2016 Jieun Lee Final Thesis

Transcription

Carbon nanotube-enhanced membrane
for advanced water treatment
A thesis submitted in fulfillment of the requirements for the
degree of Doctor of Philosophy
By
Jieun Lee
School of Chemical and Biomolecular Engineering
Faculty of Engineering and Information Technologies
The University of Sydney
Mar 2016
Declaration
I hereby declare that this submission is my own work and that, to the best of my knowledge and
belief, it contains no material previously published or written by another person nor material
which to a substantial extent has been accepted for the award of any other degree or diploma of
the university or other institute of higher learning, except where due acknowledgment has been
made in the text.
Jieun Lee
March 2016
i
“For at the proper time we will reap a harvest if we do not give up” Galathian 6:9
愚公移山, 孔子
ii
Acknowledgement
I would like to express my sincere gratitude to those who supported me throughout my Ph.D.
program. I sincerely appreciate the Australian Government of an opportunity to do Ph.D.
research in Australia with 2012 Kapyong Commemorative Scholarship postgraduate Award. It
has been such a fantastic and unforgettable research journey like a slow growing eucalypt tree. I
would like to thank my scholarship manager Ms. Rachel Tyne and Ms. Monika Marzec in Scope
global for their wholehearted support and cares for 4 years.
I appreciate my supervisor A/Prof. Zongwen Liu of his invaluable support for research to
continue the Ph.D. program. I am also thankful to Dr.Soryong Chae for his guidance so that I
could build up my research niche. I sincerely appreciate Prof. Saravanamuth Vigneswaran at
UTS and Prof. Vicki Chen at UNSW for their support on research. Without their big help, I
would not have completed my thesis. I also would like to express my appreciation to Dr.Yun Ye
at UNSW for her big support for research. Further, I would like to thank my sincere friend, Dr.
Sanghyun Jeong at KAUST for his unconditional support for research. I appreciate A/Prof
Andrew Minett for his help and guidance. I would like to express my appreciation to Dr.Jeffrey
Shi, Dr. Tino Kausmann and Ms. Nancy Xie for their technical assistance. I am also grateful to
Dr. Cuifeng Zhou at USyd and Dr.Gayathri Gnaidu and Mr. Mohammed Johir at UTS for their
technical support. I would like to express my gratitude to Dr.Marie Mcinnes for her English
editing of my literature review in the thesis.
I would like to thank my group members- Dr. Tahereh, Noeiaghaei, Mr. Murat Cakici and Ms.
Farideh Heidarpour for their help. I also appreciate Mr. Peter Newman, Dr. Xiaoshuang Yang at
USyd, and Dr. Suwan Meng at UNSW for their help and support.
Lastly, I would like to express my gratitude to my family for their unconditional love, support
and patience over the entire period of Ph.D program.
iii
Publications
Research papers included in the thesis
1. Jieun Lee, Yun Ye, Antony J. Ward, Cuifeng Zhou, Vicki Chen, Andrew.I.Minett, Sanghyup
Lee, Zongwen Liu, Soryong Chae, Jeffrey Shi, “High flux and high selectivity carbon nanotube
composite membrane for natural organic matter removal”, Separation and purification
technology, 2016(163), 109-119. This article is presented in Chapter 4.
2. Jieun Lee, Sanghyun Jeong, Gayathri Naidu, Yun Ye, Vicki Chen, Zongwen Liu,
Saravanamuth Vigneswaran, “Carbon nanotube membrane can be alternative for sustainable
seawater pretreatment? ”, Journal of Industrial and Engineering Chemistry. 2016(38), 123-131.
This article is presented in Chapter 5.
3. Jieun Lee, Sanghyun Jeong, Yun Ye, Vicki Chen, Saravanamuth Vigneswaran, Zongwen Liu,
“Is Carbon nanotube-enhanced membrane feasible for bioprocess? - Protein fouling behavior of
positively
charged
membrane
under
different
ionic
strength
and
solution
pH”,
Chemical Engineering Journal, (Under review). This article is presented in Chapter 6.
Review paper included in the thesis
Jieun Lee, Sanghyun Jeong, Zongwen Liu, (2016). Progress and challenges of carbon nanotube
membrane in water treatment, Critical reviews in Environmental Science and Technology (In
press). This article is presented in Chapter 2.
iv
Relavant Papers not included in the thesis
1. Jieun Lee, Yongshun Zhang, Ramireddy S.P. Raj Reddy, Zongwen Liu (2016), “Effective
natural organic matter removal in pond water by carbon nanotube enhanced membrane”, Journal
of Chemical Engineering, (To be Submitted)
2. Xiaoshuang Yang, Jieun Lee, Lixian Yuan, Soryong Chae, Vanessa K. Peterson, Andrew I.
Minett, Yongbai Yin, Andrew T. Harris., “Removal of natural organic matter in water using
functionalized carbon nanotube bucky paper”, Carbon, 2013(59), 160-166.
Conference Proceedings
1. Jieun Lee, Ramireddy S.P. Raj Reddy,Yongshun Zhang, Jeffrey Shi, Zongwen Liu, “High
flux and high selectivity carbon nanotube membrane-coagulation system for pondwater
treatment”, World Water Congress Exhibition, IWA, 2016.
2. Jieun Lee, Sanghyun Jeong, Gayathri Naidu, Cuifeng Zhou, Zongwen Liu, Saravanamuthu
Vigneswaran, “Carbon nanotube-enhanced membrane for seawater pretreatment”, accepted as a
poster presentation at the IDA World Congress on Desalination and Water Reuse, Aug 2015.
3. Jieun Lee, Peter Newman, Andrew I. Minett, Yun Ye, Vicki Chen, Zongwen Liu, “Carbon
nanotube enhanced membrane for water purification’, accepted as a poster presentation at the
Nanotechnology Entrepreneuship Workshop for Early Career Researchers, June 2015
4. Jieun Lee, Peter Newman, A I. Minett, Andrew T. Harris, Sanghyup Lee, Soryong Chae,
“Carbon nanotube/Polyaniline/Polyethersulfone membrane for enhanced removal of Natural
Organic Matter in water”, accepted as an oral presentation at The 10th International Congress on
Membranes and Membrane Processes (ICOM), 2014.
v
Abstract
Membrane technology has been extensively used in advanced water treatment as a response to
the global demand for drinking water due to the global water shortage. Driven by its separation
performance, membrane technology is able to apply to municipal/wastewater treatment, drinking
water treatment and seawater desalination. Despite its benefits such as energy efficiency and
high performance in separation, the challenging issue in currently used polymeric membrane is a
trade-off between permeate flux and separation efficiency, and vulnerability to fouling. Such
drawbacks made progress on developing membrane material for the purpose of advancing
membrane process. For example, incorporation of nanomaterials into a membrane matrix has
brought attention to its favorable characteristics. There have been many studies on the
development of membrane materials using nanomaterials such as carbon nanotubes (CNTs),
metal organic frameworks (MOFs) or silver. Among those candidates, CNTs possess advantages
such as high adsorption capacity for contaminant removal and hydrophobicity for slippage effect
in the water channel. Further, CNTs as a nanofiller in the polymer matrix are able to engineer the
membrane structure to the way in which is favorable to the flux and selectivity. However,
dispersion in a polymer matrix to maximize the benefits of CNTs still remains a challenge which
prevents its wide application to the membrane process.
First, we synthesized multiwall carbon nanotubes/polyaniline (MWCNTs/PANI) complex by insitu polymerisation and successfully introduced it to the membrane matrix. Doping process
during MWCNTs/PANI complex synthesis via chemical oxidation transformed the charge of
PES polymer to a positive one. The MWCNTs/PANI enhanced membrane exhibited remarkably
enhanced performance compared to the conventional polyethersulfone (PES) membrane. The
vi
significance of this research is that the membrane removed natural organic matter 4 fold (80 %)
which cannot be effectively removed by currently used low-pressure driven membranes, as well
as delivering 30 times higher water product (1400 LMH/bar) than the PES membrane. Further
research on physical/chemical properties of the membrane revealed that the hydroxyl
functionalized MWCNTs/PANI complex increased the hydrophilicity, porosity and slip length of
the membrane by improving the membrane structure. It also altered the surface charge of the
membrane, which contributed to the increase electrostatic interaction between the membrane
surface and the negatively charged organic matter, compared that currently used polymeric
membranes are negatively charged. Due to the positively charged surface, this membrane can
open
up
an
opportunity
for
recovery
of
valuable
cationic
macromolecules
in
bioprocess/pharmaceutical industries. Examination of removal mechanism and fast water flux
would be of great importance for enhancing the performance of UF membrane which is widely
used in water treatment. The membrane exhibited 100 % UPW permeability recovery and 65 %
total fouling ratio after acid/base cleaning.
A second part of the research targeted to the application for enhanced organic matter removal in
seawater pre-treatment. The research was aimed to enhance UF membrane coupled with
adsorption system which can prevent performance decline in RO membrane desalination plant.
By enhanced natural organic matter removal and delivering high permeate flux, the MWCNTs
enhanced membrane contributed to minimizing sludge volume generated from seawater
desalination plant as removing organic matter in seawater. Compared that conventional PES
membrane required 1.5 g/L PAC, the MWCNTs membrane coupled PAC adsorption system
exhibited enhanced removal efficiency at 0.5 g/L PAC, and 4 fold increased permeate flux. The
vii
ionic strength was found to influence reduced organic removal and permeate flux enhancement
in MWCNTs membrane.
Lastly, protein fouling behavior in the MWCNTs membrane was examined under different
solution chemistry (pH and ionic strength) for the wide application of MWCNTs membrane in
wastewater reclamation and recovery of bioprocess industry. MWCNTs membrane showed high
permeate flux at different pH and ionic strength compared to the commercial PES membrane.
Four different fouling models were used to explain the flux decline with protein deposition
during membrane filtration. For MWCNTs membrane, two dominant fouling mechanismsstandard blocking and cake filtration operated simultaneously under different pH and increasing
ionic strength throughout the entire filtration. In a single protein filtration test, Lys filtration
caused more strong fouling potential than BSA filtration due to its comparable size to the pore
diameter of the membrane. Low pH condition alleviated standard blocking and cake layer
formation on the membrane due to increased electrostatic repulsion between the membrane
surface and Lys. Further, increasing ionic strength lessened the protein deposition because of ion
shielding effect on the model protein. For oppositely charged BSA filtration, membrane fouling
potential was weakened at pH 4.7 and 10.4 due to the decreased electrostatic interaction at these
points. Increasing ionic strength seemed to contribute to fouling mitigation, but not comparable
to the pH effect. For mixed protein filtration, severe pore blocking and cake layer formation were
observed at their IEP (4.7 and 10.4), presumably due to the enhanced bridging effect of
neutralized proteins with their counter proteins, compared to the intermolecular interaction (LysBSA) at pH 7. Consequently, protein fouling of MWCNTs membrane was found to be alleviated
by controlling electrostatic interaction with charged proteins via solution chemistry (pH and
ionic strength).
viii
Table of Contents
Chapter 1. Introduction.…………………………………………………………………………….1
1. Introduction …………………………………………………………………………………….1
1.1 Background…………………..………………………………………………………………..1
1.2 Scope and purpose of the study…………………………………….…………………………4
1.3 Thesis overview ………………………………………………………………………………6
Chapter 2.Literature Review .....……………………...………….……………………………….11
2.1 Introduction…………………………………………….…………………………………..………….11
2.2 CNT materials …….………………………………………………………………………....13
2.3 Fabrication of CNT membranes……………………………………………………………...16
2.3.1 Mixed matrix membrane – CNT/polymer composite membrane ………………………..16
2.3.2 Horizontally aligned carbon nanotube (HACNT) membrane……………………………..19
2.3.3 Vertically aligned carbon nanotube (VACNT) membrane ………………………………20
2.3.4 Key factors affecting the ultrafast water permeability of VACNT membrane …………..24
2.3.4.1 Slip length………………………………………………………………………………..24
2.3.4.2 Modification of CNT membrane for channels of water and target molecules…………..32
2.3.4.3 Challenges in VACNT membrane……………………………………………………….38
2.4 Application of CNT membranes in water treatment …….…………………………………..39
2.4.1 NOM removal for surface water treatment and seawater pretreatment ………………….39
2.4.1.1 Surface water treatment
……………………………………………………………….39
2.4.1.2 Seawater pretreatment …………………………………………………………………...43
2.4.2 Salt rejection: from brackish water treatment to MD/RO concentrate recycling………….44
2.4.3 Anti-fouling effect…………………………………………………………………………49
2.4.4 How the CNTs enhanced membrane contributed to the membrane system? ……………..51
ix
2.4.5 Hybrid material with CNT as a new generation of membrane material …………………..53
2.4.5.1 CNT/GO and CNT/Graphene …………………………………………………………..53
2.4.5.2 CNT/TiO2 ……………………………………………………………………………….55
2.4.5.3 CNT/PANI and CNT/PEDOT …………………………………………………………..56
2.5 Current challenges and future perspectives …………………………………………………57
Chapter 3. Materials and Method ………………………………………….…………………..76
3.1. Materials ……………………………………………………………………………………76
3.2. Fabrication of MWCNTs/PANI/PES membranes ………………………………………….76
Chapter 4. Fabrication of carbon nanotube-enhanced membrane for natural
organic matter removal …………………………………..………………………………….80
4.1. Introduction …………………………………………………………………………………80
4.2. Materials and methods ……………………………………………………………………...83
4.2.1. Materials ………………………………………………………………………………….83
4.2.2. Fabrication of membranes ………………………………………………………………..83
4.2.3 Characterization of MWCNTs/PANI/PES membranes …………………………………..83
4.2.4 Evaluation of filtration performance ………………………………………………………84
4.2.5 Chemical cleaning efficiency ……………………………………………………………..86
4.3. Results ………………………………………………………………………………………88
4.3.1 Significance of in-situ polymerization as a membrane fabrication procedure ……………88
4.3.1.1 Dispersity of MWCNTs …………………………………………………………………88
4.3.1.2 Effects of in situ polymerization fabrication on enhanced membrane performance ……92
4.3.2 Membrane performance of MWCNTs/PANI/PES membranes by MWCNTs
concentration……………………………………………………………………........................94
4.3.3 Physical and chemical properties of membrane …………………………………………...96
4.3.3.1 Physical properties: porosity, pore size and its size distribution ………………………..96
x
4.3.3.2 Chemical properties: hydrophilicity and zeta potential ………………………………..102
4.4. Discussion …………………………………………………………………………………105
4.4.1 Why fast water flux through the MWCNTs/PANI/PES membrane? ……………………105
4.4.2 NOM removal mechanism by the MWCNTs/PANI/PES membrane ……………………109
4.5. Flux recovery by chemical cleaning ………………………………………………………111
4.6. Conclusion ………………………………………………………………………………...114
Chapter 5. Application of carbon nanotube-enhanced membrane for seawater
pretreatment and examination of salinity on the membrane performance ……..…121
5.1 Introduction ………………………………………………………………………………..121
5.2. Materials and methods …………………………………………………………………….124
5.2.1 Materials …………………………………………………………………………………124
5.2.1.1 Seawater ………………………………………………………………………………..124
5.2.1.2 Membranes ……………………………………………………………………………..124
5.2.1.3 Powder activated carbon (PAC) ………………………………………………………..125
5.2.2 Membrane filtration test ………………………………………………………………….125
5.2.3 Dissolved organic matter measurement ………………………………………………….125
5.2.4 Surface charge of membrane ...…………………………………………………………..126
5.2.5 Fouled membrane characterization ...…………………………………………………….126
5.3. Results and discussion …………………………………………………………………….127
5.3.1 Seawater filtration performance (without PAC addition) ...……………………………...127
5.3.1.1 Permeate flux …………………………………………………………………………..127
5.3.1.2 Organic reduction ………………………………………………………………………128
5.3.2 Effect of PAC addition …………………………………………………………………..129
5.3.3 Effect of salinity and divalent cation on HA rejection ………………………………......133
5.3.4 Factors affecting membrane performance ……………………………………………….137
xi
5.3.4.1 Surface charge of membrane under different ionic strength – organic removal ……….137
5.3.4.2 Chemical precipitation on membrane – filtration behavior ……………………………138
5.3.4.3 Membrane hydrophilicity (water contact angle) ……………………………………….141
5.4. Conclusion ………………………………………………………………………………...143
Chapter 6. Protein fouling in carbon nanotubes enhanced ultrafiltration membrane:
fouling mechanism as a function of pH and ionic strength……………………………147
6.1. Introduction ………………………………………………………………………………..147
6.2. Materials and methods …………………………………………………………………….149
6.2.1 Materials …………………………………………………………………………………149
6.2.1.1 Model protein foulants and chemicals …………………………………………………149
6.2.1.2 Membranes ……………………………………………………………………………..149
6.2.2 Fouling experiments ……………………………………………………………………...150
6.2.3 Membrane characterization ………………………………………………………………150
6.2.4 Membrane fouling model ………………………………………………………………...150
6.3. Results and discussion …………………………………………………………………….152
6.3.1 Effects of solution chemistry on surface charge of the membranes ……………………..152
6.3.2 Effects of pH on permeate flux with single proteins …………………………………….154
6.3.2.1 Permeate flux behavior in positively charged protein (Lys) …………………………..154
6.3.2.2 Permeate flux behavior in negatively charged protein (BSA) …………………………160
6.3.3 Effect of ionic strength on permeate flux with single protein …………………………...164
6.3.3.1 Permeate flux pattern with positively charged protein (Lys) ………………………….164
6.3.3.2 Permeate flux behavior with negatively charged protein (BSA) ………………………166
6.3.4 Effect of solution chemistry on permeate flux with binary mixture ……………………..169
6.3.4.1 Permeate flux pattern as a function of pH ……………………………………………..169
6.3.4.2 Permeate flux pattern as a function of ionic strength ………………………………….170
xii
6.3.5 Fouling mechanisms under solution chemistry ………………………………………….173
6.3.5.1 MWCNT membrane …………………………………………………………………...173
6.3.5.2 PES-UF membrane …………………………………………………………………….178
6.4. Conclusion ………………………………………………………………………………...181
Chapter 7. Conclusion …………………………………………………………………......181
xiii
Table of Figures
Fig 1.1. (a) Comparison of water demand in the future and current supply [18] and (b) Diagram
of alternative water resource [19]…………………………………………………………………4
Fig 1.2 Application of MWCNTs-enhanced membrane to the membrane processes…………….5
Fig 2.1 Structure of CNTs: (a) a schematic representation of a graphene sheet and CNT roll-up
vector, (b) a 3-D model of single-wall CNT (SWCNTs), and (c) hybridization states of carbonbased nanomaterials [21, 22]…………………………………………………………………….13
Fig 2.2 Schematic diagrams of (a) single-wall carbon nanotubes (SWCNTs), (b) multi-wall
carbon nanotubes (MWCNTs), (c) double wall carbon nanotubes, and (d) peapod nanotubes
consisting of SWCNTs filled with fullerenes (e.g. C60) [25]…………………………………...14
Fig 2.3 Schematic illustrations of the CNTs composite membrane fabrication procedure: (a)
CNTs UF membrane by phase inversion [44] and (b) CNTs/PA NF/RO membrane by interfacial
polymerization…………………………………………………………………………………...18
Fig 2.4 SEM images of (a) HACNT [45] and (b) VACNT arrays [46]…………………………19
Fig 2.5 Schematic of the synthesis and detachment of a VACNTs array [53]…………………..20
Fig 2.6 Fabrication procedure of VACNTs membrane………………………………………….21
Fig 2.7 Slip length flow past a stationary surface [65]…………………………………………..25
Fig 2.8(a) Concept of tip functionalization of VACNT membrane and (b) oxidation process of
CNTs using acid and oxidative gas [31]…………………………………………………………33
Fig 2.9 Classification of fabrication method in membrane process…………………………......51
Fig 3.1 Synthesis of (a) MWCNTs/PANI nanocomposite and (b) fabrication procedure of
MWCNTs/PANI/PES composite membrane. ……………………………………………………78
Fig 3.2 Preparation of polymer membrane by phase inversion………………………………….78
Fig 4.1 Snapshot shows dispersed MWCNTs/PANI in NMP over 24 h………………………...88
Fig 4.2 UV-Vis spectra of (a) the MWCNTs/PANI complex and MWCNTs in NMP, and (b)
absorbance at 295 nm for 10 hours………………………………………………………………90
Fig 4.3 SEM cross-sectional images of MWCNTs dispersion in PES polymer matrix: Crosssectional images of (a) MWCNTs/PANI/PES membrane: MWCNTs are well-dispersed into the
xiv
PES polymer and (b) MWCNT/PES-1 (simply blended) membrane: MWCNTs are aggregated in
the PES matrix…………………………………………………………………………………...91
Fig 4.4 Top down views of MWCNTs/PES-1 (simply blended) membrane show that MWCNTs
are protruded from PES matrix………………………………………………………………….92
Fig 4.5 Effect of MWCNTs/PANI complex on the membrane performance – (a) water
permeability, (b) HA removal (1- CNTs simply blended, 2 – membrane). MWCNTs/PANI
membrane exhibited 30 fold enhanced water permeability and effective HA removal up to 80 %,
compared to PES and PANI/PES membrane. MWCNTs simply blended in PES membrane did
not show enhanced separation properties………………………………………………………..93
Fig 4.6 Effect of MWCNT proportion on (a) water permeability and (b) HA removal efficiency
as a function of MWCNT concentration (%). ……………………………………………………95
Fig 4.7 SEM images of the PES membrane, PANI membrane and MWCNTs/PANI/PES
membrane (a) Top down view, (b) Overall cross-sectional structure, (c) Cross section of top
layer………………………………………………………………………………………………98
Fig 4.8 Pore size distribution of membranes. (PANI/PES membrane: [A], 0.25, 0.5, 1, 1.5 and 2
wt% MWCNT/PANI/PES membranes: [B-1], [B-2], [B-3], [B-4] and [B-5])………………...101
Fig 4.9 Effect of MWCNTs concentration on the hydrophilicity of the MWCNTs/PANI/PES
membrane………………………………………………………………………………………102
Fig 4.10 Zeta potential of the membrane by streaming potential at pH 5.6 at a background
electrolyte concentration of 0.001 M KCl (Zeta potential of 5 ppm HA solution = - 7.5 mV).
Insertion of MWCNTs in the membrane altered to positively-surface charged membrane……104
Fig 4. 11 Water permeability as a function of porosity by Kozeny-Carmen relationship……..106
Fig 4.12 (a) HA permeability decline behaviour and (b) HA removal decline behaviour of both
the B-4 and the S membranes (684.9 L/m2 filtration)…………………………………………110
Fig 4.13 HA permeate flux decline and its removal rate behaviour of [B-4] membrane with
chemical cleaning during 3 cycles of filtration test: HA permeability was recovered for 3 cycles
with the chemical cleaning (method 4 presented in Table 5). (The arrow indicates the time when
chemical cleaning was conducted)……………………………………………………………...113
Fig 5.1 Permeability pattern of MWCNTs and PES-UF membrane in UPW and seawater
filtration…………………………………………………………………………………………128
Fig 5.2 Permeability decline pattern in MWCNTs and PES membrane filtration of seawater with
different PAC doses…………………………………………………………………………….131
xv
Fig 5.3 Permeability of membranes with HA 5mg/L (ppm) under the different NaCl
concentrations…………………………………………………………………………………..135
Fig 5.4 HA rejection (%) of membranes with HA 5mg/L (ppm) under the different NaCl
concentrations…………………………………………………………………………………..136
Fig 5.5 Permeability and HA rejection (%) of membranes with HA 5 ppm under 0.01M of
CaCl2……………………………………………………………………………………………136
Fig 5.6 Zeta potential of membranes under different ionic strengths by streaming potential at pH
7.8 and different KCl (background electrolyte) concentrations (0.001, 0.005 and 0.010 M)…..138
Fig 5.7 SEM images and elements of the membrane surface by EDS analysis before and after
seawater filtration (Filtered volume = 684.9 L/m2): (a) MWCNTs enhanced membrane, and (b)
PES membrane………………………………………………………………………………….141
Fig 6.1 Zeta potential of MWCNT membrane and PES membrane (a) at different pHs (4.7, 7.0
and 10.4) and (b) different ionic strengths by adding different concentrations of KCl (background
electrolyte) (0.001, 0.005 and 0.010 M) at pH 7.0……………………………………………...153
Fig 6.2 Comparison of permeate flux pattern of MWCNT and PES-UF membrane filtration with
positively charged protein (Lys) and negatively charged protein (BSA) under different pHs (4.7,
7.0 and 10.4)……………………………………………………………………………………155
Fig 6.3 Comparison of permeability pattern with positively charged protein (Lys) and negatively
charged protein (BSA) fouling under different ionic strength (concentration, monovalent/divalent
ionic strength) (0.001, 0.01, 0.1 M NaCl)………………………………………………………168
Fig 6.4 Comparison of permeability pattern with binary mixture fouling under different pH (4.7,
7.0, 10.4) and ionic strength (NaCl 0.001, 0.01 and 0.1 M)……………………………………172
Fig 6.5 Fouling mechanisms in the MWCNT membrane under different solution chemistry (a)
Lys protein filtration, (b) BSA filtration and (c) Mixed protein filtration……………………..177
Fig 6.6 Fouling mechanisms in the PES membrane under different solution chemistry (a) Lys
protein filtration, (b) BSA filtration and (c) Mixed protein filtration………………………….180
xvi
Lists of Tables
Table 2.1 Physical properties of different carbon materials [24]………………………………..14
Table 2.2 Summary of fabrication methods of VACNTs membrane…………………………...22
Table 2.3 Theory of slip length calculation……………………………………………………...27
Table 2.4 Methods of slip length measurement…………………………………………………29
Table 2.5 Slip length of CNTs membrane……………………………………………………….31
Table 2.6 Summary of hydrophilic treatments of CNTs………………………………………...36
Table 2.7 Summary of HA removal efficiency by CNTs membranes…………………………..42
Table 2.8 Summary of VACNT in water treatment……………………………………………..47
Table 2.9 Summary of salt removal efficiency by CNTs membranes…………………………..48
Table 2.10 Enhanced performance by the CNTs insertion in polymer membrane……………...53
Table 4.1 Parameters in the Kozeny-Carmen equation………………………………………….86
Table 4.2 Chemical cleaning conditions………………………………………………………...87
Table 4.3 Porosity and pore size of membranes…………………………………………………99
Table 4.4 Enhancement factor and slip length of membranes…………………………………108
Table 4.5 The results shows a comparison of water flux discovery, total fouling ratio, and
removal efficiency by cleaning method. By applying method 4, ultra-pure water flux recovery
reached to 100 %, and its total fouling ratio was 65 %. Based on the results, cleaning method (4)
is the most effective method for fouling mitigation…………………………………………….112
Table 5.1 Characteristics of the membranes used in this study………………………………..124
Table 5.2 Removal efficiencies of HS in seawater and 5 mg/L of HA in UPW by MWCNTs and
PES-UF membrane filtration (HS and HA concentrations were measured using LC-OCD and
TOC analyzer, respectively)……………………………………………………………………128
Table 5.3 Concentration and removal efficiency (R) of organic fractions by MWCNTs enhanced
and PES-UF membrane with and without PAC adsorption……………………………………132
Table 5.4 Changes of water contact angle of the membrane before and after filtration (filtered
volume = 684.9 L/m2)………………………………………………………………………….142
xvii
Table 6.1 Information of the membranes used in this chapter…………………………………150
Table 6.2 Fouling mechanism equations……………………………………………………….151
Table 6.3 Comparison of the correlation coefficient (R2) in the MWCNT membrane filtration
with proteins under different solution chemistry at (a) the initial stage of filtration (2h) and (b)
final stage of filtration (2 to 12h)………………………………………………………………156
Table 6.4 Comparison of fouling potential in the MWCNT membrane filtration with proteins
under different solution chemistry at (a) the initial stage of filtration (2h) and (b) final stage of
filtration (2 to 12h)……………………………………………………………………………..158
Table 6.5 Comparison of the correlation coefficient (R2) in the PES membrane filtration with
proteins under different solution chemistry at (a) the initial stage of filtration (2h) and (b) final
stage of filtration (2 to 12h)…………………………………………………………………….160
Table 6.6 Comparison of fouling potential in the PES-UF membrane filtration with proteins
under the different solution chemistry (a) at the initial stage (2h) and (b) later stage of
filtration…………………………………………………………………………………………163
xviii
Chapter 1 Introduction
1. Introduction
1.1 Background
Water shortage is an emerging issue due to a significant gap between available water supply
(4,200 billion m3) and water demand (6,900 billion m3) in 2030 as shown in Fig 1.1(a). To
address the issue for the future, it is imperative to find new water resources and recycle current
one by increasing water quality. In such a way, seawater/brackish water desalination and
recycling of waste/municipal water will be the critical resources to produce clean water (Fig
1.1(b)).
Membrane technology has been widely applied for the water/wastewater treatment and
desalination due to its separation performance and energy intensiveness over the last decade.
Low pressure driven membrane such as microfiltration (MF) and ultrafiltration (UF) is an
effective process for drinking water and membrane bioreactor in municipal/wastewater treatment
at a relatively low cost. Reverse osmosis membrane is increasingly adopted in a large number of
desalination plants due to its highly competitive performance, energy efficiency and cost saving
among various desalination technologies [1, 2].
Despite its benefits, such membrane processes still have several drawbacks in the application to
water treatment. Natural organic matter (NOM) is the major pollutant which causes serious
membrane fouling and performance decline. Further, NOM generates disinfection by-products
(DBPs), carcinogen when it reacts with chlorine residual in tap water. However, conventional
MF/UF membranes such as polyethersulfone (PES), polysulfone (PSf) and polyvinylidene
1
fluoride (PVDF) is limited to NOM removal due to its relatively larger pore size than
humic/fulvic acid (<1 nm), exhibiting only 20 -50 % rejection. Moreover, RO membrane in
desalination plants such as polyamide (PA) and cellulose acetate (CA) has two challenges to be
overcome. The first challenge is concentrate recycling. RO desalination plants generate
increasing amount of concentrate from the process, which contains a high concentration of salts,
process chemicals and effluent organic matter [3]. Another issue on RO membrane is the need
for the capital and energy-intensive pretreatment due to the vulnerability in membrane fouling [4,
5]. Seawater contains low molecular organic matter such as humic/fulvic acids, carboxylic acid
and extracellular polymeric substances (EPS). Such contaminants result in serious membrane
performance decline by irreversible membrane fouling. While MF/UF membrane hybrid system
coupled with coagulation/adsorption seems to have effective removal efficiency, it generates a
large amount of chemical sludge due to a usage of the high amount of coagulant/adsorbent.
