employing seawater-driven forward osmosis to pre-concentrate organic matter and
nutrients in sludge centrate is vital to render sludge centrate anaerobically digestible and less
chemical-intensively precipitable.
This thesis aimed to demonstrate for the first time the efficiency of organic matter enrichment
in sludge centrate using a seawater-driven forward osmosis process. In particular, this
research elucidated the impacts of membrane materials, prewetting procedures and draw
solution on the performance of the seawater-driven forward osmosis process. The results
indicated that the cellulose triacetate membrane (CTA) offered better performance than the
polyamide membrane (PA) in terms of organic materials enrichment, fouling resistance and
membrane cleaning efficiency. The prewetting protocol using 70% alcohol solution led to the
significantly increased pure water flux of the PA membrane, while this step appeared to be
ineffective to improve that of the CTA membrane. Membrane fouling decreased the
enrichment efficiency of organic matter since the deposition of suspended particulate matter
on the membrane surface caused fouling and loss of organic matter from the concentrated
sludge centrate. The results showed that increasing the draw solution concentration increased
flux but did not aggravate membrane fouling, however, it could reduce the efficiency of
physical flushing to recover the flux. Seawater showed comparable forward osmosis
performance to that of analytical grade NaCl as draw solutes in terms of flux and organic
enrichment. The results also showed that seawater as the draw solution resulted in more
membrane fouling and lower flux recovery compared to NaCl.
This study also proposed and evaluated techniques to mitigate membrane fouling during
nutrient enrichment in sludge centrate by forward osmosis. Phosphorus precipitation in the
bulk feed solution and on the membrane surface was systematically quantified and compared
with respect to solution pH and operation time. The results indicated that low efficiency of
nutrients enrichment when using seawater as the draw solution was attributed to formation of
phosphorus precipitates. Moreover, increase in pH during the filtration also caused low
iv
efficiency of ammonium enrichment. Formation of inorganic precipitates (i.e. calcium
phosphate and struvite) on the membrane surface led to severe fouling. The results also
showed a relationship among pH, duration of filtration, fouling and phosphorus precipitation.
This study demonstrated that membrane fouling could be minimized by encouraging
phosphorus precipitation in the bulk solution (avoid feed solution mixing) and increasing the
membrane surface area to shorten operation time. Furthermore, proper ratio of membrane area
over permeate volume to enrich nutrients was 175 m-1. The reduction in membrane fouling
also resulted in a proportional increase of phosphate and ammonia in the final concentrated
feed solution.
v
ACKNOWLEDGEMENTS
I would like to express the deepest gratitude to Prof. Long Nghiem in the School of Civil and
Environmental Engineering at the University of Technology Sydney for his supervision and
support. Prof. Long Nghiem has been a very dedicated and thoughtful mentor. He has
provided me with massive helpful advice on my research. He has always taught me right from
wrong. I am grateful for all opportunities he has given me. Without him, my thesis would not
be successfully completed.
I would like to acknowledge my principal supervisor, Assoc. Prof. Faisal I. Hai in the School
of Civil, Mining and Environmental Engineering at the University of Wollongong for his
supervision and encouragement during my study here.
My special thanks are sent to my co-supervisor, Dr. Ashley J. Ansari for her support and
encouragement during the difficult times. As for me, she is like a friend rather than a
supervisor. Her stories in terms of research really inspire me and keep me motivated to pursue
my research career.
I gratefully acknowledged the University of Wollongong Australia for financial scholarship
support to assist me to carry out my research.
I would like to thank all my friends in the environmental engineering lab. They are so
friendly, kind-hearted and trustworthy. They have helped me settle down my life as well as
get on well with the new life in Australia. I will never forget what they have done for me.
Words are powerless to express my gratitude to my parents and relatives for their love and
encouragement throughout my life. From the bottom of my heart, I wish them health and
happiness.
Finally, I would like to spend the most sincere and special thanks to my wife, Ngoc (Emily),
especially for her patience, encouragement and love. Without her, my study could be never
possible.
vi
CONTENTS
ABSTRACT ................................................................................................................... iii
ACKNOWLEDGEMENTS ............................................................................................ v
CONTENTS ................................................................................................................... vi
LIST OF FIGURES ....................................................................................................... ix
LIST OF TABLES .......................................................................................................... x
ACRONYMS ................................................................................................................ xii
CHAPTER 1. INTRODUCTION ................................................................................... 1
1.1. Background .............................................................................................................. 2
1.2. Objectives and scope of thesis ................................................................................. 3
1.3. Thesis outline ........................................................................................................... 4
CHAPTER 2. LITERATURE REVIEW ........................................................................ 6
2.1. Forward osmosis....................................................................................................... 7
2.1.1. Osmotic process ................................................................................................. 7
2.1.2. Transport phenomena ........................................................................................ 7
2.1.2.1. Mechanisms of mass transfer ...................................................................... 7
2.1.2.2. External concentration polarization ............................................................ 8
2.1.2.3. Internal concentration polarization ........................................................... 10
2.1.2.4. Reverse solute flux .................................................................................... 12
2.1.3. Materials .......................................................................................................... 14
2.1.4. Forward osmosis operation .............................................................................. 15
2.1.4.1. Temperature .............................................................................................. 15
2.1.4.2. Feed solution pH ....................................................................................... 16
2.1.4.3. Cross-flow velocities and spacer designs .................................................. 17
2.1.4.4. Draw solution ............................................................................................ 18
2.1.4.5. Membrane orientation ............................................................................... 21
2.1.4.6. Module configurations .............................................................................. 21
2.2. Fouling phenomena in the FO membrane operation .............................................. 22
2.2.1. FO membrane fouling ...................................................................................... 22
2.2.2. Colloidal and organic fouling .......................................................................... 23
vii
2.2.2.1. Characteristics of colloids and macromolecules ....................................... 23
2.2.2.2. Colloidal interactions ................................................................................ 26
2.2.3. Inorganic fouling.............................................................................................. 26
2.2.4. Biofouling ........................................................................................................ 27
2.2.5. Factors affecting FO membrane fouling .......................................................... 28
2.2.5.1. Feedwater characteristics .......................................................................... 28
2.2.5.2. Hydrodynamic conditions ......................................................................... 29
2.2.5.3. Membrane properties ................................................................................ 30
2.2.5.4. The composition of draw solution ............................................................. 31
2.2.5.5. Membrane orientation ............................................................................... 32
2.2.6. Effects of membrane fouling on the performance of FO process.................... 32
2.2.7. Mitigation of fouling in the FO process .......................................................... 33
2.2.7.1. Fouling characterization ............................................................................ 33
2.2.7.2. Reduced-fouling methods ......................................................................... 34
2.2.7.3. Membrane fouling control ......................................................................... 38
CHAPTER 3 ................................................................................................................. 42
PERFORMANCE OF A SEAWATER-DRIVEN FORWARD OSMOSIS PROCESS
FOR PRE-CONCENTRATING SLUDGE CENTRATE: ORGANIC ENRICHMENT
AND MEMBRANE FOULING ................................................................................... 42
3.1. Introduction ............................................................................................................ 43
3.2. Materials and methods ........................................................................................... 44
3.2.1. Forward osmosis system and membranes ....................................................... 44
3.2.2. Feed solution and draw solution ...................................................................... 45
3.2.3. Membrane prewetting ...................................................................................... 46
3.2.4. Experimental protocols .................................................................................... 46
3.2.5. Calculations ..................................................................................................... 46
3.2.6. Analytical methods .......................................................................................... 47
3.3. Results and discussions .......................................................................................... 47
3.3.1. Pure water flux under different conditions ...................................................... 47
3.3.2. Pre-concentration of sludge centrate by forward osmosis .............................. 49
3.3.3. Factors affecting fouling behaviour ................................................................ 51
viii
3.3.3.1. Membrane properties and orientations ...................................................... 51
3.3.3.2. Draw solution ............................................................................................ 52
3.3.3.3. Fouling layer characteristics ..................................................................... 53
3.3.4. Cleaning and flux reversibility ........................................................................ 54
3.4. Conclusions ............................................................................................................ 57
CHAPTER 4 ................................................................................................................. 58
SEAWATER-DRIVEN FORWARD OSMOSIS PRE-CONCENTRATION OF
NUTRIENTS IN SLUDGE CENTRATE: PERFORMANCE AND IMPLICATIONS ..
................................................................................................................................ 58
4.1. Introduction ............................................................................................................ 59
4.2. Materials and methods ........................................................................................... 60
4.2.1. Forward osmosis system.................................................................................. 60
4.2.2. Feed solution and draw solution ...................................................................... 61
4.2.3. Experimental protocols .................................................................................... 61
4.2.4. Analytical methods .......................................................................................... 62
4.3. Results and discussions .......................................................................................... 62
4.3.1. Performance of FO for enriching nutrients in sludge centrate ........................ 62
4.3.1.1. Water flux .................................................................................................. 62
4.3.1.2. Nutrient enrichment .................................................................................. 63
4.3.2. Performance of FO for enriching nutrients in sludge centrate with buffered
seawater ..................................................................................................................... 66
4.3.2.1. Effects of using buffered seawater as the DS on water flux ..................... 66
4.3.2.2. Improvement in nutrient enrichment using buffered seawater as the DS . 67
4.4. Conclusions ............................................................................................................ 68
CHAPTER 5 ................................................................................................................. 70
CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK ................ 70
5.1. Conclusions ............................................................................................................ 71
5.2. Recommendations for Future Work ...................................................................... 71
REFERENCES ............................................................................................................. 73
ix
LIST OF FIGURES
Figure 1. The outline of thesis ........................................................................................ 5
Figure 2. Structure of an FO membrane. ........................................................................ 8
Figure 3. (a) ECP phenomena and dilutive ICP in AL-FS orientation and (b) ECP
phenomena and concentrative ICP in AL-DS orientation in the FO process. ................ 9
Figure 4. Geometric structure of ladder (a) and diamond (b) configurations of spacer 17
Figure 5. Flow direction in forward and reverse flushing in FO mode ........................ 38
Figure 6. The permeate flow direction during backwashing in PRO mode ................. 39
Figure 7. Schematic diagram of the lab-scale FO system. ........................................... 45
Figure 8. Pure water fluxes of CTA and PA FO membranes using seawater as DS in
AL-FS and AL-DS orientations with and without prewetting. ..................................... 47
Figure 9. The performance of FO for enrichment of COD in sludge centrate using
seawater as the DS. The theoretical COD concentration factor was calculated assuming
100% COD retention by the FO membrane.................................................................. 49
Figure 10. The performance of FO for enrichment of COD in sludge centrate using the
NaCl solution as the DS and the CTA membrane in the AL-FS orientation. Note: The
theoretical COD concentration factor was calculated assuming 100% COD retention
by the FO membrane. .................................................................................................... 50
Figure 11. Fouling behaviour of the FO membranes in (A) AL-FS orientation and (B)
AL-DS orientation using seawater as DS. .................................................................... 51
Figure 12. Comparison of fouling behaviour towards CTA membranes in the AL-FS
orientation using seawater and the NaCl solutions at different concentration as DSs. 52
Figure 13. SEM images of the AL of (A) the fouled CTA membrane and (B) the
fouled TFC PA membrane in the AL-FS orientation using seawater DS..................... 53
Figure 14. Flux recovery of the fouled CTA, PA and prewetted PA membranes using
seawater DS after physical flushing. ............................................................................. 54
Figure 15. Comparison of pure water fluxes and flux recovery of CTA membranes in
the AL-FS orientation using seawater and NaCl DS. ................................................... 55
Figure 16. SEM images of (A) the AL of the cleaned CTA membrane in the AL-FS
orientation using seawater as DS, (B) the AL of the cleaned CTA membrane in the
AL-FS orientation using the 0.5 M NaCl solution as DS, (C) the SL of the cleaned
CTA membrane in the AL-DS orientation using seawater as DS and (D) the AL of the
cleaned PA membrane in the AL-FS orientation using seawater DS. .......................... 56
x
Figure 17. Water flux profile as a function of water recovery at different Am/VP ratios
with and without FS stirring during the pre-concentration of nutrients in sludge
centrate by FO. Seawater was used as the DS. ............................................................. 62
Figure 18. Effects of stirring up the FS on sludge centrate pH and nutrients enrichment
during the filtration. The theoretical values of phosphate and ammonium concentration
factor were calculated assuming 100% nutrient retention by the FO membrane. The
Am/VP ratio in this experiment was 175 m-1. Seawater was used as the DS. ................ 63
Figure 19. Phosphate and ammonium rejection towards different membrane areas and
stirring conditions at water recovery of 70%. Seawater was used as the DS ............... 64
Figure 20. Correlation between duration of filtration and pH of the FS towards
different experimental conditions at water recovery of 70%. Raw seawater was used as
the DS. ........................................................................................................................... 65
Figure 21. Changes in water flux at different ratios of Am/VP without FS stirring when
seawater buffered at pH 4.4 was used as the DS. ......................................................... 66
Figure 22. The performance of FO for enrichment of nutrients in sludge centrate
without FS stirring at different ratios of Am/VP. The theoretical values of phosphate
and ammonia concentration factor were calculated assuming 100% nutrients retention
by the FO membrane. Buffered seawater was used as the DS. .................................... 67
Figure 23. Differences in mass balance of phosphate between with and without FS
stirring, as well as between using raw seawater and buffered seawater as the DSs. The
used ratio of Am/VP was 175 m-1. .................................................................................. 68
LIST OF TABLES
Table 1. Comparison between CTA and TFC-PA FO membranes regarding
performance parameters and stability ........................................................................... 15
Table 2. Typical non-responsive draw solutes for FO applications ............................. 19
Table 3. Typical responsive draw solutes for FO applications ..................................... 20
Table 4. Advantages and disadvantages of different membrane module configurations
used for water applications ........................................................................................... 22
Table 5. Properties of some typical colloids in membrane fouling .............................. 25
Table 6. Characteristics of typical scalants ................................................................... 27
Table 7. Properties of several typical bio-foulants in FO operation ............................. 28
Table 8. FO pretreatment technologies in fouling control ........................................... 35
xi
Table 9. Several modification techniques of the FO membrane surface to mitigate the
membrane fouling ......................................................................................................... 37
Table 10. Several typical cleaning chemicals according to the type of foulant. .......... 40
Table 11. Key properties of the AL of the FO membranes [100, 101]......................... 44
Table 12. Composition of seawater and sludge centrate (Values indicated average ±
standard deviation of three samples). ........................................................................... 45
Table 13. Composition of seawater and sludge centrate (values indicated average ±
standard deviation of at least five samples). ................................................................. 61
xii
ACRONYMS
AL Active layer
AHL Acyl homoserine lactone
ATP Adenosine triphosphate
CECP Concentrative external concentration polarization
CICP Concentrative internal concentration polarization
COD Chemical oxygen demand
CP Concentration polarization
CTA Cellulose triacetate
DECP Dilutive external concentration polarization
DI Deionized
DICP Dilutive internal concentration polarization
DS Draw solution
ECP External concentration polarization
EDTA Ethylenediaminetetraacetic acid
EPSs Extracellular polymeric substances
FO Forward osmosis
FS Feed solution
HTI Hydration technology innovation
ICP Internal concentration polarization
IEP Isoelectric point
NF Nanofiltration
PA Polyamide
PRO Pressure retarded osmosis
PZC Point of zero charge
RO Reverse osmosis
xiii
SEM Scanning electron microscopy
SL Supporting layer
TFC Thin film composite
WWTP Wastewater treatment plant
1
CHAPTER 1. INTRODUCTION
2
1.1. Background
In wastewater treatment facilities, anaerobic treatment is extensively employed to assimilate
biosolids released during wastewater treatment. These processes aim to decrease the volume
of the solid waste, thereby reducing transportation and disposal costs. Moreover, anaerobic
digestion of sewage sludge can illustrate the enormous potential of energy recovery via biogas
production. Biogas can be burnt to generate heat and electricity so as to compensate the
amount of consumed energy in wastewater treatment plants (WWTPs). Ultimate products of
anaerobic digestion processes are composed of biogas and anaerobically digested sludge
which is afterwards dewatered to separately create biosolids and a liquid fraction called
sludge centrate. Biosolids can be considered as a source of fertilizer for agriculture production
or be discarded, while sludge centrate is returned to the head of work and mixed with the
influent raw sewage to be treated [1, 2]. Sludge centrate is nutrient-rich wastewater containing
>150 mg/L orthophosphate, >1100 mg/L ammonia and around 400 mg/L biological oxygen
demand [3]. The recirculation of raw sludge centrate results in negative impacts on biological
treatment processes and water quality of the final effluent due to the significant addition of
organic and nutrient load. Among methods to overcome the challenges posed by sludge
centrate, pre-concentration of sludge centrate can be a promising approach. The reduced
volume of sludge centrate with high content of organic matter and nutrients such as ammonia
and phosphorus after pre-concentrated can lay the foundation of subsequent nutrient and
energy recovery.
Phosphorus is an essential fertilizer ingredient for agriculture production [4, 5]. These days,
the depletion of phosphorus reservation due to the over-exploitation of minable phosphate
rocks for agricultural production has threaten food security [6]. Phosphorus preservation is
particularly important in Australia, where the soils are mostly phosphorous deficient [7].
Besides, the excess of phosphorus in aquatic environment can cause ecological disasters such
as eutrophication and bloom algae [8]. Thus, recycling and recovery of phosphorus from
nutrient-rich sources such as municipal wastewater, urine and especially from sludge centrate
to produce fertilizers can be promising approaches to ensure the sustainable development of
agriculture and the environmental protection. Furthermore, nutrient recovery from sludge
centrate can help WWTPs reduce the treatment costs and comply with the regulation of
stringent phosphorus and ammonia concentrations in the effluent.
Nutrients such as nitrogen and phosphorus can be reportedly extracted from nutrient-rich
sources under the forms of precipitation, namely calcium phosphate (Ca3(PO4)2) and struvite
3
(MgNH4PO4) [1, 9]. Recently, membrane technologies have exhibited as perspective
candidates to pre-concentrate nutrients prior to phosphate mineral precipitation [1, 9].
Nanofiltration (NF) and reverse osmosis (RO) have been effectively applied for nutrient
retention from wastewater such as urine and sludge centrate [10, 11]. However, these
processes have a high membrane fouling tendency and consume much energy [12]. Compared
to pressure-driven membrane technologies such as RO, NF used to recover nutrients from
nutrient-rich sources, forward osmosis (FO) filtration is supposed to overcome the challenges
in terms of operational cost, energy consumption and membrane fouling [13]. The high
potential of using FO for nutrient recovery from treated municipal wastewater and sludge
centrate has been reported in some previous studies [1, 9, 13-16]. The high rate of nutrient
recovery from the effluent of a membrane bioreactor using a flat sheet FO membrane was
reported [13, 16]. A combined FO/RO membrane system using NaCl solution as a draw
solution exhibited the effective production of clean water from anaerobic digester centrate
[14]. A hybrid FO/MD membrane system using MgCl2 solution as a draw solution was
developed to effectively extract phosphorus and clean water from sludge centrate [9].
However, these systems require an extra system to re-generate draw solutes because both
NaCl and MgCl2 are costly. The requirement of draw solute regeneration system is one of
obstacles to full scale deployment of FO. In this case, seawater, which is cheap and readily
available in coastal areas is considered as a potential draw solution (DS). The diluted seawater
after FO filtration can be discharged into the ocean. Thus, regeneration of DS is unnecessary.
A seawater-driven FO process was applied to directly extract phosphorus via calcium
phosphate precipitation from sludge centrate. However, the content of phosphorus inside
obtained precipitate was low [1].
1.2. Objectives and scope of thesis
This research thesis is conducted to comprehensively evaluate a seawater-driven FO system
for resource recovery from sludge centrate. The feasibility of this system is considered based
on the efficiency of organic matter and nutrients enrichment and the magnitude of decline in
water flux of FO membranes as well as flux reversibility after physical cleaning. Expected
results of this thesis thus entail:
• Demonstrating the potential of using a seawater-driven FO process for organic matter and
nutrients enrichment in sludge centrate and identifying possible solutions to address key
challenges associated with the seawater-driven FO process for organic matter and nutrients
enrichment.
4
• Comprehensively evaluating factors affecting the stable performance of a seawater-driven
FO process for pre-concentrating organic matter and nutrients in sludge centrate for
subsequent resource recovery. These factors include membrane material properties,
prewetting protocols, membrane orientation, DS and operating parameters such as stirring
conditions and effective membrane area.
• Characterizing membrane fouling during operation of a seawater-driven FO system for
pre-concentration of sludge centrate to understand more about the nature and mechanism of
fouling phenomenon.
• Evaluating efficiency of fouling preventative methods to find out a feasible cleaning
method as well as assessing impacts of cleaning techniques on the performance of the system.