UF membranes have been increasingly applied to secondary effluent from municipal and
wastewater plant. However, effluent organic foulants containing proteins, humic/fulvic
substances cause serious membrane fouling and flux decline via pore blocking and/or pore
constructions. Such drawbacks then lead to frequent membrane replacement and increase in
chemical cleaning, which will have a negative impact on the environment. The protein fouling is
strongly affected by solution chemistry such as solution pH, ionic strength and its concentration
[6-8]. For this reason, UF membrane in municipal/wastewater treatment is severely fouled by
feed water pH and ionic strength.
The attempt to address those challenging issues makes progress on developing membrane
material by nanomaterial incorporation [9-12]. Of nanomaterials, carbon nanotubes (CNTs) are
2
one of the favored candidates to the membrane for advanced water treatment due to its unique
properties such as an excellent adsorption capacity by high specific surface area, electroconductivity, slippage effect on permeate flux by the molecular smoothness of nanotube wall
(hydrophobicity) and easiness to functionalization [13-16]. In addition to its intrinsic
characteristics, incorporation of CNTs alters physical/chemical properties of the polymeric
membrane, which is the key parameter to influence significantly separation performance such as
porosity, pore size, membrane structure, surface roughness, hydrophilicity and surface charge of
the membrane [17].
However, the main challenge to inhibit its benefit from outperforming conventional polymeric
membrane performance is CNTs aggregation in the polymer matrix. Moreover, CNTs
nanocomposite membrane fouling remains as challenges due to the trade-off between high
removal efficiency and adhesion of foulants on the membrane. For the successful application of
the robust CNTs membrane to the municipal/wastewater treatment, it is suggested to examine
protein fouling behavior under different solution chemistry (pH, ionic strength).
3
(a)
(b)
Fig 1.1. (a) Comparison of water demand in the future and current supply [18] and (b) Diagram
of alternative water resource [19].
1.2 Scope and purpose of the study
This research highlights on the fabrication of multi walled carbon nanotubes (MWCNTs)
enhanced membrane which can overcome current issues on UF membranes such as low NOM
removal, a large volume of sludge generation in seawater pretreatment. For the purpose of
advancing membrane process, the focus is on in-depth of study in protein fouling behavior under
different solution chemistry which represents various conditions in municipal/wastewater
treatment and seawater desalination. Fig 1.2 presents an overview of the research.
This study has three principal objectives:
1. Fabrication of MWCNTs membrane for efficient NOM removal and examination of the
removal mechanism and its fast water flux
4
2. Application to the UF membrane hybrid system in the seawater pretreatment and examination
of salinity effect on membrane performance
3. Examination of protein fouling behavior under different solution chemistry (pH, protein
charge, and ionic strength)
(a) MWCNTs-enhanced membrane for UF membrane process.
(b) MWCNTs membrane coupled with PAC adsorption with a high permeate flux, organic
reduction in seawater filtration to lessen the sludge volume in pretreatment.
Fig 1.2 Application of MWCNTs-enhanced membrane to the membrane processes.
5
1.3 Thesis overview
The thesis is set out in 6 chapters:
Chapter 1 introduces the background and establishes objectives of this research.
Chapter2 provides a comprehensive review of the progress and challenges of carbon nanotube
membranes for water treatment. A potential of carbon nanotubes (CNTs) membranes in the water
treatment has highly been strengthened during last decade. The performance of CNTs membrane
is highly likely dependent on the fabrication method. The intrinsic properties of CNTs could be a
key factor of applicability to membrane process. This chapter provides an explicit and systematic
review on CNTs membranes addressing the current epidemic-whether the CNTs membranes can
tackle current challenges in the pressure and thermally driven membrane process, and efficacy of
CNTs hybrid nanocomposite membrane to complement current CNTs enhanced membrane.
This chapter was submitted to the Critical Reviews in Environmental Science and Technology.
Chapter 3 presents the experimental procedure of the membrane fabrication.
Chapter 4 builds up 1) the fabrication of MWCNTs enhanced UF membrane and 2) investigates
mechanism of its NOM removal efficiency and fast water flux.
The aim of Chapter 4 is to develop ultrafiltration (UF) membrane material by incorporating
carbon nanotubes (CNTs) into the polymer matrix that can effectively remove natural organic
matter (NOM) in water. Its physical/chemical properties are thoroughly analyzed for the purpose
of determining the critical factor of fast water flux and NOM removal mechanism.
The significance of the CNTs-enhanced membrane is to overcome the challenge in conventional
UF membrane that cannot remove NOM efficiently and has relatively low water flux. Compared
6
that the conventional membrane can only remove 30 – 50 % of NOM, the CNTs-enhanced
membrane in this research can remove up to 80 % of NOM. Further, the membrane achieved
greatly enhanced water permeability (1400 LMH/bar), which is 30 times faster than the
conventional PES UF membrane. The significant achievement is attributed to the synergetic
effect of the increased porosity, narrow pore size distribution and hydrophilicity. Relatively
narrow pore size on much thinner skin later and well developed finger like structure led to the
slippage effect. Positively charged surface charge of the membrane also contributes significantly
to the high removal efficiency on NOM. The membrane contributes to the progress of the UF
membrane material development by engineering physic-chemical properties of membrane via
nanomaterial incorporation. This chapter was published in the Separation and Purification
Technology (Elsevier).
Chapter 5 discusses 1) the application of MWCNTs-enhanced UF membrane for seawater
pretreatment in RO desalination and 2) examines the effect of salinity on the membrane
performance.
Since RO membrane technology has been increasingly applied to the seawater desalination plant,
attention has been on the energy-intensive pretreatment prior to the RO membrane system due to
the performance decline driven by RO membrane fouling. Currently, low-pressure membrane
coupled with adsorption/coagulation system seems to address the issue, but a large amount of
sludge and the limitation on the poor removal efficiency on the low molecular weight organic
matter remains challenges.
MWCNTs-enhanced membrane coupled with PAC adsorption achieved significantly enhanced
permeability with an organic reduction in seawater pretreatment. The study in this chapter
considerably contributes to lessening the sludge volume in the seawater pretreatment in the RO
7
desalination plant. Another significance of the research is an examination of the effect of ionic
strength on the organic removal efficiency in seawater filtration. This in-depth study determines
the feasibility of CNTs membrane and delivers the sustainable separation technology employing
membrane in the seawater desalination for the future. This chapter was published to the Journal
of Industrial Engineering and Chemistry (Elsevier).
Chapter 6 examines protein fouling behavior under different solution chemistry (pH, protein
charge, and ionic strength) and fouling mechanism.
1) The protein fouling behavior was investigated in the filtration of the multiwall carbon
nanotube (MWCNT) composite membrane and commercial polyethersulfone ultrafiltration
(PES-UF) membrane. 2) The effect of solution chemistry such as pH and ionic strength on the
protein fouling mechanism was systematically examined using filtration model such as complete
pore blocking, intermediate pore blocking and cake layer formation.
This research demonstrates that MWCNT membrane fouling can be alleviated by changing pH
condition and ionic strength based on the foulant-foulant interaction and the electrostatic
interaction between the membrane and foulant. This chapter is currently under Review in the
Chemical Engineering Journal.
Chapter 7 presents conclusions of this research which summarize results and conclusion derived
from each of journal published chapters and suggests future work.
8
Reference
1.
Malaeb, L. and G.M. Ayoub, Reverse osmosis technology for water treatment: State of the art
review. Desalination, 2011. 267(1): p. 1-8.
2.
Fritzmann, C., J. Loewenberg, T. Wintgens, and T. Melin, State-of-the-art of reverse osmosis
desalination. Desalination, 2007. 216(1-3): p. 1-76.
3.
Tularam, G.A. and M. Ilahee, Environmental concerns of desalinating seawater using reverse
osmosis. Journal of Environmental Monitoring, 2007. 9(8): p. 805-813.
4.
Sheikholeslami, R. and J. Bright, Silica and metals removal by pretreatment to prevent fouling of
reverse osmosis membranes. Desalination, 2002. 143(3): p. 255-267.
5.
Greenlee, L.F., D.F. Lawler, B.D. Freeman, B. Marrot, and P. Moulin, Reverse osmosis
desalination: Water sources, technology, and today's challenges. Water Research, 2009. 43(9): p.
2317-2348.
6.
Miao, R., L. Wang, N. Mi, Z. Gao, T. Liu, Y. Lv, X. Wang, X. Meng, and Y. Yang, Enhancement
and Mitigation Mechanisms of Protein Fouling of Ultrafiltration Membranes under Different
Ionic Strengths. Environmental Science & Technology, 2015.
7.
Wang, Y.-N. and C.Y. Tang, Fouling of Nanofiltration, Reverse Osmosis, and Ultrafiltration
Membranes by Protein Mixtures: The Role of Inter-Foulant-Species Interaction. Environmental
Science & Technology, 2011. 45(15): p. 6373-6379.
8.
Yu, Y., S. Lee, and S. Hong, Effect of solution chemistry on organic fouling of reverse osmosis
membranes in seawater desalination. Journal of Membrane Science, 2010. 351(1-2): p. 205-213.
9.
Choi, J.-H., J. Jegal, and W.-N. Kim, Fabrication and characterization of multi-walled carbon
nanotubes/polymer blend membranes. Journal of Membrane Science, 2006. 284(1-2): p. 406-415.
10.
Zhu, X., R. Bai, K.-H. Wee, C. Liu, and S.-L. Tang, Membrane surfaces immobilized with ionic
or reduced silver and their anti-biofouling performances. Journal of Membrane Science, 2010.
363(1-2): p. 278-286.
11.
Wu, H.Q., B.B. Tang, and P.Y. Wu, Novel ultrafiltration membranes prepared from a multiwalled carbon nanotubes/polymer composite. Journal of Membrane Science, 2010. 362(1-2): p.
374-383.
12.
Zodrow, K., L. Brunet, S. Mahendra, D. Li, A. Zhang, Q. Li, and P.J.J. Alvarez, Polysulfone
ultrafiltration membranes impregnated with silver nanoparticles show improved biofouling
resistance and virus removal. Water Research, 2009. 43(3): p. 715-723.
9
13.
Yang, K., X. Wang, L. Zhu, and B. Xing, Competitive sorption of pyrene, phenanthrene, and
naphthalene on multiwalled carbon nanotubes. Environmental Science & Technology, 2006.
40(18): p. 5804-5810.
14.
Corry, B., Water and ion transport through functionalised carbon nanotubes: implications for
desalination technology. Energy & Environmental Science, 2011. 4(3): p. 751-759.
15.
Majumder, M., N. Chopra, R. Andrews, and B.J. Hinds, Nanoscale hydrodynamics: Enhanced
flow in carbon nanotubes. Nature, 2005. 438(7064): p. 44-44.
16.
Kim, S.W., T. Kim, Y.S. Kim, H.S. Choi, H.J. Lim, S.J. Yang, and C.R. Park, Surface
modifications for the effective dispersion of carbon nanotubes in solvents and polymers. Carbon,
2012. 50(1): p. 3-33.
17.
Yin, J. and B. Deng, Polymer-matrix nanocomposite membranes for water treatment. Journal of
Membrane Science, 2015. 479(0): p. 256-275.
18.
Water 2030 Global Water Supply and Demand model ; Economic frameworks to inform decisionmaking. 2009.
19.
GWI DesalData / IDA (2011-2012).
10
Chapter 2 Literature Review
2.1
Introduction
Membrane technology has been extensively applied for water and wastewater treatment, and
water desalination due to its separation performance and energy intensiveness over the last
decade [1]. Low pressure driven membranes such as microfiltration (MF) and ultrafiltration
(UF) are an effective method in drinking water and membrane bioreactors for municipal and
wastewater treatment at a relatively low cost. Reverse osmosis (RO) membrane is
increasingly adopted in a large number of desalination plants since it shows highly
competitive performance, energy efficiency and cost saving [2].
Despite their many benefits, such membrane processes still have several drawbacks in the
application to water treatment. For example, natural organic matter (NOM) is one of the
major pollutants that contribute to serious membrane fouling and performance decline. NOM
can stimulate microbial regrowth in the distribution system [3]. Further, NOM generates
disinfection by-products (DBPs), and carcinogen when it reacts with chlorine residual present
in water. However, conventional MF/UF membranes such as polyethersulfone (PES),
polysulfone (PSf) and polyvinylidene fluoride (PVDF) are limited to NOM removal due to its
relatively larger pore size than humic/fulvic acid (<1 nm), exhibiting only 20 - 50 %
rejections. Polyamide (PA) and cellulose acetate (CA) membranes for RO desalination have
several challenges to be overcome. They generate an increasing amount of concentrate from
the process, which contains a high concentration of salts, process chemicals, and wastewater.
In addition, PA-RO membranes are vulnerable to degradation by chlorine. Another issue on
RO membrane is the need for capital and energy-intensive pretreatment due to the poor
quality of feed water [4]. Seawater contains relatively low molecular weight (LMW) organic
matter such as humic/fulvic acids, and carboxylic acid. Such contaminants result in serious
11
membrane performance decline by irreversible membrane fouling (i.e. organic and biological
fouling).
For
these
reasons,
MF/UF
membrane
hybrid
system
coupled
with
coagulation/adsorption seems to have effective pretreatment with high removal efficiency of
LMW organic matter [5]. However, it generates a large amount of chemical sludge due to the
usage of the high amount of coagulant/adsorbent.
The attempt to address those challenging issues is making progress on developing membrane
material by incorporating nanomaterials [6, 7]. Of nanomaterials, carbon nanotube (CNT) is
one of the favored candidates to the membrane for advanced water treatment due to its unique
properties such as an excellent adsorption capacity by high specific surface area [8, 9],
electro-conductivity [10], slippage effect on permeate flux by the molecular smoothness of
nanotube wall (hydrophobicity) [11] and easiness to functionalization [12]. In addition to its
tunable physical, chemical and electrical properties, incorporation of CNT enables the
alteration of physical/chemical properties of the polymeric membrane, which is the key
parameter influencing significantly separation performance such as porosity, pore size,
surface roughness, hydrophilicity and surface charge of the membrane [13, 14].
Current research on the CNT membrane is focusing on fabrication/modification of pressure
driven membrane by incorporating CNT into the conventional polymer matrix in order to
enhance permeate flux, rejection efficiency, and fouling resistance [15, 16]. However, the
main challenge to inhibiting its benefit from outperforming conventional membrane
performance is inconsistencies of performance due to CNT aggregation in the polymer matrix.
Moreover, CNT nanocomposite membrane fouling remains a challenge due to the trade-off
between high removal efficiency and adhesion of foulants on the membrane.
Although currently reported critical reviews focus on the summary of membrane fabrication,
CNT functionalization for better adsorption capacity and performance evaluation in the
12
context of its properties for water treatment [17-19], very few studies have provided an indepth insight on applicability of CNT membranes and the impact of CNT nanohybrid
membranes on the membrane processes.
This review, therefore, highlights firstly, the applicability of CNT enhanced membranes for
current systems from pressure to thermally driven membrane processes, secondly potential
impacts of the CNT hybrid nanocomposite membranes, and finally currently encountered
challenges and further research directions in the point of view of membrane material and
fabrication.
2.2
CNT materials
Carbon nanotube (CNT) is allotropes of carbon with a cylindrical carbon network, and the
one-dimensional analogs of zero-dimensional fullerene molecules. A nanotube visualizes a
micrometer-scale graphene sheet rolled into a cylinder of nano-scale diameter and capped
with a spherical fullerene (Fig 2.1). Graphene sheets are composed of a monolayer of sp2bonded carbon atoms in the x-y plane. Due to the presence of delocalized π- electron in the zaxis, CNTs have unique electrical properties [20].
(a)
(b)
(c)
Fig 2.1 Structure of CNTs: (a) a schematic representation of a graphene sheet and CNT roll-up vector,
(b) a 3-D model of single-wall CNT (SWCNTs), and (c) hybridization states of carbon-based
nanomaterials [21, 22].
13
CNT is classified into two types: single-wall carbon nanotube (SWCNT) and multi-wall
carbon nanotube (MWCNT) (Fig 2.2). SWCNTs, which are a cylinder of a single graphene
sheet, consist of a planar array of benzene molecules with hexagonal rings with double and
single carbon-carbon bonding. The MWCNT is multi-layers of rolled graphene sheets. The
SWCNT is revealed as metallic or semiconducting nanowires according to the chirality and
diameter [10, 23]. Table 2.1 summarizes the characteristics of carbon materials [24]. The
electrical conductivity of SWCNT is in the wider range (102-106 S/cm) than that of
MWCNTs. SWCNT have 3-fold thermal conductivity than MWCNT.
Fig 2.2 Schematic diagrams of (a) single-wall carbon nanotubes (SWCNTs), (b) multi-wall carbon
nanotubes (MWCNTs), (c) double wall carbon nanotubes, and (d) peapod nanotubes consisting of
SWCNTs filled with fullerenes (e.g. C60) [25].
Table 2.1 Physical properties of different carbon materials [24].
Carbon materials
Property
Graphite
Diamond
Fullerene
SWCNTs
MWCNTs
Specific gravity (g/cm3)
1.9-2.3
3.5
1.7
0.8
1.8
Electrical conductivity (S/cm)
4000p, 3.3c
10-2 10-15
10-5
102-106
103-105
Electron mobility (cm2/(Vs))
2.0 ×104
1800
0.5-6
~105
104-105
Thermal conductivity (W/(m K))
298p, 2.2c
900-2320
0.4
6000
2000
p: in-plane, c: c-axis
14
The high polarized π-electron clouds in CNT result in strong attractive forces, so-called van
der Waals, between CNT and the weak inter-planar interactions of the graphene sheet. The
physical properties of individual nanomaterials, such as the size, shape, and surface area of
carbonaceous nanomaterials are highly dependent on aggregation state and solvent chemistry.
The physicochemical properties attributed to secondary structures of nanomaterial aggregates
are highly variable and poorly characterized [21]. Resolving these characteristics is
imperative for application of carbonaceous nanomaterials to water purification-membrane
filtration [26].
CNT is intrinsically hydrophobic, and such property has benefits to water treatment. Firstly,
hydrophobicity and capillarity contribute to the adsorption behavior and orientation of
sorbates in microporous carbons. Physio-sorption is the dominant mechanism of sorption for
un-functionalized nanomaterials. In environmental applications, adsorptive capacity has
broad implications for contaminant removal and hydrogen storage. For example, CNT for
conventional drinking water treatment strongly depends on physicochemical sorption
processes for the removal of an organic and inorganic contaminant, such as NOM and heavy
metals [27-29]. Carbonaceous nanosorbents, which have high surface area to volume ratio,
controlled pore size distribution, and manipulable surface chemistry, outperform traditional
sorbents due to their rapid equilibrium rates, high adsorption capacity, and effectiveness over
a broad pH range and consistency with Brunauer, Emmett and Teller (BET), Langmuir, or
Freundlich isotherms. Secondly, the superhydrophobic CNT channel can also contribute to
ultrafast water fluid transport. This will be elucidated in section 3.4.1.
15
2.3
Fabrication of CNT membranes
2.3.1 Mixed matrix membrane – CNT/polymer composite membrane
Mixed matrix CNT membrane was initially designed to advance the separation performance
of polymeric membranes. Hence, the incorporation of CNT as inorganic fillers to polymeric
membranes was introduced to tune the physical/chemical properties that significantly control
permeability, rejection efficiency, and fouling resistance [13]. Recently, a good fabrication
method of the CNT incorporation to UF membrane was summarized systemically [30]. The
mixed matrix membrane led to fabrication with the relatively easy process and without
extensive efforts regarding membrane formation and its separation. This fabrication process
can be divided into two steps: i) CNT functionalization, and ii) UF membrane preparation by
phase inversion. Fig 3.3a presents the procedure of CNT composite UF membrane
fabrication via phase inversion. CNT is difficult to be incorporated uniformly into the
polymer matrix due to the intrinsical aggregation (hydrophobicity) by van der Waals force
between nanotubes. Van der Waals forces between nanotubes hinder the CNT dispersion in
the polymer matrix as CNT is formed as tightly aggregated conformation and low dispersed
in both polar and apolar solvents. Thus, dispersion of CNT is a necessary step in the
fabrication. The dispersed CNT by hydrophilic functionalization enables to maximize the
effect of nanomaterial and to have a large surface area for removal efficiency and to enhance
the flux of membrane. There have been many efforts to improve CNT dispersion. Chemical
oxidation by acid treatment is the most commonly used technique that attaches
hydroxyl/carboxylic groups to CNT wall [31, 32]. Sonication is the physical treatment that
breaks van der Waals force temporarily by physical shock. However, these methods could not
completely disperse the CNTs in the polymer matrix.
The second stage of the membrane fabrication is the formation of support layer by UF
membrane via phase inversion method (Fig 3.3a). The physical properties of the membrane
16
such as pore size, porosity, and thickness strongly influence the permeate flux. Such
properties are strongly affected by the parameters that can be controlled and determined by
the fabrication procedure: insertion of nanofiller, polymer and solvent proportion, etc [33, 34].
Thus, researchers have focused on nanomaterial incorporation into the UF membrane to
enhance the membrane performance [16, 35, 36].
In addition, the fabrication process of CNT composite RO membrane can be divided into
three steps: i) CNT functionalization, ii) UF membrane preparation by phase inversion and iii)
the formation of polyamide (PA) thin film layer by interfacial polymerization. The highlight
in this fabrication is the formation of selective layer (PA thin layer) on the top of UF
membrane support layer via interfacial polymerization (Fig 3.3b). Less than 1 nm diameter of
the pore on a 0.5 µm-sized support layer is formed by the interaction of m-phenylenediamine
(MPD) with trimesoyl chloride (TMC), leading to the greatly enhanced selectivity and
mechanical strength to high pressure [37]. The thin film composite (TFC) membrane is
suitable for RO and forward osmosis (FO) membrane system due to the greatly enhanced
selectivity for salts and minerals by the rejection layer on top of the membrane[37-39]. Here,
CNT can be incorporated into both aqueous phase (MPD) [40] and an organic phase (TMC in
Hexane or Isopar-G) [41], or deposited on UF membrane support layer before PA layer
formation by interfacial polymerization [42] [43].
17
(a) CNTs UF membrane by phase inversion [44]
(b) CNTs/PA NF/RO membrane by interfacial polymerization
Fig 2.3 Schematic illustrations of the CNTs composite membrane fabrication procedure: (a) CNTs UF
membrane by phase inversion [44] and (b) CNTs/PA NF/RO membrane by interfacial polymerization.
18
2.3.2 Horizontally aligned carbon nanotube (HACNT) membrane
Self-assembly CNTs membrane can be divided into two types of the fabrication method,
determined by the structure and arrangement of CNT in the membrane: horizontally aligned
CNT (HACNT) membrane (Fig 2.4a) and vertically aligned CNT (VACNT) membrane (Fig
2.4b).
(a)
(b)
Fig 2.4 SEM images of (a) HACNT [45] and (b) VACNT arrays [46].
HACNTs membrane has a randomly arranged CNT structure. Its fabrication procedure is
followed by i) CNT functionalization for homogenous CNT dispersion and ii) vacuum
filtration for the proper membrane structure. Hydroxyl/carboxylic-functionalized CNT is
dissolved in solvents (ethanol, propa-2-ol) by ultra-sonication. This homogenous suspension
is vacuum-filtered through a membrane filter and then dried in the oven to remove all
solvents and moistures from the film [47-49]. Due to the arrangement and fabrication
procedure, its pore size varies from few nanometers to micrometer. It can be determined by
adjusting the volume of the CNT suspension and selecting the types of CNT [50].
The advantages of the HACNT membrane are highly microporous structure, high adsorption
capacity [51] and superior inactivation E-coli cells [52]. However, due to poor mechanical
strength, which cannot be overcome during the fabrication process, this membrane may be
inadequate for the high-pressure process.
19
2.3.3 Vertically aligned carbon nanotube (VACNT) membrane
In the VACNT membrane matrix, CNT is vertically aligned on the membrane substrate. The
gap between nanotubes is filled with fillers such as epoxy, polystyrene, and/or parlyene for
proper membrane structure.
VACNT membrane is prepared in two stages: i) growth of an aligned array of CNT and ii)
infiltration with a matrix material in the gap between nanotubes. VACNT arrays are usually
synthesized by chemical vapor deposition (CVD) method [53-55]. As shown in Fig 2.5, CNT
is grown by the reaction between ethylene [C2H4(g)] and ferrocene [Fe(C5H5)2] as a catalyst
precursor and further grown in the furnace. Ferrocene is decomposed with Fe by heating, and
CNT is grown on the Si wafer substrate coated with aluminum oxide (AAO). In this step,
aluminum acts as a template for the growth of aligned CNT. In order to remove substrate,
water etching and vacuum extraction are performed. Here, CNT diameter can be controlled
by modulating feeding rate of catalyst-precursor in a floating system. This has opened a way
for VACNT arrays to be a potential membrane for water treatment.
Fig 2.5 Schematic of the synthesis and detachment of a VACNTs array [53].
However, CNT array cannot be employed to water treatment directly since it does not have a
proper nanoporous structure for water transport, nor can it introduce water molecule into the
20
nanochannel due to its hydrophobicity. Thus, gap infiltration needs to be carried out for water
to pass through only the channel of CNT, after growth of aligned CNT array. Table 2.2
summarizes the fabrication procedures of VACNT membrane. It provides a classification of
CNT synthesis and post-treatment for fine ordered membrane structure by substrate removal
and tip opening. Fig 2.6 presents the typical fabrication procedure of VACNT membrane:
Firstly, gap infiltration is done with polymers (epoxy, polystyrene, and parlyene, etc.) under
the vacuum pressure. Then, the substrate is removed by NaOH etching and finally, a tip
opening is carried out by mechanical polishing, ion etching or plasma etching.
Fig 2.6 Fabrication procedure of VACNTs membrane.
21
Table 2.2 Summary of fabrication methods of VACNTs membrane.
Type of
VACNT
SW (6,6)
functionalized
with 8COO(5,5) armchair
type tube
(6,6) armchair
type tube
(7,7) armchair
type tube
(8,8) armchair
type tube
Double-wall
Multi-wall
Internal
Diameter
Length
Contact
angle
1.1 nm
1.3 nm
-
0.32 nm
~1.4 nm
-
0.47 nm
~1.4 nm
-
0.59 nm
~1.4 nm
-
0.75 nm
~1.4 nm
-
1.3 - 2 nm
1-5 μm
-
Operating
pressure
Flux
1.51
LMH/bar
1.2
LMH/bar
2.1
LMH/bar
3.2
LMH/bar
4.9
LMH/bar
2.2×105
LMH/bar
Test material
Reactive ion
etching in an
oxygen
containing
plasma
[56]
Catalytic chemical vapor
deposition/
Silicon nitride (Si3N4)
encapsulation method
Reactive ion
etching in an
oxygen
containing
plasma
[57]
deposition/
Silicon nitride (SiN4)
Reactive ion
etching in an
oxygen
containing
plasma
deposition/
Reactive ion
etching in an
oxygen
containing
plasma
NaCl
100
55 bar
NaCl
100
55 bar
NaCl
95
55 bar
NaCl
58
0.82 bar
Ru2+(bipyr)3
1.3 nm of
particle size
-
1-5 μm
-
0.9×10
LMH/bar
0.82 bar
Single-wall
1.3 - 2 nm
1-5 μm
-
1.3×105
LMH/bar
0.82 bar
Colloidal Au
5 ± 0.75 nm of
particle size
-
deposition/
55 bar
5
-
Ref.
100
1.3 - 2 nm
1-5 μm
Method of tip
opening
NaCl
Double-wall
1.3- 2 nm
Method of CNT
synthesis/fabrication
26.4 bar
Colloidal Au
2 ± 0.4 nm of
particle size
Double-wall
Re. (%)
-
Na4PTS
96 %
K3Fe(CN)6
90 %
K2SO4
80 %
CaSO4
37 %
KCl
37 %
0.69 bar
22
[26]
[58]
Hinds’ group has firstly introduced CNT membrane fabrication using the polymer infiltration
[59]. CNT was grown for 30 min on a quartz substrate by a CVD process. Then, polystyrene
and toluene solution encapsulated CNT arrays in order to form a well-ordered nanoporous
membrane structure. Lastly, H2O plasma-enhanced oxidation process was performed to
remove a thin layer of excess polymer from the top surface and open the tips for membrane
structure. Interestingly, plasma process attached carboxylic groups on the tips of CNT, which
could be the basis for gatekeeper-controlled chemical separations or ion channel mimetic
sensors.
An encapsulation of VACNT with low-stress Si3Nx was developed by using a low-pressure
CVD process [60]. After encapsulation, the membrane is etched in order to remove an
excessed Si3Nx from the tips of the CNT, followed by oxygen plasma to uncap the CNT. The
VACNT membrane has less than 2 nm pores in diameter and consistent with diameters of asgrown nanotubes.
Another approach was the fabrication of hierarchically organized vertical CNT arrays by
plasma-enhanced CVD (PECVD) with self-assembled block copolymer nano templates [61].
Conventional thermal CVD does not produce highly oriented vertical architecture due to
thermal fluctuation and gas flow established within the chamber during the growth process.
However, PECVD can enhance the vertical alignment of CNT by the electric field in the
PECVD chamber due to the high electrical polarizability of the structures. The large dipole
moments induced in CNT by the external electric field impose aligning torques on the
growing CNT, resulting in the growth of highly oriented VACNT arrays. Polystyrene-blockpoly (methyl methacrylate) (PS-b-PMMA) was used as a self-assembling polymeric material
to form cylindrical nanostructure with center-to-center distanced between them of 72-34 nm
as nano templates for the hierarchical organization of CNT. The cylinder cores were etched
by UV radiation and wet method for polymeric nano templates with hexagonally packed
23
cylindrical nanopores. The iron catalysts were deposited, and lift-off process was conducted
to control the size and lateral distribution of the catalyst particles. Next, PECVD process was
applied to the growth of CNT. Heat treatment at 750°C for 1 min was carried out for size
decrease of the catalyst particles from 21 to 14 nm. As a result, CNT diameter became half of
the agglomerated catalyst particle size. In this research, the diameter of CNT is dependent on
the size of the catalysts particles. The nanopores’ size (D) and center to center distance
between nanopores (L) of the block copolymer template depend on the molecular weight of
the block copolymer (M): D, L α M⅔
Beside polymer infiltration method, thermally infiltration was applied to fill the gap between
the nanotubes [62]. For the controlled embedment of nanotube into polymer films, aligned
CNT film grown on the SiO2/Si substrate was held on the hot plate and heated up to a
temperature above the melting point (Tm) and below decomposition temperature (Tc). The top
of a CNT array was fully covered with a thin film of polystyrene (~50μm). The melted
polystyrene (PS) layer was gradually infiltrated into the aligned nanotube forest.
VACNT membrane has several benefits for the advanced water treatment due to its intrinsic
properties. Firstly, the membrane provides well-defined narrow pore size, due to the role of
CNT channel as a water passage. In addition, the pore size of VACNT membrane can be
controlled in the CNT synthesis step by selecting proper catalysts such as Cu and Al [63]. Its
diameter can be tuned from few hundreds nanometer (nm) to less than 1 nm size, indicating
that VACNT membrane could be an ideal candidate for membrane distillation (MD) and RO
membrane system that requires narrow pore size and its size distribution.
2.3.4 Key factors affecting the ultrafast water permeability of VACNT membrane
2.3.4.1 Slip length
Super ultra-fast water flux is mainly induced by the slippage effect in CNT channel [56, 64].
The reduced viscosity near the superhydrophobic wall generates the slippage effect for the
24
fluid transport in the CNT channel. Here, this fluid slip is so called, “slip length.” Slip length
is the distance inside the wall at which the extrapolated fluid velocity would be equal to the
velocity of the wall. In other words, it is the ratio of the slip speed to the mean shear rate.
Thus, slip length is proportional to shear rate. In order to explain water transport within CNT,
the Navier-Stokes equations can be used, which are continuum fluid mechanics along with a
slip boundary condition. As can be seen in Fig 2.7, the slip boundary condition occurs when
liquid is interfaced on the non-wetting surfaces (e.g., water on hydrophobic surfaces).
Perfect
slip
Partial
slip
No slip
b
b=0
b=∞
0<b<∞
Fig 2.7 Slip length flow past a stationary surface [65].
The theory of calculation is well explained in Table 2.3.
Slip length measurement can be defined as the ratio of slip velocity to shear rate at the wall. It
can be expressed as follow:
𝜈𝑏 = 𝑏
Where,
νb = the slip velocity of the fluid on the surface
b = Slip length
dν
dz
= Velocity gradient
25
𝑑𝜈
𝑑𝑧
Table 2.4 summarizes the measurement of slip length in the channel.
Rheometry experiment
Slip length is obtained by measuring torque through a rheological experiment. In the
rheological system with the plate-and plate arrangement, the driving torque applied to the
clamp is recorded at each shear rate.
Even though this method is the most popular geometry because it produces the uniform shear
rate over the sample [66], there is a difference between the apparent viscosities in the
rheological experiment for CNT and the intrinsic viscosity due to the slippage effect [67].
Hence, calculated slip length is over-estimated, and the influence of the edge effect should be
removed in the slip length calculation.
Particle image velocimetry (PIV)
This is based on the measurement of the velocity of fluorescent particles embedded water.
First, the velocity profile of the liquid is measured, and the slip length is extracted from the
equation in Table 2.3. It is a direct technique to measure the liquid flow close to solid
surfaces. It allows direct access to the velocity profile and extraction of the slip length with
high accuracy [66].
Pressure-drop experiments
This is the measurement of pressure drop as a function of the flow rate in Navier boundary
condition. However, light exposed particles make it difficult to achieve good spatial
resolution in the PIV measurement normal to the plane of the ultra-hydrophobic surface (z
direction).
26
Table 2.3 Theory of slip length calculation.
Equation
Note
Philip – flow parallel to the stripe on the planar surface with onedimensional features
Philip [68]
Langua &Stone [69]
When R →∞, the effective slip length is
𝑏𝑃ℎ𝑖𝑙𝑖𝑝 =
𝜋𝜙𝑔
𝐿
𝑙𝑛 (𝑐𝑜𝑠 (
))
𝜋
2
Lauga & Stone – Flow perpendicular to stripes stripe on planar
Where, L = the pitch of the pattern
surface with one-dimensional features
ϕg= the gas fraction area = (L-a)/L
a = stripe width with zero slip length
𝑏𝐿𝑎𝑔𝑢𝑎−𝑆𝑡𝑜𝑛𝑒 =
Cottin-Bizonne et al.
[70]
𝜋𝜙𝑔
𝐿
𝑙𝑛 (𝑐𝑜𝑠 (
))
2𝜋
2
𝑏=
𝑉𝑠
𝛾̇
Microstructured superhydrophobic surfaces
Flow perpendicular to the slip
Where, Vs = effective velocity slip on the surface
𝛾̇ = Shear rate
Calculated by Navier-Stokes equation
Agreement with Philip and Langua & Stone
27
A surface with two-dimensional patterning, regular lattice
Ybert et al. [71]
Linear reg between ression
𝑏𝑒𝑓𝑓 = 𝐿 (
0.325
√𝜙𝑠
− 0.44)
Where, ϕs=solid fraction
(Surface area where the lip length vanishes)
𝜙𝑠 = 1 − 𝜙𝑔 = 𝑎/𝐿
Davis & Lauga [72]
𝑏𝑒𝑓𝑓 = 𝐿 (
3 𝜋
3
0.332
𝑙𝑛(1 + √2)) ≈ 𝐿 (
− 0.421)
√ −
16 𝜙𝑠 2𝜋
√𝜙𝑠
28
Table 2.4 Methods of slip length measurement.
Measurement
Direct
Measurement
Equation
𝑏 = 𝜐𝑠 /𝜕𝜐𝑠 /𝜕𝑧
Where, υs = the velocity
∂υs/∂z = shear rate
Particle image velocimetry
[73]
Cone and Plate Rheometer
𝑀=
Indirect
Measurement
Rheometry experiment
[67, 74, 75]
2
3
3𝑏𝑒𝑓𝑓
3𝑏𝑒𝑓𝑓 3𝑏𝑒𝑓𝑓
𝑅𝜃0 + 𝑏𝑒𝑓𝑓
2𝜋𝜂𝛺𝑅3
[1 −
+ 2 2−
𝑙𝑛 (
)]
3𝜃0
2𝑅𝜃0 𝑅 𝜃0
𝑅3
𝑏𝑒𝑓𝑓
Where, M= Torque (Applied), R = radius
Ω = angular velocity, η = liquid viscosity
Where, ηa = apparent viscosity
measured in the rheometer
ηe =apparent viscosity measured in the
final wetting state (Wenzel state)
Parallel Plate Rheometer
29
Note
- Measurement of the velocity of
fluorescent particles in water
- Torque measurement with rheometer
- Uniform shear rate
- Questions about the accuracy
- Experimental error in the apparent
viscosity with one measuring process
𝐻3
𝑑𝑃 1
𝑏
(− ) [ +
]
4𝜂
𝑑𝑍 3 𝐻 + 𝑏
b : effective slip length
dP/dz : pressure gradient
𝑞=
Pressure-drop experiment
[76, 77]
𝛥𝑃𝑛𝑜−𝑠𝑙𝑖𝑝 − 𝛥𝑃𝑆𝐻
𝛥𝑃𝑛𝑜−𝑠𝑙𝑖𝑝
∆PSH : experimentally measured pressure drop
∆Pno-slip : Pressure drop for flow over a no-slip surface
𝐷𝑅 =
- Measurement of pressure drop as a
function of the flow rate in Navier
boundary condition
- μ-PIV method for detailed flow
kinematics
Surface force apparatus
experiment [78, 79]
Damping versus distance
Atomic force microscope
experiment [80, 81]
Dynamic mode AFM experiment
High accuracy
Controlling the gap of the flow
Indirect
Measurement
1+
𝛾𝐻
𝐴0
𝑄0
= − 𝑠𝑖𝑛(𝜑)
𝐴
𝛾0
𝐴
√1 + 𝑄02 + 2 ( 0 ) 𝑐𝑜𝑠(𝜑) + 𝐴20 /𝐴0 √1 + 𝑄02
𝐴
𝛾𝐻 =
6𝜋𝑅 2 𝜂
𝐷
𝑓∗
1
6𝐷
4
4𝑏
𝑓 ∗ = {1 +
30
[(1 +
𝐷
4𝑏
) 𝑙𝑛 (1 +
4𝑏
𝐷
) − 1]}
There are three possibilities that can explain large slip length in CNT channel: Narrow pore
diameter, reduced viscosity, and hydrophobicity.
First, molecular dynamic simulation by Holt et al. [26] demonstrates that CNT membrane has
significantly higher slip length than the other polymer membranes due to its relatively smaller
pore diameter compared to the pore length. Table 2.5 compares slip length of nanochannel
with different sizes of diameter. The largest slip length of CNT membrane is 1400 nm, which
is almost three orders of magnitude greater than the pore size and is on the order of the
overall size of the system (pore length). In contrast, it reveals that the polycarbonate
membrane with a pore size of 15 nm has a much smaller slip length of 5 nm. This clear
comparison suggests that the slip flow through sub-2 nm CNT happens possibly due to the
length scale confinement [82] or partial wetting of CNT’s surface [83]. The result raises the
question whether Newtonian fluid could be used for determining water flow in CNT.
Table 2. 5 Slip length of CNTs membrane.
Membrane
MWCNT [84]
DWNT 1 [26]
DWNT 2 [26]
DWNT3 [26]
Polycarbonate [26]
Transport
Viscosity
(Pa s)
Water
Expected flow
Slip length
(mm)
8.94×10-3
25 cm/s bar
39 – 54
Ethanol
IsoPropanol
1.074 ×10-3
4.5 cm/s bar
28
1.12 cm/s bar
13
Hexane
294 ×10-3
5.6 cm/s bar
9.5
Decane
920 ×10-3
0.67 cm/s bar
3.4
8.94×10-3
1500 – 8400
680 – 3800
560 – 3100
3.7
380 – 1400
170 – 600
140 – 500
5.1
Water
Pore diameter
(nm)
1.945×10-3
7
1.3 – 2.0
15
*Feed temperature derived from viscosity: 20 ℃
Another factor to control the slip length is viscosity. Frictionless surface at CNT wall
(hydrophobic) was found to make fluid velocities high [84]. This result can explain why the
slip length of CNT is remarkably long. The slip length decreases as solvents become more
31
hydrophobic, which indicates stronger interaction with the CNT’s wall. It is suggested that
higher contact angle results in greater slip and more reduction in effective friction [85]. On
the other hand, lower ɛr (potential well depth ← relative energy parameter) and lower σr
(Atom size radius) tend to increase contact angle.
Another explanation for the slip length is based on the fact that CNT is hydrophobic. The
strength of attraction between water molecules is greater than the attraction between the
hydrophobic solid and the water. Myers suggested the reduced viscosity model that is
analogous to a slip model to provide the explanation [86]. The model provides a physical
interpretation of the classical Navier slip condition and explains why slip length may be
greater than the tube radius.
2.3.4.2 Modification of CNT membrane for channels of water and target molecules
Even though CNT channel has large slip length due to its frictionless hydrophobic
nanochannel structure, it requires high pressure for water transport in the superhydrophobic
channel in the pressure driven membrane process. In this way, hydrophobic VACNT channel
requires being modified by attaching a specific hydrophilic functional group for enhanced
selectivity and electrical and thermal property [87-90] (Fig 2.8a). Further, CNT
functionalization can lead to highly selective separations caused by forcing a chemical
interaction between permeate and functional molecules as the formation of hydroxyl and
carboxyl groups on the surface of nanotubes [59, 91, 92].
32
(a)
(b)
Fig 2.8(a) Concept of tip functionalization of VACNT membrane and (b) oxidation process
of CNTs using acid and oxidative gas [31].
Table 2.6 summarizes methods of hydrophilic functionalization in VACNT membrane.
Oxidation with acid treatment is one of the widely used methods that directly attach
carboxylic groups and other oxygen-bearing groups, such as hydroxyl (-OH), carbonyl (C=O),
ester, and nitro groups to ends and/or defect sites in the sidewalls of CNT [93-95] (Fig 2.8b).
This method was employed in order to disperse nanotubes and purify amorphous carbon. This
procedure is straight-forward, and commonly used for hydrophilic functionalization of CNT.
33
Plasma oxidation (during the membrane fabrication process) was reported to add carboxylic
acid groups at the entrance of CNT [92]. These functional groups introduced water molecules
to the CNT channels, which acted as a passage of water through the membrane. Therefore,
water flux was enhanced by columbic attraction between surface modified CNT and water
molecules. This approach increased not only water flux but also selectivity of the separation
process. The charge state of carboxyl groups at the entrance to the CNT tips stimulated the
attraction of ions into them [96]. It is due to the fact that the presence of polar water
molecules, the cation-π effect by the ions, or an external electric field can alter the electronic
structure of CNT and hence cause field-induced polarization. The presence of functional
groups could also alter the electronic structure of the CNT. This coupled density functional
theory of molecular dynamics approach resulting in an improved force field could explain
such interactions.
A previous study on the relationship between functionalization and membrane performance
was conducted by means of molecular dynamic simulation [97]. The concept of CNT
functionalization was inspired by biomimetic modification – Aquaporin-4-channels in which
four amino acids are added. Location of functional groups (CONH2, NH3+, COO-, and OH) in
CNT channels was set out to the interior and or at the entrance of the channel. The result
revealed that CNT with a wide diameter (1.35 nm) completely removed salt, delivering 150 –
170 % enhanced water flux compared to the unmodified (8,8) CNT [64]. Simulation of
functional group location elucidated that the membrane exhibited a high rejection without
sacrificing water flux when one functional group is attached to the tip and other four groups
are added to the inside of the CNT channel.
Surface modification can be employed by an electrical method that offers a powerful,
nondestructive, and selective way to achieve water transport through CNT membrane [98, 99].
Polarity and voltage-dependent wetting method could be used for various applications of
34
nanotube membranes. For example, VACNT membranes act as an anode (positive potential),
and the Pt wire as the cathode (negative potential) [100]. Once critical voltage (1.7 V) is
reached, the membrane is transferred from superhydrophobic to a hydrophilic state, and the
water droplet rapidly passes through the nanotube arrays. The mechanism is that the sinking
of the droplet is dependent on the presence of electrolysis in this system due to the
electrochemical oxidation of the nanotube.
35
Table 2.6 Summary of hydrophilic treatments of CNTs.
Method
Plasma etching
[101] [92]
H2O plasma-enhanced
oxidation [59]
Functional
group
COOH
COOH
Effect
Removal of amorphous carbon
and one opened CNT tip
- Opening CNT tip
- Attaching COOH group
- Etching polymer matrix
(CNT 10 -50 nm above than PS
matrix)
- Increasing ionic flux
(Ru(NH3)63+ )
0.07 → 0.9 μmol/cm2hr-1
Condition
- For 30 min for removing
amorphous carbon,
- For 80 min for opening tip
at 250 kHz, 30 W
PH2O=0.62 Torr
PH2O= 60 mtorr
2.5 W/cm2 for 7 min
Results
(a) Before
(b) After 80 min
70 % of CNT tip open
-Immersing membrane in
concentrate HCl for 24 hr at RT
-Rinsing with DI water
HCl Treatment [59]
COOH
Reactive Ion etching [26]
COOH
- Opening CNT tip for gas and
water permeability
Reactive ion etching in an oxygen
containing plasma
Ag Catalyze oxidation
[102]
Oxidized
Carbon
Attaching Oxygen containing
group - Epoxidation
Oxidation of 5 % Ag/CNTs for 105
min at 300 °C
CNT channels are wet by hydrophilic solvents
Opening CNT tips
Selective oxidation by CO2, O2
above 600 °C
No results for hydrophilic group attachment on
CNT tips
- Opening CNT tips,
- Attaching hydrophilic group to
CNT walls
Oxidation of CNTs by boiling
mixture of H2SO4 and HNO3
Applied to HACNTs
Wet chemistry method
[103, 104]
Oxidation method [31]
COOH, OH
36
Acid oxidative
[105, 106]
(Sonochemistry)
COOH
- Opening CNT tips,
- Attaching hydrophilic group to
CNT walls
- Oxidation by H2SO4 /30% H2O2
solution
- Sonication for 5 min
Acid sonicated CNTs show 1590 cm-1 band of
COO- groups increase
37
2.3.4.3 Challenges in VACNT membrane
VACNT membrane has many advantages for water treatment as discussed in the previous
sections. The slippage effect derived from frictionless wall leads to the several orders of
magnitude higher water flow than conventional fluid transport. However, it is possible only
in the case that pore density is as compact as (6 ± 3) × 1010 cm-2 [59]. It indicates that lowdensity CNT array may not guarantee ultra-fast water flux. Moreover, it is imperative to
synthesize highly ordered oriented vertical structure of CNT arrays, which may not be
assured in conventional thermal CVD due to the thermal fluctuation and inconsistent gas flow
in chamber. Even if VACNT membrane has many possibilities of tackling current challenges
in membranes for water treatment, difficult fabrication, and synthesis of CNT arrays with
optimal density for ultra-fast water flux procedure hinder its wide application to water
treatment. In particular, further study on post-treatment such as hydrophilic functionalization
on CNT tip would be a critical step for rejecting salt completely and realizing high water
permeability. In addition, a scaled-up issue may be a key to making the progress of practical
application of the VACNT membrane.
38
2.4
Application of CNT membranes in water treatment
2.4.1 NOM removal for surface water treatment and seawater pretreatment
2.4.1.1 Surface water treatment
NOM is a complex of particulate and soluble materials in surface water with varying
molecular weight, mainly produced from the decomposition of plant and animal residues. It
contains a heterogeneous mixture of humic substances, hydrophilic acid, protein lipids,
carbohydrates, carboxylic acids, amino acids, and hydrocarbons. NOM in dissolved organic
carbon can be divided into three segments: Hydrophobic (49 %), hydrophilic (30 %) and
transphilic (21 %) [107]. Here, Humic substances are composed of 60 - 70 % of TOC in soil
and 60 – 90 % of DOC in natural water. Due to the carboxylic and phenolic moieties, NOM
carries a negative charge in the natural environment [108]. NOM generates carcinogen such
as Trihalomethanes (THMs), Haloacetic Acids (HA) when it reacts with chlorine in
distribution system [109]. Furthermore, it causes both severe reversible/irreversible
membrane fouling in the presence of calcium [110], thus, leading to flux decline [111]and
decreased rejection efficiency [112]. Such a foulant is hard to remove by low pressure driven
membranes due to their relatively larger pore size.
An attempt to overcome this issue encouraged the enhancement of adsorption capacity of the
membrane by developing nanocomposite membrane. In this way, CNT membrane can be
extensively applied in the removing NOM due to its unique properties such as large surface
area and hydrophobicity. CNT removes NOM by its adsorption capacity of organic chemicals
[8]. The interaction between CNT and NOM includes π- π attractions, hydrophobic
interaction, electrostatic interaction, and hydrogen bond. For example, π- π attractions appear
between bulk π systems on CNT surfaces and organic molecules with C=C double bonds or
39
benzene rings. Hydrogen bonds occur between functional groups on CNT surfaces and NOM.
Electrostatic interactions occur when CNT surfaces are charged.
HACNTs membrane removes NOM by adsorption mechanism. CNT directly contact to the
organic contaminants due to their arrangement in membrane matrix. Such a fabrication
facilitates hydrophobic interactions, π- π interaction between CNT and organic chemicals. ππ systems on CNT attracts with a cross-linked aromatic network of molecules on organics
such as humic acids [113]. Thus, the adsorption capacity of CNT is proportional to the
accessible surface area.
CNT composite membrane. Removal mechanism of CNT nanocomposite membranes is
determined by system configuration (or application). It is noted that HA/FA with mean size
less than 1 nm easily pass through low-pressure membranes such as MF/UF, exhibiting only
20 – 50 % rejection efficiency. However, NF/RO membrane can reject more than 90% NOM
by the size exclusion. To increase removal efficiency in UF membrane, its adsorption
capacity needs to be enhanced. Recently, MWCNT enhanced UF membrane has been
reported to remove 80% HA removal rate by increasing electrostatic interaction between
negatively charged HA and positively charged membrane surface. Compared that CNT
incorporated membrane exhibits negative zeta potential, the CNT/PANI complex
incorporated PES-UF membrane possesses positive zeta potential, which contributes greatly
enhanced NOM rejection.
VACNT membrane. Currently fabricated VACNT membranes are classified in the UF or
NF/RO membrane by their pore size (1 – 7 nm), rejection and applied pressure (2 and 20 bar)
[114, 115]. Despite that application to NOM removal has not been reported, it is expected
that VACNT membrane would have strong efficacy for the high NOM removal due to its
mechanism driven by steric hindrance and adsorption capacity.
40
Table 2.7 compares HA removal efficiency of recently published CNT membranes with
conventional polymeric membranes. In case of synthesized NOM filtration test, CNT
composite membrane outperformed conventional polymeric membranes in UF membrane
process, which cannot completely remove HA due to its separation performance (larger pore
size than HA). As shown in Table 2.7, the CNT/PANI/PES membrane with 10 kDa MWCO
exhibited 80 % of SRHA removal rate, compared that the ceramic and PES membrane (50
and 150 kDa, respectively) could remove 60 – 70 % SRHA. CA-UF membrane in the range
of 3- 15 kDa removed 90 % of HA effectively. The notable thing is that the feed
concentration in CNT/PANI/PES membrane is low (5 mg/L) than the one in ceramic and PES
membranes (10 - 20 mg/L). Further, recently fabricated HACNT membrane reached to 90 %
HA removal by facilitating adsorption capacity of CNT to the membrane matrix. However, in
case of natural surface water filtration test, removal efficiency of PSf UF membrane (50 %)
was comparable to CNT membrane (45 %). Based on current researches on NOM removal,
MWCNT membrane seemed to have superior performance in synthetic NOM feed water
compared to the conventional UF membrane, but did not outperform it in the low range of
concentration of NOM such as natural surface water. The conventional UF membrane with
small MWCO (< 10 kDa) showed high removal efficiency in the presence of high
concentration HA filtration test.
41
Table 2.7 Summary of HA removal efficiency by CNTs membranes.
Membrane Type
System
Composition
Proportion of
CNTs (%)
MWCO (kDa)
or
Pore size
Water
permeability
(LMH/bar)
Model
compound
Rejection
efficiency (%,
UVA 254 nm)
Ref.
UF
Polysulfone (PSf)
-
68
-
NOM
(From natural
surface water)
50
[116]
UF
Ceramic
-
50, 150
-
Polymer
[117]
UF
PES
-
100
-
UF
Cellulose Acetate (CA)
-
3- 15
10 - 60
HA 9 -10 mg/L
90
[118, 119]
HACNTs
100 %
(41 ± 10 nm)
-
HA 5 mg/L
90
[51]
RO
CNT-Polyamide
0–1
0.71 – 0.76
HA
10 mg/L
90
[120]
UF
CNT- PES
0–4
25 - 27
2.4 – 21.8
NOM
(From natural
surface water)
45
[121]
UF
CNTs/PANI/PES
1.5
10
(4.6 nm)
500 - 1400
SRHA*
5 mg/L
80
[44]
HACNTs
CNT nanocomposite
60
SRHA*
20 mg/L
VACNTs**
*SRHA: Suwanee river humic acid, **: No data reported.
42
70
2.4.1.2 Seawater pretreatment
Removal of low molecular weight dissolved organic matters (LMW-DOMs) in seawater is a
critical issue due to the decline in the membrane performance by irreversible membrane fouling.
Colloidal issues such as humic substances and biopolymers (3-20 nm) mainly make membrane
fouled through adsorption on the membrane surface or in the membrane pores. However, direct
MF and UF membrane filtration with higher than 100 kDa MWCO cannot remove LMW-DOM
(less than 350 Da) [122, 123]. An option to overcome the limit in pretreatment for seawater
reverse osmosis (SWRO) is suggested to be the integration of filtration system with the
adsorption or coagulation process [124, 125]. Recently developed membrane hybrid system
coupled with adsorption with addition of powder activated carbon (PAC) demonstrated enhanced
performance in the removal of LMW-DOM containing 75% of organic matter in seawater [126],
mostly due to the fact that PAC (as an adsorbent) removes LMW-DOM with its affinity (covalent bonding) to the LMW organic compounds [127, 128].
Despite the outperformance of UF membrane hybrid system, a relatively high amount of
adsorbent causes a large volume of chemical sludge from SWRO desalination plant, which
brings serious environmental issues to the aquatic condition. Therefore, further study is
suggested to focus on the reduction of chemical dosage by developing membrane material with
enhanced removal efficiency and higher water flux.
For this reason, CNT composite membranes could be one of the favored approaches due to its
unique characteristic of excellent adsorption capacity for organic matter. However, there are few
existing research papers on the application of CNT membrane to seawater pretreatment for
LMW-DOM removal. Thus, examination of CNT membrane for effective LMW-DOM removal
in seawater pretreatment would considerably contribute to addressing the limit of MF/UF
43
membrane system for seawater pretreatment. Based on the recent progress of CNT
nanocomposite membrane in water treatment, CNT membrane with greatly enhanced adsorption
capacity is highly likely to contribute to the reduction of sludge volume by decreasing chemical
dosage.
2.4.2 Salt rejection: from brackish water treatment to MD/RO concentrate recycling
VACNT applicability for seawater desalination has been investigated using molecular dynamics
simulation. For example, Corry determined the optimal diameter of carbon nanotube for water
desalination by molecular dynamic simulation [57]. Depending the effect of the different CNT’s
sizes on the water flux and ion rejection, the optimal internal diameter of CNT for desalination
was determined. Table 2.8 summarizes VACNT’s application in water treatment. CNT with up
to 10 Å [(7, 7) armchair tube type] of C-C diameter removed nearly 100% of NaCl, with
maintaining improved water flux. This result is comparable with FILMTEC SW30HR-380
commercial RO membrane [129]. It is notable that Na+ tends to pass through the pores of (6,6)
membrane more readily than Cl-. Cl- is not seen to pass through the (7, 7) membrane and five
passes through the (8, 8) membrane in 25 ns compared to 23 Na+. This result can be similar to
carbon nanotubes of different chirality as well as double or multi-walled ones. Donnan
membrane equilibrium supports this phenomenon because electrochemical equilibrium is
established when an ionic solution contacts a charged membrane and equilibrium partitioning of
ions between the solution and the membrane phase under the constraints of electro-neutrality
[58]. Electrostatic force repulses anions and attracts cations from the negatively charged CNT
tips (-COOH group). The condition of the electro-neutrality prevents an independent migration
of anions and cations. Then, two opposite electrostatic forces are balanced, thus leads to the
overall rejection.
44
The ion rejection mechanism in CNT was assumed to be both steric hindrance and electrostatic
repulsion. In order to understand the relative importance of these driving forces, Fornasiero et al.
examined ion exclusion and selectivity as a function of solution concentration, pH, ion valence,
and ion size [58]. Ion rejection in nanopores is mostly controlled by the electrostatic effect,
which is demonstrated by Donnan membrane equilibrium model [130, 131]. CNT functionalized
with COOH group were tested at two solutions having different pH values, one above the pKa of
COOH group (pKa=5.5) and another below it. COOH groups are protonated (COO-) at high pH
while neutral (COOH) at low pH. The research reported that there was more than 90 % ion
rejection above pKa of COOH (pH=7.2), indicating that major driving force for ion rejection is
electrostatic interaction. In addition to pH effect, the valency of cationic (z +) and anionic (z-)
species in solution affects ion rejection according to the Donnan equilibrium model. K3Fe(CN)6
having higher ion valance (3:1) was shown to have the highest rejection (> 90%), which is
considered that overall rejection is determined by a valence between two opposite electrostatic
forces and anion size. This is mainly due to the fact that the larger the anion valence relative to
the cation valence the stronger the net repulsive force, and, therefore, the salt rejection.
Table 2.9 summarizes performance evaluation in NaCl filtration test in three types of CNT
membrane. Overall, CNT composite membranes outperformed NaCl removal and permeate flux
of commercial PA-RO membrane in molecular dynamic simulation. Compared that VACNTs
membrane was proved to have high NaCl rejection and deliver several magnitude higher
permeate flux than polymeric membranes by molecular dynamic simulation, the results from lab
scale filtration has yet to be reported. A recent research demonstrated the applicability of
HACNT membrane for NaCl rejection by much enhanced adsorption capacity of plasma
modified long CNT [132]. This may contribute to point-of-use potable water purification in the
45
future. However, it is questionable for this type of membrane to be applied to recycling
concentrate in high pressure driven process. It is suggested that the following be developed: 1)
the HACNT membrane which can bear for high pressure, and 2) a suitable process for the
practical application.
46
Table 2.8 Summary of VACNT in water treatment.
Type of
VACNT
SW (6,6)
functionalized
with 8COO(5,5) armchair
type tube
(6,6) armchair
type tube
(7,7) armchair
type tube
(8,8) armchair
type tube
DW
MW
DW
SW
DW
Internal
Diameter
Length
Contact
angle
Flux
4.2
(× 10-12 m
s-1 Pa-1)
1.1 nm
1.3 nm
-
0.32 nm
~1.4 nm
-
0.47 nm
~1.4 nm
-
0.59 nm
~1.4 nm
-
0.75 nm
~1.4 nm
-
1.3 - 2 nm
1.3 - 2 nm
1.3 - 2 nm
1.3- 2 nm
1-5 μm
1-5 μm
1-5 μm
1-5 μm
-
1.2
LMH/bar
2.1
LMH/bar
3.2
LMH/bar
4.9
LMH/bar
5 (×1013
molecules
cm-2 s-1)
Operating
pressure
Method of CNT
synthesis/fabrication
Method of tip
opening
Ref.
deposition/
Reactive ion
etching in an
oxygen
containing
plasma
Corry
(2011)
[56]
deposition/
Reactive ion
etching in an
oxygen
containing
plasma
Corry
(2008)
[57]
deposition/
Reactive ion
etching in an
oxygen
containing
plasma
deposition/
Reactive ion
etching in an
oxygen
containing
plasma
NaCl
100
55 bar
NaCl
100
55 bar
NaCl
100
55 bar
NaCl
95
55 bar
NaCl
58
Ru2+(bipyr)3
1.3 nm of
particle size
Colloidal Au
2 ± 0.4 nm
of particle
size
Colloidal Au
5 ± 0.75 nm
of particle
size
Na4PTS
K3Fe(CN)6
K2SO4
CaSO4
KCl
0.82 bar
-
0.82 bar
-
< 3 (×108
molecules
cm-2 s-1)
0.82 bar
-
Re.
(%)
26.4 bar
< 2 (×109
molecules
cm-2 s-1)
-
Test
material
0.69 bar
47
-
-
96 %
90 %
80 %
37 %
37 %
Holt et
al.
(2006)
[26]
Fornasi
ero et
al. [58]
Table 2. 9 Summary of salt removal efficiency by CNTs membranes.