1.3. Thesis outline
The thesis outline is schematically described in Figure 1. This thesis consists of five chapters.
Chapter 1 presents the introduction. Chapter 2 – Literature review summarizes fundamentals
of FO and membrane fouling. Chapter 3 investigates effects of membrane materials,
membrane orientation, prewetting protocols and DS on the performance of seawater-driven
FO for organic matter enrichment in sludge centrate. The performance of the system is
evaluated through the efficiency of chemical oxygen demand enrichment, membrane fouling
and flux reversibility after physical cleaning. In addition, the potential and challenges of using
seawater for pre-concentrating organic materials in sludge centrate are highlighted via a
comparison between seawater and analytical grade NaCl as DSs. In Chapter 4, key challenges
associated with seawater-driven FO for enriching nutrients in sludge centrate are identified
and potential solutions are proposed to address these challenges. Furthermore, effects of
operating conditions such as stirring and effective membrane area on the efficiency of nutrient
enrichment and membrane fouling are examined and optimized for the first time. Chapter 5
summarizes key findings of this research thesis before suggesting further work in the future.
5
Figure 1. The outline of thesis
6
CHAPTER 2. LITERATURE REVIEW
7
2.1. Forward osmosis
2.1.1. Osmotic process
Forward osmosis has attracted significant attention both academically and commercially
because of its potentials in water purification and nutrient recovery [17]. Osmosis is the
transport of a solvent (normally water) across a semi-permeable membrane from a solution of
high water chemical potential referred to as a feed solution (lower osmotic pressure) to
another solution of low water chemical potential referred to as a draw solution (high osmotic
pressure) [18]. This semi-permeable membrane selectively allows for water permeation, while
restricting the passage of dissolved solutes [18]. Several recent studies have demonstrated the
potentials of FO for the treatment of difficult wastewaters such as fracking fluids [19], landfill
leachate [18], sludge centrate [1, 18] and for resource recovery from these waste solutions [1,
20, 21].
2.1.2. Transport phenomena
2.1.2.1. Mechanisms of mass transfer
In FO, the transport of water, contaminants, and draw solutes is governed by diffusion [17,
18, 22]. Due to the chemical potential difference of water between the feed and DS, water is
diffused from the feed solution (FS) (low concentration solution) to the DS (high
concentration solution) [18]. In practice, FO membranes are not completely impermeable to
all dissolved molecules, thus, some contaminants can permeate through the FO membrane to
the DS [20]. Similarly, the diffusion of draw solutes from the DS to the FS can also occur
[23].
The structure of FO membranes resembles that of RO membranes but without the backing
layer. Most RO and FO membranes are from an advanced membrane generation called thin-
film composite (TFC) membranes [17, 18]. These membranes are composed of a dense
ultrathin active layer and a porous supporting layer as illustrated in Figure 2 [24]. The dense
ultrathin active layer is responsible for separation and is made of polyamide. The porous
support layer is made of polysulfone and provides mechanical support to the FO membrane
[17]. Unlike RO membranes, FO membranes do not have a backing layer because they are
operated at a low hydraulic pressure and this backing layer can exacerbate the concentration
polarization during FO operation [17].
8
Figure 2. Structure of an FO membrane.
In addition to the diffusion of water through the membrane, there are several other mass
transfer phenomena, namely concentration polarization (CP) and reverse solute diffusion [17,
18]. CP refers to the high concentration of solutes at the membrane surface compared to that
in the bulk solution [25]. The reason for CP is that during the membrane separation processes,
draw solutes are retained on the membrane surface. Moreover, the forward solute diffusion
rate from the bulk solution to the membrane surface is significantly higher than that in the
reverse direction [25]. Thus, draw solutes become increasingly accumulated on the membrane
surface, which results in increase in the concentration of draw solutes near the FO membrane
surface [25]. In the FO process, there are two types of CP, namely external CP (ECP) and
internal CP (ICP) [17, 18]. Reduction of permeate flux, membrane fouling aggravation, and
decrease in membrane rejection can be significant consequences of CP phenomena on FO
membrane performance [17, 18, 26, 27]. In addition to CP, reverse draw solute diffusion
should also be taken into consideration during the FO process because this phenomenon
results in harmful effects on FO performance such as the salt accumulation in the FS, and loss
of draw solutes in the DS [17, 20, 22, 27].
2.1.2.2. External concentration polarization
The ECP phenomenon in FO is similar to that encountered in the RO process [18]. In RO, CP
implies the build-up of solute at the surface of the membrane dense active layer under an
applied pressure, leading to higher concentration of solute in the bulk feed compared to solute
concentration on the membrane surface [18, 25]. Similarly, during osmotic-driven membrane
processes, ECP phenomena can occur on both sides of the membrane without dependence on
membrane orientation, as described in Figure 3 [18]. In Figure 3, Cb is the concentration of
the bulk feed; Cal and Csl are the concentrations of draw solute on the active layer and
supporting layer, respectively; Ci is the solute concentration at the active layer – supporting
9
layer interface; Cds is the concentration of DS [18]. When the FS flows on, either the active
layer (AL) (Figure 3a, AL-FS orientation), or the supporting layer (SL) of the FO membrane
(Figure 3b, AL-DS orientation), draw solutes can accumulate and be concentrated near the
surface of membrane layers [18]. It is called concentrative ECP (CECP) [18]. On the other
hand, the solute concentration near the membrane surface on the draw side is diluted because
of the permeate flux [18]. This is called dilutive ECP (DECP) [18].
Figure 3. (a) ECP phenomena and dilutive ICP in AL-FS orientation and (b) ECP phenomena
and concentrative ICP in AL-DS orientation in the FO process.
Both CECP and DECP can cause negative impacts on the FO membrane performance [18].
Firstly, the accumulation of solutes in CECP increases the osmotic pressure of FS near the
AL, resulting in reducing the osmotic pressure difference across the membrane. This outcome
also results from the dilution of DS near the permeate side in DECP. Thus, both CECP and
DECP contribute to the reduction of the effective osmotic driving force which causes the
water flux decline [18]. Secondly, similar to scaling fouling in RO, inorganic scaling can
occur in the FO membrane process due to either the deposition and growth of salt
crystallization, or precipitation of sparingly salt ions on the membrane surface [26]. The
extent of scaling-fouling is greatly dependent on the concentration of solutes [26]. According
to the thermodynamic theory, solutes will commence crystalizing when their concentration is
higher than the saturation limit, or metal ions in the FS can be able to precipitate if the product
of ions is greater than the soluble product [26]. Thus, ECP with the high increasing solute
concentrations near the membrane surface can contribute to the enhancement of scaling-
fouling. Finally, in FO mode with the occurrence of ECP, the solution on the permeate side
10
near the active-support layer interface is less diluted due to flux decline [18]. This can be a
favorable condition for the diffusion of solutes from the DS to the SL, thus aggravating ICP
[17]. Consequently, solutes can permeate through the membrane to the FS, leading to worse
reverse solute flux.
The above adverse effects of ECP on osmotic-driven membrane process can be tackled by
several solutions. General principle to reduce ECP is to abolish the accumulation of solutes on
the membrane surface. The first solution is that increase in cross-flow velocities and
turbulence can minimize ECP because greater cross-flow velocity causes greater shear force
on the membrane surface, resulting in lower accumulation of solutes [18, 26]. Secondly, a
study indicated that using spacers, especially in wastewater applications, could alleviate ECP
as well as fouling. The inserted spacer in the feed channel can promote mixing, the effective
cross-flow velocity, turbulence and mass transport near the membrane surface [17, 26].
However, in most flux models for FO, ECP is assumed to be neglected because of low fluxes
and a high mass transfer [17].
2.1.2.3. Internal concentration polarization
The occurrence of ICP is exclusive to FO because of typical characteristics of the FO process
[17]. In general, ICP is a type of CP phenomena taking place within the porous SL of the TFC
FO membrane [17]. In the FO process, there are always two solutions of different
concentration flowing on the both sides of the membrane. Moreover, the hydraulic pressure
and water flux are low in FO processes [17]. Therefore, ICP always occurs that is independent
of membrane orientation. Different membrane orientations will establish different ICP
phenomena. If the porous SL of the FO membrane faces the DS in AL-FS orientation (Figure
3a), the polarized layer of draw solutes will be diluted by water flux from the FS across the
membrane [18]. This case can be called as dilutive ICP (DICP) [18]. On the other hand, when
the FS streams on the SL in AL-DS orientation (Figure 3b), and water permeates the AL,
draw solutes within the porous layer will be concentrated [18]. This is referred to as
concentrative ICP (CICP) [18]. Both DICP and CICP phenomena occur within the membrane
porous SL, so they cannot be mitigated by conventional hydrodynamic conditions [18].
The detrimental impacts of ICP on the FO membrane process are reportedly more severe than
that of ECP in terms of water flux decline, increase in membrane clogging phenomena and
reverse draw solute diffusion [17, 18, 26]. First of all, either the concentrated draw solute flow
in AL-DS orientation, or the diluted DS in the SL in AL-FS orientation due to ICP leads to
the reduction of the discrepancy of osmotic pressure across the FO membrane [17, 18, 26].
11
This causes significant loss of the driving force of the mass transfer, which results in water
flux decline [17]. Secondly, the membrane fouling deriving from ICP is similar to that from
ECP, except that it happens within the porous layer behaving as an unstirred layer, and
therefore, is almost irreversible [18]. This phenomenon causes increase in overall membrane
resistance for water transport [26]. Thus, the ultimate consequence of ICP is still a decrease of
water flux. In the end, ICP can exacerbate the reverse solute flux [17]. In AL-FS orientation,
the accumulation of draw solutes within the porous layer leads to significant increase in the
difference of solute concentration across the membrane. This enhances the driving force of
reverse solute diffusion. In addition, CICP occurs in AL-DS orientation that can aggravate
DECP [17]. Thus, the possibility of reverse solute transport can be higher because of the
higher solute concentration near the active layer on the draw side.
The modifications in fabrication of FO membrane, along with optimizing operating conditions
and changes in membrane orientation and materials can be applied to alleviate ICP. The
extent of ICP in FO processes is associated with the degree of the resistance to solute
diffusion (K) within the membrane porous SL (Equation 1) [17, 18]. A larger value of K
implies more severe ICP [18]. Thus, by modifying the structural parameter to reduce the K
value, ICP can be minimized. As can be seen from Equation 1 that a thinner, more porous and
less tortuous SL is indicative of decrease of K, resulting in the less ICP. In addition, the pore
clogging can result from the deposition of the foulants into the membrane porous SL [26].
This contributes to the reduced porosity of the SL, which enhances ICP. This phenomenon
can be deteriorated in AL-DS orientation when the FS containing many different
contaminants flows on the porous SL. Therefore, the selection of appropriate membrane
orientation and hydrodynamic adjustments such as increase in cross-flow velocities and more
frequent flushing can help diminish the foulant accumulation, so alleviate ICP [17, 18, 26].
Moreover, temperature is referred as an important operating condition in FO processes
because of its effects on the diffusion coefficient of the draw solute (Ds) [26]. The increase in
temperature can lead to the intensification of the solute diffusivity because of the reduced
viscosity of the solution [26]. As a result, ICP can be alleviated. Lastly, the membrane
materials can influence the degree of ICP [17]. Hydrophilic SLs can improve water flux and
decrease ICP because of increasing the wetting of small pores within the SL [17].
12
Equation 1
s
KD
=
In which K is the resistance of solute diffusion; ℓ, τ and ε are the membrane thickness,
tortuosity and porosity, respectively; Ds is the diffusion coefficient of the solute [17, 18].
2.1.2.4. Reverse solute flux
Reverse solute leakage is a unique phenomenon which can be mathematically quantified in
the FO process. In fact, no FO membranes can remove all solutes completely [27]. Thus, both
forward and reverse solute diffusion can simultaneously occur in the FO process. The reverse
permeation of solutes from the DS into the FS is called as reverse solute diffusion [17, 27]. In
contrast, the transport of dissolved solute through the membrane from the feed side to the
draw side is referred as the forward solute diffusion [17]. The low hydraulic pressure and
water flux in FO process can facilitate these phenomena. To determine the extent of reverse
permeation of solute, Js/Jw ratio is introduced to depict the amount of draw solute reversely
diffusing into the FS per unit volume of water flux into the DS (Equation 2) [17, 26, 28]. It
can be seen clearly from Equation 2 that the ratio is constant for a certain membrane, and only
dependent on temperature and the membrane transport properties of the AL, namely A and B
values. A membrane with the superior selectivity which means a higher value of A and a
lower value of B can mitigate the reverse solute flux. Theoretically, Js/Jw is independent of the
properties of the SL, membrane orientation, operating conditions, and hydrodynamic
conditions [17, 26, 28]. In practice, however, the membrane separation properties can vary
with the high concentration of the DS in contact with the dense AL [26, 29], and various types
of solution [26].
Equation 2
w
s
g
J B
J A R T=
Where Js is the reverse solute flux; Jw is the permeate flux; A is the pure water permeability
constant; B is the solute permeability constant; β is the Van’t Hoff coefficient; Rg is the ideal
gas constant; T is the absolute temperature.
The solute in the draw side diffusing into the feed side can lead to disadvantageous effects on
the FO process. The first adverse impact is that the reverse solute leakage causes the loss of
draw solute in the DS [17, 28]. This leads to the reduced osmotic pressure of the DS, and the
13
enhanced osmotic pressure of the FS [17]. Thus, the osmotic driving force of mass transport
will decrease significantly, which results in the acceleration of water flux decline. To solve
this problem, the replenishment of draw solute to maintain the effective osmotic pressure of
the DS can be necessary [27]. This increases the cost of treatment process [17, 27]. Secondly,
the accumulation of draw solute in the FS can harm microorganisms in osmotic membrane
bioreactors because these species are susceptible to the high salt concentration [17]. When
using some types of DS such as NH3-CO2 solution, the discharge of the FS after accumulating
the draw solute can lead to damage of the environment, namely the eutrophication
phenomenon [28]. Thirdly, the reverse draw solute permeation can be responsible for the
enhancement of CP in the FO process [26]. The deteriorated CECP and CICP will occurs with
increase in the concentration of draw solute near the AL and the porous layer due to the
reverse solute leakage in AL-FS orientation and AL-DS orientation, respectively. Finally,
reverse solute diffusion not only can aggravate CP, but also may contribute to changes in
chemical composition of the FS, which may detrimentally affect the membrane fouling
behavior [26]. The draw solutes such as divalent cations can play a role as fouling promoters
[26]. The appearance of these solutes in the FS can exacerbate the organic and colloidal
fouling which results from the deposition of colloidal particles on the membrane [26].
Moreover, some draw solutes, namely Ca2+, Mg2+, SO42-, CO3
2- and PO43- are considered as
scaling precursors [26]. Thus, the accelerated inorganic fouling in the feed side can happen
due to reverse diffusion of these scaling precursor ions [26]. In many cases, the draw solutes
can behave as nutrients, and the reverse diffusion of these solutes can lead to the biofilm
development which causes biofouling [26].
Reverse solute diffusion can be minimized by changing the membrane intrinsic properties,
appropriate draw solute selection, and hydrodynamic conditions. At first, as mentioned above,
the extent of reverse draw solute diffusion is dependent on the membrane intrinsic separation
properties [17, 26]. A FO membrane with a dense AL of higher water permeability and lower
solute permeability can result in less solute diffusing from the DS into the FS [26]. Thus,
some advanced modifications of the AL such as sealing defects which leads to the reduced B
value will help alleviate the reverse solute leakage. Another factor affecting the rate of reverse
solute leakage is the DS composition [17, 26]. In FO process, the high concentration of DS is
vital for the osmotic driving force of water transport across the membrane [17, 18]. However,
higher DS concentration can lead to greater rate of reverse draw solute diffusion [26]. Thus,
an appropriate value of DS concentration to ensure the harmony between water flux and
14
reverse solute flux should be taken into consideration. In addition to the DS concentration, the
draw solute type can also influence the draw solute reverse diffusion noticeably [17, 26]. A
study indicated that using monovalent ions as draw solutes demonstrates the higher solute
leakage despite releasing higher permeate fluxes compared to using bivalent ions [17]. The
reason is that the size exclusion and electrostatic repulsion of monovalent ions is low [17].
Meanwhile, bivalent ions illustrate lower reverse solute fluxes because of their higher steric
hindrance and electrostatic repulsion [17]. At last, it is reported that the reverse solute
diffusion can be deteriorated by CP [17]. Thus, some adjustments of hydrodynamic conditions
such as increase in cross-flow velocities to decrease CP can lead to the reduced reverse solute
diffusion.
2.1.3. Materials
FO membranes are similar to those used in RO applications regarding materials and
fabrication methods [24]. Currently, most FO and RO membranes are commercially made of
two typical polymeric materials including cellulose triacetate (CTA) and polyamide (PA)
[24]. Particularly, PA FO and RO membranes are made on a support of polysulfone. The
methods of FO membrane fabrication are dependent on the specific materials. CTA FO
membranes are fabricated by phase inversion, followed by heat annealing treatment [24]. In
this method, the characteristics of these CTA FO membranes can be significantly affected by
polymer concentration, type of solvent, evaporation time, annealing temperature, casting
substrate and coagulant bath [24]. For example, a more dense CTA FO membrane can result
from using higher concentration of polymeric solution [25]. When it comes to PA TFC FO
membranes, these membranes are composed of a porous SL and a selective AL [24]. The
porous SL made of polysulfone is fabricated via phase inversion with/without a thin non-
woven fabric [24]. Meanwhile, the dense AL made of crosslinked aromatic PA are generally
synthesized via interfacial polymerization reaction between meta phenylenediamine monomer
in water and trimesoyl chloride monomer in organic solvent [24]. In this process, factors such
as concentrations of monomers and solvents, types of monomers and solvents, and
polymerization conditions need to be optimized to create high performance PA TFC
membranes [24].
Membrane materials can directly influence the FO performance in terms of separation
properties, permeability, fouling tendency and chemical and biological stability. Firstly, using
FO membranes with superior intrinsic separation properties which are indicative of higher
water permeability and selectivity can lead to greater water flux and minimize the negative
15
impacts of draw solute leakage [26, 28]. When utilizing these membranes, it is more
compatible with using small and fast-diffusing draw solutes such as sodium chloride,
magnesium chloride and ammonia-carbon dioxide [28]. The reason is that these draw solutes
can establish a high osmotic driving force and alleviate ICP, which results in higher water
flux, but they are also susceptible to draw solute diffusion due to their small size [28].
Secondly, the dense AL of smaller permeability and lower rejection can cause the enhanced
forward and reverse draw solute diffusion. As a consequence, this may deteriorate the risk of
fouling as mentioned above [26]. A study indicated that hydrophobic membranes are more
vulnerable to adsorptive fouling than hydrophilic membranes such as CTA because of higher
repulsive acid-base interaction that deters the accumulation of colloidal particles on the
membrane surface [17, 26]. Also, the chemical and biological endurance of FO membranes
can be noticeably impacted by used materials [17, 26]. This depends on the interaction
between chemical components of the FO membrane material and contaminants in wastewater
[17]. Table 1 compares the performance and properties of TFC PA and CTA membranes [17,
18, 26, 28].
Table 1. Comparison between CTA and TFC-PA FO membranes regarding performance
parameters and stability.
Parameters CTA TFC-PA
Water permeability Low High
Rejection Low High
Membrane fouling High fouling propensity Low fouling tendency
Chemical stability
A narrow range of pH
tolerance (3-8)
Highly resistant to chlorine
Stability at broad pH values (2-12)
Sensitive to chlorine
Biological stability High biodegradable tendency Low biodegradable propensity
2.1.4. Forward osmosis operation
2.1.4.1. Temperature
The feed and draw solution temperature and transmembrane temperature difference can
significantly affect the performance of FO process [30, 31]. Regarding membrane properties,
an increase in the FS temperature can increase the solute diffusivity and decrease the viscosity
of water, thereby raising the pure water and solute permeability coefficients (A and B
coefficients, respectively) of the FO membrane [31]. In addition, the membrane polymeric
matrix can be deformed at excessively elevated temperature [17]. When it comes to water and
16
reverse solute fluxes, a rise in the DS temperature can lead to decrease in DS viscosity and
increase in osmotic pressure as well as the draw solute diffusivity [30-32]. Thus, water and
reverse salt fluxes can be enhanced with the high temperature of the DS [30, 31]. This is
consistent with the increase in the A and B values of the FO membrane mentioned above.