Type of
membrane
CNTs TFC
membrane
HACNTs
membrane
with support
layer
VACNTs
membrane
Filler matrix
Fabrication
procedure
System
Pure water
permeability
(LMH/bar)
Feed water
(mg/L)
Salt rejection
(%)
Removal
mechanism
Ref
MWCNTs/polyamide
(PA)
Polymer grafting
process
RO
0.71± 0.11
NaCl 4000
76 ± 1.10
Size exclusion
[120]
PA/CNTs
Interfacial
polymerization
RO
3
(15.5)
NaCl 2000
95
Size exclusion
[40]
20% CNTs/PA
Zwitter ion
functionalization of
CNTs + Interfacial
polymerization
RO
0.31 → 1.33
(36.5)
NaCl 1000
97.60 → 98.60
Steric hindrance +
electrostatic
repulsion
[42]
PA/CNT/PVA
Interfacial
polymerization
RO
4.15 ± 0.99
(15.5)
NaCl 2000
98.40 ± 0.51
Size exclusion
[133]
CNT/PES
Phase inversion
NF
0.2
(4)
NaCl 500
17
Size exclusion
[134]
Self- supported (Bucky
paper)
Vacuum filtration
DCMD*
52.9
(With 22.7 kPa**)
NaCl 3500
99
Vapor pressure
difference
[135]
Plasma modified ultralong CNTs on the MCE
support layer
Vacuum filtration
-
NaCl 3500
400% w/w
Adsorption
[132]
Polystyrenepolybutadiene (PS-PB)
copolymer
In-situ polymerization
612
-
-
-
[136]
UF
1100
-
-
-
[115]
Polyamide
RO
2 (20)
Size exclusion
[137]
Cellulose triacetate
FO
0.3
Epoxy infiltration
Polymeric
membrane
*DCMD: Direct contact membrane distillation, ** Vapor pressure
48
NaCl 580
90
100
2.4.3 Anti-fouling effect
Toxicity to bacteria/virus. CNT has toxicological effects, and its mechanism is to create
reactive oxygen species (ROS) production by chemical interactions with CNT. As free
radicals are formed by the interaction between CNT and virus and induce in vitro apoptosis,
peroxidative products deplete cell antioxidants [138]. This toxicological perspective of CNT
was introduced to antibacterial effects. Another research reported that diameter of CNT is the
key factor for antibacterial effects [139]. SWCNT, which have higher surface area and
shorter length than MWCNTs, were revealed to have higher toxicity by providing greater
interaction with the outer membrane of cells.
Antifouling. Incorporation of CNT into polymer also influenced antifouling properties of
CNT composite membrane by altering physical/chemical properties of the membrane. For
example, CNT/PES composite membrane was reported to exhibit a lower irreversible fouling
ratio for bovine serum albumin (BSA) and ovalbumin (OVA) and higher flux recovery ratio
compared to pure PES membrane [140]. This research also indicated that protein fouling on
CNT composite membrane could be overcome by simple washing. Recently, a publication on
the HACNTs membrane has presented fouling resistance and enhanced permeability [141].
Iron oxide doped CNT membrane fabricated by sintering at higher temperature achieved 90%
of sodium alginate (SA) as a model compound for EPS and minor flux decline after 4h of
operation probably due to iron doped CNT’s role as increasing hydrophilicity. Fouling
resistant CNTs/PES NF membrane by surface modification of CNT – amine functionalization
were developed recently[142]. This paper demonstrated that decreased roughness, negatively
charged membrane surface and increased hydrophilicity by insertion of amine functionalized
CNT into polymer matrix led to enhanced fouling recovery ratio. In another study, MWCNT
incorporated PES membrane had less amount of desorbed foulant than pure PES membrane,
indicating that CNT/PES membrane could alleviate membrane fouling [121].
49
However, the trade-off between high adsorption and membrane fouling was not completely
solved in the CNT enhanced membrane. For this reason, electrochemical system was
introduced to the membrane process in which electrical potential is applied to remove foulant
layer. Developing CNT membrane, which has enhanced electrical conductivity in the CNT
synthesis step, was attempted for removing foulant layer from the membrane [143]. The
mechanism of defouling/antifouling in electrically driven CNT membrane system is the
removal of foulant layer by giving a shock to the foulant layer via applying electricity.
Further, the electrically conductive membrane could contribute to decreasing foulant electromigration by applying electrical potential [144]. Interestingly, recent studies applied
electrochemistry for anti-biofouling (electrostatic interaction between charged CNT
membrane and bacteria) by CNT membrane. First, an anti-biofouling test was applied to the
electrically conducting HACNT membrane in a dead-end filtration by applying electric
potential. It showed 5.8 – 7.4 log removal value when 2 to 3 voltages were applied to CNT
filter [145, 146]. As the electrical potential is applied, a positively charged CNT filter acted
as an anode while negatively charged viral particles were oxidized on the MWCNT surface,
and then it resulted in direct oxidation. The fouling behavior of HACNT membrane was
investigated from large biomolecules to small molecules aromatics [147]. BSA and
naphthalene were found to be major foulants due to solution coagulation and π-π stacking.
SWCNT was found to be strong fouling resistant by size exclusion with BSA. The fouling
mechanisms were oxidation and electrochemical etching of the CNT. It removed oxidized
foulants by applying oxidation or reduction bias to the membrane as a working electrode.
However, applying positive bias shortened CNT due to the etching effect.
In addition, the applicability of electrochemical system was expanded to the CNT composite
membrane. Electrically conductive polymer RO and NF membrane was fabricated in
CNT/PA nanocomposite by interfacial polymerization and applied to cross flow
50
configuration [148]. Due to the directly contacting rejection layer of CNT/PA on the top of
PES support, the fabrication procedure conveyed conductivity of CNT to the PA membrane
and showed up to 400 S/cm (20 orders of magnitude greater than PA polymer). The
membrane achieved more than 95% of NaCl rejection and delivered a high water flux.
To sum up, anti-biofouling and fouling mitigation were attributed to electrical conductivity
and toxicology of CNT. Thus, how to introduce these properties to membrane matrix without
a loss is challenging in CNT membrane fabrication and its application.
2.4.4 How the CNTs enhanced membrane contributed to the membrane system?
This section discusses the suitability of CNT membrane to the membrane processes – MF/UF,
and RO. The membrane processes are classified via its performance and fabrication method
as shown in Fig 2.9.
Since dispersion of CNT in a polymer matrix to overcome its
aggregation was achieved through many studies, incorporation of CNT into the polymer has
been highly adopted to the membrane fabrication. From low-pressure (MF and UF) to highpressure (NR and RO) driven membrane process, CNTs incorporation into the polymer
membrane led to the superior outperformance, compared to that of conventional membranes
which are presented in Fig 2.9.
Fig 2.9 Classification of fabrication method in membrane process.
51
UF membrane system. Altering physical/chemical properties of the membrane during
membrane fabrication process (functionalized CNT incorporation into the polymer matrix)
made considerable performance enhancement in low pressure driven membrane (i.e. UF
membrane). It contributed to the flux enhancement by increasing pore size, porosity and/or
hydrophilicity [149, 150]. Another benefit was enhanced anti-biofouling properties driven by
increased surface roughness and decreased negative surface charge [121].
NF/RO membrane system. Progress of the CNT composite membranes in NF/RO system
were made via the insertion of CNT into i) the selective layer, either MPD (aqueous phase) or
TMC (organic solvent layer) via interfacial polymerization [40] and ii) UF support layer by
dispersing in a polymer matrix [151]. The CNT inserted into the PA selective layer led to
approximately 2 to 4-fold flux increase, enhanced salt rejection, antifouling effect and the
durability and chemical resistance against salty water [41, 133]. In addition, CNT
incorporation into the polymer matrix contributed to enhancing mechanical strength bearable to high pressure (more than 10- 20 bar) [152]. Such outperformance appeared
mainly due to the increased hydrophilicity and antimicrobial properties of CNT. CNT
inserted into UF support layer mainly increased porosity, water uptake capacity and
hydrophilicity of support layer, reducing the salt concentration polarization effect.
52
Table 2.10 Enhanced performance by the CNTs insertion in polymer membrane.
System
Membrane type
MF
CNT bridged 3D GO membrane
Enhancement in performance
Ref.
- Disinfection of pathogenic bacteria
[153]
- Toxic heavy metal removal
- Enhanced conductivity
MF
CNTs/PVDF hollow fiber membrane
UF
Hydrophilic CNTs coated membrane
[154]
- Increased permeability
- Improved antifouling properties
[15]
- High permeability
UF
Dual layer of CNTs/PVDF membrane
- Electrical conductivity
[143]
- Improved antifouling properties
Polyelectrolyte complex/MWCNTs
- Enhanced NF membrane performance
hybrid membrane
- Water softening
NF
[155]
2.4.5 Hybrid material with CNT as a new generation of membrane material
CNT composites have been tested for multi-functional effects for separation performance. In
the attempts, CNT was blended with other nanomaterials such as graphene oxide (GO) and
titanium dioxide (TiO2) and conductive polymers such as PANI and Poly (3,4ethylenedioxythiophene) (PEDOT).
2.4.5.1 CNT/GO and CNT/Graphene
CNT/GO complex was designed to supplement the limit of CNT membrane such as difficulty
in assembling VACNT array in the polymer matrix and its aggregation due to Van der Waals
force between CNT by incorporating with GO. Graphene-based material has many
advantages such as two-dimensional structure, mono-atomic thickness, antibacterial
properties, and high mechanical strength [156-158]. Several studies revealed that layered
two-dimensional GO bridged with one-dimensional CNT could deliver the synergetic effect
to the water treatment membrane. Research showed that 5:5 CNT/GO composite UF
53
membrane could contribute to enhancement in permeability and antifouling, compared to the
CNT and GO membrane itself [159]. This enhanced dispersion of CNT/GO complex
compared to individual one altered physical/chemical properties such as increased porosity
and hydrophilicity, resulting in a 2.5-fold increase in permeate flux. Further, this membrane
showed the enhanced antifouling effect and 98.8 % flux recovery due to the fact that the
affinity of hydrophilic low-dimensional CNT material hinders protein adsorption. In addition,
CNT/GO incorporated RO membrane was reported to have much enhanced water flux,
chlorine resistance, mechanical strength, durability compared to the pure PA-RO membrane
and GO PA membrane (without CNTs) [160]. CNT/GO complex was inserted in MPD
solution during the interfacial polymerization step. The enhanced performance could be
mostly attributed to the effect of GO insertion to the increased dispersion of CNT in the
aqueous solution and polymer matrix. A recent study also showed the progress in NF
membrane by the synergetic effect of GO (for molecular sieving) and CNT (for expanding
interlayer spaces between graphene sheets) [161]. These intercalated CNT/GO complex NF
membrane achieved increased water flux (100%) and high dye rejection (> 96%) in
comparison that GO membrane exhibited high permeate flux but low rejection efficiency.
However, the membrane remains inferior BSA fouling resistance, suggesting that an in-depth
study on engineering nanostructured membrane is required to overcome it. Furthermore, a 3D
GO bridged with CNT membrane showed a high removal efficiency of toxic metal and E.
coli O157: H7 bacteria simultaneously [153]. In this research, firstly CNT was conjugated
with PGLa antimicrobial peptide and glutathione for the purpose of selective removal of toxic
metals and then it was bridged with 2D GO. Finally, the 3D porous graphene-based
membrane was developed.
Besides the effect of CNT/GO, the effort on the synthesis of CNT/Graphene hybrid material
contributed to developing supercapacitor materials by the synergetic effect of both
54
nanomaterials [162-165]. For example, synthesis of CNT/Graphene led to the 3-dimensional
hybrid nanomaterial with excellent electrical conductivities and high specific surface area (>
2000 m2/g) due to the covalent transformation of sp2 carbon between the planar graphene and
SWCNTs [166]. CNT/graphene hybrid material was designed for supercapacitor due to its
highest energy density (188 Wh/kg) and high power density (200 kW/kg) [167]. Such
properties of CNT/Graphene hybrid material may be employed for developing the next
generation material of water desalination.
2.4.5.2 CNT/TiO2
Photo-catalyst such as TiO2 has been employed to the membrane for water purification with
photocatalytic degradation of many organic pollutants such as persistent toxic substances,
dyes, pesticides and pharmaceutical and personal care products (PPCPs) [168]. TiO2
nanotube arrays having photo-catalysts property showed much enhanced photocatalytic
activity due to the longer distance of electron transport and larger specific surface area [169171]. Furthermore, CNT/TiO2 has been developed in order to complement the limit from both
nanomaterials and to have the synergetic effect of high surface area and high photoreactivity
on decontamination process [172, 173]. Several studies focused on that CNT/TiO2 hybrid
materials have much enhanced photocatalytic activity compared to the TiO2 nanotubes by
minimizing recombination of photo induced electrons and holes [174-176]. TiO2 was limited
to the light absorption and rapid combination of electron-hole pairs. However, TiO2
incorporated with CNT can be considered to be the robust hybrid material due to the
excellent transporting electron properties and large surface area. Leveraged by the
development of CNT/TiO2 hybrid materials, a multifunctional membrane having superior
adsorption capacity coupled with photocatalytic degradation has been thoroughly studied to
aim at completely overcoming the fouling issue from the degradation of organic foulants. For
example, 0.5/0.5 (wt%) of TiO2 and MWCNT exhibited superior antifouling property and
55
nearly 60 % TOC rejection, in comparison that pure PSf, and individually inserted TiO2 and
MWCNT to PSf membrane showed only less than 40 % TOC rejection [177]. TiO2/MWCNT
membrane exhibited the lowest reversible/irreversible fouling ratio in both low (2 ppm) and
high (700 ppm) HA filtration. This research reported that the superior performance was
attributed to the synergetic effects of CNT and TiO2 nanoparticle to engineer
physical/chemical properties of membrane: 300 % increase in mean pore size, increased
hydrophilicity and surface roughness.
An attempt to highlight the synergetic effect of photocatalytic degradation with sieving
mechanism is stimulated by the recent research on the fabrication of core web structured
CNT/ZnO/TiO2 nanocomposite membrane [178]. CNT/ZnO/TiO2 nanocomposite was
synthesized
via
hydrothermal
and
acid
treatment
coupled
with
ultrasonication.
CNT/ZnO/TiO2 membrane outperformed A07 degradation rate and permeability of single
CNT and TiO2 membrane by multi-channeled pores and increased specific surface area for
photocatalytic degradation.
2.4.5.3 CNT/PANI and CNT/PEDOT
CNT/PANI nanocomposite has been developed for the purpose of enhancing the conductivity
of both CNT and PANI [179-181], which can contribute to delivering the conductivity of
CNT. Besides conveying the conductivities, PANI has been used as nanofiller to increase
porosity and hydrophilicity of the membrane [182, 183]. Recent publications on CNT/PANI
membrane demonstrate much enhanced water permeability and increased removal efficiency
by altering physical/chemical properties of the membrane [44]. Further, SWCNT/PANI was
synthesized by a flash welding method, and this membrane exhibited 10 factors of
permeability and more than 90 % of SiO2 nanoparticle and around 20% of BSA rejection
[184]. In addition, MWCNT/PANI magnetic composites fabricated by plasma induced
56
polymerization were applied for removal of Pb (II) due to its enhanced adsorption capacity of
MWCNT [185].
2.5
Current challenges and future perspectives
The future research on membrane technology for water treatment is suggested to focus on
overcoming trade-off between high removal efficiency and vulnerability to membrane fouling.
CNT incorporation to the polymer matrix has been adopted to develop the nano-engineered
membrane that can overcome the current challenges of polymeric membranes.
Despite the considerable contribution of CNT membrane to complementing the limitation of
the conventional polymeric membranes, CNT membrane remains some challenging issues to
be overcome. CNT nanocomposite membrane has significantly contributed to enhance
separation performance in pressure driven membrane processes such as UF, NF and RO.
However, it is only limited to the functionalized CNT properties. Further, there is lack of
studies on the application of CNT enhanced membrane to seawater pretreatment that
challenges in low pressure driven membrane system in desalination plant.
The attempts on complementing individual nanomaterials have inspired further research
development on hybrid nanomaterials. These hybrid complexes could be an ideal candidate
for a multifunctional membrane with enhanced fouling resistance, which could achieve
superior performance without trade-off between high selectivity and highly adsorbed foulant
in the membrane. For example, CNT/TiO2 membrane driven by photocatalytic degradation
could be an option of alleviating membrane fouling as delivering high performance.
Electrically enhanced CNT/Graphene complex would be employed to the electrochemical
membrane system, which removes foulant layer by applying electrical potential. Further, as a
newly developed membrane material, CNT/Graphene has a potential to RO and/or MD for
water desalination besides its benefits for supercapacitors in CDI. However, efforts on
57
developing membrane structure with such hybrid nanomaterial are required to a first place in
order to facilitate their synergetic effect of two properties while simply blending those
nanohybrid complexes with polymers.
Based on the research gap in conventional polymeric membranes and nanomaterial
incorporated membranes, the research scope in this thesis is focused on the development of
CNT composite membrane that can tackle the challenge in low pressure driven membrane
such as low organic matter removal rate and permeate flux. Furthermore, lack of studies on
the protein fouling in the various conditions of solution chemistry will be covered in depth so
as that it will contribute to the understanding of fouling mechanisms in waste water treatment.
58
References
1.
Pendergast, M.M. and E.M.V. Hoek, A review of water treatment membrane
nanotechnologies. Energy & Environmental Science, 2011. 4(6): p. 1946-1971.
2.
Malaeb, L. and G.M. Ayoub, Reverse osmosis technology for water treatment: State
of the art review. Desalination, 2011. 267(1): p. 1-8.
3.
Alborzfar, M., G. Jonsson, and C. Gron, Removal of natural organic matter from two
types of humic ground waters by nanofiltration. Water Research, 1998. 32(10): p.
2983-2994.
4.
Greenlee, L.F., D.F. Lawler, B.D. Freeman, B. Marrot, and P. Moulin, Reverse
osmosis desalination: Water sources, technology, and today's challenges. Water
Research, 2009. 43(9): p. 2317-2348.
5.
Kim, J., Z. Cai, and M.M. Benjamin, NOM fouling mechanisms in a hybrid
adsorption/membrane system. Journal of Membrane Science, 2010. 349(1-2): p. 35-43.
6.
Ng, L.Y., A.W. Mohammad, C.P. Leo, and N. Hilal, Polymeric membranes
incorporated with metal/metal oxide nanoparticles: A comprehensive review.
Desalination, 2013. 308: p. 15-33.
7.
Lee, K.P., T.C. Arnot, and D. Mattia, A review of reverse osmosis membrane
materials for desalination-Development to date and future potential. Journal of
Membrane Science, 2011. 370(1-2): p. 1-22.
8.
Pan, B. and B. Xing, Adsorption Mechanisms of Organic Chemicals on Carbon
Nanotubes. Environmental Science & Technology, 2008. 42(24): p. 9005-9013.
9.
Hyung, H. and J.-H. Kim, Natural organic matter (NOM) adsorption to multi-walled
carbon nanotubes: Effect of NOM characteristics and water quality parameters.
Environmental Science & Technology, 2008. 42(12): p. 4416-4421.
10.
Wildöer, J.W.G., L.C. Venema, A.G. Rinzler, R.E. Smalley, and C. Dekker,
Electronic structure of atomically resolved carbon nanotubes. Nature, 1998.
391(6662): p. 59-62.
11.
Joseph, S. and N.R. Aluru, Why are carbon nanotubes fast transporters of water?
Nano Letters, 2008. 8(2): p. 452-458.
12.
Xie, X.L., Y.W. Mai, and X.P. Zhou, Dispersion and alignment of carbon nanotubes
in polymer matrix: A review. Materials Science & Engineering R-Reports, 2005. 49(4):
p. 89-112.
59
13.
Yin, J. and B. Deng, Polymer-matrix nanocomposite membranes for water treatment.
Journal of Membrane Science, 2015. 479(0): p. 256-275.
14.
Qiu, S., L. Wu, X. Pan, L. Zhang, H. Chen, and C. Gao, Preparation and properties
of functionalized carbon nanotube/PSF blend ultrafiltration membranes. Journal of
15.
Bai, L.M., H. Liang, J. Crittenden, F.S. Qu, A. Ding, J. Ma, X. Du, S.D. Guo, and
G.B. Li, Surface modification of UF membranes with functionalized MWCNTs to
control membrane fouling by NOM fractions. Journal of Membrane Science, 2015.
492: p. 400-411.
16.
Rahimpour, A., M. Jahanshahi, S. Khalili, A. Mollahosseini, A. Zirepour, and B.
Rajaeian, Novel functionalized carbon nanotubes for improving the surface properties
and performance of polyethersulfone (PES) membrane. Desalination, 2012. 286: p.
99-107.
17.
Gupta, V.K., O. Moradi, I. Tyagi, S. Agarwal, H. Sadegh, R. Shahryari-Ghoshekandi,
A.S.H. Makhlouf, M. Goodarzi, and A. Garshasbi, Study on the removal of heavy
metal ions from industry waste by carbon nanotubes: Effect of the surface
modification: a review. Critical Reviews in Environmental Science and Technology,
2016. 46(2): p. 93-118.
18.
Kumar, R., M.A. Khan, and N. Haq, Application of Carbon Nanotubes in Heavy
Metals Remediation. Critical Reviews in Environmental Science and Technology,
2014. 44(9): p. 1000-1035.
19.
Goh, P.S., A.F. Ismail, and B.C. Ng, Carbon nanotubes for desalination:
Performance evaluation and current hurdles. Desalination, 2013. 308(0): p. 2-14.
20.
Terrones, M., Science and Technology of the Twenty-First Century: Synthesis,
Properties, and Applications of Carbon Nanotubes, in Annual Review of Materials
Research. 2003. p. 419-501.
21.
Mauter, M.S. and M. Elimelech, Environmental applications of carbon-based
nanomaterials. Environmental Science & Technology, 2008. 42(16): p. 5843-5859.
22.
Savage, N., M. Diallo, J. Duncan, A. Street, and R. Sustich, Nanotechnology
Applications for Clean Water. 2009, William Andrew Publishing. p. 80.
23.
Odom, T.W., J.L. Huang, P. Kim, and C.M. Lieber, Atomic structure and electronic
properties of single-walled carbon nanotubes. Nature, 1998. 391(6662): p. 62-64.
60
24.
Ma, P.-C., N.A. Siddiqui, G. Marom, and J.-K. Kim, Dispersion and functionalization
of carbon nanotubes for polymer-based nanocomposites: A review. Composites Part
a-Applied Science and Manufacturing, 2010. 41(10): p. 1345-1367.
25.
Dresselhaus, M.S., Y.M. Lin, O. Rabin, A. Jorio, A.G. Souza, M.A. Pimenta, R. Saito,
G.G. Samsonidze, and G. Dresselhaus, Nanowires and nanotubes. Materials Science
& Engineering C-Biomimetic and Supramolecular Systems, 2003. 23(1-2): p. 129-140.
26.
Holt, J.K., H.G. Park, Y.M. Wang, M. Stadermann, A.B. Artyukhin, C.P.
Grigoropoulos, A. Noy, and O. Bakajin, Fast mass transport through sub-2nanometer carbon nanotubes. Science, 2006. 312(5776): p. 1034-1037.
27.
Lu, C.S., Y.L. Chung, and K.F. Chang, Adsorption of trihalomethanes from water
with carbon nanotubes. Water Research, 2005. 39(6): p. 1183-1189.
28.
Yang, K., X. Wang, L. Zhu, and B. Xing, Competitive sorption of pyrene,
phenanthrene, and naphthalene on multiwalled carbon nanotubes. Environmental
29.
Mubarak, N.M., J.N. Sahu, E.C. Abdullah, and N.S. Jayakumar, Removal of Heavy
Metals from Wastewater Using Carbon Nanotubes. Separation & Purification
Reviews, 2014. 43(4): p. 311-338.
30.
Goh, P.S., B.C. Ng, W.J. Lau, and A.F. Ismail, Inorganic Nanomaterials in Polymeric
Ultrafiltration Membranes for Water Treatment. Separation and Purification Reviews,
2015. 44(3): p. 216-249.
31.
modifications for the effective dispersion of carbon nanotubes in solvents and
polymers. Carbon, 2012. 50(1): p. 3-33.
32.
de Lannoy, C.-F., E. Soyer, and M.R. Wiesner, Optimizing carbon nanotubereinforced
polysulfone
ultrafiltration
membranes
through
carboxylic
acid
functionalization. Journal of Membrane Science, 2013. 447: p. 395-402.
33.
Zheng, Q.-Z., P. Wang, Y.-N. Yang, and D.-J. Cui, The relationship between porosity
and kinetics parameter of membrane formation in PSF ultrafiltration membrane.
Journal of Membrane Science, 2006. 286(1-2): p. 7-11.
34.
Barth, C., M.C. Goncalves, A.T.N. Pires, J. Roeder, and B.A. Wolf, Asymmetric
polysulfone and polyethersulfone membranes: effects of thermodynamic conditions
during formation on their performance. Journal of Membrane Science, 2000. 169(2):
p. 287-299.
61
35.
Choi, J.-H., J. Jegal, and W.-N. Kim, Fabrication and characterization of multiwalled carbon nanotubes/polymer blend membranes. Journal of Membrane Science,
2006. 284(1-2): p. 406-415.
36.
Wu, H., B. Tang, and P. Wu, MWNTs/Polyester Thin Film Nanocomposite Membrane:
An Approach To Overcome the Trade-Off Effect between Permeability and Selectivity.
Journal of Physical Chemistry C, 2010. 114(39): p. 16395-16407.
37.
Yip, N.Y., A. Tiraferri, W.A. Phillip, J.D. Schiffman, and M. Elimelech, High
Performance Thin-Film Composite Forward Osmosis Membrane. Environmental
38.
Cheng, Z.L., X. Li, Y.D. Liu, and T.-S. Chung, Robust outer-selective thin-film
composite polyethersulfone hollow fiber membranes with low reverse salt flux for
renewable salinity-gradient energy generation. Journal of Membrane Science. 506: p.
119-129.
39.
Ghosh, A.K. and E.M.V. Hoek, Impacts of support membrane structure and chemistry
on polyamide-polysulfone interfacial composite membranes. Journal of Membrane
Science, 2009. 336(1-2): p. 140-148.
40.
Kim, H.J., K. Choi, Y. Baek, D.-G. Kim, J. Shim, J. Yoon, and J.-C. Lee, HighPerformance Reverse Osmosis CNT/Polyamide Nanocomposite Membrane by
Controlled Interfacial Interactions. Acs Applied Materials & Interfaces, 2014. 6(4): p.
2819-2829.
41.
Baroña, G.N.B., J. Lim, M. Choi, and B. Jung, Interfacial polymerization of
polyamide-aluminosilicate SWNT nanocomposite membranes for reverse osmosis.
Desalination, 2013. 325(0): p. 138-147.
42.
Chan, W.F., H.Y. Chen, A. Surapathi, M.G. Taylor, X.H. Hao, E. Marand, and J.K.
Johnson, Zwitterion Functionalized Carbon Nanotube/Polyamide Nanocomposite
Membranes for Water Desalination. Acs Nano, 2013. 7(6): p. 5308-5319.
43.
Kim, E., G. Hwang, M.G. El-Din, and Y. Liu, Development of nanosilver and multiwalled carbon nanotubes thin-film nanocomposite membrane for enhanced water
treatment. Journal of Membrane Science, 2012. 394: p. 37-48.
44.
Lee, J., Y. Ye, A. Ward, C. Zhou, V. Chen, A.I. Minett, S. Lee, Z. Liu, S. Chae, and J.
Shi, High flux and High selectivity carbon nanotube enhanced membrane for natural
organic matter removal. Separation and Purification Technology, Accepted.
45.
Dumee, L., V. Germain, K. Sears, J. Schuetz, N. Finn, M. Duke, S. Cerneaux, D.
Cornu, and S. Gray, Enhanced durability and hydrophobicity of carbon nanotube
62
bucky paper membranes in membrane distillation. Journal of Membrane Science,
2011. 376(1-2): p. 241-246.
46.
Yuan, D.J., W. Lin, R. Guo, C.P. Wong, and S. Das, The fabrication of vertically
aligned and periodically distributed carbon nanotube bundles and periodically
porous carbon nanotube films through a combination of laser interference ablation
and metal-catalyzed chemical vapor deposition. Nanotechnology, 2012. 23(21).
47.
Hussein, L., G. Urban, and M. Kruger, Fabrication and characterization of
buckypaper-based nanostructured electrodes as a novel material for biofuel cell
applications. Physical Chemistry Chemical Physics, 2011. 13(13): p. 5831-5839.
48.
Kim, Y.A., H. Muramatsu, T. Hayashi, M. Endo, M. Terrones, and M.S. Dresselhaus,
Fabrication of high-purity, double-walled carbon nanotube buckypaper. Chemical
Vapor Deposition, 2006. 12(6): p. 327-+.
49.
Endo, M., H. Muramatsu, T. Hayashi, Y.A. Kim, M. Terrones, and N.S. Dresselhaus,
'Buckypaper' from coaxial nanotubes. Nature, 2005. 433(7025): p. 476-476.
50.
Sears, K., L. Dumée, J. Schütz, M. She, C. Huynh, S. Hawkins, M. Duke, and S. Gray,
Recent Developments in Carbon Nanotube Membranes for Water Purification and
Gas Separation. Materials, 2010. 3(1): p. 127-149.
51.
Yang, X., J. Lee, L. Yuan, S.-R. Chae, V.K. Peterson, A.I. Minett, Y. Yin, and A.T.
Harris, Removal of natural organic matter in water using functionalised carbon
nanotube buckypaper. Carbon, 2013. 59(0): p. 160-166.
52.
Kang, S., M. Pinault, L.D. Pfefferle, and M. Elimelech, Single-walled carbon
nanotubes exhibit strong antimicrobial activity. Langmuir, 2007. 23(17): p. 86708673.
53.
Yang, X., L. Yuan, V.K. Peterson, A.I. Minett, Y. Yin, and A.T. Harris, Facile
Preparation of Free-Standing Carbon Nanotube Arrays Produced Using Two-Step
Floating-Ferrocene Chemical Vapor Deposition. Acs Applied Materials & Interfaces,
2012. 4(3): p. 1417-1422.
54.
Barreiro, A., S. Hampel, M.H. Rummeli, C. Kramberger, A. Gruneis, K. Biedermann,
A. Leonhardt, T. Gemming, B. Buchner, A. Bachtold, and T. Pichler, Thermal
decomposition of ferrocene as a method for production of single-walled carbon
nanotubes without additional carbon sources. Journal of Physical Chemistry B, 2006.
110(42): p. 20973-20977.
63
55.
Prasek, J., J. Drbohlavova, J. Chomoucka, J. Hubalek, O. Jasek, V. Adam, and R.
Kizek, Methods for carbon nanotubes synthesis-review. Journal of Materials
Chemistry, 2011. 21(40): p. 15872-15884.
56.
Corry, B., Water and ion transport through functionalised carbon nanotubes:
implications for desalination technology. Energy & Environmental Science, 2011.
4(3): p. 751-759.
57.
Corry, B., Designing carbon nanotube membranes for efficient water desalination.
Journal of Physical Chemistry B, 2008. 112(5): p. 1427-1434.
58.
Fornasiero, F., H.G. Park, J.K. Holt, M. Stadermann, C.P. Grigoropoulos, A. Noy,
and O. Bakajin, Ion exclusion by sub-2-nm carbon nanotube pores. Proceedings of the
National Academy of Sciences of the United States of America, 2008. 105(45): p.
17250-17255.
59.
Hinds, B.J., N. Chopra, T. Rantell, R. Andrews, V. Gavalas, and L.G. Bachas, Aligned
multiwalled carbon nanotube membranes. Science, 2004. 303(5654): p. 62-65.
60.