When the FS temperature increases, the FS viscosity decreases, thereby enhancing the
diffusivity of water through the membrane [30]. Moreover, the improved mass transfer
coefficient of the FS can result from the increase in the diffusion coefficient due to high
temperature, resulting in alleviating ECP [30, 31]. Notably, the temperature difference
between the feed and draw solution can cause certain effects on the rejection of neutral
contaminants [31]. When the temperature of the FS is higher than that of the DS, the
decreased rejection of neutral contaminants can be attributed to an increase in their diffusivity
at an elevated temperature [31]. By contrast, the rejection of neutral pollutants can be
improved when the DS temperature is higher than the FS temperature [31]. This is because
the increased reverse salt flux which is induced by an increase in the draw solute diffusivity
can hinder the forward diffusion of contaminants [31]. Another reason is the increase in water
flux can directly account for an increase in rejection, which is similar to that observed in
nanofiltration and reverse osmosis processes [31].
2.1.4.2. Feed solution pH
The FS pH can influence water flux, reverse salt flux in the FO process through its effects on
FO membrane structure [20, 26, 33, 34]. Firstly, the increased water flux can be a result of
increasing the FS pH [33, 34]. This phenomenon can be attributed to changes in cross-linked
membrane polymer structure and membrane wettability [33]. It is assumed that an increase in
the FS pH can accelerate the electrostatic repulsion among ionizable functional groups of the
membrane polymeric structure, leading to an increased average pore size and higher permeate
flux [33]. Moreover, more dissociation of carboxyl functional group (COO-) of the AL can
occur at higher FS pH, so the FO membrane surface can become more hydrophilic [33, 35]. A
more hydrophilic membrane can facilitate water transport [33]. Secondly, the reverse salt flux
can decrease as the FS pH increases [33]. An increase in the FS pH can not only cause a rise
in water flux, but also make the FO membrane more negatively charged [33]. Both of these
consequences can suppress the salt flux [33]. Finally, the FO membrane damage can occur at
either excessively low or high pH values because they can break peptide bonds in polymer
chains of the membrane material such as polyamide [17, 18]. As a consequence, the
membrane separation properties such as permeability and selectivity can be affected.
17
2.1.4.3. Cross-flow velocities and spacer designs
The cross-flow velocity and spacer design can significantly impact on the level of CP and
mass transfer near the membrane surface in the FO process [26]. Increased cross-flow
velocities can encourage the back diffusion of those solutes and particles as well as enhance
mas transfer coefficient, resulting in mitigating the accumulation of retained solutes and
colloids near the membrane surface [26]. As a result, CP and membrane fouling, especially
external fouling can be significantly alleviated [26]. In practice, increase in cross-flow
velocities aims to reduce the membrane fouling, not to improve water flux [18]. However, the
intensified cross-flow velocities are mostly beneficial for the reduction of external fouling and
ECP, but may not be effective to reduce internal fouling and ICP occurring within the SL
[26]. In addition to the increase of cross-flow velocities, the use of spacers can also affect
cross-flow conditions [17]. In particular, spacers are necessary in spiral-wound membrane
configurations [36]. They are alternately positioned among membrane flat sheets to form the
flow channels [36]. Therefore, spacers can be utilized in both the feed and permeate channel
[36]. Commercial spacer consists of two layers of cylindrical filaments creating a ladder-type
or diamond-type spacer geometry (Figure 4) [36]. In operation, these spaces can cause the
enhanced turbulence of liquid streams, which can encourage mixing and improve mass
transfer [17]. This can help to restrict the formation of the membrane boundary layer and CP,
especially ECP [17]. Thus, membrane fouling can be also alleviated, and water flux can be
enhanced.
Figure 4. Geometric structure of ladder (a) and diamond (b) configurations of spacer
Top layer
Bottom layer
Top layer
Bottom layer
(a) (b)
18
2.1.4.4. Draw solution
Draw solution is a highly concentrated solution which has low chemical potential of water on
the permeate side of the FO membrane to induce the transport of water through the membrane
[18, 37]. A suitable DS needs to satisfy the following requirements [17, 18, 37, 38]: (i)
generating higher osmotic pressure compared to the FS; (ii) causing no damage to the
membrane structure of the active and support layers; (iii) minimal reverse diffusion of the
draw solute; (iv) chemical stability for re-use; (v) non-toxicity with no harmful impacts on
human and the environment; (vi) low cost; (vii) easy re-concentration of the diluted DS; (viii)
low viscosity and small molecular weight. Herein commonly-used types of draw solute will
be characterized and categorized into non-responsive and responsive solutes.
Non-responsive draw solutes including inorganic salts, non-functionalized polymers and
organic molecules, and seawater and RO brine indicate draw solutes that cannot change their
properties with external stimulants such as pH, temperature, electro-magnetic field or light
[17, 37, 38] (Table 2). Responsive draw solutes can include nanoparticles, hydrogels,
metathesis precipitable salts, soluble gases and volatile liquids, NH3-CO2 draw solutes, highly
charged compounds and polarity-switchable draw solutes. They respond to external stimuli
(Table 3) [17, 37, 38]. In practical, the most common draw solute is sodium chloride (NaCl)
because it has high solubility, creates high osmotic pressure and is somewhat simple to re-
concentrate to high concentration with RO without risk of scaling [17, 26].
The dilution of DS, the accumulation of dissolved contaminants in the DS and reverse draw
solute leakage can simultaneously occur during the FO operation [17, 18, 20, 26]. In the FO
process, water is continually transferred from the feed side to the draw side over time, thereby
leading to the dilution of DS [18]. This can diminish the osmotic driving force of mass
transfer. Thus, a re-concentration system to extract water from the diluted DS is required [20,
26]. Currently, draw solutes can be re-concentrated by hybrid membrane systems combining
the FO membrane and other types of membrane such as RO membranes and membrane
distillation membranes [20]. In addition to the water transfer through the FO membrane,
pollutants especially trace organic compounds can also pass through the membrane and enter
the DS because FO membranes are not completely impermeable to all dissolved solutes [20].
The accumulation of these contaminants can result in the degradation of product water and
even membrane fouling in the draw solute recovery system [20, 26]. Therefore, the removal
of these contaminants should be considered. Meanwhile, the reverse diffusion of draw solute
from the draw side to the feed side can cause the deficit of draw solute which reduces the
19
osmotic pressure of the DS [17, 26]. Thus, a replenishment of draw solute needs to be
periodically conducted to ensure the necessary osmotic driving force of mass transfer [17, 20].
Table 2. Typical non-responsive draw solutes for FO applications
Draw
solute Examples Advantages Disadvantages
Non-r
espo
nsi
ve
solu
tes
Inorganic salts
(e.g. NaCl, MgCl2,
Na2SO4, (NH4)2SO4,
Ca(NO3)2)
• High solubility
• Low cost
• High osmotic
pressure
potential
• Sensitive to scaling/clogging
• Higher reverse draw solute flux
with small draw solutes
• Difficult to recover in the re-
concentration system
Polymers/ macro-
organic molecules (e.g.
poly-sodium-acrylate,
glycine)
• Reduce reverse
diffusion
• High water flux
• Aggravate CP
• Difficulties in regeneration
process and circulation
• Limited storage time due to
biodegradation
Seawater and RO brine • Abundant source • Sensitive to fouling
• May need to be pre-treated
20
Table 3. Typical responsive draw solutes for FO applications
Draw
solute Examples Advantages Disadvantages
Res
ponsi
ve
dra
w s
olu
tes
Nanoparticles such as Fe3O4
• Very low reverse draw solute flux
• High osmotic pressure potential at low
concentration
• Easy to re-concentrate using a magnet
• Agglomeration during magnetic separation
• Ultrasonication weakens magnetic properties
• Viscosity of solution diminishes effective
driving force and the flux
Hydrogels • No draw solute reverse leakage • Low water flux
• Poor water recovery
Metathesis precipitable salts such
as CuSO4, MgSO4, Al2(SO4)3
• Only be applicable for brackish water
treatment
• Not suitable for drinking water production
• Use of large amount of chemicals and formation
of byproducts during generation
Soluble gases and volatile liquids
such as SO2, CO2, NH3 • High osmotic pressure
• Loss of draw solutes
• Need external pressure to aid the dissolution of
the draw solute
• Complication on actual implementation
NH3-CO2 draw solutes • High solubility in water
• Recovery by moderate heat
• Toxic thermolytic product
• High diffusive loss
Polarity-switchable draw solutes
such as hydrophobic amines • High osmotic pressure
• Tend to swell and damage the AL
• Difficulty in regeneration process
Highly-charged compounds such
as EDTA
• High water flux
• Low reverse diffusion
• Less energy consumption during re-
concentration
• High cost
• pH dependency
21
2.1.4.5. Membrane orientation
Membrane orientation is indicative of the position of an FO membrane relative to the DS and
the FS [17]. As noted in section 2.1.2, the FO membrane comprises of a dense selective AL
and a porous SL [17, 18]. Thus, the FO membrane can be positioned with the AL facing either
the feed side (AL-FS or FO mode), or the draw side (AL-DS or PRO mode) [17, 18, 39].
Because of the characteristics of the AL and the SL, membrane orientation can directly
influence the performance of FO process [17].
Permeate flux and salt rejection of the FO process should be taken into comprehensive
account during the selection of membrane orientation for FO applications [17, 18, 33, 39].
Previous studies reveal that the permeate flux of the FO membrane in PRO mode is higher
than that in FO mode as the DS concentration is high. This is because CICP can be less severe
in PRO mode [17, 39]. Salt rejection in FO mode is usually higher than that in PRO mode
[33, 39]. This can be attributed to the smoother AL surface that can mitigate the possibility of
salt accumulation on the membrane surface in FO mode, thereby reducing the salt diffusion
from the feed side to the draw side [26, 33]. Moreover, the diminished turbulence of the FS
can accelerate ICP in PRO mode, resulting in the reduced salt rejection [26, 33].
2.1.4.6. Module configurations
In water purification and wastewater treatment applications, FO module configurations share
similarities with those used in RO. Module configurations imply the ways of packing or
holding membranes into a module so as to maximize the surface to volume area and reduce
particles deposition by sufficient cross-flow [17]. In practice, there are three prevalent FO
membrane module configurations including spiral-wound, hollow fiber and plate and frame
employed to meet various treatment requirements [17, 18]. Each module configuration has
different benefits and drawbacks in terms of economic considerations, the complication of
fabrication into the module design, the flexible applicability, fouling tendency and CP
phenomena (Table 4) [17, 18, 40].
22
Table 4. Advantages and disadvantages of different membrane module configurations used
for water applications
Modules Advantages Disadvantages
Spir
al-w
ound
• High packing density
• Easy cleaning of fouling
agents
• Limited pressures in membrane envelop
and clogging of spacers
• Inability to induce flow in the permeate
channels (inner side of the envelopes) of
the membrane elements for cleaning or
backwashing
Holl
ow
fib
er
• Self-supported
characteristics
• Appropriate flow
• Simplicity of fabrication
• High packing density
• Limited mixing at membrane surface
• In open channels, the CP film grows
undisturbed in the channel
Pla
te a
nd f
ram
e
• Well-suited to wastewater
applications
• Less complicated in design
• Better backwashing
• Higher cross-flow velocities
• Lack of adequate membrane supports
and problems with internal and external
sealing
• Low packing density (leads to a larger
system footprint, higher capital costs,
and higher operating costs
2.2. Fouling phenomena in the FO membrane operation
2.2.1. FO membrane fouling
Similar to RO, fouling is inevitable during FO membrane operation. Membrane fouling in the
FO process refers to the deposition and accumulation of constituents such as solutes, particles
or microorganisms on the FO membrane surface, which can deteriorate the performance of
the membrane [17, 26, 34]. Flux decline, changes in intrinsic separation properties and even
damage to the FO membrane can be consequences of membrane fouling [17, 26, 41]. It
appears that FO fouling is more reversible than that in RO due to the lack of hydraulic
pressure [17, 26].
In the FO process, membrane fouling can be classified into four groups in terms of foulant
types [26, 29]. Colloidal fouling is indicative of the deposition of colloids such as silica and
hydroxides on the membrane surface [26, 42]. Organic fouling refers to the deposition and
23
adsorption of organic macromolecules such as polysaccharides, humic acid and protein [26,
42]. In addition, once the precipitation or crystallization of dissolved inorganic substances
takes place on the membrane, inorganic scaling can be established [26, 29]. Biofouling
implies the adhesion and accumulation of microorganisms forming biofilm on the membrane
[26, 29]. In practice, all fouling types can occur during FO operation [17, 26].
Membrane fouling in the FO process can also be divided into surface fouling (external
fouling) and internal fouling in terms of the fouling places [26, 29]. Both types of fouling can
be likely to occur on the dense AL and within the porous SL at the same time. In FO mode,
once the dense AL faces the feed side, numerous foulants in the FS can accumulate on the
surface of this layer, thereby building up a cake layer which resembles fouling in RO.
Membrane fouling in this case can be referred to as external fouling [26, 29]. In the meantime,
the possibility of membrane fouling within the porous SL can exist under the form of
inorganic fouling in DS side due to forward diffusion of fouling precursors [26]. Membrane
fouling behavior can be more complicated in PRO mode as the porous SL keeps into contact
with the feed side. If the size of foulants is small enough, they can go into the SL through
convection [26]. Some foulants can be absorbed on the pore wall [26]. Others can be finally
retained by the dense AL [26]. As a result, the pore can be filled up with foulants over time.
This phenomenon is called as internal fouling [26]. Under the real wastewater condition,
foulants still keep depositing on the outer surface of the porous layer, which can cause the
formation of external fouling after the pore is replete [26]. Thus, both external fouling and
internal fouling can be significant in PRO mode.
2.2.2. Colloidal and organic fouling
2.2.2.1. Characteristics of colloids and macromolecules
Colloids can be divided into two groups namely inorganic colloids and organic
macromolecules [26, 42]. The former include some common inorganic substances such as
silica, iron hydroxide and clay minerals [42]. The latter include rigid biopolymers (large
molecular weight polysaccharides), fulvic compounds (fulvic and humic acid), proteins and
other macro-organic molecules [42]. In FO operation, both colloidal fouling and organic
fouling can be affected by similar factors as well as explained by the same fouling
mechanisms because inorganic colloids and organic macromolecules share many similarities
in terms of their properties [26].
Colloids can be fully described through their size and conformation and their charge in the
feed water (Table 5) [26, 42]. In terms of inorganic colloids, the size and shape of inorganic
24
particles are various, and they are mostly rigid [42]. In water environment, most inorganic
colloids are negatively charged at around neutral pH [42].
Regarding organic colloids, polysaccharides are large macromolecules with average
molecular weight range from hundred to a few thousand kDa [42]. They are weakly
negatively charged [42]. One of typical polysaccharides causing fouling is alginate which has
the charge density of 3 meq/g (up to 6 meq/g) [42]. In particular, polysaccharides can exist
under as a gel-like structure (a three-dimensional cross-linked substance in the presence of
multivalent metal ions) [42]. Another organic foulant in wastewater is fulvic substances,
which are produced by biodegradation of dead organic matter [26, 42]. Fulvic substances are
abundant in wastewater, thus are a major cause of organic fouling during FO operation.
Compared to polysaccharides, the molecular weight of humic acid is lower (just from several
kDa to a few hundred kDa) [42]. This acid is highly polydispered [42]. Moreover, humic acid
is a negatively charged compound at a broad range of pH due to the deprotonation of
carboxylic and phenolic functional groups, and its charge density will increase as pH
increases [42]. Most proteins are amphoteric because of the presence of both carboxylic (-
COO-) and amine (-NH3+) groups in their chemical structure. When pH is lower than
isoelectric point pH (pHIEP), protein molecules are positively charged. By contrast, these
molecules are negatively charged once pH of the solution is higher than pHIEP. In aqueous
solution, protein are globular macromolecules.
25
Table 5. Properties of some typical colloids in membrane fouling
Colloids Examples Molar mass
(kDa) Shape Charge in water pHIEP or pHzpc
Inorganic fouling
am-SiO2 Flexible Round Negative 3-3.5
am-Al2O3 Flexible Angular Negative 7
am-FeOOH Flexible Crystalline or amorphous Positive 7.9-8.1
Polysaccharide
Alginate 200 - 2000 Extended random coil Negative
3-6 meq/g 5.4
Xanthan and Gellan 100 - 2500 Single-stranded or double-helical Negative 4.5
Schizophyllan 400 - 500 Rigid rod-like Neutral -
Humic substance
Humic acid in International
Humic Substance Society
(IHSS) standards
Of a few kDa to a
few hundred kDa
globular molecule (linear under
high pH, low ionic strength and
low concentration)
Negative
5-10 meq/g 3
Protein
Bovine serum albumin 67 Flexible Negative
1.5 meq/g 4.7
Bovine immunoglobulin G 155 Flexible Negative 6.6
Bovine hemoglobin 68 Flexible Negative 7.1
Bovine pancreas ribonuclease
A 13.7 Flexible Negative 7.8
Lysozyme 14.4 Flexible Positive 11
Note: “-” indicates not available; “am” stands for amorphous; pHpzc stands for the point of zero charge.
26
2.2.2.2. Colloidal interactions
Membrane fouling can be governed by colloidal-colloidal and colloidal-surface interactions
[26, 42]. The colloidal-surface interaction can enable colloids to approach the membrane
surface, and then deposit on it, while colloidal-colloidal interactions will allow foulants to
attach the deposited foulants and develop into a cake layer on the membrane surface [26, 42].
According to a classical theory, particle-particle and particle-surface interactions can be
dominated by synergic impacts of Van der Waals force and electrostatic interaction force
known as the electrical double layer force [42]. The sum of these forces can establish an
interaction energy called an energy barrier [26, 42]. This energy barrier must be overcome if
any particles wish to attach onto the membrane surface [26]. Once a colloidal system is stable
(i.e. it has a large energy barrier), it can be difficult for particles to affix onto membrane
surface. This interaction energy is dependent on physico-chemical parameters such as particle
size, surface charge and surface roughness as well as solution chemistry such as pH, ionic
strength and types of ions [26, 42]. For example, electrostatic interaction can decrease at high
solution ionic strength. In addition to the classical theory, a state-of-the art theory called the
Lewis acid-base theory, has been introduced to explain for the aggregation of colloids [26,
42]. In accordance with this novel approach, a colloid in aqueous solution can interact with
others through an electron-acceptor-electron-donor interaction [26]. There will be a repulsion
among hydrophilic colloids because of hydration force, or an attraction among hydrophobic
colloids [26]. In summary, both theories can be utilized to characterize the colloidal
interactions entirely [26].
2.2.3. Inorganic fouling
The occurrence and intensity of scaling is dependent on the concentration and solubility of
scalants in the solutions [26]. Scaling is the deposition of sparingly soluble inorganic salts on
the membrane surface at certain conditions [26, 41]. In general, scaling happening in the FO
process can be a consequence of CP which causes increase in the concentration of insoluble
salts as well as the possibility of precipitation [41]. When the product of concentration of
minerals is higher than the solubility product, the precipitation of these minerals will occur.
The accumulation of these crystals can lead to scale occurrence on the membrane surface.
Mineral salt crystal deposition from the bulk solution, along with the direct crystallization
onto membrane surface can form a scaling cake layer [26]. Unlike colloidal and organic
fouling that can happen at low concentration, inorganic scaling only take place as the
27
supersaturation exceeds a critical value [26]. Several typical scalants can be listed and
characterized in Table 6 [26, 41].
Table 6. Characteristics of typical scalants
Examples Size and shape Alkalinity
/Acidity
Factors affecting the
formation
Calcium sulphate
(Gypsum) Needle-like or plate-like Neutral
Concentrations of ions
and temperature
Calcium carbonate
(CaCO3)
Calcite: rhombohedral
Aragonite clusters:
outward oriented needles
Vaterite: spherical
Alkaline
Calcium hardness,
bicarbonate alkalinity,
temperature and pH
Calcium phosphate
(Ca3(PO4)2) Amorphous Alkaline
Phosphate concentration,
temperature and pH
Silicate scale Crystalline or amorphous Weakly acid trivalent metal cations and
pH
2.2.4. Biofouling
Biofouling is associated with the development of biofilm on the FO membrane surface [26,
29, 41]. During FO operation, live microbes such as bacteria, fungi and microalgae attached
on the membrane surface can utilize nutrients existing in the FS to propagate [26, 29]. Over
time, these microorganisms can quickly proliferate and form the microbe layer colonizing the
membrane surface [26, 29]. This layer can be referred to as biofilm. Basically, two essential
components of biofilm consist of microorganisms and extracellular polymeric substances
(EPSs) comprising of polysaccharides, proteins, lipid and nucleic acid [26, 29]. EPSs are
excreted from the metabolism process of microbes to protect them from the external harsh
environment as well as enhance the adhesion to the membrane surface [26, 29, 41]. Compared
to the other fouling types, biofouling is more complex due to the interaction of live matters
[29]. Characteristics of representative foulants causing biofouling can provide further insights
into the nature of biofilm (Table 7) [26, 29, 41, 43].