Noy, A., H.G. Park, F. Fornasiero, J.K. Holt, C.P. Grigoropoulos, and O. Bakajin,
Nanofluidics in carbon nanotubes. Nano Today, 2007. 2(6): p. 22-29.
61.
Lee, D.H., D.O. Shin, W.J. Lee, and S.O. Kim, Hierarchically organized carbon
nanotube arrays from self-assembled block copolymer nanotemplates. Advanced
Materials, 2008. 20(13): p. 2480-+.
62.
Qu, L.T., W. Chen, L.M. Dai, A. Roy, and T.B. Tolle, Polymer and aligned carbon
nanotube nanocomposites and nanodevices. Sampe Journal, 2007. 43(6): p. 38-46.
63.
Yamada, T., T. Namai, K. Hata, D.N. Futaba, K. Mizuno, J. Fan, M. Yudasaka, M.
Yumura, and S. Iijima, Size-selective growth of double-walled carbon nanotube
forests from engineered iron catalysts. Nature Nanotechnology, 2006. 1(2): p. 131136.
64.
Song, C. and B. Corry, Intrinsic Ion Selectivity of Narrow Hydrophobic Pores.
Journal of Physical Chemistry B, 2009. 113(21): p. 7642-7649.
65.
Eijkel, J., Liquid slip in micro- and nanofluidics: recent research and its possible
implications. Lab on a Chip, 2007. 7(3): p. 299-301.
66.
Maali, A. and B. Bhushan, Measurement of slip length on superhydrophobic surfaces.
Philosophical Transactions of the Royal Society a-Mathematical Physical and
Engineering Sciences, 2012. 370(1967): p. 2304-2320.
64
67.
Zhou, M., J. Li, C. Wu, X. Zhou, and L. Cai, Fluid drag reduction on
superhydrophobic surfaces coated with carbon nanotube forests (CNTs). Soft Matter,
2011. 7(9): p. 4391-4396.
68.
Philip, J.R., Flows satisfying mixed no-slip and no-shear conditions. Zeitschrift Fur
Angewandte Mathematik Und Physik, 1972. 23(3): p. 353-372.
69.
Lauga, E. and H.A. Stone, Effective slip in pressure-driven Stokes flow. Journal of
Fluid Mechanics, 2003. 489: p. 55-77.
70.
Cottin-Bizonne, C., C. Barentin, E. Charlaix, L. Bocquet, and J.L. Barrat, Dynamics
of simple liquids at heterogeneous surfaces: Molecular-dynamics simulations and
hydrodynamic description. European Physical Journal E, 2004. 15(4): p. 427-438.
71.
Ybert, C., C. Barentin, C. Cottin-Bizonne, P. Joseph, and L. Bocquet, Achieving large
slip with superhydrophobic surfaces: Scaling laws for generic geometries. Physics of
Fluids, 2007. 19(12): p. 123601-.
72.
Davis, A.M.J. and E. Lauga, Hydrodynamic friction of fakir-like superhydrophobic
surfaces. Journal of Fluid Mechanics, 2010. 661: p. 402-411.
73.
Joseph, P., C. Cottin-Bizonne, J.M. Benoit, C. Ybert, C. Journet, P. Tabeling, and L.
Bocquet, Slippage of water past superhydrophobic carbon nanotube forests in
microchannels. Physical Review Letters, 2006. 97(15): p. 156104-.
74.
Lee, C., C.H. Choi, and C.J. Kim, Structured surfaces for a giant liquid slip. Physical
Review Letters, 2008. 101(6).
75.
Lee, C. and C.-J.C.J. Kim, Maximizing the Giant Liquid Slip on Superhydrophobic
Microstructures by Nanostructuring Their Sidewalls. Langmuir, 2009. 25(21): p.
12812-12818.
76.
Ou, J. and J.P. Rothstein, Direct velocity measurements of the flow past dragreducing ultrahydrophobic surfaces. Physics of Fluids, 2005. 17(10): p. 103606-.
77.
Ou, J., B. Perot, and J.P. Rothstein, Laminar drag reduction in microchannels using
ultrahydrophobic surfaces. Physics of Fluids, 2004. 16(12): p. 4635-4643.
78.
Steinberger, A., C. Cottin-Bizonne, P. Kleimann, and E. Charlaix, High friction on a
bubble mattress. Nature Materials, 2007. 6(9): p. 665-668.
79.
Steinberger, A., C. Cottin-Bizonne, P. Kleimann, and E. Charlaix, Nanoscale flow on
a bubble mattress: Effect of surface elasticity. Physical Review Letters, 2008. 100(13):
p. 134501-.
65
80.
Bhushan, B., Y. Wang, and A. Maali, Boundary Slip Study on Hydrophilic,
Hydrophobic, and Superhydrophobic Surfaces with Dynamic Atomic Force
Microscopy. Langmuir, 2009. 25(14): p. 8117-8121.
81.
Jai, C., T. Cohen-Bouhacina, and A. Maali, Analytical description of the motion of an
acoustic-driven atomic force microscope cantilever in liquid. Applied Physics Letters,
2007. 90(11).
82.
Cottin-Bizonne, C., S. Jurine, J. Baudry, J. Crassous, F. Restagno, and E. Charlaix,
Nanorheology: An investigation of the boundary condition at hydrophobic and
hydrophilic interfaces. European Physical Journal E, 2002. 9(1): p. 47-53.
83.
Craig, V.S., C. Neto, and D.R. Williams, Shear-dependent boundary slip in an
aqueous Newtonian liquid. Physical Review Letters, 2001. 87(5): p. 054504-054504.
84.
Majumder, M., N. Chopra, R. Andrews, and B.J. Hinds, Nanoscale hydrodynamics:
Enhanced flow in carbon nanotubes. Nature, 2005. 438(7064): p. 44-44.
85.
Voronov, R.S., D.V. Papavassiliou, and L.L. Lee, Slip length and contact angle over
hydrophobic surfaces. Chemical Physics Letters, 2007. 441(4-6): p. 273-276.
86.
Myers, T.G., Why are slip lengths so large in carbon nanotubes? Microfluidics and
Nanofluidics, 2011. 10(5): p. 1141-1145.
87.
H.S. Philip Wong, D.A., Carbon nanotube and graphene device physics. 2011, United
Kingdom Cambridge university press.
88.
Zheng, J., E.M. Lennon, H.K. Tsao, Y.J. Sheng, and S.Y. Jiang, Transport of a liquid
water and methanol mixture through carbon nanotubes under a chemical potential
gradient. Journal of Chemical Physics, 2005. 122(21): p. 214702-.
89.
Allen, T.W., S. Kuyucak, and S.H. Chung, The effect of hydrophobic and hydrophilic
channel walls on the structure and diffusion of water and ions. Journal of Chemical
Physics, 1999. 111(17): p. 7985-7999.
90.
Gulotty, R., M. Castellino, P. Jagdale, A. Tagliaferro, and A.A. Balandin, Effects of
Functionalization on Thermal Properties of Single-Wall and Multi-Wall Carbon
Nanotube-Polymer Nanocomposites. Acs Nano, 2013. 7(6): p. 5114-5121.
91.
Lee, S.B., D.T. Mitchell, L. Trofin, T.K. Nevanen, H. Soderlund, and C.R. Martin,
Antibody-based bio-nanotube membranes for enantiomeric drug separations. Science,
2002. 296(5576): p. 2198-2200.
92.
Majumder, M., N. Chopra, and B.J. Hinds, Effect of tip functionalization on transport
through vertically oriented carbon nanotube membranes. Journal of the American
Chemical Society, 2005. 127(25): p. 9062-9070.
66
93.
Nednoor, P., V.G. Gavalas, N. Chopra, B.J. Hinds, and L.G. Bachas, Carbon
nanotube based biomimetic membranes: mimicking protein channels regulated by
phosphorylation. Journal of Materials Chemistry, 2007. 17(18): p. 1755-1757.
94.
Datsyuk, V., M. Kalyva, K. Papagelis, J. Parthenios, D. Tasis, A. Siokou, I. Kallitsis,
and C. Galiotis, Chemical oxidation of multiwalled carbon nanotubes. Carbon, 2008.
46(6): p. 833-840.
95.
Shen, J., W. Huang, L. Wu, Y. Hu, and M. Ye, Study on amino-functionalized
multiwalled carbon nanotubes. Materials Science and Engineering: A, 2007. 464(1–2):
p. 151-156.
96.
Joseph, S., R.J. Mashl, E. Jakobsson, and N.R. Aluru, Electrolytic transport in
modified carbon nanotubes. Nano Letters, 2003. 3(10): p. 1399-1403.
97.
Li, Q., D. Yang, J. Shi, X. Xu, S. Yan, and Q. Liu, Biomimetic modification of large
diameter carbon nanotubes and the desalination behavior of its reverse osmosis
membrane. Desalination, 2016. 379: p. 164-171.
98.
Dokyoung Yoon, C.L., Jongju Yun, Wonjae Jeon, Bong Jun Cha, Seunghyun Baik,
Enhanced condensation, agglomeration, and rejection of water vapor by
superhydrophobic aligned multiwalled carbon nanotube membranes. Acs Nano, 2012.
6(7): p. 5980-5987.
99.
Zhan, X., J. Wu, Z.Q. Chen, and B.J. Hinds, Single-step electrochemical
functionalization of double-walled carbon nanotube (DWCNT) membranes and the
demonstration of ionic rectification. Nanoscale Research Letters, 2013. 8: p. 279-.
100.
Wang, Z., L. Ci, L. Chen, S. Nayak, P.M. Ajayan, and N. Koratkar, Polaritydependent electrochemically controlled transport of water through carbon nanotube
membranes. Nano Letters, 2007. 7(3): p. 697-702.
101.
Huang, S.M. and L.M. Dai, Plasma etching for purification and controlled opening of
aligned carbon nanotubes. Journal of Physical Chemistry B, 2002. 106(14): p. 35433545.
102.
Wang, C.F., S.J. Guo, X.L. Pan, W. Chen, and X.H. Bao, Tailored cutting of carbon
nanotubes and controlled dispersion of metal nanoparticles inside their channels.
Journal of Materials Chemistry, 2008. 18(47): p. 5782-5786.
103.
Tsang, S.C., P.J.F. Harris, and M.L.H. Green, Thinining and opening of carbon
nanotubes by oxidation using carbon-dioxide. Nature, 1993. 362(6420): p. 520-522.
67
104.
Ajayan, P.M., T.W. Ebbesen, T. Ichihashi, S. Iijima, K. Tanigaki, and H. Hiura,
Opening carbon nanotubes with oxygen and implications for filling. Nature, 1993.
362(6420): p. 522-525.
105. Liu, J., A.G. Rinzler, H.J. Dai, J.H. Hafner, R.K. Bradley, P.J. Boul, A. Lu, T. Iverson,
K. Shelimov, C.B. Huffman, F. Rodriguez-Macias, Y.S. Shon, T.R. Lee, D.T. Colbert,
and R.E. Smalley, Fullerene pipes. Science, 1998. 280(5367): p. 1253-1256.
106.
Zhao, W., C.H. Song, and P.E. Pehrsson, Water-soluble and optically pH-sensitive
single-walled carbon nanotubes from surface modification. Journal of the American
107.
Hedges, J.I., Organic Geochemistry of Natural Waters: E. M. Thurman. Martinus
Nijhoff/Dr. W. Junk Publishers, 1985, 497 p., $39.50. Geochimica Et Cosmochimica
Acta, 1986. 50(9): p. 2119.
108.
Summers, R.S. and P.V. Roberts, Activated carbon adsorption of humic substances. 2.
Size exclusion and electrostatic interactions. Journal of Colloid and Interface Science,
1988. 122(2): p. 382-397.
109.
Kabsch-Korbutowicz, M., Application of ultrafiltration integrated with coagulation
for improved NOM removal. Desalination, 2005. 174(1): p. 13-22.
110.
Schafer, A.I., A.G. Fane, and T.D. Waite, Nanofiltration of natural organic matter:
Removal, fouling and the influence of multivalent ions. Desalination, 1998. 118(1-3):
p. 109-122.
111.
Li, Q.L. and M. Elimelech, Synergistic effects in combined fouling of a loose
nanofiltration membrane by colloidal materials and natural organic matter. Journal
of Membrane Science, 2006. 278(1-2): p. 72-82.
112.
Schafer, A.I., A.G. Fane, and T.D. Waite, Fouling effects on rejection in the
membrane filtration of natural waters. Desalination, 2000. 131(1-3): p. 215-224.
113.
Saleh, N.B., L.D. Pfefferle, and M. Elimelech, Aggregation Kinetics of Multiwalled
Carbon Nanotubes in Aquatic Systems: Measurements and Environmental
Implications. Environmental Science & Technology, 2008. 42(21): p. 7963-7969.
114.
Park, S.M., Y.K. Choi, S. Lee, Y. Baek, J. Yoon, D.K. Seo, and Y.H. Kim,
Experimental analysis of transport characteristics for vertically aligned carbon
nanotube membranes. Desalination and Water Treatment, 2013. 51(25-27): p. 53495354.
115.
Baek, Y., C. Kim, D.K. Seo, T. Kim, J.S. Lee, Y.H. Kim, K.H. Ahn, S.S. Bae, S.C.
Lee, J. Lim, K. Lee, and J. Yoon, High performance and antifouling vertically aligned
68
carbon nanotube membrane for water purification. Journal of Membrane Science,
2014. 460: p. 171-177.
116.
Zularisam, A.W., A.F. Ismail, M.R. Salim, M. Sakinah, and H. Ozaki, The effects of
natural organic matter (NOM) fractions on fouling characteristics and flux recovery
of ultrafiltration membranes. Desalination, 2007. 212(1-3): p. 191-208.
117.
Lee, S.J. and J.H. Kim, Differential natural organic matter fouling of ceramic versus
polymeric ultrafiltration membranes. Water Research, 2014. 48: p. 43-51.
118.
Lowe, J. and M.M. Hossain, Application of ultrafiltration membranes for removal of
humic acid from drinking water. Desalination, 2008. 218(1-3): p. 343-354.
119.
Domany, Z., I. Galambos, G. Vatai, and E. Bekassy-Molnar, Humic substances
removal from drinking water by membrane filtration. Desalination, 2002. 145(1–3): p.
333-337.
120.
Shawky, H.A., S.R. Chae, S.H. Lin, and M.R. Wiesner, Synthesis and
characterization of a carbon nanotube/polymer nanocomposite membrane for water
treatment. Desalination, 2011. 272(1-3): p. 46-50.
121.
Celik, E., H. Park, and H. Choi, Carbon nanotube blended polyethersulfone
membranes for fouling control in water treatment. Water Research, 2011. 45(1): p.
274-282.
122.
Bonnelye, V., L. Guey, and J. Del Castillo, UF/MF as RO pre-treatment: the real
benefit. Desalination, 2008. 222(1-3): p. 59-65.
123.
Voutchkov, N., Considerations for selection of seawater filtration pretreatment
system. Desalination, 2010. 261(3): p. 354-364.
124.
Jeong, S.H., T.V. Nguyen, and S. Vigneswaran, Submerged membrane coagulation
hybrid system as pretreatment to organic matter removal from seawater. Water
Science and Technology-Water Supply, 2011. 11(3): p. 352-357.
125.
Campinas, M. and M.J. Rosa, Assessing PAC contribution to the NOM fouling control
in PAC/UF systems. Water Research, 2010. 44(5): p. 1636-1644.
126.
Tansakul, C., S. Laborie, and C. Cabassud, Adsorption combined with ultrafiltration
to remove organic matter from seawater. Water Research, 2011. 45(19): p. 6362-6370.
127.
Duan, J.M., F. Wilson, N. Graham, and J.H. Tay, Adsorption of humic acid by
powdered activated carbon in saline water conditions. Desalination, 2003. 151(1): p.
53-66.
128.
Tomaszewska, M. and S. Mozia, Removal of organic matter from water by PAC/UF
system. Water Research, 2002. 36(16): p. 4137-4143.
69
129.
http://www.dow.com/liquidseps/prod/sw30hr_380htm.
130.
Donnan, F.G., The theory of membrane equilibria. Chemical Reviews, 1924. 1(1): p.
73-90.
131.
Donnan, F.G., Theory of membrane equilibria and membrane-potentials in the
presence of non-dialyzing electrolytes - A contribution to physical-chemical
physiology Reprinted from zeitshrift fur electrochemie und angewandte physikalische
chemie, Vol 17, PG 572, 1911). Journal of Membrane Science, 1995. 100(1): p. 4555.
132.
Yang, H.Y., Z.J. Han, S.F. Yu, K.L. Pey, K. Ostrikov, and R. Karnik, Carbon
nanotube membranes with ultrahigh specific adsorption capacity for water
desalination and purification. Nature communications, 2013. 4: p. 2220-2220.
133.
Kim, H.J., Y. Baek, K. Choi, D.G. Kim, H. Kang, Y.S. Choi, J. Yoon, and J.C. Lee,
The improvement of antibiofouling properties of a reverse osmosis membrane by
oxidized CNTs. Rsc Advances, 2014. 4(62): p. 32802-32810.
134.
Vatanpour, V., S.S. Madaeni, R. Moradian, S. Zinadini, and B. Astinchap,
Fabrication and characterization of novel antifouling nanofiltration membrane
prepared
from
oxidized
multiwalled
carbon
nanotube/polyethersulfone
nanocomposite. Journal of Membrane Science, 2011. 375(1-2): p. 284-294.
135.
Dumee, L.F., K. Sears, J. Schutz, N. Finn, C. Huynh, S. Hawkins, M. Duke, and S.
Gray, Characterization and evaluation of carbon nanotube Bucky-Paper membranes
for direct contact membrane distillation. Journal of Membrane Science, 2010. 351(12): p. 36-43.
136.
Kim, S., F. Fornasiero, H.G. Park, J. Bin In, E. Meshot, G. Giraldo, M. Stadermann,
M. Fireman, J. Shan, C.P. Grigoropoulos, and O. Bakajin, Fabrication of flexible,
aligned carbon nanotube/polymer composite membranes by in-situ polymerization.
Journal of Membrane Science, 2014. 460: p. 91-98.
137.
Heo, J., L.K. Boateng, J.R.V. Flora, H. Lee, N. Her, Y.G. Park, and Y. Yoon,
Comparison of flux behavior and synthetic organic compound removal by forward
osmosis and reverse osmosis membranes. Journal of Membrane Science, 2013. 443: p.
69-82.
138.
Oberdörster, G., E. Oberdörster, and J. Oberdörster, Nanotoxicology: An emerging
discipline evolving from studies of ultrafine particles. Environmental Health
Perspectives, 2005. 113(7): p. 823-839.
70
139.
Kang, S., M. Herzberg, D.F. Rodrigues, and M. Elimelech, Antibacterial effects of
carbon nanotubes: Size does matter! Langmuir, 2008. 24(13): p. 6409-6413.
140.
Celik, E., L.
Liu,
and H. Choi,
Protein fouling
behavior of
carbon
nanotube/polyethersulfone composite membranes during water filtration. Water
Research, 2011. 45(16): p. 5287-5294.
141.
Ihsanullah, A.M. Al Amer, T. Laoui, A. Abbas, N. Al-Aqeeli, F. Patel, M. Khraisheh,
M.A. Atieh, and N. Hilal, Fabrication and antifouling behaviour of a carbon
nanotube membrane. Materials & Design, 2016. 89: p. 549-558.
142.
Vatanpour, V., M. Esmaeili, and M. Farahani, Fouling reduction and retention
increment of polyethersulfone nanofiltration membranes embedded by aminefunctionalized multi-walled carbon nanotubes. Journal of Membrane Science, 2014.
466: p. 70-81.
143.
Wang, S., S. Liang, P. Liang, X. Zhang, J. Sun, S. Wu, and X. Huang, In-situ
combined dual-layer CNT/PVDF membrane for electrically-enhanced fouling
resistance. Journal of Membrane Science, 2015. 491: p. 37-44.
144.
Liu, L.F., J.D. Liu, B. Gao, F.L. Yang, and S. Chellam, Fouling reductions in a
membrane bioreactor using an intermittent electric field and cathodic membrane
modified by vapor phase polymerized pyrrole. Journal of Membrane Science, 2012.
394: p. 202-208.
145.
Rahaman, M.S., C.D. Vecitis, and M. Elimelech, Electrochemical Carbon-Nanotube
Filter Performance toward Virus Removal and Inactivation in the Presence of
Natural Organic Matter. Environmental Science & Technology, 2012. 46(3): p. 15561564.
146.
Vecitis, C.D., M.H. Schnoor, M.S. Rahaman, J.D. Schiffman, and M. Elimelech,
Electrochemical Multiwalled Carbon Nanotube Filter for Viral and Bacterial
Removal and Inactivation. Environmental Science & Technology, 2011. 45(8): p.
3672-3679.
147.
Sun, X., J. Wu, Z. Chen, X. Su, and B.J. Hinds, Fouling Characteristics and
Electrochemical Recovery of Carbon Nanotube Membranes. Advanced Functional
Materials, 2013. 23(12): p. 1500-1506.
148.
de Lannoy, C.-F., D. Jassby, K. Gloe, A.D. Gordon, and M.R. Wiesner, Aquatic
Biofouling Prevention by Electrically Charged Nanocomposite Polymer Thin Film
Membranes. Environmental Science & Technology, 2013. 47(6): p. 2760-2768.
71
149.
Wu, H.Q., B.B. Tang, and P.Y. Wu, Novel ultrafiltration membranes prepared from a
multi-walled carbon nanotubes/polymer composite. Journal of Membrane Science,
2010. 362(1-2): p. 374-383.
150.
Yu, Z.X., G.Y. Zeng, Y. Pan, L. Lv, H. Min, L. Zhang, and Y. He, Effect of
functionalized
multi-walled
carbon
nanotubes
on
the
microstructure
and
performances of PVDF membranes. Rsc Advances, 2015. 5(93): p. 75998-76006.
151.
Dumee, L., J. Lee, K. Sears, B. Tardy, M. Duke, and S. Gray, Fabrication of thin film
composite poly(amide)-carbon-nanotube supported membranes for enhanced
performance in osmotically driven desalination systems. Journal of Membrane
Science, 2013. 427: p. 422-430.
152.
Coleman, J.N., U. Khan, W.J. Blau, and Y.K. Gun’ko, Small but strong: A review of
the mechanical properties of carbon nanotube–polymer composites. Carbon, 2006.
44(9): p. 1624-1652.
153. Nellore, B.P.V., R. Kanchanapally, F. Pedraza, S.S. Sinha, A. Pramanik, A.T. Hamme,
Z. Arslan, D. Sardar, and P.C. Ray, Bio-Conjugated CNT-Bridged 3D Porous
Graphene Oxide Membrane for Highly Efficient Disinfection of Pathogenic Bacteria
and Removal of Toxic Metals from Water. Acs Applied Materials & Interfaces, 2015.
7(34): p. 19210-19218.
154.
Wei, G., S. Chen, X. Fan, X. Quan, and H. Yu, Carbon nanotube hollow fiber
membranes: High-throughput fabrication, structural control and electrochemically
improved selectivity. Journal of Membrane Science, 2015. 493: p. 97-105.
155.
Zhao, F.-Y., Q.-F. An, Y.-L. Ji, and C.-J. Gao, A novel type of polyelectrolyte
complex/MWCNT hybrid nanofiltration membranes for water softening. Journal of
Membrane Science, 2015. 492: p. 412-421.
156.
Liu, S., T.H. Zeng, M. Hofmann, E. Burcombe, J. Wei, R. Jiang, J. Kong, and Y.
Chen, Antibacterial Activity of Graphite, Graphite Oxide, Graphene Oxide, and
Reduced Graphene Oxide: Membrane and Oxidative Stress. Acs Nano, 2011. 5(9): p.
6971-6980.
157.
Koenig, S.P., L. Wang, J. Pellegrino, and J.S. Bunch, Selective molecular sieving
through porous graphene. Nature Nanotechnology, 2012. 7(11): p. 728-732.
158.
Garaj, S., W. Hubbard, A. Reina, J. Kong, D. Branton, and J.A. Golovchenko,
Graphene as a subnanometre trans-electrode membrane. Nature, 2010. 467(7312): p.
190-U73.
72
159.
Zhang, J., Z. Xu, M. Shan, B. Zhou, Y. Li, B. Li, J. Niu, and X. Qian, Synergetic
effects of oxidized carbon nanotubes and graphene oxide on fouling control and antifouling mechanism of polyvinylidene fluoride ultrafiltration membranes. Journal of
160.
Kim, H.J., M.Y. Lim, K.H. Jung, D.G. Kim, and J.C. Lee, High-performance reverse
osmosis nanocomposite membranes containing the mixture of carbon nanotubes and
graphene oxides. Journal of Materials Chemistry A, 2015. 3(13): p. 6798-6809.
161.
Han, Y., Y. Jiang, and C. Gao, High-Flux Graphene Oxide Nanofiltration Membrane
Intercalated by Carbon Nanotubes. ACS Applied Materials & Interfaces, 2015. 7(15):
p. 8147-8155.
162.
You, B., L.L. Wang, N. Li, and C.L. Zheng, Improving the Energy Storage
Performance of Graphene through Insertion of Pristine CNTs and Ordered
Mesoporous Carbon Coating. Chemelectrochem, 2014. 1(4): p. 772-778.
163.
Kim, Y.S., K. Kumar, F.T. Fisher, and E.H. Yang, Out-of-plane growth of CNTs on
graphene for supercapacitor applications. Nanotechnology, 2012. 23(1): p. 015301-.
164.
You, B., J.H. Jiang, and S.J. Fan, Three-Dimensional Hierarchically Porous AllCarbon Foams for Supercapacitor. Acs Applied Materials & Interfaces, 2014. 6(17):
p. 15302-15308.
165.
Li, N., H.W. Song, H. Cui, G.W. Yang, and C.X. Wang, Self-assembled growth of
Sn@CNTs on vertically aligned graphene for binder-free high Li-storage and
excellent stability. Journal of Materials Chemistry A, 2014. 2(8): p. 2526-2537.
166.
Zhu, Y., L. Li, C. Zhang, G. Casillas, Z. Sun, Z. Yan, G. Ruan, Z. Peng, A.-R.O. Raji,
C. Kittrell, R.H. Hauge, and J.M. Tour, A seamless three-dimensional carbon
nanotube graphene hybrid material. Nature Communications, 2012. 3.
167.
Cheng, Q., J. Tang, N. Shinya, and L.-C. Qin, Polyaniline modified graphene and
carbon nanotube composite electrode for asymmetric supercapacitors of high energy
density. Journal of Power Sources, 2013. 241: p. 423-428.
168.
Bet-moushoul, E., Y. Mansourpanah, K. Farhadi, and M. Tabatabaei, TiO2
nanocomposite based polymeric membranes: A review on performance improvement
for various applications in chemical engineering processes. Chemical Engineering
Journal, 2016. 283: p. 29-46.
169.
Jennings, J.R., A. Ghicov, L.M. Peter, P. Schmuki, and A.B. Walker, Dye-sensitized
solar cells based on oriented TiO2 nanotube arrays: Transport, trapping, and
73
transfer of electrons. Journal of the American Chemical Society, 2008. 130(40): p.
13364-13372.
170.
Tachikawa, T. and T. Majima, Exploring the Spatial Distribution and Transport
Behavior of Charge Carriers in a Single Titania Nanowire. Journal of the American
171.
Quan, X., S.G. Yang, X.L. Ruan, and H.M. Zhao, Preparation of titania nanotubes
and their environmental applications as electrode. Environmental Science &
Technology, 2005. 39(10): p. 3770-3775.
172.
Wei, Q., D. Yang, M. Fan, and H.G. Harris, Applications of Nanomaterial-Based
Membranes in Pollution Control. Critical Reviews in Environmental Science and
Technology, 2013. 43(22): p. 2389-2438.
173.
Yang, M.-Q., N. Zhang, and Y.-J. Xu, Synthesis of Fullerene-, Carbon Nanotube-,
and Graphene-TiO2 Nanocomposite Photocatalysts for Selective Oxidation: A
Comparative Study. Acs Applied Materials & Interfaces, 2013. 5(3): p. 1156-1164.
174.
Qu, X., L. Cao, and F. Du, Fabrication of ordered arrays of CNT/TiO2 nanotubes and
their photocatalytic properties. Rsc Advances, 2015. 5(27): p. 20976-20980.
175.
Yu, H., X. Quan, S. Chen, and H. Zhao, TiO2-multiwalled carbon nanotube
heterojunction Arrays and their charge separation capability. Journal of Physical
Chemistry C, 2007. 111(35): p. 12987-12991.
176.
Dai, K., T. Peng, D. Ke, and B. Wei, Photocatalytic hydrogen generation using a
nanocomposite of multi-walled carbon nanotubes and TiO2 nanoparticles under
visible light irradiation. Nanotechnology, 2009. 20(12).
177. Esfahani, M.R., J.L. Tyler, H.A. Stretz, and M.J.M. Wells, Effects of a dual nanofiller,
nano-TiO2 and MWCNT, for polysulfone-based nanocomposite membranes for water
purification. Desalination, 2015. 372: p. 47-56.
178.
Bai, H., X. Zan, L. Zhang, and D.D. Sun, Multi-functional CNT/ZnO/TiO2
nanocomposite membrane for concurrent filtration and photocatalytic degradation.
Separation and Purification Technology, 2015. 156: p. 922-930.
179. Sainz, R., A.M. Benito, M.T. Martinez, J.F. Galindo, J. Sotres, A.M. Baro, B. Corraze,
O. Chauvet, A.B. Dalton, R.H. Baughman, and W.K. Maser, A soluble and highly
functional polyaniline-carbon nanotube composite. Nanotechnology, 2005. 16(5): p.
S150-S154.
180.
Liao, Y.Z., C. Zhang, Y. Zhang, V. Strong, J.S. Tang, X.G. Li, K. Kalantar-zadeh,
E.M.V. Hoek, K.L. Wang, and R.B. Kaner, Carbon Nanotube/Polyaniline Composite
74
Nanofibers: Facile Synthesis and Chemosensors. Nano Letters, 2011. 11(3): p. 954959.
181.
Konyushenko, E.N., J. Stejskal, M. Trchova, J. Hradil, J. Kovarova, J. Prokes, M.
Cieslar, J.Y. Hwang, K.H. Chen, and I. Sapurina, Multi-wall carbon nanotubes coated
with polyaniline. Polymer, 2006. 47(16): p. 5715-5723.
182.
Guillen, G.R., T.P. Farrell, R.B. Kaner, and E.M.V. Hoek, Pore-structure,
hydrophilicity, and particle filtration characteristics of polyaniline-polysulfone
ultrafiltration membranes. Journal of Materials Chemistry, 2010. 20(22): p. 46214628.
183.
Zhao, S., Z. Wang, J. Wang, S. Yang, and S. Wang, PSf/PANI nanocomposite
membrane prepared by in situ blending of PSf and PANI/NMP. Journal of Membrane
Science, 2011. 376(1-2): p. 83-95.
184.