The occurrence and severity of biofouling can be controlled by biofouling potential which is
represented by microorganisms and nutrient concentration in the aqueous environment [26].
Each factor plays a crucial role in each stage of the development of biofilm on the membrane
surface. The biofilm development can be divided into three stages, namely induction,
28
logarithmical growth and plateau [26, 29]. The induction stage can be corresponding to the
attachment of microbes onto membrane surface [29]. Therefore, the concentration and type of
microbes can be pivotal in this stage [26, 29]. The second stage can imply the reproduction
and proliferation of microorganisms by consuming nutrients in the FS [29]. The third stage
can be indicative of detachment of bacteria due to lack of nutrients and increase of population
density [29]. Thus, the nutrient concentration and type can strongly influence the second and
third stages. To decrease biofouling, the reduction of nutrients supply is necessary [26].
Conventionally, nutrients existing in the FS can stem from three sources including pollutants
in the original bulk FS, biodegradable membrane materials and the reverse diffusion of draw
solutes as nutrients from the DS [26]. To conclude, the control of microorganisms and
nutrients in the FS can contribute to the biofouling mitigation.
Table 7. Properties of several typical bio-foulants in FO operation
Foulants Examples Size and shape Charge
Bacteria
Pseudomonas, E. coli,
Corynebacterium, B.
subtilis
Rod-shaped Neutral
Fungi Penicillium Brush-like and flask-shaped
- Trichoderma Divergent and flask-shaped
Microalgae Chlorella sorokiniana
Cell diameter ~ 5μm Negative Chlorella pyrenoidosa
Note: “-” indicates not available.
2.2.5. Factors affecting FO membrane fouling
2.2.5.1. Feedwater characteristics
The type and concentration of foulants and their physicochemical properties are pivotal
factors affecting the membrane fouling. As mentioned in Section 2.2.1, each type of foulant
will be responsible for a single fouling phenomenon. However, membrane fouling can be
deteriorated by inter-foulant-species interactions among mixed-foulants in the real FS due to
their physicochemical properties [26, 44]. The coexistence of both alginate and gypsum
foulants can result in the enhancement of membrane fouling because of the accelerated
gypsum crystal nucleation and growth in the presence of alginate [45]. Another example is
that the fouling behavior can increase in the appearance of alginate and lysozyme due to the
electrostatic interaction between the oppositely charged foulants [44]. In addition, the high
29
concentration of foulants can contribute to the enhanced extent of initial fouling [26, 29, 42].
The reason is the increased foulant concentration can enhance the frequency of foulant
collision with the membrane, thereby increasing the rate of foulant aggregation on the
membrane surface [26].
The FS chemistry such as pH, ionic strength and ionic composition can cause significant
effects on fouling behavior [26, 42]. Similar to effects of pH and ionic strength on membrane
fouling in pressure-driven membrane processes, the FO membrane fouling caused by
negatively charged alginate, or humic acid can be aggravated at low pH and high ionic
strength [26, 42]. Also, protein fouling in the FO process can be more severe at pHIEP [26,
42]. The negative impacts of pH and high ionic strength on the FO membrane fouling can be
attributed to the reduced electrostatic repulsion among foulants and that between foulants and
the membrane surface [26]. Furthermore, the membrane fouling can be affected by the
existence of divalent ions such as Ca2+ and Mg2+ in the bulk FS due to foulant-ion and
membrane-ion interactions [26, 42, 46, 47]. Multivalent cations such as Ca2+ and Mg2+ can
bind to macromolecules strongly because these ions have an affinity for -COO- groups in the
molecular structure of organic compounds such as alginate, protein and organic acids [26, 42,
46]. This can partially neutralize the negative charge of these substances, thereby facilitating
their aggregation [26, 46]. Furthermore, divalent ions such as Ca2+ can connect two
macromolecules through the bridging effect [26, 42, 46]. For example, Ca2+ can play a role as
a bridging agent which can bind to carboxyl groups in extracellular polymeric substances to
link molecules [46]. Finally, these cations can influence the charge of polyamide – based
membrane surface through their interaction with -COO- groups in the membrane structure
[26].
2.2.5.2. Hydrodynamic conditions
Initial permeate flux and cross-flow velocities can influence fouling behavior through changes
in the structure of fouling layer, the influences of CP and mass transfer near the membrane
surface [26, 42, 47-49]. In general, the more severe membrane fouling can occur at higher
initial water permeate flux [26, 47, 49]. Three assumptions can be referred to explain for this
phenomenon [26, 42, 48]. Firstly, higher water permeate flux can increase the volume of
water permeating through the membrane, thereby increasing a number of foulants towards the
membrane surface due to water convection [26, 42]. Secondly, the exacerbated fouling can be
a consequence of the enhanced CP at high initial permeate flux [26, 48]. In the end, increase
in permeate flux can lead to greater hydrodynamic drag force towards the membrane surface,
30
which can enhance the aggregation and compaction of foulants [26, 48]. In addition to initial
permeate flux, the increase of cross-flow velocity can mitigate the fouling propensity [26, 42,
48]. The reason is a high cross-flow velocity can increase the shear force and back diffusion
of foulants from the membrane surface to the bulk solution [26, 42, 48]. This can promote
mass transfer as well as reduce CP and the accumulation of colloids and salts on the
membrane surface [26, 42, 48].
Temperature can alter thermodynamic characteristics of feed and draw solutions as well as
hydrodynamic conditions, thereby causing impacts on the membrane fouling tendency [26,
50]. The increase of DS temperature which lead to the increased the osmotic pressure can
deteriorate the membrane fouling due to increase in the initial permeate flux [26, 48, 50]. In
the meantime, the decreased membrane fouling can be observed when heating the FS [26, 42,
48, 50]. This can be ascribed to the improvement of back diffusion of colloids from the
membrane surface as well as the increased foulant solubility at elevated temperature [26, 42,
48, 50]. Furthermore, temperature can be also an important parameter affecting the solubility
product of sparingly soluble salts in the scaling phenomenon [26, 48].
2.2.5.3. Membrane properties
The membrane intrinsic properties can indirectly affect membrane fouling via changes in
initial permeate flux and solution chemistry [17, 18, 26, 46]. A FO membrane with superior
separation properties (high water permeability and rejection) can induce high initial permeate
flux [17, 18]. As mentioned in section 2.2.5.1, this can exacerbate the membrane fouling due
to the increased hydrodynamic drag force. On the other hand, a membrane with inferior
intrinsic properties (low water permeability and rejection) can lead to increase in forward and
reverse draw solute diffusion between the FS and the DS [17, 18] . In many cases, the draw
solutes from the DS can be considered as nutrients for microorganisms [51]. This can
facilitate the biofouling development. Moreover, the draw solutes such as Ca2+ and Mg2+ can
raise the risk of scaling if they accumulate in the FS through the reverse solute diffusion [26,
46]. Also, effects of these ions on organic fouling are specified in section 2.2.5.1. Besides,
forward diffusion can result in the enhanced membrane scaling potential near the membrane
surface on the draw side [26].
In addition to the membrane separation properties, the surface properties of the FO membrane
such as morphology, wettability, surface functional groups and charge can directly impact
fouling behavior through the influence of foulant-surface interactions [26, 29, 42, 52, 53].
Membrane morphology implies the roughness of membrane surface. It is indicated that a
31
smoother membrane surface may be less sensitive to fouling for two reasons [52]. The first
reason is the surface area of a smooth membrane for colloid-membrane interaction is
presumably smaller, compared to that of a rough membrane [26, 52]. Another reason is the
reduction of the shear force caused by the frictional resistance of the rough membrane can be
greater than that of the smooth membrane [26, 52]. Thus, the rate of foulants swept away from
the rough membrane surface can be low. In addition, the less severe fouling of the FO
membrane can be a result of using hydrophilic membrane materials [26, 29]. This can be
because of the large repulsive acid-base interaction that can prevent the attachment of colloids
on the membrane surface [26, 29]. Furthermore, functional groups such as –COO- groups
within the material structure of the FO membrane can bind to divalent cations such as Ca2+ to
form complexes [26, 29, 42, 46]. Consequently, a significant increase in the concentration of
calcium ions can promote the growth of gypsum on the membrane surface [42, 46]. Similarly,
the adhesion of alginate and the membrane surface can be intensified due to the bridging
effect of Ca2+ ions [42, 46]. Thus, organic fouling can become more severe in this case.
Noticeably, once most colloids existing in the FS are negatively charged, negatively charged
membranes can facilitate the removal of fouling because of the electrostatic repulsion [26, 29,
53]. However, these membranes can also be prone to potential ions such as Ca2+ or positively
charged colloids in the FS [26]. In summary, a FO membrane with a smooth, hydrophilic and
neutral surface and without carboxyl groups should be developed for the fouling prevention.
2.2.5.4. The composition of draw solution
The effects of DS composition on membrane fouling are via the impact of flux and changes in
the chemistry of the FS [26, 27, 38, 54]. The high concentration of DS can increase initial
permeate flux, which can lead to the increased membrane fouling as explained in section
2.2.5.2 [26, 38]. Moreover, the reverse draw solute diffusion can be more severe at a high DS
concentration, thereby potentially exacerbating the fouling tendency [26, 27, 38]. In addition
to the DS concentration, the type of draw solute can also elevate the membrane fouling via the
degree of reverse draw solute diffusion [26, 27, 38]. For example, a FO system operated with
CaCl2 solution as a DS can be more susceptible to biofouling, compared to that using a DS of
MgCl2 [55]. This is because of the complexation of calcium ions to EPSs of bacteria [55].
Another example is the utilization of glucose or ammonia carbonate as a DS can be a cause of
the enhanced biofouling because these draw solutes can play roles as nutrients for the microbe
development [26, 54].
32
2.2.5.5. Membrane orientation
Membrane orientation can influence fouling behavior through the synergic effects of surface
properties, hydrodynamic conditions and CP [26, 56-58]. Many studies pointed out that the
less austere fouling was observed in AL-FS orientation compared to that in AL-DS
orientation [56-58]. The reason is the foulant deposition on the membrane surface can be
alleviated as the smooth dense AL faces the FS in the AL-FS orientation [17, 26]. In addition,
the disappearance of cross-flow velocity within the pores of the porous SL can aggravate the
internal fouling as well as ICP in AL-DS orientation [47, 56, 58]. Therefore, AL-FS
orientation can be superior compared to AL-DS orientation regarding the fouling mitigation.
However, AL-FS orientation can experience the more severe ICP, thereby reducing permeate
flux when compared with AL-DS orientation under otherwise identical conditions [59]. Thus,
the selection of appropriate membrane orientation should be taken account of the synergic
effects of membrane fouling and ICP rather than any single impact.
2.2.6. Effects of membrane fouling on the performance of FO process
The occurrence of membrane fouling can noticeably affect the FO membrane properties and
reverse solute flux [26, 60]. The membrane surface can be more negatively charged in the
presence of a humic acid fouling layer because of the plentiful carboxylic and hydroxyl
functional groups in humic acid molecules [60]. This fouled membrane can hinder the
negative ions dissociated from the draw solute molecules because of the electrostatic
repulsion [60]. Therefore, the reverse draw solute diffusion can be alleviated. Moreover, the
formation of this fouling layer can decrease the membrane solute permeability coefficient, but
insignificantly influence the membrane water permeability coefficient [60]. The reason is the
diffusion of water through the membrane is not significantly affected by the hydrated humic
acid layer on the membrane surface, while the solute diffusion can be restricted by steric
hindrance due to this layer [60]. The relieved reverse solute diffusion in the existence of the
membrane fouling layer can also be explained based on this mechanism [26]. In some cases,
membrane damage can result from the high concentration of acidic by-products at the
membrane surface [61, 62]. For example, cellulose acetate used to fabricate FO and RO
membranes can be biodegraded by fungi and bacteria [61].
In addition to the effects on the membrane properties, the membrane fouling can cause
impacts on water flux [26, 48, 60]. Two adverse consequences of membrane fouling are the
additional hydraulic resistance and the cake enhanced CP phenomenon and are considered as
major causes of flux decline [26, 41]. The former comes from a foulant cake layer established
33
by the accumulation and deposition of foulants on the membrane surface [26, 60]. This can
reduce the overall water permeability of the fouled membrane [26]. The latter can also be
considered a result of this cake layer because this layer can hinder back diffusion of solutes
within it [26]. Thus, the solute concentration near the membrane surface can significantly
increase, which can enhance CP [26, 42]. As a result, the net driving force of water transport
is reduced due to the increased osmotic pressure of the solution near the membrane surface at
the feed side [26].
The accumulation of foulants on the FO membrane surface can either improve or decrease the
removal of pollutants and solutes due to different mechanisms [26, 49, 60, 63]. Several
explanations for the increased membrane rejection because of fouling can be elucidated as
follows. The first hypothesis is the membrane fouling can result in the pore blockage or seal
the molecular-scale defects of the FO separation layer [49, 63]. These changes can vary the
effective pore size of the FO membrane and impede the passage of foulants and solutes [49,
63]. Another hypothesis is related to the presence of an additional barrier known as a cake
layer on the membrane surface [49, 63]. This layer can enhance the steric hindrance to
solutes, resulting in the increase of membrane rejection [60]. On the other hand, the
membrane fouling can also decrease the membrane rejection due to the deteriorated fouling-
enhanced CP [26, 49]. Xie et al observed that the membrane rejection of trace organic
compounds witnessed a decline when the FO system was operated at high initial permeate
flux [49]. It is possible that the more compact and adhesive membrane fouling layer at high
initial permeate flux can exacerbate fouling-enhanced CP [49].
2.2.7. Mitigation of fouling in the FO process
2.2.7.1. Fouling characterization
Similar to fouling detection in RO, calculation of fouling potentials of the FS can help
forecast fouling during FO membrane operation. Foulant adhesion forces measurement can be
effective to evaluate the potential of organic fouling [64]. In addition, the determination of
carboxylic acidity can expose the concentration of carboxylic functional groups in foulants
such as alginate, humic acid and proteins [65]. As a result, the extent of interaction between
these foulants and calcium which can potentially aggravate the membrane fouling, can be
predicted [65]. Furthermore, the biofouling potential can be assessed by using bioassays that
include various techniques to determine the microbial growth potential of the FS [66].
However, once membrane fouling takes place on the FO membrane surface, ex-situ and off-
line methods can be essential for specifying the nature of fouling cake, facilitating the further
34
preventative methods [17, 67]. Zeta potential and contact angle measurements can be used to
characterize the chemical properties of a fouled membrane such as surface charge and
wettability [67]. In addition, the morphology of the fouled membrane surface can be
determined by using ex-situ methods such as scanning electron microscopy and atomic force
microscopy [17, 67]. Also, some techniques namely adenosine triphosphate measurements,
extracellular polymeric substances quantification and confocal laser scanning microscopy can
be useful for biofilm characterization [17, 67].
Online and non-invasive monitoring of the performance of FO process is one of effective
methods to detect early signs of fouling in real-time [67, 68]. In-situ monitoring techniques
for membrane fouling can be divided into two groups [69]. The first group includes optical
methods such as direct observation technique, laser-based techniques, photo interrupt sensor,
particle image velocimetry, fluorescence-based technique, optical coherence tomography, and
membrane fouling simulator [67-69]. The other is associated with non-optical techniques
namely nuclear magnetic resonance imaging, X-ray micro-imaging, ultrasonic time-domain
reflectometry, electrical impedance spectroscopy and streaming potential measurement [67-
70].
2.2.7.2. Reduced-fouling methods
2.2.7.2.1. Pretreatment of the FS
Pretreatment methods applied to RO membrane system can be applicable to pre-treat the FS
during FO membrane operation [26]. These methods either pay attention to the removal of
foulants from the FS, or focus on the control of the feedwater chemistry such as pH, ionic
strength and specific ion composition to reduce the risk of membrane fouling in the FO
process [26, 71]. Many conventional and advanced technologies such as screening,
coagulation, flocculation, sedimentation, media filtration, microfiltration and ultrafiltration
can be applied to extract particles and microorganisms from the feedwater [26, 29, 71]. In
addition, the potential of membrane fouling can also be alleviated through the adjustment of
pH, disinfection with chlorine and the supplement of anti-scalants into the feedwater [26, 29,
71]. However, the use of anti-scalants can enhance the development of biofilm on the
membrane surface in some cases [71]. For example, polyphosphate-based anti-scalants can
play a role as phosphorous suppliers for the growth of microbe, thereby promoting biofouling
[72]. Thus, using anti-scalants in the FO system should be cautiously considered.
Characteristics and roles of each pretreatment method can be illustrated in Table 8 [26, 29, 71,
72].
35
Table 8. FO pretreatment technologies in fouling control
Pretreatment
techniques Typical chemicals used Major roles
Screening No chemicals
Only mechanically separate
Remove large particles and
suspended solids
Disinfection Chlorine, sodium hypochloride
Prevent the growth of
microorganisms
Control biofouling
Coagulation-
flocculation and
sedimentation
Aluminum and ferric salts
Remove colloids, suspended solids
and dissolved organic carbon
Prevent organic and colloidal fouling
and biofouling
Media filtration Activated carbon, anthracite,
coal, glass fiber, expanded clay
Remove suspended solids and
dissolved organic carbon
Prevent organic and colloidal fouling
Membrane
filtration
Using microfiltration and
ultrafiltration
Can use coagulation and
flocculation simultaneously in
this technique
Remove suspended solids, colloids,
microorganisms, and
macromolecules
Prevent organic fouling, colloidal
fouling and biofouling
Anti-scalants
Polyacrylates,
organophosphonates, sodium
hexametaphosphate
Control scaling
2.2.7.2.2. Optimization of hydrodynamic conditions
Appropriate adjustments of initial flux, cross-flow velocity and membrane orientation can
alleviate risk of fouling. As elucidated in section 2.2.5.2, the membrane fouling propensity
can be more severe at high initial permeate flux. Meanwhile, the membrane fouling was
reported not to occur when the flux was under a certain value [56]. This initial flux value can
be referred to as a critical flux value [26]. Thus, the FO system should be operated at this
critical permeate flux. Furthermore, the intensification of cross-flow velocity is also proved to
reduce the membrane fouling and CP in section 2.2.5.2, and thus is preferred in the operation.
In addition to the increased cross flow velocity, the presence of spacers can also encourage
the effective cross flow velocity, turbulence and mass transfer near the membrane surface,
36
thereby alleviating the membrane fouling [26]. As mentioned in section 2.2.5.5, the
membrane fouling occurred at FO mode is less severe than that at PRO mode. Nevertheless,
the initial water permeate flux attained at PRO mode is higher than that at FO mode. Thus, the
tradeoff between the reduction of membrane fouling and the achievement of high initial water
flux should be evaluated before selecting the most appropriate membrane orientation for
practical applications.
2.2.7.2.3. The use of relevant draw solution
Based on effects of the type and concentration of draw solute on fouling behavior, selection of
a suitable DS for the FO process can be beneficial for fouling control. As discussed in section
2.2.5.4, the increased concentration of DS can deteriorate the fouling tendency due to the
higher permeate flux and reverse draw solute flux. Thus, the introduction of critical DS
concentration which is corresponding to critical permeate flux can be useful for mitigating the
membrane fouling [73]. The critical DS concentration is indicative of the threshold DS
concentration below which fouling is not noticeable [73]. In addition, selected draw solutes
should not behave as promoters, fouling precursors and nutrients which can facilitate the
enhancement of fouling. Furthermore, draw solutes should possess low solute permeability to
minimize the reverse draw solute diffusion through the membrane, which can potentially
exacerbate the membrane fouling propensity.
2.2.7.2.4. Modification of membrane properties and structure
Two aspects including surface modification and renovation of membrane materials can be
considered to create a superior anti-fouling FO membrane [17, 26, 29, 48]. To date, most
studies have focused on modifying the dense AL of FO membrane, while very few studies
have paid attention to the modification of the porous SL as well as found out novel membrane
materials [17, 26, 48]. The attempts to alter the AL can be listed as increasing surface
hydrophilicity, reducing the surface roughness, decreasing charge density, reducing carboxyl
groups in the material structure, embedding nanoparticles to prevent biofouling and forming a
second nanofiltration layer on the support layer side [17, 26, 48]. Details of some membrane
alteration techniques can be found in Table 9 [74-79].
37
Table 9. Several modification techniques of the FO membrane surface to mitigate the membrane fouling
Membrane Modification technique Characteristics of the modified membrane surface
Cel
lulo
se
Tri
acet
ate
The porous side of the membrane is coated by poly amino acid 3-(3,4-
Dihydroxyphenyl)-L-alanine, a zwitterionic polymer Increase the membrane hydrophilicity
Silver particles are deposited on the membrane surface by a photoinduced
approach before charge-driven self-assembly of titanium dioxide (TiO2)
nanoparticles are attached onto the layers of silver nanoparticles
Increase anti-biofouling
TF
C p
oly
amid
e
In-situ grafting Jeff amine (an amine-terminated poly-ethylene glycol
derivative) to dangling acyl chloride surface groups on the nascent PA AL
More hydrophilic surface
Decrease the membrane roughness
Decrease the number of carboxyl groups
Less negative surface charge
Functionalize the AL of a TFC PA membrane with poly-ethylene glycol
(PEG)
Increase the membrane hydrophilicity
The PEG layer can impede foulant adsorption.
Reduce the number of carboxyl groups
Grafting polyzwitterions onto the PA membrane via click chemistry
Specific binding sites such as carboxyl groups on the
membrane surface are shielded
Enhance hydrophilic and steric repulsion
Using a dip coating method to bind free carboxyl groups of the membrane
surface with super hydrophilic silica nanoparticles which are
functionalized
Super hydrophilicity
Neutralized carboxylic groups
Covered by positively charged nanoparticles
38
2.2.7.3. Membrane fouling control
2.2.7.3.1. Physical cleaning
Flushing and osmotic backwashing are the most widely-studied methods to physically clean
the fouled FO membrane. The basis of flushing is to utilize the high shear force generated by
increasing cross-flow velocity of cleaning solution to dislodge and remove fouling layers
from the membrane surface [67, 80, 81]. The introduction of air bubbles into the cleaning
solution for air scouring may also help to enhance the efficiency of cleaning process [26, 80,
81]. Schematic diagram of flushing process can be depicted in Figure 5 [80]. The rate of
cross-flow is an important parameter affecting the effectiveness of flushing. Flux recovery can
reach up to 98% after 15 min of flushing with deionized water (DI water) at the increased
cross-flow velocity from 8.5 cm/s to 21 cm/s without any chemicals used [81]. Apart from
cross-flow velocity, the efficiency of the flushing process can be influenced by other factors
such as flushing solution, cleaning duration, use of air bubble and cleaning frequency of the
FO system [80, 81]. However, flushing can only remove foulants on the membrane surface,
except for those within the porous support layer [26]. In this case, backwashing can be
effective to clean the fouled membrane [26].
Figure 5. Flow direction in forward and reverse flushing in FO mode
Osmotic backwashing has justified its efficiency of membrane cleaning in lab-scale tests [26,
82]. Osmotic backwashing can be performed by replacing the original FS with a high salinity
solution and the DS with a low salinity solution [80, 82]. This act aims to create a reverse
permeate flow from the draw side to the feed side to detach foulants both on the membrane
39
surface and within the porous SL and move them away from the membrane (Figure 6) [26, 40,
80, 82]. It is reported that the implementation of osmotic backwashing can allow recovery of
more than 85% of the initial flux [82]. The cleaning efficiency of the osmotic backwash
process is dependent on the backwashing frequency and duration as well as intensity of
backwash and cross-flow velocity [40, 80, 82]. In addition to osmotic backwashing, hydraulic
backwashing is supposed to remove the foulant layer effectively. A recent study demonstrated
that the permeate flux restoration could reach nearly 100% when hydraulic backwash was
applied to clean gypsum scaling within the porous SL [83]. However, the FO membrane is not
designed for high-pressure operation. Therefore, further studies need to be conducted to verify
the applicability of hydraulic backwashing.
Figure 6. The permeate flow direction during backwashing in PRO mode
2.2.7.3.2. Chemical cleaning
Chemical agents such as acids, bases, chelators, surfactants and disinfectants can be utilized
to remove and prevent the FO membrane fouling effectively [26, 67, 80]. Although physical
cleaning methods can be justified having the promising efficiency of fouling mitigation
during FO operation, they still face difficulties in removal of the irreversible fouling [26, 67,
80]. The foulant-surface interaction such as adhesion force is quite high in the occurrence of
organic fouling and biofouling [26, 67]. Thus, it is difficult for physical cleaning to break this
interaction and completely dislodge foulants from the membrane surface. In this case,
membrane fouling can be overcome by chemical cleaning. In this type of cleaning, the choice
of cleaning agent is critical and dependent on the membrane material and type of foulant [80].
An appropriate chemical agent must ensure both the effectiveness of removal of fouling and
40
no damage to the membrane surface [80]. In practice, the cleaning agents for FO can resemble
that for RO. Some acids such as nitric, phosphoric, hydrochloric, sulfuric and citric can be
used to remove precipitate and scaling, while some bases namely NaOH, KOH can be useful
for removing organic fouling [26, 67, 80]. In addition, some oxidants and disinfectants such
as NaOCl, ozone, H2O2, KMnO4 and EDTA can be suitable for biofouling removal [26, 67,
80]. Several common cleaning agents according to fouling type can be illustrated in Table 10
[26, 67, 80].
Table 10. Several typical cleaning chemicals according to the type of foulant.
Type of foulant Cleaning agents
Colloids NaOH, chelating agents and surfactants
Organic compounds NaOH, chelating agents and surfactants
Metal oxides Citric acid or Na2S2O4
Silica NaOH
Scalants such as CaCO3, CaSO4 Citric acid, HCl and EDTA
Biofilm NaOH, chelating agents, EDTA, surfactants and
disinfectants
2.2.7.3.3. Novel fouling prevention approaches
Three novel solutions of fouling control including quorum quenching, enzymatic cleaning and
energy un-coupling can be adopted to mitigate the membrane fouling in FO operation [26, 84-
88]. Even though physical and chemical cleaning methods can remove reversible foulants
effectively, they can also cause changes in physicochemical properties of the membrane as
well as even reduce its mechanical stability. For example, oxidizing agents such as chlorine
can damage polyamide TFC FO membranes [89]. Moreover, the implementation of flushing,
backwashing and chemical cleaning require the use of clean water, energy consumption as
well as cause negative effects of used chemicals on the environment [84-86]. Therefore, three
above-mentioned advanced cleaning methods can be expected to overcome these detrimental
impacts. Quorum quenching is a technique to control biofouling through the interruption of
quorum sensing [84, 86]. Quorum sensing is a biological communication mechanism among
microorganisms via signal molecules such as Acylhomoserine lactone (AHL) [84, 86]. These
molecules are produced from microbe’s metabolism [84, 86]. This mechanism can facilitate
microbe’s ties and thus enhance the formation of biofilm on the membrane surface [84, 86].
Therefore, quorum quenching techniques to interrupt or eliminate these signal molecules can
be effective to prevent biofouling [26, 86]. Several approaches to disrupt the process of
41
quorum sensing can be listed as: (i) suppressing the production of signal molecules; (ii)
degrading AHL; (iii) reducing the activity of AHL protein or AHL synthase; (iv) mimic the
signal molecules by synthetic or natural compounds [86]. For example, natural inhibitors such
as AHL-lactonase from Bacillus cereus and AHL-acylase from Tenacibaculum can degrade
AHLs successfully [86].
In addition to quorum quenching, enzymatic cleaning can be applied to control the membrane
fouling, especially organic fouling. According to this method, various species of enzyme such
as protease, lipase and α-chymotrypsin can be employed to decompose organic compounds
such as proteins, lipids and humic acid on the membrane surface [85, 88]. The type of
enzyme, enzyme concentration and cleaning time are three major factors affecting the
efficiency of enzymatic cleaning [85, 88]. Furthermore, energy uncoupling method is also
useful for biofouling control through the disruption of microbial metabolism energy [87]. It
appears that the development of biofilm is closely connected to bioenergy existing under the
form of adenosine triphosphate (ATP) [87]. This means biofilm can be eliminated by using
chemical un-couplers such as 2,4-dinitrophenol to inhibit ATP synthesis [87]. The halted ATP
synthesis can not only stop any microbial activities, but also contribute to the disruption of
signal molecules which are means of biological communication [87]. As a result, the
membrane biofouling can be controlled effectively [87].
42
CHAPTER 3
PERFORMANCE OF A SEAWATER-DRIVEN FORWARD OSMOSIS
PROCESS FOR PRE-CONCENTRATING SLUDGE CENTRATE:
ORGANIC ENRICHMENT AND MEMBRANE FOULING
43
3.1. Introduction
Forward osmosis is a robust separation platform capable of treating highly complex solutions
that are not suitable for conventional membrane processes [9, 17]. In FO, mass transfer
through the membrane is osmotically driven. Thus, when a DS is readily available, the FO
process can occur with very low energy input [17, 18, 90, 91]. The absence of external
hydraulic pressure can also explain the low membrane fouling tendency and excellent fouling
reversibility of FO. In recent years, the potential of FO to treat many complex streams has
been demonstrated in the literature. These complex solutions include drilling and fracking
fluids from oil and gas exploration [92-95], sludge [96, 97], sludge centrate [1, 9, 14], and
municipal wastewater [13, 91, 98].
A major obstacle to full scale deployment of FO is the lack of a suitable DS [17]. Issues
associated with cost of the draw solutes, regeneration, and loss of draw solutes due to reverse
diffusion can increase the operating cost, thereby hindering the feasibility of FO applications
[17, 28, 37]. In this context, seawater, which is abundant and cheaply available in coastal
areas, has been increasingly considered as a potential DS [17, 37]. The diluted seawater
released from the process can be returned to the sea, and thus, DS regeneration is not
necessary.
In a typical wastewater treatment plant, the sludge is anaerobically digested. The digested
sludge is then dewatered to obtain biosolids for land application. The liquid from this
dewatering process is called sludge centrate, which has a high content of suspended solids,
nutrients, and organic matter [14]. Due to the difficulties associated with the treatment of this
sludge centrate, in most cases, it is returned to the headworks of the treatment plant. The
recirculation of untreated sludge centrate to the headworks leads to additional organic and
nutrient loading, and deprive the plant from any opportunities for energy and nutrient
recovery [1, 2].
The use of FO to pre-concentrate sludge centrate has been investigated in several recent
studies [9, 14]. However, draw solutes such as MgCl2 and NaCl are expensive and must be
regenerated. On the other hand, seawater appears to be a particularly promising DS for pre-
concentrating sludge centrate. Ansari et al. [1] has recently demonstrated a seawater-driven
FO process for phosphorus recovery from sludge centrate with a specific focus on evaluating
the efficiency of nutrient recovery. Results from the previous study suggest that identifying
the most suitable membrane and orientation is necessary to ensure the best performance of
seawater-driven FO process for pre-concentrating sludge centrate. More importantly,
understanding of the fouling process and developing strategies to control fouling need to be
discussed to guarantee the long-term operation of the FO system. It is also necessary to
consider all other factors affecting water flux and membrane fouling, such as membrane pre-
wetting and DS.
This study aims to elucidate the effects of membrane materials, prewetting procedures and DS
on the performance of seawater-driven forward osmosis for pre-concentrating organic matter
44
in the sludge centrate. The performance of the FO process is observed in terms of chemical
oxygen demand (COD) enrichment, membrane fouling and flux recovery by physical
cleaning. Comparison between seawater and analytical grade NaCl as the DS was made to
highlight the potential and challenges of using seawater for enriching COD in sludge centrate.
3.2. Materials and methods
3.2.1. Forward osmosis system and membranes
A lab-scale cross-flow FO system (Figure 7) was used. The FO system included a membrane
cell, two variable speed gear pumps (Micropump, Vancouver, Washington, USA),
conductivity and temperature controllers, and a digital balance to measure the flux. The FO
membrane cell consisted of two symmetric rectangular chambers for the feed and draw
solutions, respectively. The internal dimensions of each chamber were 10 cm in length, 5 cm
in width and 0.2 cm in height. The system was operated in the counter-current mode. The FO
membranes were positioned either in AL-FS orientation, or AL-DS orientation.
Flat sheet FO membranes were obtained from Hydration Technology Innovations (HTI,
Albany, OR) and Porifera, Inc. (Hayward, California, USA). The HTI membrane had an
asymmetric structure and was made of CTA with an embedded polyester mesh for mechanical
support. The Porifera membrane was a TFC membrane consisting of a PA layer on a
microporous polysulfone SL. Key properties of the HTI and Porifera membranes were
summarised in Table 11.
Table 11. Key properties of the AL of the FO membranes [99, 100].
Membrane HTI-CTA Porifera-PA
Water permeability (A) (L/m2.h.bar) 0.84 3.2
Salt (NaCl) permeability (B) (L/m2.h) 0.32 0.41
Structural parameter (S) (mm) 0.57 0.46
Contact angle (o) 61 49.5
Surface roughness (nm) 3.8 57.4
Zeta potential at pH = 7 (mV) -5 -16
45
Figure 7. Schematic diagram of the lab-scale FO system.
3.2.2. Feed solution and draw solution
In this study, seawater and NaCl solutions were used as DSs. Seawater was collected from
North Wollongong Beach (Wollongong, New South Wales, Australia) and was filtered using
filter paper with a pore size of 1 μm prior to the experiments. NaCl solutions of 0.25, 0.5, 0.75
and 1 M were prepared using analytical grade NaCl and deionised (DI) water. Anaerobically
sludge centrate was collected from a high-speed centrifuge at the Shellharbour Wastewater
Treatment Plant (Shellharbour, New South Wales, Australia). The sludge centrate was pre-
filtered using a 0.2 mm plastic screen to remove any large objects. The compositions of
seawater and sludge centrate were summarised in Table 12.
Table 12. Composition of seawater and sludge centrate (Values indicated average ± standard
deviation of three samples).
Parameters Unit Seawater Sludge centrate
pH - 7.3 ± 0.2 7.1 ± 0.1
Electrical conductivity mS/cm 44.2 ± 0.3 6.8 ± 0.1
Osmotic pressure bar 28.1 ± 0.6 6.5 ± 0.1
Total solids g/L 31.7 ± 2.8 1.6 ± 0.1
COD mg/L - 420.3 ± 15.5
Ammonia mg/L - 520 ± 2.6
Phosphorus mg/L - 371.5 ± 1.4
46
3.2.3. Membrane prewetting
Prewetting was conducted by soaking in a solution containing 70% ethanol and 30% water for
45 min. Following soaking, the membrane was rinsed and preserved in DI water overnight
prior to filtration experiments.
3.2.4. Experimental protocols
All FO experiments were conducted in four steps. In the first step, DI water was used as the
FS for 30 min to determine the pure water flux. In the second step, DI water was substituted
with sludge centrate and the FO experiment was conducted until a water recovery of 55% had
been achieved. Duration of this second step varied from experiment to experiment ranging
from 24 to 120 hours. At specified time intervals, 10 mL samples were collected from the
feed for analysis. In the third step, the draw and feed solutions were replaced by DI water to
facilitate membrane cleaning by flushing at a cross-flow velocity of 24 cm/s for 20 min. In the
last step, the pure water flux was determined again using DI water under identical
experimental conditions as in the first step. In all experiments, initial volumes of the feed and
draw solutions were 1 and 3 L, respectively. The circulation flow rate of the feed and the draw
solution was 0.8 L/min (i.e., cross flow velocity of 13 cm/s).
3.2.5. Calculations
Water flux (Jw) was calculated based on the change in weight of DS, and expressed as in
Equation 3:
( )1
1
i it t iw
i i m i m
m m mJ
t t A t A −
−
− = =
− (Equation 3)
In which: Δmi: the change in weight of DS over a time interval (g)
Δti: a time interval (h)
ρ: the solution density (g/cm3)
Am: the effective membrane area (m2)
Water recovery was determined based on the ratio of the cumulative permeate volume and the
initial volume of FS, and presented as in Equation 4:
Water recovery (%) = 0 0
0100
t
m w
initial
A J dt
V
(Equation 4)
In which: Jw: the observed water flux at time t
Vinitial: the initial volume of FS
47
The draw solute flux (Js) was calculated by Equation 5:
2 1, 2 , 1f t f f t f
s
m
C V C VJ
t A
− =
(Equation 5)
In which: Cf,t2: The concentration of draw solute in FS at time t2 (g/L)
Vf2: The volume of FS at time t2 (L)
Cf,t1: The concentration of draw solute in FS at time t1 (g/L)
Vf1: The volume of FS at time t1 (L)
Δt: a certain period of filtration time (h)
3.2.6. Analytical methods
Key water quality parameters of sludge centrate and seawater were measured according to
standard methods. COD was determined using a Hach DRB200 COD Reactor and Hach
DR3900 spectrophotometer following the US-EPA Standard Method 5220. Temperature and
pH of solutions were measured by an Orion 4-Star Plus pH/conductivity meter (Thermo
Scientific, Waltham, MA).
The surface characteristics of the FO membranes were characterized using scanning electron
microscopy (SEM) (JOEL, JSM-6400LV, Japan). Prior to taking SEM images, coupon
membrane samples were coated with a thin layer of gold.
3.3. Results and discussions
3.3.1. Pure water flux under different conditions
Figure 8. Pure water fluxes of CTA and PA FO membranes using seawater as DS in AL-FS
and AL-DS orientations with and without prewetting.
48
Under the same experimental conditions, the TFC PA membrane showed a higher pure water
flux than that of the CTA membrane (Figure 8). This observation could be explained by the A
value (water permeability under a hydraulic pressure) and the structure parameter (S value) of
these two membranes (Table 11). Indeed, the A value of the TFC PA membrane was
approximately four times higher than that of the CTA membrane. The structural parameter
value of the TFC PA membrane (460 μm) was slightly lower than that of the CTA membrane
(570 μm) [99]. Previous studies have demonstrated that a smaller structural parameter resulted
in less severe ICP and thus higher water flux [17, 99]. It is noteworthy that in the FO process,
CP could also influence the water flux. Thus, the difference in pure water flux between the
TFC PA and CTA membranes was not necessarily proportional to the difference in their A
and S values.
The AL-DS orientation exhibited a higher pure water flux compared to the AL-FS orientation.
The difference in water flux between these two orientations was considerably less than the
comparison between the TFC PA and CTA membranes discussed above. Indeed, the higher
water flux under the AL-DS orientation compared to the AL-FS orientation could be solely
attributed to a less severe ICP condition [17, 33, 101].
Prewetting significantly improved the pure water flux of the TFC PA membrane but had a
negligible impact on the CTA membrane. As a result of prewetting, water flux of the PA
membrane increased by 29.7% and 59.7% under the AL-FS and AL-DS orientation,
respectively (Figure 8). There were two possible reasons for this notable increase in flux by
the TFC PA membrane after prewetting, namely swelling of the PA skin layer and prewetting
of the polysulfone SL. The PA skin layer could swell in alcohol causing an increase in the
effective pore size, and thus, increased the water flux [100]. However, there was no
discernible increase in the reverse salt flux due to prewetting. The reverse draw solute fluxes
of the PA membrane and the prewetted PA membrane were all 3.38 g/m2.h. Thus, swelling
was not expected to be a major reason for the improvement in water flux observed here. The
polysulfone SL was hydrophobic and could not be completely wetted by water [102].
Compared to water, ethanol had a lower surface tension, thus, could easily penetrate into the
porous structure and prewet the pores of the polysulfone SL for subsequent water permeation
[102, 103]. On the other hand, the CTA membrane was asymmetric and readily hydrophilic
because of the presence of hydroxyl groups. Thus, the membrane can be fully hydrated by
soaking in the DI water overnight (Figure 8) [104].
49
3.3.2. Pre-concentration of sludge centrate by forward osmosis
Figure 9. The performance of FO for enrichment of COD in sludge centrate using seawater as
the DS. The theoretical COD concentration factor was calculated assuming 100% COD
retention by the FO membrane.
As the sludge centrate was concentrated by FO, COD concentration increased proportionally
(Figure 9). In all cases, the COD concentration factor was lower than the theoretical value
assuming complete COD retention. The observed difference between the COD concentration
factor and theoretical value could possibly be attributed to the deposition of particulate COD
materials on the membrane surface. In fact, there was a correlation between membrane
fouling and the COD enrichment results in Figure 9 as discussed further in section 3.3.3.
The best performance in terms of COD enrichment was from the CTA membrane with the
AL-FS orientation (Figure 9). The AL of the CTA membrane had a lower surface roughness
than its own SL as well as that of the TFC PA membrane [26, 52, 100, 105]. Thus, due to the
hydrodynamic drag force from the cross flow, the deposition of organic substances on the
CTA membrane was expected to be less compared to the TFC PA membrane. In addition,
surface chemistry interaction between organic matter and the membrane surface could also be
a reason for the lower organic enrichment of using the PA membrane. The PA membrane with
the considerable number of highly polar carboxylic functional groups on its surface had
significant affinity with organic colloids in the FS [100]. As a result, the accumulation of
organic matter on the PA membrane surface could be enhanced, thus reducing COD
concentration in the bulk feed. It is noted that this combination (CTA under AL-FS
orientation) also had the lowest initial water flux. However, it appears that water flux did not
affect COD enrichment performance as demonstrated below.