Liao, Y.Z., D.G. Yu, X. Wang, W. Chain, X.G. Li, E.M.V. Hoek, and R.B. Kaner,
Carbon nanotube-templated polyaniline nanofibers: synthesis, flash welding and
ultrafiltration membranes. Nanoscale, 2013. 5(9): p. 3856-3862.
185.
Shao, D., C. Chen, and X. Wang, Application of polyaniline and multiwalled carbon
nanotube magnetic composites for removal of Pb(II). Chemical Engineering Journal,
2012. 185: p. 144-150.
75
Materials and Method
Chapter 3. Materials and Methods
3.1. Materials
The following chemicals were used for preparation of membrane casting solution: Aniline as a
monomer in the PANI polymerization (VWR International), Ammonium peroxidisulfate (APS) as
an oxidant in the PANI polymerization (Calbiochem, USA), PES (58,000 g/mol, Goodfellow
Cambridge Ltd, UK), N-methyl-2-pyrrolidone (NMP) from Merck and 37 % HCl from VWR
International. Suwannee River humic acid (SRHA) standard II was used as a model NOM
compound (International Humic Substances Society). Latex particles (hydraulic diameter = 30 nm)
were purchased from Sigma-Aldrich (L5155). Hydroxylated-MWCNT was supplied from
BuckyUSA, and characterized by the company as follows: purity of 98 wt% a diameter of 5 - 15
nm and lengths ranging from 1 to 5 μm.
3.2. Fabrication of MWCNTs/PANI/PES membranes
Membrane fabrication was divided into two steps. The first step was a synthesis of the
MWCNTs/PANI complex by in-situ polymerization; the second step was a fabrication of the
MWCNTs/PANI/PES membrane by the phase inversion method. The MWCNTs/PANI complex
was fabricated into MWCNTs/PANI/PES membrane
The first step is a synthesis of MWCNTs/PANI nanocomposite by in-situ polymerization, as shown
in Fig 3.1 (a) [1, 2]. A solution of 3 mM aniline monomer and 0.8 mM APS was prepared in 1 M
HCl and 99.5% NMP. MWCNTs were dispersed in 99.5% NMP solution by sonication (500 W)
for one h. Three substances (i.e., aniline, APS and MWCNTs) were mixed in a glass vessel and
stirred for 48 h at 4 oC.
76
The dispersity of the MWCNTs/PANI composite was confirmed by observing UV absorbance
(UV-2450, Shimadzu) as a function of time. MWCNTs/PANI complex was dispersed in an organic
solvent (NMP). The absorbance at 295 nm was monitored every hour for 24 hours in order to
quantify the dispersity of MWCNTs in the MWCNTs/PANI complex. The corresponding
absorbance of MWCNTs dispersing in NMP without PANI was also monitored for comparison. In
addition, synthesis of the MWCNTs/PANI charge transfer complex was proved by the peaks
around 280-300 nm and 400-500 nm corresponding to the benzoid and quinoid band via UV/visible
spectrophotometry [2]. Two peaks at 295, 450 nm correspond to the interaction between nanotubes
and planar PANI [1, 3], corresponding to the π- π transition on the benzoid and quinoid excitation
bands [4, 5].
In the second step, the MWCNTs/PANI/PES composite was blended in the PES solution and the
membrane was fabricated by phase inversion (Fig 3.1 (b)) [6]. 15% PES was mixed with NMP for
1 h and then blended with the MWCNTs/PANI composite for 24 h to prepare the membrane casting
solution. The resultant solution was cast on a glass plate using an applicator with 300 μm thickness.
The casting film on the glass plate was immersed in the pure water bath, which allows the polymer
to be transformed from a liquid to solid state by phase inversion (Fig 3.2). The membranes were
stored in a deionized water bath for one day to remove residual solvents.
77
(a)
(b)
Fig 3.1 Synthesis of (a) MWCNTs/PANI nanocomposite and (b) fabrication procedure of
MWCNTs/PANI/PES composite membrane.
Fig 3.2 Preparation of polymer membrane by phase inversion
78
REFERENCES
1.
Sainz, R., A.M. Benito, M.T. Martinez, J.F. Galindo, J. Sotres, A.M. Baro, B. Corraze, O.
Chauvet, and W.K. Maser, Soluble self-aligned carbon nanotube/polyaniline composites.
Advanced Materials, 2005. 17(3): p. 278-+.
2.
Zengin, H., W.S. Zhou, J.Y. Jin, R. Czerw, D.W. Smith, L. Echegoyen, D.L. Carroll, S.H.
Foulger, and J. Ballato, Carbon nanotube doped polyaniline. Advanced Materials, 2002.
14(20): p. 1480-+.
3.
Leng, J.M., R.P. McCall, K.R. Cromack, Y. Sun, S.K. Manohar, A.G. Macdiarmid, and
A.J. Epstein, Photoexcited solutions and polarons in pernigraniline-base polymers.
Physical Review B, 1993. 48(21): p. 15719-15731.
4.
Quillard, S., G. Louarn, S. Lefrant, and A.G. Macdiarmid, Vibrational analysis of poly
aniline - A comparative study of leucoemraldine, emeraldine, and pernigraniline bases.
5.
Sun, Y., S.R. Wilson, and D.I. Schuster, High dissolution and strong light emission of
carbon nanotubes in aromatic amine solvents [1]. Journal of the American Chemical
Society, 2001. 123(22): p. 5348-5349.
6.
Zhao, S., Z. Wang, J. Wang, S. Yang, and S. Wang, PSf/PANI nanocomposite membrane
prepared by in situ blending of PSf and PANI/NMP. Journal of Membrane Science, 2011.
376(1-2): p. 83-95.
79
Chapter 4 Fabriation of carbon nanotube enhanced membrane for natural organic matter removal
Chapter 4. Fabrication of carbon nanotube-enhanced
membrane for natural organic matter removal
4.1. Introduction
In order to tackle the drawbacks of conventional polymeric UF membranes, nanomaterials have
been recently introduced into the membrane fabrication [1-3]. Among various nanomaterials,
CNTs possess excellent adsorption capacity for organic matter. Compared to the traditional
adsorbents such as activated carbon, CNTs are superior adsorbents due to their mesoporous
structure and less negative surface charge [4]. Furthermore, the π - π interaction between the
aromatic group in NOM and CNTs contributes to the better adsorption behaviour of the CNTs
[5]. Moreover, CNTs demonstrate excellent electrical and mechanical properties [6-8]. However,
due to the van der Waals forces between carbon nanotubes, they are tightly bundled and
insoluble in organic solvent. In order to maximize the use of CNTs over a large surface area and
with high conductivity, their aggregation within the membrane casting solution has to be
prevented [9-11]. There have been many studies on CNT dispersion in polymers and solvents
[12-14]. The methods of CNT dispersion in these studies varied from physical treatment to
chemical functionalization, which enhances compatibility between the polymer matrix and
nanotubes. However, sonication disperses CNTs by breaking van der Waals forces temporarily
and thus does not guarantee all CNTs are homogenously dispersed within the polymer matrix.
Functionalization by chemical oxidation of CNTs is the most commonly used method, which
breaks sp2 hybrid carbon bonds on the sidewalls, and attaches carboxyl/hydroxyl groups to the
CNTs [15]. Functionalized CNTs with the hydrophilic group have been reported as attributing to
increased flux because hydrophilic modified CNTs can make membrane more permeable [16,
80
17]. Further, such increase to hydrophilicity is also a well-known approach to reducing
membrane fouling [18, 19].
Another approach to enhance the separation properties of membranes is the incorporation of
nanofillers [20, 21]. PANI is one of the favoured candidates due to its good environmental
stability, miscibility with PES polymer and ease of synthesis [22]. Its synthesis procedure is
considered as simple nonredox doping/dedoping chemistry under an acid/base reaction. PANI
has been examined for potential applications for separation membranes [23], and biosensors [24]
as a conductive nanofiller due to its electrical conductivity. PANI/PSf UF membranes were
fabricated by an in-situ blending method and non-solvent induced phase inversion [25, 26], and
exhibited the enhancement of water permeability due to the increase in hydrophilicity and pore
size of the membrane.
Several studies reported that MWCNTs/PANI can form an electron transfer complex via donoracceptor interaction by in-situ polymerization [27, 28]; this complex showed enhanced electrical
and mechanical properties, and dispersions of CNT in polymer matrices [29]. These two
integrated materials were found out to have strong potential for application in functional
membranes [30], conductive films, capacitors [31] and biosensors. Such attempts to the
integration of an MWCNTs/PANI composite with polymer matrix have opened up the new
opportunities for advanced membrane materials for water treatment by overcoming MWCNTs
aggregation and tuning membrane structure and its physicochemical properties which influences
permeate flux and selectivity [32].
Even though the MWCNTs/PANI complex has a potential as a novel membrane material, the
performance of the MWCNTs/PANI membrane for NOM removal and mitigation of fouling is
81
largely unknown. Therefore, it is desirable to develop MWCNTs/PANI composite membrane,
which combines advantages of these two materials to complement the limitations of conventional
UF membrane.
While currently a UF membrane fouling under different NOM foulant model has been
investigated by the modification of membrane with functionalized MWCNTs [33], this study
highlights CNTs composite membrane fabrication for advancing membrane process. We
fabricated novel MWCNTs/PANI/PES membrane that showed enhanced removal efficiency and
remarkably high flux compared to a PES membrane. An MWCNTs/PANI composite was
synthesized by in-situ polymerization and introduced to a PES matrix. The significance of this
research is the development of CNTs enhanced UF membrane for effective NOM removal with
high water product that requires less operating pressure. In-situ polymerization was introduced to
the fabrication procedure to overcome nanomaterial aggregation and tune the membrane
structure and its physical/chemical properties. The study was aimed to determine key factors
contributing to the enhanced performance, and examine the NOM removal mechanism. The
chemical/physical properties of the membranes were thoroughly characterized regarding
membrane morphology, pore size and its size distribution, porosity, zeta potential, and
hydrophilicity. Flux decline with humic acid (HA) was conducted, and suitable cleaning methods
were examined.
82
4.2. Materials and methods
4.2.1. Materials
Suwannee River humic acid (SRHA) standard II was used as a model NOM compound
(International Humic Substances Society). Latex particles (hydraulic diameter = 30 nm) were
purchased from Sigma-Aldrich (L5155).
4.2.2. Fabrication of membranes
The membranes were fabricated according to Chapter 3 Materials and methods - membrane
fabrication.
4.2.3 Characterization of MWCNTs/PANI/PES membranes
Porosity (ε) was determined by the gravimetric method [34]. Each membrane was initially
immersed in the propan-2-ol for 30 min and then deionized water for another 30 min at room
temperature. Mass loss of wet membrane after drying was measured. The porosity of membrane
was calculated by the following equation 1:
𝜀 = (𝑊
(𝑊𝑤 −Wd )/ρwater
𝑤 −Wd )/ρwater +Wd /ρPES,PANI
(1)
Where Ww and Wd are the weight of wet membrane and dry membrane, respectively, and ρwater
and ρPES,PANI are the water density (1 g/cm3), PES and PANI density (1.36 g/cm3), respectively.
The samples were degassed under vacuum before measurement.
Pore size and its size distribution were analysed using N2 adsorption/desorption at –K
(Micromeritics), and calculated by the Brunauer–Emmett–Teller (BET) method [35]. Contact
angle measurements were performed using 4 – 8 μL sessile droplets of Milli-Q water with a
Krüss Easy Drop goniometer. The zeta potential of membrane surface was measured by using
83
the streaming potential technique (Anton Paar electrokinetic analyzer). The experiment was run
at an ionic strength of 0.001 M KCl and pH ranging from 2 to 10. Membrane morphology (crosssectional image and MWCNTs dispersion within PES matrix) was obtained by field emission
scanning electron microscopy (FE-SEM, Hitachi S4500 FE-SEM). To obtain cross-sectional
images, samples were soaked in n-Hexane to preserve pore structure and fractured in liquid
nitrogen. Mounted samples were sputter coated with gold before examined. Top down view of
membranes was observed by FE-SEM (Zeiss ULTRA Plus).
4.2.4 Evaluation of filtration performance
Pure water flux was evaluated in a high-pressure stirred cell (HP 4750, Sterlitech) under 1- 2 bar
at a room temperature (23ºC).
In order to examine key parameters of flux enhancement, flux was calculated by the HagenPoiseuille equation, which indicates the no-slip condition. The Hagen-Poiseuille equation well
describes water flux (J) as a function of physical properties of the membrane. Thus, the effect of
average pore size on water flux can be demonstrated by the theoretical flux calculated by the
Hagen-Poiseulle equation (Eq.3).
𝐽=
𝜀𝑟 2 Δ𝑃
8𝜂𝜏 𝐿
(3)
Where ε is porosity, r is radius (m), ΔP is the applied pressure (kPa), η is the viscosity of water
(1mP.s), and L is the membrane thickness(m).
The Enhancement factor was calculated using the measured flux over calculated one by the
Hagen-Poiseuille equation:
84
𝑄
𝐸 = 𝑄 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑 (4)
𝑛𝑜−𝑠𝑙𝑖𝑝 𝐻𝑃
where E is the enhancement factor, Qmeasured is the flux measured by experiment, and Qno-slip is
the flux calculated by the Hagen-Poiseuille equation on the no-slip condition.
Slip length was calculated using the slip flow Hagen-Poiseuille equation.
𝑏=
𝑄
𝑑(( 𝑚𝑒𝑎𝑠𝑢𝑟𝑒𝑑 )−1)
𝑄𝑛𝑜−𝑠𝑙𝑖𝑝 𝐻𝑃
(5)
8
where b is the slip length (m).
The impact of porosity on flux was examined by the Kozeny-Carmen equation.
𝜀3
Δ𝑃
𝐽𝐾−𝑐 = 𝐾𝜂𝑆2 (1−𝜀)2 Δ𝑥 (6)
Where Jk-c is flux calculated by Kozeny-Carmen equation (LMH). Jk-c was plotted as a function
of porosity of membranes. Δx as a membrane thickness was measured with a digital caliper. All
parameters except for porosity of membranes was fixed and presented in Table 4.1.
The removal efficiency of latex particle was measured in the HP stirred cell. Latex nanoparticles
which have a uniform structure with a 30 nm diameter were selected for the removal test. A
stock solution with latex particles (10 ppm) was prepared to confirm that membranes can retain >
90 % of the solute. The concentration of latex particles in the feed and membrane permeate was
measured using a fluorometer (Turner Biosystems), and removal efficiency (R) was calculated
by the equation (7) presented in supporting information. HA removal efficiency of the membrane
was evaluated using a 5 ppm SRHA solution as a feed. The concentration of HA in the feed and
85
membrane permeate was measured using a UV/visible spectrophotometry (UV-2450, Shimadzu)
at 254 nm. (All equations are explained in the supporting information.)
Table 4.1 Parameters in the Kozeny-Carmen equation.
Parameter
Value
ε = volume fraction of pores (porosity)
Variable
K = Kozeny-Carman constant (shape of pores, tortuosity)
10.58
S= internal surface area
33m2/g
η= viscosity
1 mPa.s
ΔP
1 bar
Δx
293 μm
4.2.5 Chemical cleaning efficiency
An efficient cleaning method was evaluated by 100 % recovery of flux and HA removal rate, and
total fouling ratio. First, after HA filtration test, the fouled membrane was cleaned using five
different cleaning methods presented in Table 4.2. The operation of cleaning is: (1) rinsing with
5 g/L NaOCl for 1 hr: (2) rinsing with 0.1 M NaOH for 1 hr: (3) rinsing with 0.1 M HCl for 1 hr:
(4) rinsing with 0.1M NaOH, 1M HCl for 1 hour each: (5) rinsing with 0.1 M citric acid for 1 hr.
After determining the efficient chemical cleaning method, flux decline behaviour was examined
by the HA removal test for approximately 200 minutes in total. The selected chemical cleaning
was carried out at the point when HA permeability and removal rate dramatically dropped. The
filtration test was repeated three cycles.
The flux recovery (FR) and total fouling ratio (Rt) were calculated as follows:
𝐽
𝐹𝑅 (%) = 𝐽𝑤,2 × 100
𝑤,1
(7)
86
𝑅𝑡 (%) = (1 −
𝐽𝑝
𝐽𝑤,1
) × 100 (8)
Where Jw,1 is the pure water permeability, Jw,2 is the water permeability of cleaned membrane
with acid/base solution, and Jp is the water permeability with 5 ppm HA solution. Here, high
flux recovery (close to 1) and low total fouling ratio indicate a suitable cleaning method to
improve the better antifouling properties of the membrane.
Table 4.2 Chemical cleaning conditions.
Type
Mechanism
(1)
5g/L NaOCl for 1 h
Oxidation, Disinfection
(2)
0.1 M NaOH for 1 h
Hydrolysis, solubilization,
saponification
(3)
0.1 M HCl for 1 h
Hydrolysis, Solubilization
(4)
0.1 M NaOH for 1 h + 0.1 M HCl for 1 h
Hydrolysis, Solubilization
(5)
0.1 M Citric acid for 1 h
Chelation
87
4.3. Results
4.3.1 Significance of in-situ polymerization as a membrane fabrication procedure
We introduced in-situ polymerization to the membrane fabrication procedure to advance
membrane process by overcoming MWCNTs aggregation. Its effects on the membrane
performance enhancement were investigated.
4.3.1.1 Dispersity of MWCNTs
MWCNTs/PANI complex synthesized by in situ polymerization were able to contribute to
MWCNTs dispersion in the polymer matrix. PANI is well known for being miscible with a
polymer such as PES and PSf. Further, MWCNTs forms a charge transfer complex with PANI.
Consequently, during the membrane fabrication, MWCNTs/PANI complex is well mixed with
PES polymer, which results in well dispersion of MWCNTs in PES polymer. After casting and
phase inversion process, MWCNTs can be well dispersed in PES membrane structure. The effect
of MWCNTs/PANI complex on reducing MWCNTs aggregation was evaluated by quantifying
MWCNTS dispersion. The snapshot in Fig 4.1 shows the dispersed MWCNTs/PANI complex
into NMP and its change in dispersity over 24 hours.
0 hr
30 min
3 hr
1 hr
6 hr
Fig 4.1 Snapshot shows dispersed MWCNTs/PANI in NMP over 24 h.
88
24 hr
The MWCNTs/PANI complex maintained dispersion for 24 hours without any settlement. The
stable dispersity of MWCNTs was approved by UV spectra of the MWCNTs/PANI complex at
200 - 700 nm wavelength and SEM images. For the MWCNTs/PANI complex, narrow and
strong peaks appear with maxima at around 280 - 300 nm, and a broad band is seen at 450 nm.
These two peaks at 295, 450 nm corresponds to the π- π transition on the benzoid and quinoid
excitation bands [29, 36, 37], indicating an interaction between nanotubes and planar PANI [38,
39]. In comparison, not any sharp and strong peak were observed in the range of 280 – 300 nm
and 450 nm when dispersing MWCNTs in NMP (Fig 4.2 (a)). The result suggests that
MWCNTs formed a charge transfer complex with PANI, which results in evenly dispersed
MWCNTs within the PES matrix. Here, MWCNTs act as an electron acceptor, and aniline as an
electron donor. The absorbance at 295 nm was examined to quantify the dispersity of MWCNTs
in the MWCNTs/PANI complex. As shown by the UV spectra of the MWCNTs/PANI complex
in Fig 4.2 (b), the absorbance attributable to the CNTs remained constant over the course of 10
hours. However, for the MWCNTs dispersed in NMP without PANI, the absorbance dropped in
1 hour.
89
(a)
(b)
Fig 4.2 UV-Vis spectra of (a) the MWCNTs/PANI complex and MWCNTs in NMP, and (b)
absorbance at 295 nm for 10 hours.
To confirm the effect of MWCNTs/PANI complex, dispersion of MWCNTs into PES matrix was
also observed by SEM. Fig 4.3 presents the SEM cross-sectional images taken at 3,000×, 5,000×
and 10,000× magnification in both (a) MWCNTs/PANI/PES membrane and (b)simply blended
MWCNTs/PES-1 membrane. In the SEM images of the MWCNTs/PANI/PES membrane (Fig
4.3 (a)), individual MWCNT (pointed by red arrows) is well dispersed into PES membrane
matrix without any agglomeration. On the other hand, simply blended MWCNTs without PANI
(pointed by circle and arrows) were aggregated in PES matrix, as shown in Fig 4.3 (b).
Moreover, Fig 4.4 shows that MWCNTs were exposed on the top layer of simply blended
MWCNTs/PES-1 membrane while no MWCNTs were observed on the surface of
MWCNTs/PANI/PES membranes. It indicates that simply blended MWCNTs cannot overcome
aggregation. Therefore, MWCNTs/PANI complex can contribute to MWCNTs dispersion within
a polymer matrix as it reacts with MWCNTs in an electron transfer complex by in-situ
polymerization.
90
Mag
3000 x
Mag
5000 x
Mag
10000 x
(a)
(b)
Fig 4.3 SEM cross-sectional images of MWCNTs dispersion in PES polymer matrix: Crosssectional images of (a) MWCNTs/PANI/PES membrane: MWCNTs are well-dispersed into the
PES polymer and (b) MWCNT/PES-1 (simply blended) membrane: MWCNTs are aggregated in
the PES matrix.
91
Mag 3000 X
Mag 5000 X
Fig 4.4 Top down views of MWCNTs/PES-1 (simply blended) membrane show that MWCNTs
are protruded from PES matrix.
4.3.1.2 Effects of in situ polymerization fabrication on enhanced membrane performance
Ultrapure water flux behaviour and NOM removal efficiency of MWCNTs/PANI/PES
membrane were compared with simply blended MWCNTs membrane (MWCNT/PES-1).
Fig 4.5(a) shows the flux behaviour of PES membrane with two MWCNTs composite
membranes. No benefit was observed for the MWCNT/PES-1 membrane where MWCNTs were
simply blended. Its water permeability was in the range of PES membrane. By contrast, the
MWCNTs/PANI/PES membrane showed greatly enhanced water permeability. As can be seen,
water permeability jumped from 20.4 to 490.8 LMH/bar by insertion of the MWCNT/PANI
complex into PES matrix. By simply added PANI into PES matrix, the water permeability
increased to 265.4 LMH/bar. Significantly enhanced removal efficiency of NOM (80 %) was
observed in the MWCNT/PANI/PES-2 membrane which MWCNTs/PANI complex by in-situ
polymerization was incorporated (Fig 4.5(b)). However, for simply blended MWCNT membrane
(i.e., MWCNT/PES-1), its removal efficiency of NOM remained at 10 %. It is well known that
92
MWCNTs can only increase adsorption capacity of the membrane when it is well dispersed.
When agglomerated, it only slightly modifies membrane property with less consistent pore size,
resulting in lower HA removal efficiency. Further, selectivity to NOM is governed by the
interfacial surface area of MWCNTs [5, 9]. Thus, MWCNTs in PES matrix by simply blending
was not able to maximize adsorption ability of MWCNTs which is required for high removal
efficiency.
As shown by the greatly enhanced performance of the MWCNTs/PANI/PES membranes
compared
to
the
poor
removal/low
flux
of
the
MWCNTS/PES
membrane,
the
MWCNTs/PANI/PES membrane did not only overcome nanomaterial aggregation but also
contributed to significantly enhanced performance - water permeability and NOM removal.
(a)
(b)
Fig 4.5 Effect of MWCNTs/PANI complex on the membrane performance – (a) water
permeability, (b) HA removal (1- CNTs simply blended, 2 – membrane). MWCNTs/PANI
membrane exhibited 30 fold enhanced water permeability and effective HA removal up to 80 %,
compared to PES and PANI/PES membrane. MWCNTs simply blended in PES membrane did
not show enhanced separation properties.
93
4.3.2 Membrane performance of MWCNTs/PANI/PES membranes by MWCNTs concentration
MWCNTs concentration in MWCNTs/PANI complex was increased, aiming to enhance water
permeability and investigation of potential mechanism to the flux and HA removal.
Fig 4.6 indicates that water permeability significantly increases with the increase of MWCNTs
concentration in MWCNTs/PANI complex (PANI was adjusted to 50
MWCNTs/PANI/PES membranes). As 0.25
wt%
wt%
in all range of
MWCNTs/PANI complex was incorporated
into the polymer matrix, water permeability slightly increased from 265.4 to 376.9 LMH/bar. As
MWCNTs increased up to 2 wt%, the water permeability dramatically increased from 376.9 up to
1498.1 LMH/bar. It is 15 - 30 times more permeable than the PES membrane while the results
from the literature were reported to be 500 LMH/bar (2 times higher than pure PSf membrane)
for CNTs blended PSf membrane and 560LH/bar for MWCNTs coated PES membrane [33, 40].
In addition to the effect on water permeability, MWCNTs/PANI complex was found to have a
critical impact on enhanced HA removal efficiency (the average HA removal rate during one hr’s
filtration). The MWCNTs/PANI/PES membrane removed up to 80 % of HA. As seen in Fig
4.6(b), the PANI/PES membrane did not overcome the low HA removal efficiency of the PES
UF membrane. However, insertion of MWCNTs as MWCNTs/PANI complex enhanced HA
removal efficiency. As MWCNTs increased to 0.25
wt
%, HA removal rate started to increase 2-
fold. By increasing the amount of MWCNTs in MWCNTs/PANI complex to 1.5
wt%,
MWCNTs/PANI/PES membranes exhibited enhanced HA removal efficiency up to 80 %,
outperforming the previously published work (45 % removal by CNTs incorporation) [16].
However, slightly decreased HA removal was observed when MWCNTs increased up to 2
wt%.
It may be presumably due to the fact that the highest permeate flux (1400 LMH/bar) causes
94
insufficient contact time of feed solution during filtration of 2 wt% MWCNTs for HA adsorption
on the membrane.
Fig 4.6 Effect of MWCNT proportion on (a) water permeability and (b) HA removal efficiency
as a function of MWCNT concentration (%).
95
4.3.3 Physical and chemical properties of membrane
It is important to understand chemical and physical properties of the MWCNTs/PANI/PES
membrane to explain the enhanced removal efficiency and high flux of the membrane. The
results from membrane characterization reveal that the MWCNTs/PANI complex alters the
physical properties of membranes: an increase in porosity, and change in pore size and its size
distribution. In addition, MWCNTs mainly transform the chemical properties of the membranes:
an increase in hydrophilicity and an alteration in the zeta potential of the membrane surface.
4.3.3.1 Physical properties: porosity, pore size and its size distribution
The insertion of hydroxyl-functionalized MWCNTs linking to PANI as pore former was found to
change the physical properties of the membrane – porosity, pore size, its size distribution and
morphology of membranes. Such parameters affecting permeation characteristics are governed
by phase inversion process-polymer-solvent-nonsolvent system. Thus, incorporation of two
material having different characteristics enables to tune physical characteristics of membranes
during the fabrication process.
First, morphologies of membranes are shown in Fig 4.7. MWCNTs/PANI/PES membranes are
observed to have well-developed finger like macrovoid while forming much narrower pores
under top layer in comparison to the PES and PANI/PES membrane. This structure is more
clearly observed as MWCNTs concentration increases. It can probably be accounted for by the
effect of incorporation of hydrophilic and higher viscosity of MWCNTs/PANI complex to the
PES casting solution in the solvent and nonsolvent exchange during phase inversion. Hydrophilic
and relatively highly viscous MWCNTs/PANI/PES casting solution than PANI/PES and PES
96
casting solution delays solvent (NMP) diffusion from nonsolvent (water) bath-demixing process
[41, 42]. It demonstrates that dense and thin skin layer and finger like macrovoid in the sub-layer
structure are formed, resulting in high porosity and increased permeate flux.
(a) Mag 500 X
(b) Mag 500 X
97
(c) Mag 3000 X
Fig 4.7 SEM images of the PES membrane, PANI membrane and MWCNTs/PANI/PES
membrane (a) Top down view, (b) Overall cross-sectional structure, (c) Cross section of top
layer.
Table 4.3 and Fig 4.8 show porosity, pore size and its size distribution. As shown in Table 4.3,
the average pore diameter of all membranes is in the range of 2 – 6 nm, and its size distribution
varies 0.3 to 30 nm (Fig 4.8). Insertion of PANI or MWCNTs/PANI complex into PES polymer
matrix resulted in different pore size and its size distribution. By the insertion of PANI (A
membrane), average pore diameter and porosity of membrane increased to 5.1 nm and 82.2 %
with broad pore size distribution (1.9 – 27 nm). This is mainly due to the fact that soluble PANI
oligomer acts as pore former, resulting slightly higher porosity and bigger pore size. As expected,
the change in physical characteristics of PES and [A] membrane is consistent with increased
permeate flux. However, the insertion of 0.5 wt% MWCNTs/PANI complex on PES matrix ([B-2]
membrane) appeared to form smaller average pore diameter (5.1 → 4.5 nm) and its relatively
98
narrow size distribution (0.9 - 1 nm) with decreased porosity, while its permeability increased.
Overall, the trend in porosity seemed to increase by increasing MWCNTs concentration in
MWCNTs/PANI complex up to 2
wt%.
Compared to the initial porosity of PES membrane
(78 %), membrane porosity increased up to 84.5 wt% by increasing amount of MWCNTs up to 2
wt%,
which is consistent with the flux behaviour. Relatively broad pore size distribution in the
range of 0.1 to 30 nm appeared when MWCNTs increased up to 2 wt%, which contributed to the
great permeate flux enhancement. However, the average pore diameter slightly decreased with
the narrowest pore size distribution (0.3 – 1 nm) when 1.5
wt%
MWCNTs/PANI complex
inserted ([B-4] membrane). The difficulty in tuning pore diameter may be due to the intrinsic
fabrication process-phase inversion. In contrast, its permeate flux increased.
Based on the physical characteristics, MWCNTs/PANI complex was found to increase porosity
and change sub-layer structure to finger-like macrovoid, resulting in enhanced permeability. Pore
size and its size distribution seemed to be changed by MWCNTs/PANI complex but were shown
to be inconsistent with flux enhancement.
Table 4.3 Porosity and pore size of membranes.