50
Figure 10. The performance of FO for enrichment of COD in sludge centrate using the NaCl
solution as the DS and the CTA membrane in the AL-FS orientation. Note: The theoretical
COD concentration factor was calculated assuming 100% COD retention by the FO
membrane.
DS concentration did not affect the enrichment of COD by FO (Figure 10). No significant
difference in COD enrichment was observed when the NaCl DS concentration increased from
0.25 to 1 M (Figure 10). The initial water flux was proportional to the DS concentration.
Thus, the results in Figure 10 also suggested that water flux did not affect COD enrichment as
discussed above.
Seawater showed comparable COD enrichment performance to that of analytical grade NaCl
as the draw solutes. The osmotic potential of seawater was similar to that of the 0.5 M NaCl
solution [106, 107]. Seawater was readily available in coastal areas and thus it was a low-cost
DS. However, in addition to NaCl, seawater contained many other salts. Some of them were
sparingly soluble and might cause membrane scaling as further discussed in section 3.3.3.2.
51
3.3.3. Factors affecting fouling behaviour
3.3.3.1. Membrane properties and orientations
Figure 11. Fouling behaviour of the FO membranes in (A) AL-FS orientation and (B) AL-DS
orientation using seawater as DS.
In all cases, membrane fouling was significant as indicated by significant water flux decline
during COD enrichment (Figure 11). The decrease in water flux could be mostly attributed to
the formation of a cake layer on the membrane surface. This cake layer caused an additional
hydraulic resistance and increased CP, thus reducing water flux.
In good agreement to the data in Figure 9, the CTA membrane was less susceptible to fouling
than the TFC PA membrane regardless of the membrane orientation. As described above, this
result was likely due to the smooth surface of the CTA membrane. The higher roughness and
prominent ridge-and-valley structure of the PA membrane surface could exacerbate the
deposition of foulants, thus more severe fouling [100]. Additionally, a high density of
carboxylic functional groups in the structure of the PA membrane could be potentially
vulnerable to fouling [108]. In the presence of Ca2+ ions, carboxyl acid functional groups
52
could lead to bridging of membrane surface and Ca2+-organic foulants, thus probably
exacerbating organic fouling.
The AL-FS orientation showed less fouling than the AL-DS orientation. There were several
possible explanations. As mentioned above, the lower roughness of the AL in the AL-FS
orientation could alleviate the accumulation of foulants on the membrane surface. FO
operation under the AL-DS orientation was susceptible to internal clogging since organic
molecules could readily penetrate the porous SL. In addition, the high reverse solute diffusion
in the AL-DS orientation could increase the osmotic potential of the feed, decreased effective
driving force, and thus, reduced the water flux [33].
3.3.3.2. Draw solution
Figure 12. Comparison of fouling behaviour towards CTA membranes in the AL-FS
orientation using seawater and the NaCl solutions at different concentration as DSs.
As discussed in the previous sections and further discussed in section 3.3.4, the CTA
membrane under the AL-FS orientation demonstrated the best suitability for enriching organic
matter in sludge centrate. Thus, only the CTA membrane under the AL-FS orientation was
used for further investigating the effect of draw solution on the performance of the process.
Increasing the DS concentration led to a higher initial water flux (Figure 15), but no
significant impacts on membrane fouling (Figure 12). The elevated initial water flux was a
result of the increased driving force of water transport due to an increase in the concentration
gradient along the membrane. However, the extent of fouling in all cases was nearly the same.
This was probably because of the smoothness of the CTA membrane surface that could
effectively minimize the accumulation and deposition of foulants. In addition, this could be
explained using a theory of ‘critical DS concentration’. According to this concept, fouling
could be less severe at below the critical DS concentration [58, 73]. It is possible that the used
53
DS concentrations in this study were lower than the critical DS concentration, thus, the
impacts of DS concentration on fouling were insignificant.
Seawater exhibited similar fouling to that from the 0.5 M NaCl solution as the DSs. This
observation was consistent with the data shown in Figure 10, and thus could also be explained
by the same reason as referred to earlier. However, it is noted that the flux profile when using
seawater was less stable than that of using the 0.5 M NaCl solution. Multivalent ions in
seawater, such as Ca2+, Mg2+, SO42- and PO4
3- could act as fouling promoters, and scaling
precursors through the reverse solute diffusion, thus increasing fouling potential.
3.3.3.3. Fouling layer characteristics
Figure 13. SEM images of the AL of (A) the fouled CTA membrane and (B) the fouled TFC
PA membrane in the AL-FS orientation using seawater DS.
A notable contrast in the morphology of the fouling layer on the CTA and TFC PA
membranes could be observed (Figure 13). The fouling layer on the AL surface of the CTA
membrane was loose and soft (Figure 13A) while that of the PA membrane was dense and
firm (Figure 13B). The observed irregular shape and size of crystals and organic cake-layer on
the membrane surface suggested the presence of both organic and inorganic foulants.
54
3.3.4. Cleaning and flux reversibility
Figure 14. Flux recovery of the fouled CTA, PA and prewetted PA membranes using seawater
DS after physical flushing.
As expected, flux recovery by flushing in the AL-FS orientation was higher than that in the
AL-DS orientation (Figure 14). Flux recovery was proportional to the efficiency of foulant
removal from the membrane surface. In the AL-FS orientation, the deposition of foulants on
the membrane was a surface phenomenon and the fouling cake could be readily removed by
shear force from flushing. On the other hand, in the AL-DS orientation, the FS was in contact
with the SL and due to pore clogging of the SL, flux recoverability was much lower than in
the AL-FS orientation (Figure 14).
The highest flux recovery (95%) was observed with the CTA membrane in the AL-FS
orientation (Figure 14). Together with the high COD enrichment performance shown in
Figure 9, this result suggested that the CTA membrane in the AL-FS orientation was the most
suitable for sludge centrate. This result was also consistent with the loose and soft fouling
layer of the CTA membrane under the AL-FS orientation previously shown in Figure 13A.
Flushing was not efficient in restoring the water flux of the PA and prewetted PA membranes.
There were two possible reasons. Firstly, foulants deposited on a rough surface of the PA
membrane could be sheltered from cross-flow shear force, thus decreasing the number of
foulants detached from the membrane surface. Secondly, the highly polar carboxylic
functional groups on the PA membrane surface were available for ionic bonding with
foulants, thus improving foulant adhesion [100]. In contrast, the CTA membrane was only
slightly negatively charged and did not have free carboxyl functional groups that could
interact with the foulants [100].
55
Figure 15. Comparison of pure water fluxes and flux recovery of CTA membranes in the AL-
FS orientation using seawater and NaCl DS.
Increasing the DS concentration resulted in a higher pure water flux but also reduced flux
recovery by flushing (Figure 15). Indeed, as the NaCl concentration increased from 0.25 to 1
M, flux recovery decreased from nearly 100% to 70%. Since the DS concentration was
proportional to water flux, results in Figure 15 showed that the extent of irreversible fouling
(by flushing) was inversely correlated to the initial flux. A denser and more compact fouling
was formed at higher initial flux, thereby, impairing the efficiency of flushing.
Seawater had a similar osmotic potential to a 0.5 M NaCl solution. However, a slightly lower
flux recovery was observed when using seawater as the DS compared to the 0.5 M NaCl
solution. The complex composition of seawater could result in a less reversible fouling layer.
Indeed, multivalent cations (such as Ca2+ and Mg2+) in seawater could exacerbate fouling and
render fouling layer more adhesive, thus lowering the flux recoverability by flushing.
56
Figure 16. SEM images of (A) the AL of the cleaned CTA membrane in the AL-FS
orientation using seawater as DS, (B) the AL of the cleaned CTA membrane in the AL-FS
orientation using the 0.5 M NaCl solution as DS, (C) the SL of the cleaned CTA membrane in
the AL-DS orientation using seawater as DS, (D) the AL of the cleaned PA membrane in the
AL-FS orientation using seawater as DS and (E) the SL of the cleaned PA membrane in the
AL-DS orientation using seawater as DS.
The complex composition of seawater and the membrane characteristics were observed to
govern the cleaning efficiency of flushing (Figure 16). After flushing, the CTA membrane
active surface in the AL-FS orientation (seawater as DS) still had some salt crystals and
organic particles (Figure 16A). In contrast, when using the 0.5 M NaCl solution, the
57
membrane was mostly clean after flushing (Figure 16B). This was consistent with the data in
Figure 15 that indicated how the complex composition of seawater resulted in a more
adhesive fouling layer.
In addition, SEM images indicated that both for CTA membrane under the AL-DS orientation
(Figure 16C) and PA membrane in the AL-FS orientation (Figure 16D) flushing removed
crystals significantly, but the organic layers only partly. Also, simple flushing seemed to be
inefficient in removing both crystals and organic layers towards the PA membrane in the AL-
DS orientation. These observations resulted from synergistic effects of membrane roughness
and chemical structure of these membranes on fouling as discussed above.
3.4. Conclusions
Results from this study demonstrated the potential of the FO process to enrich COD in sludge
centrate. Compared to the TFC PA membrane, the CTA membrane in the AL-FS orientation
showed a much better COD enrichment efficiency, lower fouling, and higher flux recovery by
simple flushing. There was a correlation between membrane fouling and COD enrichment
efficiency. In other words, COD enrichment efficiency decreased when organic matter
accumulated on the membrane surface, causing fouling. The results also showed that
membrane fouling was not affected by the DS concentration (or initial water flux) possibly
because of the low initial flux in this study, however, flux recovery by membrane flushing
decreased as the initial water flux increased. Seawater was a potentially low-cost and effective
DS for COD enrichment. However, compared to NaCl, seawater as the DS led to more severe
membrane fouling and lower flux recovery by flushing.
58
CHAPTER 4
SEAWATER-DRIVEN FORWARD OSMOSIS PRE-CONCENTRATION
OF NUTRIENTS IN SLUDGE CENTRATE: PERFORMANCE AND
IMPLICATIONS
59
4.1. Introduction
Phosphorus is an essential element for all living organisms on earth [109]. It is a vital
component of biological molecules, such as adenosine triphosphate, deoxyribonucleic acid
and ribonucleic acid [109]. Also, phosphorus can be found in all living beings, especially in
bones and teeth [109]. Human beings intake phosphorus via daily meals with agricultural
products predominantly cultivated using phosphate fertilizers [1]. However, minable
phosphorus will be exhausted in the near future [110]. This threatens food security and a
source of renewable phosphorus supply is essential for future generations. On the other hand,
phosphorus, along with ammonia are also potential pollutants at their excessive concentration
in water bodies [109]. Thus, the phosphorus and ammonia concentrations in the effluent of
WWTPs are strictly regulated. Consequently, environmental concerns, regulatory need for
phosphorus and ammonia removal and the severe depletion of global phosphorus reserves
lead to the idea of recovery of nutrients from wastewater and its by-products in WWTPs [1,
111, 112].
In a typical WWTP, sludge centrate contains phosphate and ammonia concentrations in the
ranges of 75 – 371 mg/L and 250 – 532 mg/L, respectively [109, 111, 113]. This stream is
usually returned to the head of work in WWTPs, resulting in gradually building up
phosphorus and ammonia in the treatment process [114]. This not only decreases the
efficiency of biological treatment, but also leads to gradual struvite precipitation causing
blockages in the pipelines and equipment scaling [115]. Thus, recycling and recovery of
phosphorus from sludge centrate can simultaneously help WWTPs abide by the stringent
regulations of nutrient concentration in effluent, reduce troubleshooting costs caused by
blockages and create an input for the subsequent production of fertilizers. Accordingly, sludge
centrate is likely to be a promising candidate for nutrients recovery.
Although nutrients recovery from sludge centrate obviously brings many benefits,
conventional recovery technologies have been still facing challenges regarding economical
and environmentally-friendly feasibility. Recovery of phosphorus from wastewater is
commonly achieved by precipitation and crystallization processes [109]. These processes are
chemically-intensive and expensive because of the requirement of addition of magnesium and
calcium salts from external sources [109]. Typical products of these processes are in the forms
of calcium phosphate and magnesium ammonium phosphate (struvite) [1, 109]. The
efficiency of precipitation as well as phosphorus content inside precipitate which defines the
value of fertilizer are controlled by the initial concentration of phosphate. Therefore,
phosphorus recovery by precipitation can be improved by firstly enriching phosphate in
sludge centrate [9, 116]. In summary, preconcentrating sludge centrate can simultaneously
increase the precipitation kinetics of nutrients recovery and decrease energy consumption in
WWTPs.
Several advanced separation technologies, namely reverse osmosis and membrane distillation
have been used to pre-concentrate minerals in wastewater for subsequent recovery [117, 118].
60
However, these membranes are not appropriate to concentrate sludge centrate due to high
susceptibility to fouling and energy intensity. Forward osmosis (FO) has been recently
identified as a robust separation platform capable of treating highly complex solutions [9, 17].
FO is an osmotically-driven membrane separation process and thus, this process can occur
with less or without external energy input [17, 18]. Moreover, the absence of external
hydraulic pressure can account for low fouling tendency and high fouling reversibility in FO
[81, 119]. Therefore, the FO pre-concentration of sludge centrate would be of interest. To
date, the use of FO to preconcentrate nutrients in sludge centrate has been reported in several
studies [1, 9, 14]. Holloway et al. used a hybrid FO-RO system with NaCl solution as DS to
concentrate sludge centrate [14]. In their study, they focused on evaluating flux stability and
water recoverability from sludge centrate rather than showing efficiency of nutrients
enrichment. Ansari et al. demonstrated possibility of using a seawater-driven FO process to
recover only phosphate via direct precipitation of calcium phosphate in sludge centrate
without any chemical addition [1]. Seawater was demonstrated as a promising DS for
preconcentrating sludge centrate. However, phosphate precipitate only made up a small
percentage of total solids recovered [1]. Furthermore, these studies have yet to refer to effects
of operating parameters on the performance of a FO system for nutrients recovery. The results
from these studies, thus, suggest the need to comprehensively and deeply evaluate the use of
seawater-driven FO process for nutrient recovery from sludge centrate.
In this study, we explore challenges of using a seawater-driven FO process for pre-
concentrating sludge centrate for subsequent nutrients recovery. A novel feasible solution to
deal with these challenges is demonstrated herein. In addition, significant effects of stirring
conditions and ratio of membrane area over permeate volume on the performance of
enrichment process are discovered for the first time. These operating conditions are then
optimized, thereby rendering this study highly practical.
4.2. Materials and methods
4.2.1. Forward osmosis system
A TFC PA membrane from Porifera was used. The membrane was soaked into DI water over
night before each experiment. Key properties of this membrane have been reported elsewhere
[113].
A lab-scale cross-flow FO system was used. The details of this cross-flow FO system are
available elsewhere [113]. Two different membrane cells were employed for comparison to
evaluate the role of membrane area for enriching centrate. They have the same channel height
of 2 mm and their effective membrane areas are 50 and 123 cm2, respectively. All
experiments were operated at cross-flow velocity of 12 cm/s and in the counter-current FO
mode. In other words, the membrane AL was facing the FS.
61
4.2.2. Feed solution and draw solution
Seawater from Fairy Meadow Beach, New South Wales, Australia was used as the DS. The
collected seawater was filtered by 0.45 μm filter paper before use. A portion of this seawater
was buffered at pH 4.4 by dissolving CH3COONa and CH3COOH in seawater to obtain 0.1 M
of each buffer reagent. Sludge centrate from the centrifuge system in the Shellharbour
wastewater treatment plant was used as the FS. The composition of sludge centrate was
presented in Table 13. Composition of seawater and sludge centrate (values indicated average
± standard deviation of at least five samples). HNO3 solution (1 M) was used to dissolve
phosphorus precipitate for mass balance calculation.
Table 13. Composition of seawater and sludge centrate (values indicated average ± standard
deviation of at least five samples).
Parameters Unit Seawater Sludge centrate
pH - 7.35 ± 0.1 7.6 ± 0.2
Electrical conductivity mS/cm 38.6 ± 0.4 6.4 ± 0.2
Osmotic pressure bar 28.0 ± 0.4 6.3 ± 0.1
Total solids g/L 29.5 ± 3.2 1.7 ± 0.2
Ammonia (NH3-N) mg/L - 486.0 ± 10.3
Phosphate (PO43-) mg/L - 284.0 ± 5.6
4.2.3. Experimental protocols
The volume of water transferred through the membrane was continuously monitored using a
digital balance connected to a computer for water flux calculation. To investigate the effects
of stirring up the FS on the performance of enrichment of sludge centrate, a magnetic stirrer at
200 rpm was used to agitate the FS during the experiment.
The initial FS and DS volumes were 1 and 3 L, respective. This high DS to FS volume ratio is
to ensure minimum dilution effect of the DS. As noted in section 2.1, two membrane cells
with effective area of 50 and 123 cm2 were used in this study. Since each experiment was
conducted until 70% water recovery, the corresponding membrane surface area (m2) over
permeate volume (m3) abbreviated as Am/ VP from these two configurations were 70 and 175
m-1, respectively. This Am/VP ratio could be used to evaluate the impact of available
membrane surface area on fouling.
At the beginning of each experiment, DI water was used as the FS to determine its pure water
flux. Then, sludge centrate was substituted to commence the enrichment process until a water
recovery of 70% was achieved, or water flux was decreased to around 90% of initial water
flux. The conductivity and pH of the FS and DS were regularly measured. At specified time
interval, a 5 mL sample was taken from the FS for ammonia and phosphate analyses. At the
end of the filtration, amounts of phosphate precipitated on the membrane surface and in the
pipeline of the system were determined using 500 mL 1 M HNO3 acid solution. The amount
62
of phosphate penetrating the FO membrane to the DS was also measured. As a result, mass
balance of phosphate in the whole system was quantified.
4.2.4. Analytical methods
pH, electrical conductivity and temperature were measured using an Orion 4 – Star
pH/conductivity meter (Thermo Scientific, Waltham, MA). Ammonia (NH3-N) was analysed
using salicylate US-EPA Standard Method 10205 and Hach DR3900 spectrophotometer.
Orthophosphate (PO43-) was determined using molybdovanadate US-EPA Standard Method
10214 and Hach DR3900 spectrophotometer.
4.3. Results and discussions
4.3.1. Performance of FO for enriching nutrients in sludge centrate
4.3.1.1. Water flux
0 10 20 30 40 50 60 70
0.0
0.2
0.4
0.6
0.8
1.0
1.2
No
rma
lize
d w
ate
r flu
x
Water recovery (%)
175 m-1 & no stirring
175 m-1 & stirring
70 m-1 & no stirring
70 m-1 & stirring
Figure 17. Water flux profile as a function of water recovery at different Am/VP ratios with
and without FS stirring during the pre-concentration of nutrients in sludge centrate by FO.
Seawater was used as the DS.
During FO pre-concentration of sludge centrate, fouling appeared to be driven mostly by the
deposition of inorganic precipitate on the membrane surface (Figure 17). Results in Figure 17
also showed that precipitate deposition occurred both in the bulk solution and on the
membrane surface. Indeed, by agitating the FS, thus preventing chemical precipitation in the
bulk solution and encouraging deposition on the membrane surface, membrane fouling was
significantly more severe (Figure 17). In addition, at the same initial water flux, larger
membrane area reduced the filtration time and the deposition of inorganic precipitate, thus,
resulting in significantly less fouling (Figure 17). At water recovery of 50% without FS
stirring, when the membrane area over permeate volume ratio was 175 m-1, water flux
63
declined by just 20%. On the other hand, without FS stirring, when the membrane area over
permeate volume ratio was 70 m-1, water flux decreased by 90%. Flux decline with FS stirring
was significantly more severe. These values were 60% and 99% for membrane area over
permeate volume ratios of 175 m-1 and 70 m-1, respectively.
4.3.1.2. Nutrient enrichment
0 10 20 30 40 50 60 70-0.5
0.0
0.5
1.0
1.5
2.0
2.5
3.0
Stirring
Phosphate
Ammonium
Am
mo
nia
an
d p
ho
sp
ha
te
co
nce
ntr
atio
n f
acto
r
pH
Water recovery (%)
Theoretical values of phosphate/ammonia
No stirring
Phosphate
Ammonium
7.6
8.0
8.4
8.8
9.2
9.6
10.0
pH of FS
pH of FS
Figure 18. Effects of stirring up the FS on sludge centrate pH and nutrients enrichment during
the filtration. The theoretical values of phosphate and ammonium concentration factor were
calculated assuming 100% nutrient retention by the FO membrane. The Am/VP ratio in this
experiment was 175 m-1. Seawater was used as the DS.
The precipitation of key nutrient constituents during FO pre-concentration of sludge centrate
was also evidenced when examining their fate in the aqueous phase (Figure 18). In this study,
phosphorous and nitrogen existed in sludge centrate as soluble ortho-phosphate and
ammonium and were retained by the FO membrane by over 99% (Figure 19). If they
remained in the aqueous phase, the concentrations of phosphate and ammonium were
expected to increase as water recovery increased based on a mass balance calculation.
However, experimental results did not show any significant increase in phosphate and
ammonium concentrations in the aqueous phase (Figure 18). In fact, a notable decrease in
phosphate and ammonium concentrations was observed as water recovery increased from
50% to 70%, coinciding with a considerable increase in pH of the sludge centrate FS. The
disappearance of phosphate and ammonium from the pre-concentrated aqueous phase
indicated the formation of struvite (i.e. MgNH4PO4) and calcium phosphate (i.e. Ca3(PO4)2)
precipitates that also caused membrane fouling as discussed in section 4.3.1.1.
64
Figure 19. Phosphate and ammonium rejection towards different membrane areas and stirring
conditions at water recovery of 70%. Seawater was used as the DS
Results in this study suggested that nutrient precipitation during FO pre-concentration of
sludge centrate was both pH and time dependent. This was evidenced when examining
phosphate concentration in the aqueous phase at different experimental conditions. Without
FS stirring, a small but nevertheless discernible increase in phosphate concentration in sludge
centrate as a function of water recovery up to 50% could be observed (Figure 18). By
contrast, with FS stirring, there was a notable decline in the concentration of phosphate in
sludge centrate. As discussed in section 4.3.1.1, FS stirring caused more membrane fouling,
thus prolonging the experimental time, allowing for more phosphorous precipitation.
Moreover, FS stirring also led to a significant pH increase, thus, promoting precipitation or
the transfer of phosphate from the aqueous phase to the solid phase that caused membrane
fouling. Indeed, there was a strong correlation between pH increase in the FS and FO
filtration time to achieve 70% water recovery (Figure 20).
65
Figure 20. Correlation between duration of filtration and pH of the FS towards different
experimental conditions at water recovery of 70%. Raw seawater was used as the DS.
It is noteworthy that ammonium could be transferred away from the aqueous phase in the
sludge centrate feed via both co-precipitation with phosphate and volatilization (or air
stripping). The latter occurred when ammonium was converted to ammonia at approximately
pH 9 or higher [120]. The decrease in ammonium concentration without FS stirring was less
severe than that with FS stirring because increase in pH of the FS without FS stirring was less
significant, compared to with FS stirring (Figure 18).
66
4.3.2. Performance of FO for enriching nutrients in sludge centrate with buffered
seawater
4.3.2.1. Effects of using buffered seawater as the DS on water flux
0 10 20 30 40 50 60 700.0
0.2
0.4
0.6
0.8
1.0
1.2
No
rma
lize
d w
ate
r flu
x
Water recovery (%)
175 m-1 - buffered seawater
175 m-1 - raw seawater
70 m-1 - buffered seawater
Figure 21. Changes in water flux at different ratios of Am/VP without FS stirring when
seawater buffered at pH 4.4 was used as the DS.
Compared to raw seawater as the DS, buffered seawater significantly reduced membrane
fouling (Figure 21). At Am/VP ratio of 175 m-1 and the same water recovery of 70%, water
flux using buffered seawater decreased by only 40% instead of 90% when unbuffered
seawater was used as the DS. Similarly, at Am/VP ratio of 70 m-1, the decrease in water flux
was only 55% when using buffered seawater (Figure 21), compared to 99% when unbuffered
seawater was used as the DS (Figure 17). This phenomenon could be ascribed to the
inhibition of phosphorus precipitate at low pH due to using buffered seawater as the DS as
discussed in section 4.3.2.2. Less phosphorus precipitate led to less membrane fouling,
thereby less decline in water flux during the filtration.
67
4.3.2.2. Improvement in nutrient enrichment using buffered seawater as the DS
0 10 20 30 40 50 60 70
1.0
1.5
2.0
2.5
3.0A
m/V
P of 70 m
-1
Phosphate
Ammonium
pH
Theoretical values of phosphate/ammonia
Am/V
P of 175 m
-1
Phosphate
AmmoniumP
ho
sp
ha
te a
nd
am
mo
nia
co
nce
ntr
atio
n f
acto
r
Water recovery (%)
4
5
6
7
8
9
10
pH of FS
pH of DS
pH of FS
pH of DS
Figure 22. The performance of FO for enrichment of nutrients in sludge centrate without FS
stirring at different ratios of Am/VP. The theoretical values of phosphate and ammonia
concentration factor were calculated assuming 100% nutrients retention by the FO membrane.
Buffered seawater was used as the DS.
In contrast to the low nutrient enrichment efficiency previously described in section 4.3.1.2,
buffered seawater resulted in a substantial improvement in nutrient enrichment (Figure 22). At
the Am/VP ratio of 175 m-1, the enrichment of both phosphate and ammonium was almost
identical to the theoretical enrichment curve (Figure 22). This observation could be due to
decreased precipitation of phosphate and lower amount of ammonium transferred away from
sludge centrate via volatilization at decreased pH of the FS. Throughout the filtration process,
pH of the FS gradually declined, whereas that of buffered seawater was stable as expected
(Figure 22). This could be explained by a bidirectional transport phenomenon of ion
hydrogens in FO [24]. It is also noted that larger membrane area was more effective in
enriching nutrients as indicated by higher nutrient enrichment efficiency towards using the
ratio of Am/VP of 175 m-1, compared to using that of 70 m-1.
68
Stirring - raw seawater
No stirring - raw seawater
No stirring - buffered seawater
0
10
20
30
40
50
60
70
80
90
100
Ma
ss b
ala
nce
of
ph
osp
ha
te (
%)
Phosphate in other parts
Phosphate in concentrated centrate
Phosphate on the membrane surface
Figure 23. Differences in mass balance of phosphate between with and without FS stirring, as
well as between using raw seawater and buffered seawater as the DSs. The used ratio of
Am/VP was 175 m-1.
Mass balance calculation of the fate of phosphorus during the preconcentration process
showed that when seawater was employed as the DS and the FS was agitated, phosphate
precipitation occurred mostly on the membrane surface. This was illustrated by the highest
proportion of phosphate (75%) found on the membrane surface at the end of the filtration
(Figure 23). The results also indicated that less soluble phosphate was precipitated on the
membrane surface (60%) without the agitation of the FS (Figure 23). This observation was in
a good agreement with the results shown in section 4.3.1 since phosphorus precipitation led to
membrane fouling and decrease in concentration of nutrients in the bulk FS.
By contrast, soluble phosphate was predominantly retained in concentrated centrate instead of
being precipitated on the membrane surface when buffered seawater was used as the DS.
(Figure 23). Up to 85% of soluble phosphate has been detected in concentrated centrate at the
end of the filtration (Figure 23). These results gave strong support to low fouling propensity
and high phosphate enrichment efficiency as presented in Figure 21 and Figure 22,
respectively when buffered seawater was utilized as the DS. Once concentrated centrate was
expected to be used for subsequent nutrients recovery via outer precipitation, these results
were really promising.
4.4. Conclusions
The results from this study highlighted obstacles in the path of enriching nutrients in sludge
centrate using seawater-driven forward osmosis and a possible way to overcome them. This
study indicated that when using seawater as the DS low efficiency of nutrients enrichment and
significant membrane fouling were closely correlated to the formation and deposition of
69
precipitates in the bulk FS and on the membrane surface. Complex interactions among pH,
precipitation potential, membrane fouling and duration of filtration have been pointed out
herein. Furthermore, the results showed that no FS stirring led to less significant decrease in
concentration of nutrients and less severe membrane fouling, compared to FS stirring. Also,
the membrane area over permeate volume ratio of 175 m-1 was suitable for pre-concentration
of sludge centrate. It is noteworthy that an innovative approach of using buffered seawater as
the DS led to a remarkable improvement in enriching nutrients and mitigating fouling. At
Am/VP ratio of 175 m-1 using buffered seawater as the DS without FS stirring, the
concentrations of nutrients closely increased to theoretically calculated values, and water flux
only decreased by 40% at water recovery of 70%.
70
CHAPTER 5
CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE WORK
71
5.1. Conclusions
The results from this thesis demonstrated that the CTA membrane was superior over the TFC
PA membrane in terms of organic matter enrichment, fouling resistance and reversibility after
cleaning. Prewetting was efficient in significantly increasing the pure water flux of the TFC
PA membrane but causing negligible impacts on that of the CTA membrane. The AL-FS
orientation exhibited higher COD enrichment and lower fouling propensity compared to the
AL-DS orientation. Also, the study indicated that membrane fouling led to the decreased
efficiency of COD enrichment due to the retention and deposition of organic suspended solids
on the membrane surface. In terms of flux and the efficiency of organic enrichment, seawater
and analytical grade NaCl showed the comparable performance. However, it is noted that
seawater as DS resulted in more fouling and lower flux recovery compared to NaCl.
This thesis work showed that during operation of the FO system using seawater as the DS for
enrichment, phosphate and ammonium concentrations in the aqueous phase decreased
significantly. These observations were attributed to formation of phosphorus precipitates (e.g.
struvite and calcium phosphate) and increase in pH of the FS during the filtration. Low
efficiency of nutrients enrichment and severe membrane fouling were closely correlated to
each other via the formation and deposition of precipitates in the bulk FS and on the
membrane surface. In addition, the results indicated that FS stirring resulted in more
significant decrease in concentration of nutrients and more severe fouling than no FS stirring.
Additionally, the membrane surface area over permeate volume ratio of 175 m-1 was right for
pre-concentration of sludge centrate. A major discovery in this study is the application of
buffered seawater as DS to significantly improve nutrients enrichment and reduce membrane
fouling during the filtration.
5.2. Recommendations for Future Work
Resource recovery from waste is an inevitable pathway to ensure the sustainable development
in the future as natural resources have been excessively exploited. The application of FO
membrane to recovery resources from wastewater (e.g. sludge centrate) has been recently
considered as the future’s direction. This study demonstrated the effectiveness of using
seawater-driven FO in enriching organic matter and nutrients in sludge centrate for
subsequent recovery. However, there are still several issues to be addressed prior to the full-
scale deployment of this technology.
Fouling is inevitable in the realm of FO-based wastewater treatment. As mentioned in Chapter
3 and Chapter 4, membrane fouling hindered FO membrane performance and decreased the
efficiency of organic matter and nutrients enrichment. Thus, further research is necessary to
address the issue of membrane fouling when using seawater as the DS for COD and nutrients
enrichment in sludge centrate. A deep understanding of foulants (e.g. colloids, organic
foulants and bio-foulants) in sludge centrate needs to be developed for more effective control
of membrane fouling. In Chapter 4, the chemicals employed to buffer seawater are costly.
Therefore, further investigations need conducting to optimize proper dosages of these
72
chemicals as well as to find out other possible chemicals for the purpose of buffering
seawater. Furthermore, the results from these chapters suggest further evaluations of the
effectiveness and stability of the seawater-driven FO system for organic matter and nutrients
recovery in long-term operation.
73
REFERENCES
[1] A. J. Ansari, F. I. Hai, W. E. Price, and L. D. Nghiem, "Phosphorus recovery from
digested sludge centrate using seawater-driven forward osmosis," Separation and
Purification Technology, vol. 163, pp. 1-7, 2016.
[2] W. Mo and Q. Zhang, "Energy–nutrients–water nexus: Integrated resource recovery in
municipal wastewater treatment plants," Journal of Environmental Management, vol.
127, no. Supplement C, pp. 255-267, 2013/09/30/ 2013.
[3] G. Galvagno, C. Eskicioglu, and M. Abel-Denee, "Biodegradation and chemical
precipitation of dissolved nutrients in anaerobically digested sludge dewatering
centrate," Water Research, vol. 96, no. Supplement C, pp. 84-93, 2016/06/01/ 2016.
[4] R. J. Jackson D., English B., Holmes B., Matthews R., Dixon R., Smith P.,
MacDonald N. , "Phosphorus management of beef cattle in Northern Australia," Meat
and Livestock Australia Limited2012.
[5] R. Norton, "4R Canola Nutrition Guide," International Plant Nutrition Institution
2013.
[6] J. Elser and E. Bennett, "Phosphorus cycle: A broken biogeochemical cycle," Nature,
10.1038/478029a vol. 478, no. 7367, pp. 29-31, 10/06/print 2011.
[7] S. T. Broad and R. Corkrey, "Estimating annual generation rates of total P and total N
for different land uses in Tasmania, Australia," Journal of Environmental
Management, vol. 92, no. 6, pp. 1609-1617, 2011/06/01/ 2011.
[8] D. J. Conley et al., "Controlling Eutrophication: Nitrogen and Phosphorus," Science,
10.1126/science.1167755 vol. 323, no. 5917, p. 1014, 2009.
[9] M. Xie, L. D. Nghiem, W. E. Price, and M. Elimelech, "Toward Resource Recovery
from Wastewater: Extraction of Phosphorus from Digested Sludge Using a Hybrid
Forward Osmosis–Membrane Distillation Process," Environmental Science &
Technology Letters, vol. 1, no. 2, pp. 191-195, 2014.
[10] C. Blöcher, C. Niewersch, and T. Melin, "Phosphorus recovery from sewage sludge
with a hybrid process of low pressure wet oxidation and nanofiltration," Water
Research, vol. 46, no. 6, pp. 2009-2019, 2012/04/15/ 2012.
[11] M. Maurer, W. Pronk, and T. A. Larsen, "Treatment processes for source-separated
urine," Water Research, vol. 40, no. 17, pp. 3151-3166, 2006/10/01/ 2006.
[12] M. Xie, H. K. Shon, S. R. Gray, and M. Elimelech, "Membrane-based processes for
wastewater nutrient recovery: Technology, challenges, and future direction," Water
Research, vol. 89, pp. 210-221, 2016/02/01/ 2016.
[13] W. Xue, T. Tobino, F. Nakajima, and K. Yamamoto, "Seawater-driven forward
osmosis for enriching nitrogen and phosphorous in treated municipal wastewater:
effect of membrane properties and feed solution chemistry," Water Res, vol. 69, pp.
120-30, Feb 01 2015.
[14] R. W. Holloway, A. E. Childress, K. E. Dennett, and T. Y. Cath, "Forward osmosis for
concentration of anaerobic digester centrate," Water Res, vol. 41, no. 17, pp. 4005-14,
Sep 2007.
[15] G. Qiu, Y. M. Law, S. Das, and Y. P. Ting, "Direct and complete phosphorus recovery
from municipal wastewater using a hybrid microfiltration-forward osmosis membrane
bioreactor process with seawater brine as draw solution," Environ Sci Technol, vol. 49,
no. 10, pp. 6156-63, May 19 2015.
74
[16] W. Xue, K. Yamamoto, and T. Tobino, "Membrane fouling and long-term
performance of seawater-driven forward osmosis for enrichment of nutrients in treated
municipal wastewater," Journal of Membrane Science, vol. 499, pp. 555-562, 2016.
[17] K. Lutchmiah, A. R. Verliefde, K. Roest, L. C. Rietveld, and E. R. Cornelissen,
"Forward osmosis for application in wastewater treatment: a review," Water Res, vol.
58, pp. 179-97, Jul 01 2014.
[18] T. Cath, A. Childress, and M. Elimelech, "Forward osmosis: Principles, applications,
and recent developments," Journal of Membrane Science, vol. 281, no. 1-2, pp. 70-87,
2006.
[19] A. Altaee and N. Hilal, "Dual-stage forward osmosis/pressure retarded osmosis
process for hypersaline solutions and fracking wastewater treatment," Desalination,
vol. 350, pp. 79-85, 10/1/ 2014.
[20] A. J. Ansari, F. I. Hai, W. E. Price, J. E. Drewes, and L. D. Nghiem, "Forward osmosis
as a platform for resource recovery from municipal wastewater - A critical assessment
of the literature," Journal of Membrane Science, vol. 529, pp. 195-206, 5/1/ 2017.
[21] W. Luo et al., "Phosphorus and water recovery by a novel osmotic membrane
bioreactor–reverse osmosis system," Bioresource Technology, vol. 200, pp. 297-304,
1// 2016.
[22] B. Kim, S. Lee, and S. Hong, "A novel analysis of reverse draw and feed solute fluxes
in forward osmosis membrane process," Desalination, vol. 352, pp. 128-135, 11/3/
2014.
[23] N. T. Hancock, P. Xu, M. J. Roby, J. D. Gomez, and T. Y. Cath, "Towards direct
potable reuse with forward osmosis: Technical assessment of long-term process
performance at the pilot scale," Journal of Membrane Science, vol. 445, pp. 34-46,
10/15/ 2013.
[24] L. Shu et al., "Why does pH increase with CaCl2 as draw solution during forward
osmosis filtration," Process Safety and Environmental Protection, vol. 104, pp. 465-
471, 2016/11/01/ 2016.
[25] R. W. Baker, Membrane Technology and Applications, Third ed. John Wiley and
Sons, Ltd., 2012.
[26] Q. She, R. Wang, A. G. Fane, and C. Y. Tang, "Membrane fouling in osmotically
driven membrane processes: A review," Journal of Membrane Science, vol. 499, pp.
201-233, 2/1/ 2016.
[27] J. S. Y. William A.Phillip, and Menachem Elimelech, "Reverse Draw Solute
Permeation In Forward Osmosis: Modeling and Experiments," Environ. Sci. Technol.,
vol. 44, pp. 5170-5176, 2010.
[28] D. L. Shaffer, J. R. Werber, H. Jaramillo, S. Lin, and M. Elimelech, "Forward
osmosis: Where are we now?," Desalination, vol. 356, pp. 271-284, 1/15/ 2015.
[29] S. Jiang, Y. Li, and B. P. Ladewig, "A review of reverse osmosis membrane fouling
and control strategies," Sci Total Environ, vol. 595, pp. 567-583, Oct 01 2017.
[30] S. Phuntsho, S. Vigneswaran, J. Kandasamy, S. Hong, S. Lee, and H. K. Shon,
"Influence of temperature and temperature difference in the performance of forward
osmosis desalination process," Journal of Membrane Science, vol. 415–416, pp. 734-
744, 10/1/ 2012.
[31] M. Xie, W. E. Price, L. D. Nghiem, and M. Elimelech, "Effects of feed and draw
solution temperature and transmembrane temperature difference on the rejection of
75
trace organic contaminants by forward osmosis," Journal of Membrane Science, vol.
438, pp. 57-64, 7/1/ 2013.
[32] A. H. Hawari, N. Kamal, and A. Altaee, "Combined influence of temperature and flow
rate of feeds on the performance of forward osmosis," Desalination, vol. 398, pp. 98-
105, 11/15/ 2016.
[33] M. Xie, W. E. Price, and L. D. Nghiem, "Rejection of pharmaceutically active
compounds by forward osmosis: Role of solution pH and membrane orientation,"
Separation and Purification Technology, vol. 93, pp. 107-114, 6/1/ 2012.
[34] P. Zhao, B. Gao, Q. Yue, P. Liu, and H. K. Shon, "Fatty acid fouling of forward
osmosis membrane: Effects of pH, calcium, membrane orientation, initial permeate
flux and foulant composition," Journal of Environmental Sciences, vol. 46, pp. 55-62,
8// 2016.
[35] M. Sadrzadeh, J. Hajinasiri, S. Bhattacharjee, and D. Pernitsky, "Nanofiltration of oil
sands boiler feed water: Effect of pH on water flux and organic and dissolved solid
rejection," Separation and Purification Technology, vol. 141, pp. 339-353, 2/12/ 2015.
[36] J. Schwinge, P. R. Neal, D. E. Wiley, D. F. Fletcher, and A. G. Fane, "Spiral wound
modules and spacers: Review and analysis," Journal of Membrane Science, vol. 242,
no. 1–2, pp. 129-153, 10/15/ 2004.
[37] Q. Ge, M. Ling, and T.-S. Chung, "Draw solutions for forward osmosis processes:
Developments, challenges, and prospects for the future," Journal of Membrane
Science, vol. 442, pp. 225-237, 9/1/ 2013.
[38] Y. Cai and X. M. Hu, "A critical review on draw solutes development for forward
osmosis," Desalination, vol. 391, pp. 16-29, 8/1/ 2016.
[39] Y. Wang, C. Xie, N. Zhou, and S. Liu, "Study on the membrane orientation in FO
process," Sichuan Daxue Xuebao (Gongcheng Kexue Ban)/Journal of Sichuan
University (Engineering Science Edition), Article vol. 46, no. SUPPL.2, pp. 49-53,
2014.
[40] M. M. Motsa, B. B. Mamba, J. M. Thwala, and A. R. D. Verliefde, "Osmotic
backwash of fouled FO membranes: Cleaning mechanisms and membrane surface
properties after cleaning," Desalination, vol. 402, pp. 62-71, 2017.