Porosity
(%)
Type of membrane
Pore size
(nm)
Latex removal
rate (%)
15 % PES
78.0 ± 0.3
4.6
98.0 ± 0.2
[A]: 50 % PANI/ 15 % PES
82.2 ± 0.1
5.1
88.4 ± 2.4
[B-1] : 0.25 % MWCNTs/ 50 % PANI/ 15 % PES
80.3 ± 0.4
5.1
98.5 ± 0.3
[B-2] : 0. 5 % MWCNTs/ 50 % PANI/ 15 % PES
80.8 ± 0.4
4.5
98.4 ± 0.2
[B-3] : 1 % MWCNTs/ 50 % PANI/ 15 % PES
83.2 ± 0.1
4.7
97.9 ± 0.4
[B-4] :1.5 % MWCNTs/ 50 % PANI/ 15 % PES
83.8 ± 0.1
4.4
99.2 ± 0.1
[B-5] : 2 % MWCNTs/ 50 % PANI/ 15 % PES
84.5 ± 0.0
6.3
93.7 ± 1.3
99
100
Fig 4.8 Pore size distribution of membranes. (PANI/PES membrane: [A], 0.25, 0.5, 1, 1.5 and 2
wt%
MWCNT/PANI/PES membranes: [B-1], [B-2], [B-3], [B-4] and [B-5])
101
4.3.3.2 Chemical properties: hydrophilicity and zeta potential
Insertion of MWCNTs in the MWCNTs/PANI complex transformed the chemical properties of
membranes- hydrophilicity. In Fig 4.9, the contact angle decreased from 73.5 to 67.4 ° as 0.25 wt%
MWCNTs/PANI was inserted into the PES matrix. As MWCNTs were added up to 2
wt%,
the
contact angle of membrane decreased from 67.4° to 52.9°. It indicates that insertion of
MWCNTs/PANI complex into PES matrix significantly increased the hydrophilicity of the
membranes. This result corresponds to enhanced water permeability from 376.9 to 1498.1
LMH/bar.
Fig 4.9 Effect of MWCNTs concentration on the hydrophilicity of the MWCNTs/PANI/PES
membrane.
The surface roughness of membranes is one of the potential effects which can influence the
hydrophilicity of membrane surface. Atomic force microscopy (AFM) was used to investigate
102
the membrane surface roughness. No obvious correlation between surface roughness and
increase in MWCNTs concentration in MWCNTs/PANI complex was found in this study.
However, there might be potentially some macro surface structure difference for
MWCNTs/PANI/PES membranes with different MWCNTs concentration. The difference of
macro surface structure affects the contact angle values of the membrane, but might not be
significantly enough to be picked up by AFM analysis.
It is important to discuss the zeta potential of the membrane surface since the surface charge
property of a polymeric membrane depends on the chemical properties of the membrane and
chemistry of the solution as membrane surfaces are electrically charged in the presence of water
[43]. The inclusion of MWCNTs as an MWCNTs/PANI complex has contributed to the
increased adsorption capacity by altering the surface charge of the membrane.
The reason for changing zeta potential by the insertion of MWCNTs/PANI complex into PES
polymer matrix can be elucidated by the doping process during MWCNTs/PANI complex
synthesis. Doping is the process of transforming a polymer to its conductive form via chemical
oxidation [44-46]. Doping process can change the charge of polymer to positive or negative one.
Thus, during the MWCNTs/PANI complex synthesis by the acid doping (chemical oxidation),
MWCNTs PANI complex became positively charged (Fig 3.1(a)) [28]. For this reason,
positively charged MWCNTs/PANI complex altered negatively charged PES membrane to the
positively charged surface of the membrane.
The zeta potential of the MWCNTs/PANI/PES membrane changed from a negative value to a
positive one by insertion of the MWCNTs/PANI complex in the polymer matrix, as shown in Fig
4.10. The zeta potential of the PES membrane is –20 mV at pH 5.6. As PANI was inserted to
103
PES, it decreased to – 4.1 mV at pH 5.6, but the membrane was still negatively charged.
However as MWCNTs were inserted into the polymer matrix as the MWCNTs/PANI complex,
the membrane became positively charged, varying from +6.2 ~ +11.3 mV at pH 5.6.
Due to the electrostatic interaction between negatively charged HA and positively charged
surface of the membrane, adsorption of HA appeared on the membrane surface and pore inside
[43]. It is supported by the results showing that enhanced adsorption capacity of NOM on
carbons is governed by the cationic functional group on carbon surface since the net surface
charge of NOM molecule is negative [47, 48]. Consequently, HA removal efficiency increased
4-fold in comparison to the PES membrane and PANI/PES membrane with a negative charge.
Fig 4.10 Zeta potential of the membrane by streaming potential at pH 5.6 at a background
electrolyte concentration of 0.001 M KCl (Zeta potential of 5 ppm HA solution = - 7.5 mV).
Insertion of MWCNTs in the membrane altered to positively-surface charged membrane.
104
4.4. Discussion
4.4.1 Why fast water flux through the MWCNTs/PANI/PES membrane?
The effects of physical/chemical properties and membrane morphology were systematically
investigated to examine the potential mechanism affecting water flux behaviour.
In section 4.3.3.1, porosity was found to correspond to the flux enhancement. To systematically
investigate the contribution of porosity on water flux enhancement, water flux as a function of
porosity was calculated using the Kozeny-Carman equation, shown in Fig 4.11. Detailed
information on parameters (Table 4.1) and equation are explained in the 2.4 section. It describes
well the permeability-porosity relationship in organic membranes by phase inversion method
having asymmetric structure [41]. Here, Kozeny-Carman constant (K) represents the shape of
pores and tortuosity. Here K = k0τ2. K0 is taken to be equal to 2 for a circular capillary. In Fig
4.11, theoretical water permeability shows an exponential function according to the increase in
porosity, 2-fold. By contrast, the experimental results showed 4 – fold increase in water
permeability. The notable thing is that porosity of the [B-1] and [B-2] membranes have an
opposite trend with enhancement in water permeability. Even though [B-1] and [B-2]
membranes are 1.5 – 2 % less porous than A membrane, their water permeability exceeded that
of [A] membrane. It indicates that water permeability may also be governed by another
characteristic – hydrophilicity, in this case.
105
Fig 4. 11 Water permeability as a function of porosity by Kozeny-Carmen relationship.
The inclusion of hydroxyl-functionalized MWCNTs was observed to make the membrane
surface more hydrophilic by the strong affinity between water molecules and membranes.
Considering the slight increase in water permeability of the 0.25
wt%
MWCNTs/PANI/PES
membrane compared to the PANI/PES membrane, the decrease in porosity between the
PANI/PES membrane and 0.25
wt%
MWCNTs/PANI/PES membrane was compensated by the
increase in hydrophilicity. The decrease in contact angle resulted in a slight increase in water
permeability from 265.4 to 376.9 LMH/bar. As MWCNTs increased to 0.5
wt%,
porosity and
hydrophilicity rose to 80.8 % and 63.6 °, respectively. It correlates with a gradual increase in
water permeability. However, as MWCNTs increased to 1 wt%, water permeability doubled. This
sudden increase could be explained by the combination of porosity and hydrophilicity increase.
106
The porosity of the one
wt%
MWCNTs/PANI/PES membrane (83.2 %) exceeded that of
PANI/PES membrane (82.2 %) as its hydrophilicity increased to 56. 1 °. The increase of both
factors is consistent with a dramatically increased permeability.
The relationship between flux and physical properties such as pore size and its size distribution
was investigated for a mechanistic understanding of membrane performance. In section 3.3.3.1,
the discrepancy between flux enhancement and reduced pore size and its narrow size distribution
was observed when MWCNTs/PANI complex inserted to PES polymer ([S], [A] vs. [B-1] – [B-5]
membrane). As seen in Table 4.3, the pore size of all membranes by BET surface area is in the
range of 2-6 nm. Given the conventional fluid flow by Haagen Poiseuille equation, decreased
average pore size must result in the flux decline. However, compared to the flux calculated by
Hagen-Poiseuille equation in this pore size range, experimental flux by MWCNTs enhanced
membrane was reported 100 to 600 fold (Table 4.4). Considering the significant effect of pore
size on the flux, the pore size of the MWCNTs/PANI/PES membranes is much smaller than the
experimental water flux. Further, [B-4] membrane having high water flux was observed to have a
smaller average pore size (4.5 nm) and narrow pore size distribution (0.3 – 1 nm). Such a
discrepancy may be explained by the difference of hydrophilicity between membrane surface and
channel within the membrane. The increase in hydrophilicity of membrane surface may
introduce water molecules by a strong affinity of MWCNTs enhanced membranes. In membrane
fabrication process, hydrophilic MWCNTs were formed on top of the membrane during
diffusion/demixing process in a water bath by phase inversion due to its hydrophilicity. Thus,
middle to bottom of membrane structure consisting of PES became less hydrophilic than the top
part with hydrophilic MWCNTs. Here, slip length effect may be expected due to the smaller pore
diameter in the range of 2 – 6 nm introducing water molecule into the relatively hydrophobic
107
channel within the membrane which is in the few micrometer range (Cross-sectional structure of
each membrane are shown in Fig 4.7). It is assumed that top surface of the membrane with
stronger affinity to water may seem to enhance water permeation, and relatively hydrophobic big
channel (3 – 8 um) may attribute to dramatically enhanced flux by slippage effect. Therefore,
the extremely low pure water flux for the PES membrane was due to the relatively small average
pore diameter, broader pore size distribution, low porosity and less hydrophilicity which had
been formed in phase inversion stage.
Based on the membrane characterization and theoretical/observed results, we may assume that
the synergetic effect of porosity and hydrophilicity could mainly lead to the flux enhancement.
Hydrophilic, narrow pore size on the top layer and relatively wide finger-like voids may
contribute considerably to the enhanced permeability by inducing slippage effect.
Table 4.4 Enhancement factor and slip length of membranes.
Enhancement
factor
Slip length
(nm)
7.2
3.6
[A]: 50 % PANI/ 15 % PES
189.6
120.2
[B-1] : 0.25 % MWCNTs/ 50 % PANI/ 15 % PES
698.0
183.0
[B-2] : 0. 5 % MWCNTs/ 50 % PANI/ 15 % PES
195.5
109.4
[B-3] : 1 % MWCNTs/ 50 % PANI/ 15 % PES
374.1
219.2
[B-4] :1.5 % MWCNTs/ 50 % PANI/ 15 % PES
511.0
280.5
[B-5] : 2 % MWCNTs/ 50 % PANI/ 15 % PES
290.7
228.1
Type of membrane
15 % PES
108
4.4.2 NOM removal mechanism by the MWCNTs/PANI/PES membrane
The NOM removal mechanism can be explained regarding two factors – adsorption capacity and
size exclusion. Due to the mean size (<1 nm) of HA, it easily passes through the low-pressure
membranes such as MF and UF, but high-pressure membranes such as NF and RO can reject 90 %
HA. As previously discussed, MWCNTs/PANI complex slightly reduced the membrane surface
pore size with latex removal rate. Based on Fig 4.8, very narrow pore size distribution was
observed in the membrane B-2, B-3 and B-4, corresponding to the higher HA removal rate
(80 %). Such a narrow pore size distribution would be one of the key factors for high humic acid
removal. However, that effect alone may not be able to change the humic acid removal
efficiency due to the much smaller size of HA than the pore size of MWCNTs/PANI/PES
membrane in the range of UF membrane. Even if MWCNTs/PANI/PES membrane has reduced
pore size than PANI/PES membrane, HA removal is still beyond the separation ability of UF
membrane with 2-6 nm of average pore size. Thus, the greatly enhanced separation property can
be mainly due to the adsorption capacity of the membrane: an alteration of the zeta potential of
the membrane surface.
The shift of the zeta potential of the membrane significantly affected the interaction between
membrane and HA with a negative charge. Humic acid represents NOM, which contains
negatively charged groups: 60 – 90 % of groups are carboxylic (COO-), methoxyl carbonyls
(C=O) and phenolic (OH-) groups [49]. As a result, HA is negatively charged when dissolved in
water with the zeta potential value of -7.1 mV at pH 5.6. Such an electrostatic interaction by
altering surface charge of the membrane leads to the great enhancement of HA removal
efficiency. It is because HA removal mechanism is strongly governed by the interaction of
membrane with feed [50, 51]. In addition to the electrostatic interaction, MWCNTs themselves
109
have adsorption capacity for organic matters [52]. The π- π system on the MWCNTs interacts
with the cross-linked aromatic network of molecules on humic acid [53]. The insertion of
MWCNTs may contribute to the increased NOM removal efficiency.
A filtration test with HA solution was run for 72 hours as shown in Fig 4.12. No decrease in HA
removal (80 %) was observed in 1.5 wt% MWCNTs/PANI/PES membrane (B-4) while permeate
flux declined. It indicates that, even after 72 hour’s filtration, HA adsorption has not reached its
capacity yet. While membrane surface which was already covered by HA would potentially limit
further the HA accessing due to the repulsion force among HA-HA, this might also result in
higher HA removal rate. By comparison, the removal rates of PES membrane (S) further dropped
with the filtration test lasted even though its initial removal rate was much lower than the 1.5 wt%
MWCNTs/PANI/PES membrane (B-4).
(a)
(b)
Fig 4.12 (a) HA permeability decline behaviour and (b) HA removal decline behaviour of both
the B-4 and the S membranes (684.9 L/m2 filtration).
110
The removal mechanism is most probably the electrostatic interaction of positively charged
membrane and negatively charged HA, resulting in enhanced HA adsorption capacity.
Adsorption capacity driven by MWCNTs/PANI complex would result in higher HA rejection
rate than that of PES membrane. When it reaches to the adsorption capacity, desorption may
happen and reduce HA rejection. Some adsorption causes the membrane pores to be narrower by
pore blocking and forms a cake layer on the membrane surface. So even after it reaches to
adsorption capacity, the rejection rate only drops slightly.
4.5. Flux recovery by chemical cleaning
The HA adsorption inside the membrane pores may result in flux decline during long-term
filtration test. The influence of cleaning method on the flux recovery in the presence of HA was
examined in this section using five different cleaning methods. Table 4.5 summarizes pure water
flux recovery, total fouling ratio and change of HA removal after cleaning. The
MWCNTs/PANI/PES membrane completely recovered pure water flux after cleaning with both
acid and base chemicals (method (4)). However, when using HCl or NaOH individually, its flux
recovery did not reach 100 %. Moreover, cleaning with NaOCl without acid/base chemicals
(method (1)) was found out to be ineffective as FR remained at 51.4 %. The effect of cleaning
method (4) was confirmed by the change in HA removal before/after cleaning membranes.
Membranes cleaned by acid/base chemicals removed 80 % HA while other methods decreased
HA removal after cleaning. The mechanism of removing HA from the membrane surface can be
explained regarding hydrolysis and solubilization [54]. First, NaOH dissolves HA with
carboxylic and phenolic functional groups. HCl oxidizes HA forming soluble aromatic aldehydes
and acids at NOM functional groups.
111
Table 4.5 The results shows a comparison of water flux discovery, total fouling ratio, and
removal efficiency by cleaning method. By applying method 4, ultra-pure water flux recovery
reached to 100 %, and its total fouling ratio was 65 %. Based on the results, cleaning method (4)
is the most effective method for fouling mitigation.
Cleaning method
pH
UPW Flux
recovery
(%)
HA
Removal
(%)
Total
fouling
ratio (%)
(1)
5g/L NaOCl 1 hr
9
51.4
80 → 50
81
(2)
0.1 M NaOH 1 hr
12
93.4
80 → 50
81
(3)
0.1 M HCl 1 hr
2
81.5
85.5 →28.4
82
(4)
0.1 M NaOH 1 hr + 0.1 M HCl 1 hr
12 → 2
100
82 → 80
65
(5)
0.1 M Citric acid (1 hr)
4
79.5
86.4 → 65
76
Fig 4.13 shows HA flux decline and its removal behaviour during three cycles of the filtration
test. As can be seen, its HA permeate flux was fully recovered after the selected chemical
cleaning (method 4) for 3 cycles. 100 % HA removal recovery maintained in the second cycle.
With a combination of acid/base cleaning, the membrane was able to exhibit 100 % water flux
recovery and the lowest total fouling ratio, compared to that the previous research showed only
50 % flux recovery by incorporation of functionalized MWCNTs [2]. Even though HA was
adsorbed in the membrane pores, the majority of the foulants was able to be removed by proper
chemical cleaning.
112
Fig 4.13 HA permeate flux decline and its removal rate behaviour of [B-4] membrane with
chemical cleaning during 3 cycles of filtration test: HA permeability was recovered for 3 cycles
with the chemical cleaning (method 4 presented in Table 5). (The arrow indicates the time when
chemical cleaning was conducted).
It is to note that high NOM adsorption normally would cause more significant fouling. However,
for such a membrane, even with significant fouling or flux drop, the permeate flux is still quite
high due to the extremely high initial flux. Moreover, the membrane has good cleanability
properties. The fouling due to adsorption can be easily removed by acid/base cleaning with
nearly 100 % flux recovery and 65 % total fouling ratio. This indicates that the flux decline of
113
MWCNTs/PANI/PES membrane can be overcome when a proper cleaning method involving
acid and base is used.
4.6. Conclusion
The aim of this study is to advance the low pressure driven membrane process by designing
nano-engineered material based membrane. MWCNTs enhanced membrane by incorporation of
in-situ polymerized MWCNTs/PANI complex achieved superior performance compared to the
conventional PES UF membranes. The unique properties of MWCNTs/PANI complex
successfully tuned the membrane structure and its physical/chemical properties, leading to the
greatly enhanced water permeability and high HA removal rate which could not be effectively
removed by the low-pressure membrane. In particular, positively charged surface of the
membrane with MWCNTs/PANI complex contributed to the enhanced adsorption capacity by
electrostatic interaction, compared that currently used polymeric membranes are negatively
charged. Due to the positively charged surface, this membrane may open up an opportunity for
recovery of valuable cationic macromolecules in bioprocess/pharmaceutical industries.
Examination of removal mechanism and fast water flux would give a significant contribution to
outperforming conventional UF membranes which are widely used in water treatment. Further
research will be targeted to the application of the membrane for enhanced organic matter
removal in seawater pre-treatment, aiming to zero discharge.
114
REFERENCES
1.
Choi, J.-H., J. Jegal, and W.-N. Kim, Fabrication and characterization of multi-walled
carbon nanotubes/polymer blend membranes. Journal of Membrane Science, 2006.
284(1-2): p. 406-415.
2.
Rahimpour, A., M. Jahanshahi, S. Khalili, A. Mollahosseini, A. Zirepour, and B.
Rajaeian, Novel functionalized carbon nanotubes for improving the surface properties
and performance of polyethersulfone (PES) membrane. Desalination, 2012. 286: p. 99107.
3.
Wu, H., B. Tang, and P. Wu, MWNTs/Polyester Thin Film Nanocomposite Membrane:
An Approach To Overcome the Trade-Off Effect between Permeability and Selectivity.
Journal of Physical Chemistry C, 2010. 114(39): p. 16395-16407.
4.
Lu, C. and F. Su, Adsorption of natural organic matter by carbon nanotubes. Separation
and Purification Technology, 2007. 58(1): p. 113-121.
5.
Hyung, H. and J.-H. Kim, Natural organic matter (NOM) adsorption to multi-walled
carbon nanotubes: Effect of NOM characteristics and water quality parameters.
6.
Diao, P. and Z. Liu, Vertically Aligned Single-Walled Carbon Nanotubes by Chemical
Assembly - Methodology, Properties, and Applications. Advanced Materials, 2010.
22(13): p. 1430-1449.
7.
Dresselhaus, M.S., A. Jorio, M. Hofmann, G. Dresselhaus, and R. Saito, Perspectives on
Carbon Nanotubes and Graphene Raman Spectroscopy. Nano Letters, 2010. 10(3): p.
751-758.
8.
Bernholc, J., D. Brenner, M. Buongiorno Nardelli, V. Meunier, and C. Roland,
Mechanical and electrical properties of nanotubes. Annual Review of Materials Science,
2002. 32: p. 347-375.
9.
Gulotty, R., M. Castellino, P. Jagdale, A. Tagliaferro, and A.A. Balandin, Effects of
Functionalization on Thermal Properties of Single-Wall and Multi-Wall Carbon
Nanotube-Polymer Nanocomposites. Acs Nano, 2013. 7(6): p. 5114-5121.
115
10.
Newman, P., A. Minett, R. Ellis-Behnke, and H. Zreiqat, Carbon nanotubes: Their
potential and pitfalls for bone tissue regeneration and engineering. NanomedicineNanotechnology Biology and Medicine, 2013. 9(8): p. 1139-1158.
11.
Coleman, J.N., U. Khan, W.J. Blau, and Y.K. Gun’ko, Small but strong: A review of the
mechanical properties of carbon nanotube–polymer composites. Carbon, 2006. 44(9): p.
1624-1652.
12.
Lin, Y., B. Zhou, K.A.S. Fernando, P. Liu, L.F. Allard, and Y.P. Sun, Polymeric carbon
nanocomposites
from
carbon
nanotubes
functionalized
with
matrix
polymer.
Macromolecules, 2003. 36(19): p. 7199-7204.
13.
Hu, N.T., H.W. Zhou, G.D. Dang, X.H. Rao, C.H. Chen, and W.J. Zhang, Efficient
dispersion of multi-walled carbon nanotubes by in situ polymerization. Polymer
International, 2007. 56(5): p. 655-659.
14.
Grossiord, N., J. Loos, O. Regev, and C.E. Koning, Toolbox for dispersing carbon
nanotubes into polymers to get conductive nanocomposites. Chemistry of Materials, 2006.
18(5): p. 1089-1099.
15.
modifications for the effective dispersion of carbon nanotubes in solvents and polymers.
Carbon, 2012. 50(1): p. 3-33.
16.
Celik, E., H. Park, and H. Choi, Carbon nanotube blended polyethersulfone membranes
for fouling control in water treatment. Water Research, 2011. 45(1): p. 274-282.
17.
Qiu, S., L. Wu, X. Pan, L. Zhang, H. Chen, and C. Gao, Preparation and properties of
functionalized carbon nanotube/PSF blend ultrafiltration membranes. Journal of
18.
Knoell, T., J. Safarik, T. Cormack, R. Riley, S.W. Lin, and H. Ridgway, Biofouling
potentials of microporous polysulfone membranes containing a sulfonated polyetherethersulfone/polyethersulfone block copolymer: correlation of membrane surface
properties with bacterial attachment. Journal of Membrane Science, 1999. 157(1): p.
117-138.
19.
Zhang, H.L., Biomimetic synthesis of poly(lactic-co-glycolic acid)/multi-walled carbon
nanotubes/apatite composite membranes. Express Polymer Letters, 2012. 6(8): p. 620628.
116
20.
Pellegrino, J., The use of conducting polymers in membrane-based separations - A review
and recent developments, in Advanced Membrane Technology, N.N. Li, et al., Editors.
2003. p. 289-305.
21.
Ng, L.Y., A.W. Mohammad, C.P. Leo, and N. Hilal, Polymeric membranes incorporated
with metal/metal oxide nanoparticles: A comprehensive review. Desalination, 2013. 308:
p. 15-33.
22.
Chiang, J.C. and A.G. Macdiarmid, Polyaniline - Protonic acid doping of the emeraldine
form to the metallic regime. Synthetic Metals, 1986. 13(1-3): p. 193-205.
23.
Fan, Z., Z. Wang, N. Sun, J. Wang, and S. Wang, Performance improvement of
polysulfone ultrafiltration membrane by blending with polyaniline nanofibers. Journal of
Membrane Science, 2008. 320(1–2): p. 363-371.
24.
Ruecha, N., R. Rangkupan, N. Rodthongkum, and O. Chailapakul, Novel paper-based
cholesterol biosensor using graphene/polyvinylpyrrolidone/polyaniline nanocomposite.
Biosensors & Bioelectronics, 2014. 52: p. 13-19.
25.
Zhao, S., Z. Wang, J. Wang, S. Yang, and S. Wang, PSf/PANI nanocomposite membrane
prepared by in situ blending of PSf and PANI/NMP. Journal of Membrane Science, 2011.
376(1-2): p. 83-95.
26.
Guillen, G.R., T.P. Farrell, R.B. Kaner, and E.M.V. Hoek, Pore-structure, hydrophilicity,
and
particle
filtration
characteristics
of
polyaniline-polysulfone
ultrafiltration
membranes. Journal of Materials Chemistry, 2010. 20(22): p. 4621-4628.
27.
Zengin, H., W.S. Zhou, J.Y. Jin, R. Czerw, D.W. Smith, L. Echegoyen, D.L. Carroll, S.H.
Foulger, and J. Ballato, Carbon nanotube doped polyaniline. Advanced Materials, 2002.
14(20): p. 1480-+.
28.
Chauvet, A.B. Dalton, R.H. Baughman, and W.K. Maser, A soluble and highly functional
polyaniline-carbon nanotube composite. Nanotechnology, 2005. 16(5): p. S150-S154.
29.
Huang, J.E., X.H. Li, J.C. Xu, and H.L. Li, Well-dispersed single-walled carbon
nanotube/polyaniline composite films. Carbon, 2003. 41(14): p. 2731-2736.
30.
Sainz, R., W.R. Small, N.A. Young, C. Valles, A.M. Benito, W.K. Maser, and M.I.H.
Panhuis, Synthesis and properties of optically active polyaniline carbon nanotube
composites. Macromolecules, 2006. 39(21): p. 7324-7332.
117
31.
Zhou, Y.K., B.L. He, W.J. Zhou, J. Huang, X.H. Li, B. Wu, and H.I. Li, Electrochemical
capacitance of well-coated single-walled carbon nanotube with polyaniline composites.
Electrochimica Acta, 2004. 49(2): p. 257-262.
32.
Yin, J. and B. Deng, Polymer-matrix nanocomposite membranes for water treatment.
Journal of Membrane Science, 2015. 479(0): p. 256-275.
33.
Bai, L.M., H. Liang, J. Crittenden, F.S. Qu, A. Ding, J. Ma, X. Du, S.D. Guo, and G.B. Li,
Surface modification of UF membranes with functionalized MWCNTs to control
membrane fouling by NOM fractions. Journal of Membrane Science, 2015. 492: p. 400411.
34.
Sarbatly, R. and C.-K. Chiam, Evaluation of geothermal energy in desalination by
vacuum membrane distillation. Applied Energy, 2013. 112: p. 737-746.
35.
Lee, S., G. Park, G. Amy, S.-K. Hong, S.-H. Moon, D.-H. Lee, and J. Cho,
Determination of membrane pore size distribution using the fractional rejection of
nonionic and charged macromolecules. Journal of Membrane Science, 2002. 201(1–2): p.
191-201.
36.
Quillard, S., G. Louarn, S. Lefrant, and A.G. Macdiarmid, Vibrational analysis of poly
aniline - A comparative study of leucoemraldine, emeraldine, and pernigraniline bases.
37.
Sun, Y., S.R. Wilson, and D.I. Schuster, High dissolution and strong light emission of
carbon nanotubes in aromatic amine solvents [1]. Journal of the American Chemical
Society, 2001. 123(22): p. 5348-5349.
38.
Chauvet, and W.K. Maser, Soluble self-aligned carbon nanotube/polyaniline composites.
Advanced Materials, 2005. 17(3): p. 278-+.
39.
Leng, J.M., R.P. McCall, K.R. Cromack, Y. Sun, S.K. Manohar, A.G. Macdiarmid, and
A.J. Epstein, Photoexcited solutions and polarons in pernigraniline-base polymers.
40.
de Lannoy, C.-F., E. Soyer, and M.R. Wiesner, Optimizing carbon nanotube-reinforced
polysulfone ultrafiltration membranes through carboxylic acid functionalization. Journal
of Membrane Science, 2013. 447: p. 395-402.
118
41.
Mulder, M., Basic principles of membrane technology. 1996: Kluwer academic
publishers.
42.
Mansourpanah, Y., S.S. Madaeni, A. Rahimpour, M. Adeli, M.Y. Hashemi, and M.R.
Moradian, Fabrication new PES-based mixed matrix nanocomposite membranes using
polycaprolactone modified carbon nanotubes as the additive: Property changes and
morphological studies. Desalination, 2011. 277(1-3): p. 171-177.
43.
Elimelech, M., W.H. Chen, and J.J. Waypa, Measuring the zeta (electrokinetic) potential
of reverse-osmosis membranes by a streaming potential analyzer. Desalination, 1994.
95(3): p. 269-286.
44.
Alam, J., L.A. Dass, M.S. Alhoshan, and A.W. Mohammad, Advances in Membrane
Development Based on Electrically Conducting Polymers. Advances in Polymer
Technology, 2013. 32(S1): p. E189-E197.
45.
Macdiarmid, A.G., J.C. Chiang, A.F. Richter, and A.J. Epstein, Polyaniline: a new
concept in conducting polymers. Synthetic Metals, 1987. 18(1-3): p. 285-290.
46.
Anand, J., S. Palaniappan, and D.N. Sathyanarayana, Conducting polyaniline blends and
composites. Progress in Polymer Science, 1998. 23(6): p. 993-1018.
47.
Cheng, W., S.A. Dastgheib, and T. Karanfil, Adsorption of dissolved natural organic
matter by modified activated carbons. Water Research, 2005. 39(11): p. 2281-2290.
48.
Dastgheib, S.A., T. Karanfil, and W. Cheng, Tailoring activated carbons for enhanced
removal of natural organic matter from natural waters. Carbon, 2004. 42(3): p. 547-557.
49.
Korshin, G.V., C.-W. Li, and M.M. Benjamin, Monitoring the properties of natural
organic matter through UV spectroscopy: A consistent theory. Water Research, 1997.
31(7): p. 1787-1795.
50.
Zularisam, A.W., A.F. Ismail, M.R. Salim, M. Sakinah, and H. Ozaki, The effects of
natural organic matter (NOM) fractions on fouling characteristics and flux recovery of
ultrafiltration membranes. Desalination, 2007. 212(1-3): p. 191-208.
51.
Zularisam, A.W., A. Ahmad, M. Sakinah, A.F. Ismail, and T. Matsuura, Role of natural
organic matter (NOM), colloidal particles, and solution chemistry on ultrafiltration
performance. Separation and Purification Technology, 2011. 78(2): p. 189-200.
52.
Pan, B. and B. Xing, Adsorption Mechanisms of Organic Chemicals on Carbon
Nanotubes. Environmental Science & Technology, 2008. 42(24): p. 9005-9013.
119
53.
Saleh, N.B., L.D. Pfefferle, and M. Elimelech, Aggregation Kinetics of Multiwalled
Carbon Nanotubes in Aquatic Systems: Measurements and Environmental Implications.
54.
Porcelli, N. and S. Judd, Chemical cleaning of potable water membranes: A review.
Separation and Purification Technology, 2010. 71(2): p. 137-143.
120
Chapter 5 Application of Carbon nanotube enhanced ultrafiltration membrane for seawater pretreatment
and examination of salinity on the membrane performance
Chapter 5. Application of carbon nanotube-enhanced
membrane for seawater pretreatment and examination of
salinity on the membrane performance
5.1 Introduction
Seawater reverse osmosis (SWRO) desalination technology requires an effective pretreatment to
inhibit the decline in performance with irreversible fouling on membrane and degradation by
frequent cleaning [1-3]. Such a drawback is derived from organic matter as a serious membrane
foulant in seawater in RO desalination plants. Seawater organic matters include extracellular
polymeric substances (EPS) or biopolymers, humics, fulvic acids, carboxylic acid and other low
molecular weight dissolved organic matters (LMW-DOMs). Of these substances, colloidal
matters in the range of 3 – 20 nm such as humic substances and biopolymers are the main
membrane foulant to be adsorbed on the membrane surface or in the membrane pores. In most
cases, LMW-DOMs account for 50% of the seawater organic matter (SWOM). Organic foulants
are the precursor to the biological growth so it can also accelerate the biofouling on membrane.