[41] Y. Liu, "Fouling in FO Membrane Processes: Characterisation, Mechanisms and
Mitigation," Doctor of Philosophy, Department of Civil and Environmental
Engineering, University of Maryland, University of Maryland, 2013.
[42] C. Y. Tang, T. H. Chong, and A. G. Fane, "Colloidal interactions and fouling of NF
and RO membranes: a review," Adv Colloid Interface Sci, vol. 164, no. 1-2, pp. 126-
43, May 11 2011.
[43] F. Zhao et al., "Using axial vibration membrane process to mitigate membrane fouling
and reject extracellular organic matter in microalgae harvesting," Journal of
Membrane Science, vol. 517, pp. 30-38, 2016.
[44] Y. Gu, Y.-N. Wang, J. Wei, and C. Y. Tang, "Organic fouling of thin-film composite
polyamide and cellulose triacetate forward osmosis membranes by oppositely charged
macromolecules," Water Research, vol. 47, no. 5, pp. 1867-1874, 2013.
[45] Y. Liu and B. Mi, "Combined fouling of forward osmosis membranes: Synergistic
foulant interaction and direct observation of fouling layer formation," Journal of
Membrane Science, vol. 407-408, pp. 136-144, 2012.
76
[46] M. Zhang, H. Lin, L. Shen, B.-Q. Liao, X. Wu, and R. Li, "Effect of calcium ions on
fouling properties of alginate solution and its mechanisms," Journal of Membrane
Science, vol. 525, pp. 320-329, 2017.
[47] P. Zhao, B. Gao, Q. Yue, P. Liu, and H. K. Shon, "Fatty acid fouling of forward
osmosis membrane: Effects of pH, calcium, membrane orientation, initial permeate
flux and foulant composition," J Environ Sci (China), vol. 46, pp. 55-62, Aug 2016.
[48] P. Xiao, J. Li, Y. Ren, and X. Wang, "A comprehensive study of factors affecting
fouling behavior in forward osmosis," Colloids and Surfaces A: Physicochemical and
Engineering Aspects, vol. 499, pp. 163-172, 2016.
[49] M. Xie, L. D. Nghiem, W. E. Price, and M. Elimelech, "Impact of organic and
colloidal fouling on trace organic contaminant rejection by forward osmosis: Role of
initial permeate flux," Desalination, vol. 336, pp. 146-152, 2014.
[50] Y. Kim, S. Lee, H. K. Shon, and S. Hong, "Organic fouling mechanisms in forward
osmosis membrane process under elevated feed and draw solution temperatures,"
Desalination, vol. 355, pp. 169-177, 2015.
[51] J. R. McCutcheon, R. L. McGinnis, and M. Elimelech, "Desalination by ammonia–
carbon dioxide forward osmosis: Influence of draw and feed solution concentrations
on process performance," Journal of Membrane Science, vol. 278, no. 1–2, pp. 114-
123, 7/5/ 2006.
[52] M. Elimelech, Z. Xiaohua, A. E. Childress, and H. Seungkwan, "Role of membrane
surface morphology in colloidal fouling of cellulose acetate and composite aromatic
polyamide reverse osmosis membranes," Journal of Membrane Science, vol. 127, no.
1, pp. 101-109, 1997/04/30/ 1997.
[53] S. Qi, C. Q. Qiu, Y. Zhao, and C. Y. Tang, "Double-skinned forward osmosis
membranes based on layer-by-layer assembly—FO performance and fouling
behavior," Journal of Membrane Science, vol. 405-406, pp. 20-29, 2012.
[54] J. R. McCutcheon, R. L. McGinnis, and M. Elimelech, "A novel ammonia—carbon
dioxide forward (direct) osmosis desalination process," Desalination, vol. 174, no. 1,
pp. 1-11, 2005/04/01/ 2005.
[55] M. Xie, E. Bar-Zeev, S. M. Hashmi, L. D. Nghiem, and M. Elimelech, "Role of
Reverse Divalent Cation Diffusion in Forward Osmosis Biofouling," Environmental
Science & Technology, vol. 49, no. 22, pp. 13222-13229, 2015/11/17 2015.
[56] C. Y. Tang, Q. She, W. C. L. Lay, R. Wang, and A. G. Fane, "Coupled effects of
internal concentration polarization and fouling on flux behavior of forward osmosis
membranes during humic acid filtration," Journal of Membrane Science, vol. 354, no.
1-2, pp. 123-133, 2010.
[57] J. Zhang, W. L. C. Loong, S. Chou, C. Tang, R. Wang, and A. G. Fane, "Membrane
biofouling and scaling in forward osmosis membrane bioreactor," Journal of
Membrane Science, vol. 403-404, pp. 8-14, 2012.
[58] S. Zou et al., "Direct microscopic observation of forward osmosis membrane fouling
by microalgae: Critical flux and the role of operational conditions," Journal of
Membrane Science, vol. 436, pp. 174-185, 2013.
[59] G. T. Gray, J. R. McCutcheon, and M. Elimelech, "Internal concentration polarization
in forward osmosis: role of membrane orientation," Desalination, vol. 197, no. 1-3,
pp. 1-8, 2006.
77
[60] M. Xie, L. D. Nghiem, W. E. Price, and M. Elimelech, "Impact of humic acid fouling
on membrane performance and transport of pharmaceutically active compounds in
forward osmosis," Water Res, vol. 47, no. 13, pp. 4567-75, Sep 01 2013.
[61] A. P. Murphy, C. D. Moody, R. L. Riley, S. W. Lin, B. Murugaverl, and P. Rusin,
"Microbiological damage of cellulose acetate RO membranes," Journal of Membrane
Science, vol. 193, no. 1, pp. 111-121, 2001/10/31/ 2001.
[62] T. Nguyen, A. F. Roddick, and L. Fan, "Biofouling of Water Treatment Membranes:
A Review of the Underlying Causes, Monitoring Techniques and Control Measures,"
Membranes, vol. 2, no. 4, 2012.
[63] X. Jin, Q. She, X. Ang, and C. Y. Tang, "Removal of boron and arsenic by forward
osmosis membrane: Influence of membrane orientation and organic fouling," Journal
of Membrane Science, vol. 389, pp. 182-187, 2012.
[64] S. Lee and M. Elimelech, "Relating Organic Fouling of Reverse Osmosis Membranes
to Intermolecular Adhesion Forces," Environmental Science & Technology, vol. 40,
no. 3, pp. 980-987, 2006/02/01 2006.
[65] B. Mi and M. Elimelech, "Chemical and physical aspects of organic fouling of
forward osmosis membranes," Journal of Membrane Science, vol. 320, no. 1, pp. 292-
302, 2008/07/15/ 2008.
[66] D. Van der Kooij, Hijnen W.A.M., Comelissen E.R., "Elucidation of Membrane
Biofouling Processes Using Bioassays for Assessing the Microbial Growth Potential
of Feedwater," presented at the Membrane Technology Conference 2007.
[67] Y. Chun, D. Mulcahy, L. Zou, and I. S. Kim, "A Short Review of Membrane Fouling
in Forward Osmosis Processes," Membranes (Basel), vol. 7, no. 2, Jun 12 2017.
[68] X. Li, Y. Mo, J. Li, W. Guo, and H. H. Ngo, "In-situ monitoring techniques for
membrane fouling and local filtration characteristics in hollow fiber membrane
processes: A critical review," Journal of Membrane Science, vol. 528, pp. 187-200,
2017.
[69] L. N. Sim and A. G. Fane, "Advanced Monitoring of Membrane Fouling and Control
Strategies," in Reference Module in Chemistry, Molecular Sciences and Chemical
Engineering: Elsevier, 2017.
[70] E. W. Tow, M. M. Rencken, and J. H. Lienhard, "In situ visualization of organic
fouling and cleaning mechanisms in reverse osmosis and forward osmosis,"
Desalination, vol. 399, pp. 138-147, 2016.
[71] S. Jamaly, N. N. Darwish, I. Ahmed, and S. W. Hasan, "A short review on reverse
osmosis pretreatment technologies," Desalination, vol. 354, pp. 30-38, 2014.
[72] A. Sweity, Y. Oren, Z. Ronen, and M. Herzberg, "The influence of antiscalants on
biofouling of RO membranes in seawater desalination," Water Research, vol. 47, no.
10, pp. 3389-3398, 2013/06/15/ 2013.
[73] Q. She, Y. K. W. Wong, S. Zhao, and C. Y. Tang, "Organic fouling in pressure
retarded osmosis: Experiments, mechanisms and implications," Journal of Membrane
Science, vol. 428, pp. 181-189, 2013/02/01/ 2013.
[74] X. Lu, S. Romero-Vargas Castrillon, D. L. Shaffer, J. Ma, and M. Elimelech, "In situ
surface chemical modification of thin-film composite forward osmosis membranes for
enhanced organic fouling resistance," Environ Sci Technol, vol. 47, no. 21, pp. 12219-
28, 2013.
78
[75] A. Nguyen, S. Azari, and L. Zou, "Coating zwitterionic amino acid l-DOPA to
increase fouling resistance of forward osmosis membrane," Desalination, vol. 312, pp.
82-87, 2013.
[76] A. Nguyen, L. Zou, and C. Priest, "Evaluating the antifouling effects of silver
nanoparticles regenerated by TiO2 on forward osmosis membrane," Journal of
Membrane Science, vol. 454, pp. 264-271, 2014.
[77] S. Romero-Vargas Castrillón, X. Lu, D. L. Shaffer, and M. Elimelech, "Amine
enrichment and poly(ethylene glycol) (PEG) surface modification of thin-film
composite forward osmosis membranes for organic fouling control," Journal of
Membrane Science, vol. 450, pp. 331-339, 2014.
[78] A. Tiraferri, Y. Kang, E. P. Giannelis, and M. Elimelech, "Superhydrophilic thin-film
composite forward osmosis membranes for organic fouling control: fouling behavior
and antifouling mechanisms," Environ Sci Technol, vol. 46, no. 20, pp. 11135-44, Oct
16 2012.
[79] H.-Y. Yu, Y. Kang, Y. Liu, and B. Mi, "Grafting polyzwitterions onto polyamide by
click chemistry and nucleophilic substitution on nitrogen: A novel approach to
enhance membrane fouling resistance," Journal of Membrane Science, vol. 449, pp.
50-57, 2014.
[80] B. G.-F. Jose Miguel Arnal, Maria Sancho, "Membrane cleaning, Expanding issues in
desalination ", I. corporation, Ed., ed, 2011.
[81] B. Mi and M. Elimelech, "Organic fouling of forward osmosis membranes: Fouling
reversibility and cleaning without chemical reagents," Journal of Membrane Science,
vol. 348, no. 1-2, pp. 337-345, 2010.
[82] G. Blandin, H. Vervoort, P. Le-Clech, and A. R. D. Verliefde, "Fouling and cleaning
of high permeability forward osmosis membranes," Journal of Water Process
Engineering, vol. 9, pp. 161-169, 2016.
[83] E. Arkhangelsky, F. Wicaksana, S. Chou, A. A. Al-Rabiah, S. M. Al-Zahrani, and R.
Wang, "Effects of scaling and cleaning on the performance of forward osmosis hollow
fiber membranes," Journal of Membrane Science, vol. 415-416, pp. 101-108, 2012.
[84] T. Ergön-Can, B. Köse-Mutlu, İ. Koyuncu, and C.-H. Lee, "Biofouling control based
on bacterial quorum quenching with a new application: Rotary microbial carrier
frame," Journal of Membrane Science, vol. 525, pp. 116-124, 2017.
[85] H. Petrus, H. Li, V. Chen, and N. Norazman, "Enzymatic cleaning of ultrafiltration
membranes fouled by protein mixture solutions," Journal of Membrane Science, vol.
325, no. 2, pp. 783-792, 2008.
[86] M. F. Siddiqui, M. Rzechowicz, W. Harvey, A. W. Zularisam, and G. F. Anthony,
"Quorum sensing based membrane biofouling control for water treatment: A review,"
Journal of Water Process Engineering, vol. 7, pp. 112-122, 2015.
[87] H. Xu and Y. Liu, "Control and Cleaning of Membrane Biofouling by Energy
Uncoupling and Cellular Communication," Environmental Science & Technology, vol.
45, no. 2, pp. 595-601, 2011/01/15 2011.
[88] C. H. Yu, L. C. Fang, S. K. Lateef, C. H. Wu, and C. F. Lin, "Enzymatic treatment for
controlling irreversible membrane fouling in cross-flow humic acid-fed
ultrafiltration," J Hazard Mater, vol. 177, no. 1-3, pp. 1153-8, May 15 2010.
79
[89] R. C. M. a. C. P. B. Lucinda F. Silva, "Modification of polyamide reverse osmosis
membranes seeking for better resistance to oxidizing agents," Membrane Water
Treatment, vol. 3, pp. 169-179, 2012.
[90] A. A. Alturki, J. A. McDonald, S. J. Khan, W. E. Price, L. D. Nghiem, and M.
Elimelech, "Removal of trace organic contaminants by the forward osmosis process,"
Separation and Purification Technology, vol. 103, pp. 258-266, 2013/01/15/ 2013.
[91] A. J. Ansari, F. I. Hai, W. Guo, H. H. Ngo, W. E. Price, and L. D. Nghiem, "Factors
governing the pre-concentration of wastewater using forward osmosis for subsequent
resource recovery," Sci Total Environ, vol. 566-567, pp. 559-66, Oct 01 2016.
[92] G. Chen et al., "Treatment of shale gas drilling flowback fluids (SGDFs) by forward
osmosis: Membrane fouling and mitigation," Desalination, vol. 366, pp. 113-120,
2015/06/15/ 2015.
[93] Y. Chun, S.-J. Kim, G. J. Millar, D. Mulcahy, I. S. Kim, and L. Zou, "Forward
osmosis as a pre-treatment for treating coal seam gas associated water: Flux and
fouling behaviour," Desalination, vol. 403, pp. 144-152, 2017/02/01/ 2017.
[94] B. D. Coday et al., "The sweet spot of forward osmosis: Treatment of produced water,
drilling wastewater, and other complex and difficult liquid streams," Desalination, vol.
333, no. 1, pp. 23-35, 2014/01/15/ 2014.
[95] X.-M. Li et al., "Water reclamation from shale gas drilling flow-back fluid using a
novel forward osmosis–vacuum membrane distillation hybrid system," Water Science
and Technology, 10.2166/wst.2014.003 vol. 69, no. 5, p. 1036, 2014.
[96] S. Liyanaarachchi, V. Jegatheesan, I. Obagbemi, S. Muthukumaran, and L. Shu,
"Effect of feed temperature and membrane orientation on pre-treatment sludge volume
reduction through forward osmosis," Desalination and Water Treatment, vol. 54, no.
4-5, pp. 838-844, 2015/05/01 2015.
[97] N. C. Nguyen, H. T. Nguyen, S.-S. Chen, N. T. Nguyen, and C.-W. Li, "Application
of forward osmosis (FO) under ultrasonication on sludge thickening of waste activated
sludge," Water Science and Technology, 10.2166/wst.2015.341 vol. 72, no. 8, p. 1301,
2015.
[98] Z. Wang, J. Zheng, J. Tang, X. Wang, and Z. Wu, "A pilot-scale forward osmosis
membrane system for concentrating low-strength municipal wastewater: performance
and implications," Scientific Reports, Article vol. 6, p. 21653, 02/22/online 2016.
[99] W. Luo, M. Xie, F. I. Hai, W. E. Price, and L. D. Nghiem, "Biodegradation of
cellulose triacetate and polyamide forward osmosis membranes in an activated sludge
bioreactor: Observations and implications," Journal of Membrane Science, vol. 510,
no. Supplement C, pp. 284-292, 2016/07/15/ 2016.
[100] N. M. Mazlan et al., "Organic fouling behaviour of structurally and chemically
different forward osmosis membranes – A study of cellulose triacetate and thin film
composite membranes," Journal of Membrane Science, vol. 520, pp. 247-261, 2016.
[101] S. Zhao, L. Zou, and D. Mulcahy, "Effects of membrane orientation on process
performance in forward osmosis applications," Journal of Membrane Science, vol.
382, no. 1, pp. 308-315, 2011/10/15/ 2011.
[102] J. Kochan, T. Wintgens, R. Hochstrat, and T. Melin, "Impact of wetting agents on the
filtration performance of polymeric ultrafiltration membranes," Desalination, vol. 241,
no. 1, pp. 34-42, 2009/05/31/ 2009.
80
[103] J. R. McCutcheon and M. Elimelech, "Influence of membrane support layer
hydrophobicity on water flux in osmotically driven membrane processes," Journal of
Membrane Science, vol. 318, no. 1, pp. 458-466, 2008/06/20/ 2008.
[104] T. Y. Cath et al., "Standard Methodology for Evaluating Membrane Performance in
Osmotically Driven Membrane Processes," Desalination, vol. 312, no. Supplement C,
pp. 31-38, 2013/03/01/ 2013.
[105] E. M. V. Hoek, S. Bhattacharjee, and M. Elimelech, "Effect of Membrane Surface
Roughness on Colloid−Membrane DLVO Interactions," Langmuir, vol. 19, no. 11, pp.
4836-4847, 2003/05/01 2003.
[106] A. Poisson and A. Papaud, "Diffusion coefficients of major ions in seawater," Marine
Chemistry, vol. 13, no. 4, pp. 265-280, 1983/10/01/ 1983.
[107] L. Yuan-Hui and S. Gregory, "Diffusion of ions in sea water and in deep-sea
sediments," Geochimica et Cosmochimica Acta, vol. 38, no. 5, pp. 703-714,
1974/05/01/ 1974.
[108] M. Xie and S. R. Gray, "Gypsum scaling in forward osmosis: Role of membrane
surface chemistry," Journal of Membrane Science, vol. 513, no. Supplement C, pp.
250-259, 2016/09/01/ 2016.
[109] S. Daneshgar, A. Callegari, A. Capodaglio, and D. Vaccari, "The Potential Phosphorus
Crisis: Resource Conservation and Possible Escape Technologies: A Review,"
Resources, vol. 7, no. 2, 2018.
[110] D. C. J.J. Schroder, A.L. Smit and A. Rosemarin, "Sustainable Use of Phosphorus ",
Netherland2010.
[111] E. Desmidt et al., "Global Phosphorus Scarcity and Full-Scale P-Recovery
Techniques: A Review," Critical Reviews in Environmental Science and Technology,
vol. 45, no. 4, pp. 336-384, 2015/02/16 2015.
[112] T. Schütte, C. Niewersch, T. Wintgens, and S. Yüce, "Phosphorus recovery from
sewage sludge by nanofiltration in diafiltration mode," Journal of Membrane Science,
vol. 480, pp. 74-82, 2015/04/15/ 2015.
[113] M. T. Vu, A. J. Ansari, F. I. Hai, and L. D. Nghiem, "Performance of a seawater-
driven forward osmosis process for pre-concentrating digested sludge centrate: organic
enrichment and membrane fouling," Environmental Science: Water Research &
Technology, 10.1039/C8EW00132D vol. 4, no. 7, pp. 1047-1056, 2018.
[114] L. E. de-Bashan and Y. Bashan, "Recent advances in removing phosphorus from
wastewater and its future use as fertilizer (1997–2003)," Water Research, vol. 38, no.
19, pp. 4222-4246, 2004/11/01/ 2004.
[115] K. S. Le Corre, E. Valsami-Jones, P. Hobbs, and S. A. Parsons, "Phosphorus Recovery
from Wastewater by Struvite Crystallization: A Review," Critical Reviews in
Environmental Science and Technology, vol. 39, no. 6, pp. 433-477, 2009/06/01 2009.
[116] A. T. K. Tran et al., "P-recovery as calcium phosphate from wastewater using an
integrated selectrodialysis/crystallization process," Journal of Cleaner Production,
vol. 77, pp. 140-151, 2014/08/15/ 2014.
[117] A. Ali, A. C. Quist-Jensen, F. Macedonio, and E. Drioli, "Application of Membrane
Crystallization for Minerals’ Recovery from Produced Water," Membranes, vol. 5, no.
4, 2015.
[118] C. Quist-Jensen, A. Ali, S. Mondal, F. Macedonio, and E. Drioli, Membrane
crystallization for minerals recovery from brine. 2015.
81
[119] B. Mi and M. Elimelech, "Silica scaling and scaling reversibility in forward osmosis,"
Desalination, vol. 312, no. Supplement C, pp. 75-81, 2013/03/01/ 2013.
[120] C. Eskicioglu, G. Galvagno, and C. Cimon, "Approaches and processes for ammonia
removal from side-streams of municipal effluent treatment plants," Bioresource
Technology, vol. 268, pp. 797-810, 2018/11/01/ 2018.