Low-pressure membranes such as microfiltration (MF) and ultrafiltration (UF) can be an option
of seawater pretreatment as they can remove particulate, bacteria and large molecular weight
organic matter. However, direct MF/UF filtration can overpass the technical limits of the process,
resulting in RO membrane fouling [4]. In particular, a low-pressure membrane with cut off
higher than 100 kDa was not able to remove LMW-DOMs (less than 350 Da) which can
accelerate biofouling on RO membranes [5].
A solution to enhance the organic removal performance in MF/UF membrane system is the
integration of the physico-chemical processes such as adsorption or coagulation/flocculation [6,
121
7]. According to previous studies, ferric chloride (FeCl3) as a coagulant forms flocs with organic
matter by co-precipitation [8, 9]. Such a coagulation/precipitation system in seawater
pretreatment has shown to be effective in total DOM removal, but basically, it cannot remove
LMW-DOMs in seawater. Recently, to overcome the limitation of the physico-chemical process
with low pressure membrane, membrane hybrid system coupled with adsorption (with the
addition of powder activated carbon, PAC) has been reported to exhibit enhanced performance
on the removal of LMW-DOMs containing around half of SWOM [10]. PAC (as an adsorbent)
adsorbed organic matter by co-valence bonding [11], and removed LMW-DOMs with its affinity
to them and by biological activity of microbial community developed [12].
Even though such an enhanced DOM removal was achieved by the membrane hybrid system in
seawater pretreatment, a large amount of chemical sludge generated due to a usage of a relatively
high amount of adsorbent. In this sense, an approach to the reduction of chemical dosage must be
investigated, and a development of membrane with higher water flux and the anti-biofouling
effect is desirable for sustainable pretreatment in SWRO.
Recently, there have been many attempts on enhancing UF membrane performance by
incorporation of nanomaterials. Of the nanomaterials, carbon nanotubes (CNTs) are one of the
favored approaches due to its unique characteristic of excellent adsorption capacity for organic
matter [13]. In particular, the introduction of functionalized CNTs to the polymeric UF
membrane has contributed to delivering increased permeate flux by changing membrane surface
hydrophilicity [14-16], improved rejection of bovine serum albumin (BSA) and alleviated
membrane fouling [17-20]. Even though development of the CNTs membrane has provided
certain benefits to water treatment, there are few existing research papers on its application to
seawater pretreatment for DOM removal. Therefore, feasibility study of CNTs membrane in
122
seawater pretreatment for effective removal of LMW-DOMs would be of great help for currently
facing the limitation of a membrane system for seawater pretreatment. Based on the recent
progress of nanocomposite membrane in water treatment, development of the membrane having
an enhanced adsorption capacity can also contribute to lowering chemical dose, resulting in the
reduction of sludge volume generated.
Seawater contains high concentrations of ions, and this ionic strength has been recognized to
seriously affect DOM removal and fouling potential in membrane operation [21, 22]. Therefore,
in-depth study on the effect of ionic strength on membrane performance would be of great
importance in successful CNTs membrane application to SWRO pretreatment.
In this study, multi wall carbon nanotubes (MWCNTs) enhanced membranes fabricated by insitu polymerization, since high water flux and effective natural organic matter (NOM) removal
was previously reported with this membrane. The aims of this study were i) to test MWCNTs
enhanced membrane in the seawater pretreatment with reducing PAC adsorbent dose by
employing MWCNTs membrane with enhanced adsorption capacity of SWOM. Especially, it
was to improve rejection efficiency of LMW-DOMs with high permeate flux in MWCNTs
membrane filtration compared to the conventional UF membrane, and ii) to examine the effect of
ion strength on organic removal and performance in the MWCNTs membrane filtration. Further,
the effects of salinity and ion strength on the organic rejection efficiency and permeate flux were
examined in MWCNTs enhanced membrane system.
123
5.2.
Materials and methods
5.2.1 Materials
5.2.1.1 Seawater
Seawater was taken from Chowder Bay, Sydney in Australia. It was withdrawn 1 m below the
sea surface level and passed through 140 µm the centrifuge filtration system to remove the large
particles. Turbidity and pH of seawater used in this study were 0.5-0.7 NTU and 7.8-8.0,
respectively.
5.2.1.2 Membranes
The fabrication procedure of 1.5
wt%
of MWCNTs enhanced membrane was adopted from
Chapter 3 Materials and method. Characteristics of membranes used in this study are given in
Table 4.1. Detailed information is elucidated in Chapter 4.
Table 5.1 Characteristics of the membranes used in this study.
Pore size
UPW* permeability
Membranes
Zeta
MWCO**
Contact angle
potential***
by BET
(LMH/bar)
(kDa)
(°)
(nm)
MWCNTs
1272 ± 103.8
(mV)
5.0 ± 0.4
45.4 ± 0.1
9.8
57.6 ± 0.4
-21
12
PES
*
60± 4.3
: UPW: Ultrapure water,
**
4.6 ± 0.3
: MWCO: Molecular weight cut-off,
background electrolyte of 0.001M KCl.
124
***
: Zeta potential was measured at pH 5.6 at a
5.2.1.3 Powder activated carbon (PAC)
Coal-based PAC (MDW3545CB, James Cumming & Sons PTY LTD) was used as an adsorbent
in this study. The mean diameter and the nominal size (80 % min. finer than) of PAC were 19.7
μm and 75.0 μm, respectively. More details of PAC can be found in elsewhere [7].
5.2.2 Membrane filtration test
Performances of PES (a laboratory-prepared UF membrane) and MWCNTs enhanced membrane
were evaluated in the membrane system with adsorption (PAC addition). For the membrane
system with PAC adsorption, PAC (0.5 – 1.5 g/L based on feedwater volume) was added to
seawater in the adsorption system. Then, pretreated seawater was undergone membrane filtration.
Membrane filtration without adsorption pretreatment (PAC=0g/L) was also conducted to
evaluate the performance of membrane filtration system itself. Membrane filtration was done in
dead-end mode under 200 kPa with 0.00146 m2 of effective membrane area. A 1L of seawater
was filtered, and permeability was monitored over the time.
5.2.3 Dissolved organic matter measurement
Dissolved organic concentration (DOC) and the detailed organic fractions in feed/permeate were
measured by liquid chromatography–organic carbon detection (LC-OCD) [23]. Hydrophilic
organic fractions were separated into each fragment depending on their molecular size by the
retention time in the size exclusion chromatography (SEC) column. Dual columns were applied
in this study with 180 min retention time. Two different detectors (an ultraviolet detector (UVD)
(absorption at 254 nm) and an organic carbon detector (OCD) (after inorganic carbon purging))
were utilized to measure the concentration of separated hydrophilic organic compounds. The
eluted order of each fragment was biopolymers – humic substances/building blocks – LMW
125
organics (acids and neutrals). The amount of each organic fraction was calculated using a
software program (ChromCALC DOC-LABOR, Karlsruhe, Germany) based on the
chromatogram obtained from LC-OCD analysis. Detailed procedure is described in a previous
study [24].
Removal efficiency (%) of organic matter was calculated by the following equation:
𝑅 = (1 −
𝐶𝑝
) × 100
𝐶𝑓
Where, Cf is the concentration (µg/L) of feed water and, Cp is the concentration (µg/L) of
permeate (or membrane filtrate) from the membrane filtration system with PAC adsorption or
membrane filtration itself.
5.2.4 Surface charge of membrane
The surface charge (in terms of zeta potential, mV) of membrane surface was measured by using
the streaming potential technique (Anton Paar electrokinetic analyzer, USA). The measurement
was conducted at different ionic strengths ranged between 0.001 and 0.010 M of KCl and pH 7.8.
5.2.5 Fouled membrane characterization
After filtering 1L of feed water (684.9 L/m2 of seawater with or without PAC adsorption),
membrane samples were taken and dried for 24h prior to analyses.
The morphology and chemical composition of fouled membranes were observed using scanning
electron microscopy with an energy dispersive X-ray spectrometry (SEM-EDX, Zeiss Ultra,
Germany). For SEM-EDX analysis, membranes then were coated with thin layer of gold. The
contact angle was performed to measure the membrane hydrophilicity using 4 – 8 µL sessile
droplets of Milli-Q water with a Kruss Easy Drop goniometer (Germany).
126
5.3.
Results and discussion
5.3.1 Seawater filtration performance (without PAC addition)
5.3.1.1 Permeate flux
Fig 5.1 shows permeability (permeate flux) pattern of MWCNTs and PES-UF membrane in
UPW and seawater filtration. MWCNTs membrane shows significantly increased UPW
permeability (initial permeate flux was 1298.4 L/m2-h-bar (LMH/bar)) while initial permeate
flux of PES-UF membrane was 66.2 LMH/bar. In seawater filtration, substantially declined
permeate flux was observed with the MWCNTs membrane. In terms of initial permeate flux; it
was 321.3 LMH/bar, which was almost 4 times dropped compared to UPW permeability.
However, there was no significant change in permeate flux in seawater filtration by PES-UF
membrane.
127
Fig 5.1 Permeability pattern of MWCNTs and PES-UF membrane in UPW and seawater
filtration.
5.3.1.2 Organic reduction
A 5 mg/L of humic acid (HA) was dissolved in UPW, and it was firstly tested with two
membranes. Especially, HA removal by MWCNTs membrane was superior (82±4%). However,
in seawater filtration, there was a marginal reduction of HS (2.62 - 4.62%) by both membranes,
even though HS concentration in seawater was only 0.496±0.012 mg/L (Table 5.2).
Table 5.2 Removal efficiencies of HS in seawater and 5 mg/L of HA in UPW by MWCNTs and
PES-UF membrane filtration (HS and HA concentrations were measured using LC-OCD and
TOC analyzer, respectively).
HS in
seawater
(mg/L)
MWCNTs
0.496±0.012
PES-UF
*
HS in permeate
(mg/L)
R*
(%)
0.473±0.020
4.64±0.40
0.483±0.011
2.62±0.22
R: removal efficiency
128
HA in
UPW
(mg/L)
5.0±0.1
HA in
permeate
(mg/L)
R*
(%)
0.9±0.2
82±4
3.1±0.3
38±6
5.3.2 Effect of PAC addition
Fig 5.2 shows the permeability patterns in seawater filtration (before and after PAC adsorption)
of MWCNTs and PES-UF membranes. When raw seawater was filtered, the initial permeability
of MWCNTs membrane was to 300 LMH/bar. Interestingly, when PAC adsorption coupled, its
permeability increased up to 600 LMH/bar. However, more rapid flux decline occurred in
MWCNTs membrane filtration. It was probably due to more organic adsorption on the
membrane surface by electrostatic interaction at the initial stage of filtration with higher organic
loading by high permeate flux. When the lower dosage of PAC decreased to 0.5 g/L, the
permeability of MWCNTs membrane still maintained in a high value, and there was no
significant decline. However, no change in flux pattern was observed on PES membrane no
matter how PAC was incorporated.
The removal efficiencies of the organic fraction by MWCNTs enhanced membrane coupled with
or without PAC adsorption are given in Table 5.3. PES-UF membrane itself removed only 13%
of DOC from raw seawater with 60% of biopolymers’ removal efficiency. DOC reduction by
MWCNTs membrane (26%) was slightly higher than PES-UF membrane.
PAC showed a high removal efficiency of hydrophilic organics, especially biopolymers and
humic substances (high molecular weight organics). However, the removal of LMW neutrals was
marginal. For example, around 50% of DOC was removed by 1.5 g/L of PAC. It is mainly due to
the significant reduction of biopolymers and humic substances, which were 87% and 48%,
respectively. Reduced PAC dose (0.5g/L) lowered DOC removal efficiency to 37%.
PES-UF membrane filtration with PAC adsorption increased DOC removal efficiency to 53%
compared to PAC adsorption only (49%). In addition, DOC removal efficiency of MWCNTs
membrane filtration significantly increased to higher than 60%, when it corporates with PAC
129
adsorption (1.5g/L). In particular, the removal of LMW neutrals in MWCNTs membrane
filtration increased by PAC adsorption to 34% from 8% (without PAC adsorption). It is alarming
to note that there was an only small reduction of LMW neutrals by either MWCNTs membrane
itself or PAC adsorption itself. Moreover, DOC removal by MWCNTs membrane, even though
low PAC dose was used (0.5g/L), was comparable to that by PES-UF membrane filtration with
1.5g/L of PAC. It indicates that MWCNTs enhanced membrane had effective hydrophilic DOC
removal efficiency when PAC adsorption was coupled.
Low HS removal efficiency of MWCNTs membrane filtration could be explained by reduced
adsorption capacity of the MWCNTs membrane in feed water with high ionic strength. In
previous research, NOM removal mechanism by MWCNTs enhanced membrane was found to
be a greatly enhanced electrostatic interaction between a humic acid (HA) and positively charged
MWCNTs enhanced membrane. However, ionic strength in seawater may probably weaken
surface charge of both the membrane and humic substances. Negatively charged humic
substances form metal ion-humic precipitation and aggregation in seawater with reducing surface
charge of HA [25]. Thus, it may reduce the affinity of the MWCNTs enhanced membrane to the
humic substances by reducing electrostatic interaction between HA (negatively-charged) and
MWCNTs enhanced membrane (positively-charged). It is consistent with a previous research
done by Jermann et al. [26]. They reported that metal ions neutralize the negatively charged
foulants and the charge on the membrane. Further, salinity could weaken the surface charge of
the membrane. It will be discussed in the section 5.3.4.1 via zeta potential measurement of the
membrane under different ionic strengths. In addition to the reduced electrostatic interaction
between humic substances and MWCNTs membranes, insufficient contact time may be
considered as the main reason for low humic substances removal efficiency. Due to the rapid
130
permeate flux, MWCNTs enhanced membrane system had a short time for adsorption of humic
substances. Based on that the removal mechanism of MWCNTs enhanced membrane is the
enhanced adsorption by electrostatic interaction, contact time may also be a critical factor in the
removing of LMW organic matter with reduced surface charge by salty water. Therefore,
organics easily pass through MWCNTs membrane due to the fast permeate flux.
Fig 5.2 Permeability decline pattern in MWCNTs and PES membrane filtration of seawater with
different PAC doses.
131
Table 5.3 Concentration and removal efficiency (R) of organic fractions by MWCNTs enhanced
and PES-UF membrane with and without PAC adsorption.
Membrane/treatme
Hydrophili
Bio-
Humic
Building
LMW
LMW
c DOC
polymers
substances
blocks
Neutrals
acids
DOC
nt types
Raw
(µg/L)
1,300
1,027
179
496
36
250
65
(µg/L)
658
529
24
256
2
241
6
R (%)
49
48
87
48
94
4
91
(µg/L)
825
669
52
364
4
242
7
R (%)
37
35
71
27
89
3
89
(µg/L)
1,135
894
71
483
35
246
59
R (%)
13
13
60
3
3
2
9
PES-UF-
(µg/L)
607
474
16
204
19
233
2
PAC1.5
R (%)
53
54
91
59
47
7
97
MWCNT
(µg/L)
968
767
21
473
29
229
15
s
R (%)
26
25
88
5
19
8
77
MWCNT
(µg/L)
439
386
14
193
8
165
6
s-PAC1.5
R (%)
66
62
92
61
78
34
91
MWCNT
(µg/L)
529
452
14
226
16
189
7
s-PAC0.5
R (%)
59
56
92
54
56
24
89
seawater
PAC1.5
PAC0.5
PES-UF
132
Chapter 5 Application of Carbon nanotube enhanced ultrafiltration membrane for seawater pretreatment and
examination of salinity on the membrane performance
5.3.3
Effect of salinity and divalent cation on HA rejection
In Chapter 4, the MWCNTs enhanced membrane was found to have effective HA removal efficiency (>
80%). However, in section 5.3.1.2, low removal efficiency on humic substances was observed in
seawater filtration where high salinity water samples were treated. It is noted that the low HS removal
efficiency of MF/UF membranes in SWRO pretreatment can also cause irreversible RO membrane
fouling [27]. For this reason, it is imperative to discuss which factor can seriously affect the HS removal
efficiency of the MWCNTs enhanced membrane in seawater filtration to evaluate the performance of the
MWCNTs membrane in SWRO pretreatment.
To examine the effect of salinity on humic acid (HA) rejection, membrane filtration was conducted with
5mg/L of HA under different NaCl concentrations. Further, the effect of ionic strength on membrane
performance in HA rejection test was investigated under the presence of divalent cation (CaCl2). Fig 5.3
and 5.4 show the permeate flux and HA removal under the different NaCl concentrations, respectively.
Overall, salinity significantly contributed to the enhanced permeability and reduced HA removal
efficiency in MWCNTs membrane. In contrast, PES-UF membrane performance was marginally
affected by monovalent ion strength (NaCl).
As can be seen from Fig 5.3, permeate flux increased by 40 % when 0.01 M of NaCl was a presence
with 0.5mg/L of HA compared to 5mg/L of HA without NaCl. Permeate flux decline at the initial stage
(30 min) was overcome when NaCl increased to 0.1 M, showing the nearly same permeability as the one
under 0.01 M NaCl. By contrast, HA removal efficiency was shown to reduce considerably under
increasing concentration of ionic strength. Especially at the final stage, it decreased sharply (80 % to
30 %) as ionic strength increased. Such a result demonstrates that adsorption capacity of MWCNTs
membrane reduced under higher concentration, and increasing concentration accelerated to reach HA
adsorption capacity. It may appear due to the reduced electrostatic interaction between negatively
133
charged humic substances and positively charged membrane surface. As ionic strength increases to 0.1
M, it may weaken the surface charge of humic substances by forming a double layer on humic
substances. Such a consequence can be interpreted to contribute to the flux enhancement by hindering
adsorption of HA from the membrane surface, compared to the HA 5 ppm filtration without NaCl.
Divalent ion (CaCl2) mostly affected to significantly reduced permeability in both MWCNTs and PESUF membranes. In contrast, the addition of CaCl2 considerably enhanced HA removal efficiency in both
membranes. Permeate flux behavior and HA rejection efficiency under divalent cationic strength (CaCl2)
is shown in Fig 5.4. Severe flux decline was observed in MWCNTs enhanced membrane after 660 min
filtration, in comparison to the flux under monovalent ionic strength (NaCl). PES membrane also
exhibited much lower flux behavior (7.4 LMH/bar) from the initial stage. The removal rate of MWCNTs
enhanced membrane in the initial stage increased by 12%, and no drop was observed during whole
filtration period (filtered volume = 684.9 L/m2), compared that there was a sharp decrease in HA
removal efficiency under increasing NaCl concentration by approximately 50 %. The enhanced HA
removal efficiency is largely due to the increased HA aggregation under divalent cation such as Ca2+. It
may make a compact network between HA by neutralizing negative charge on HA under around pH 6,
resulting in larger size of HA. Even if the surface charge of MWCNTs enhanced membrane was
weakened by salt ions, the synergetic effect of Ca2+ was able to contribute to the HA removal
enhancement [28].
The synergetic effect of CaCl2 mostly contributed to PES-UF membrane, which was shown to have low
HA removal efficiency. PES-UF membrane exhibited higher removal efficiency (87 %) during the
whole period. The notable thing is its removal rate increased by final stage. It appeared probably due to
the fact that divalent cation increases HA aggregation by cross-linking of HA, which accelerated
adsorption of humic acid on the membrane. In addition to the enhanced adsorption by the larger size of
134
HA, Ca ion has a bridge between HA adsorbed on the membrane surface and the one in feed water by
reducing repulsion force among HA. Thus, relatively lower removal rate was observed at the initial
stage (100 mL filtration) and increased at the final stage. Further, sufficient contact time of HA
adsorption due to the severe flux decline by CaCl2 combining with HA led to the increased HA rejection
efficiency in PES-UF membrane.
(a) MWCNTs membrane
(b) PES-UF membrane
Fig 5.3 Permeability of membranes with HA 5mg/L (ppm) under the different NaCl concentrations.
135
(a) MWCNTs membrane
(b) PES-UF membrane
Fig 5.4 HA rejection (%) of membranes with HA 5mg/L (ppm) under the different NaCl concentrations.
(a) Permeability
(b) HA rejection
Fig 5.5 Permeability and HA rejection (%) of membranes with HA 5 ppm under 0.01M of CaCl2.
136
5.3.4
Factors affecting membrane performance
5.3.4.1 Surface charge of membrane under different ionic strength – organic removal
According to the previous study (Chapter 3), the surface charge of the membrane was found to be the
most critical factor of organic matter removal by a membrane. The surface charges of the membranes
depending on ionic strength (under different background electrolyte, KCl) were examined to verify the
effects of ionic strength on membrane performance. Fig 5.6 shows that zeta potential of MWCNTs and
PES-UF membrane under different KCl concentrations at pH 7.8. Zeta potentials of MWCNTs and PESUF membrane at 0.001M of KCl were 9.8 and -21.0 mV, respectively. It indicates MWCNTs enhanced
membrane charged positively while PES-UF membrane had negatively charged surface. Zeta potential
of MWCNTs membrane significantly decreased when KCl concentration increased from 0.001 to 0.010
M. However, there was not a significant decline in zeta potential in PES-UF membrane despite
increasing KCl concentration. It indicates that positively charged MWCNTs enhanced membrane is
more strongly influenced by the ionic strength, in comparison to the negatively charged PES-UF
membrane. Based on the results, it implies that organic matter removal efficiency of MWCNTs
enhanced membrane could be lowered by increased ionic strength. It may be due to the weakened
electrostatic interaction between membrane surfaces and mostly negatively charged SWOM such as
humic substances.
137
Fig 5.6 Zeta potential of membranes under different ionic strengths by streaming potential at pH 7.8 and
different KCl (background electrolyte) concentrations (0.001, 0.005 and 0.010 M).
5.3.4.2 Chemical precipitation on membrane – filtration behavior
SEM-EDS analysis on the fouled membranes surface was performed to observe the morphology of
membrane surface and chemical precipitation (elements). Fig 5.7 presents SEM images and EDS results
of each membrane surface (a: MWCNTs enhanced membrane and b: PES-UF membrane) before and
after seawater filtration tests. Both virgin membranes (before filtration) were mainly composed of
carbon (C), oxygen (O) and sulfur (S) since they were fabricated based on PES. As expected,
organic/inorganic foulants oriented from seawater (or PAC) were observed on the fouled membranes
with/without PAC adsorption. As shown in Fig 5.7, some crystals were found on the surface of fouled
membranes. The main components of the crystals were carbon (C) and oxygen (O) as the components of
foulants by organic matter [10]. Sodium (Na), chloride (Cl) and magnesium (Mg) were found to be other
138
components of these crystals, which are the main components of seawater. More crystals binding with
organic contaminants were observed on the surface of MWCNTs enhanced membrane than PES-UF
membrane. In particular, the proportion of inorganic foulants such as Na (7.85 %), Cl (15.53 %), S
(6.14 %), Fe (5.79 %) and Mg (2.23 %) was shown to increase in crystal form when PAC adsorption
was used, compared that only Na (3.40 %) and Cl (9.93 %) were detected on the surface MWCNTs
enhanced membrane filtered with raw seawater (without PAC adsorption). The result explains flux
behaviors in MWCNTs enhanced membrane filtration coupled with PAC adsorption. Deposition of
accelerating crystallization on the surface of the membrane may cause rapid flux decline in MWCNTs
enhanced membrane filtration with higher organic removal efficiency (due to more bonding with
organics) although improved permeate flux compared MWCNTs enhanced membrane alone. In contrast,
there was no increase of crystal bonded with organic foulants on the PES-UF membrane after PAC
adsorption was integrated. It may be due to the fact that crystallization occurs in the interactions
between crystals containing humic substances and salts, and membrane surface [11]. By the bridging
effect of hydrophilic organic contaminants on salt and positively charged membrane surface, more
crystallization deposited on the more hydrophilic surface of MWCNTs enhanced membrane, compared
to the less hydrophilic PES-UF membrane.
139
(a)
Virgin
Raw seawater
PAC 1.5g/L
treated seawater
(b)
Virgin
140
Raw seawater
PAC 1.5g/L
treated seawater
Fig 5.7 SEM images and elements of the membrane surface by EDS analysis before and after seawater
filtration (Filtered volume = 684.9 L/m2): (a) MWCNTs enhanced membrane, and (b) PES membrane.
5.3.4.3 Membrane hydrophilicity (water contact angle)
Hydrophilicity of fouled membranes was measured in terms of water contact angle. Table 5.4 shows
contact angle of fouled membrane before and after filtering of 684.9 L/m2 of seawater samples
(untreated and PAC treated). As can be seen from Table 5.4, MWCNTs enhance membrane has more
hydrophobic membrane surface. After filtration of 684.9 L/m2 of raw seawater, contact angles of PESUF and MWCNTs enhanced membranes decreased from 57.6±0.4° to 53.9±1.9° and from 45.4±0.1° to
42.7±2.6°, respectively. It demonstrates that membrane surface became more hydrophilic and it may be
due to the adsorption of hydrophilic contaminants in seawater onto the membrane surface. Such a result
may occur due to the fact that organic matter in seawater forms ion-humic precipitation and
agglomeration by the interaction between salt ions and negatively charged organic matters [25]. Such
141
precipitates are adsorbed onto the membrane surface during filtration test, resulting in contact angle
increase of the fouled membrane. However, an interesting trend on contact angle was observed on
MWCNTs enhanced membrane when it was filtered the PAC treated seawater. However, when
pretreated seawater with PAC adsorption was used in MWCNTs enhanced membrane filtration, contact
angle increased slightly to 45.9±2.1°, indicating that hydrophilic organic compounds were removed by
PAC adsorption. It is fact that SWOM contains mostly hydrophilic organic compounds, which were
effectively removed by PAC. However, for PES-UF membrane, there was not a significant change in
contact angle when PAC adsorption incorporated. Such a result was confirmed by SEM-EDX
observation indicating that there was not a significant crystallization formation on PES-UF membrane
even though PAC adsorption was coupled.
Table 5. 4 Changes of water contact angle of the membrane before and after filtration (filtered volume =
684.9 L/m2).
PES-UF membrane
MWCNTs enhanced membrane
(°)
(°)
Virgin
57.6 ± 0.4
45.4 ± 0.1
Filtration of raw seawater
53.9 ± 1.9
42.7 ± 2.6
Filtration of 1.5 g/L PAC treated seawater
54.8 ± 3.2
45.9 ± 2.1
142
5.4.
Conclusion
This study evaluated MWCNTs enhanced UF membrane for SWRO pretreatment. The results show that
MWCNTs enhanced membrane can be an alternative SWRO pretreatment option. The MWCNTs
enhanced membrane exhibited greatly enhanced permeate flux when PAC adsorption was coupled,
compared to PES-UF membrane. It is mainly due to the increased porosity, hydrophilicity and fingerlike structure having narrow pore size on top and relatively large micro-pores on sub-layer. There was
no decline in permeate flux and removal efficiency even if PAC dosage decreased from 1.5 g/L to 0.5
g/L, which can substantially contribute to reducing sludge volume generated from SWRO pretreatment
by reducing chemical usage. The salinity was found to affect greatly mainly enhanced permeate flux.
However, it contributed to the considerably reduced organic matter removal efficiency. The detrimental
effect on the organic matter removal efficiency in MWCNTs membrane could be most probably due to
the significantly reduced electrostatic interaction between positively charged membrane surface and
negatively charged SWOM. The salinity was proven to reduce surface charge of HS (negative charge)
by HA filtration test under different ionic strengths. Further, it mainly reduced the surface charge of
MWCNTs membrane (positively charge) via zeta potential measurement under different ionic strengths.
143
References
1.
Greenlee, L.F., D.F. Lawler, B.D. Freeman, B. Marrot, and P. Moulin, Reverse osmosis
desalination: Water sources, technology, and today's challenges. Water Research, 2009. 43(9): p.
2317-2348.
2.
Sheikholeslami, R. and J. Bright, Silica and metals removal by pretreatment to prevent fouling of
reverse osmosis membranes. Desalination, 2002. 143(3): p. 255-267.
3.
Abu Tarboush, B.J., D. Rana, T. Matsuura, H.A. Arafat, and R.M. Narbaitz, Preparation of thinfilm-composite polyamide membranes for desalination using novel hydrophilic surface modifying
macromolecules. Journal of Membrane Science, 2008. 325(1): p. 166-175.
4.
Bonnelye, V., L. Guey, and J. Del Castillo, UF/MF as RO pre-treatment: the real benefit.
Desalination, 2008. 222(1-3): p. 59-65.
5.
Voutchkov, N., Considerations for selection of seawater filtration pretreatment system.
Desalination, 2010. 261(3): p. 354-364.
6.
Kim, J., Z. Cai, and M.M. Benjamin,
NOM fouling mechanisms in a hybrid
adsorption/membrane system. Journal of Membrane Science, 2010. 349(1-2): p. 35-43.
7.
Jeong, S., S. Vigneswaran, and V. Nguyen, Submerged membrane coagulation hybrid system as
pretreatment to organic matter removai from seawater. Water Science and Technology, 2011.
8.
Chapman, H., S. Vigneswaran, H.H. Ngo, S. Dyer, and R. Ben Aim, Pre-flocculation of
secondary treated wastewater in enhancing the performance of microfiltration. Desalination,
2002. 146(1-3): p. 367-372.
9.
Duan, J., N.J.D. Graham, and F. Wilson, Coagulation of humic acid by ferric chloride in saline
(marine) water conditions. Water Science and Technology, 2003. 47(1): p. 41-48.
10.
Tansakul, C., S. Laborie, and C. Cabassud, Adsorption combined with ultrafiltration to remove
organic matter from seawater. Water Research, 2011. 45(19): p. 6362-6370.
11.
Duan, J.M., F. Wilson, N. Graham, and J.H. Tay, Adsorption of humic acid by powdered
activated carbon in saline water conditions. Desalination, 2003. 151(1): p. 53-66.
12.
Tomaszewska, M. and S. Mozia, Removal of organic matter from water by PAC/UF system.
Water Research, 2002. 36(16): p. 4137-4143.
144
13.
Saleh, N.B., L.D. Pfefferle, and M. Elimelech, Aggregation Kinetics of Multiwalled Carbon
Nanotubes in Aquatic Systems: Measurements and Environmental Implications. Environmental
14.
de Lannoy, C.-F., E. Soyer, and M.R. Wiesner, Optimizing carbon nanotube-reinforced
polysulfone ultrafiltration membranes through carboxylic acid functionalization. Journal of
15.
Celik, E., H. Park, and H. Choi, Carbon nanotube blended polyethersulfone membranes for
fouling control in water treatment. Water Research, 2011. 45(1): p. 274-282.
16.
Choi, J.-H., J. Jegal, and W.-N. Kim, Fabrication and characterization of multi-walled carbon
nanotubes/polymer blend membranes. Journal of Membrane Science, 2006. 284(1-2): p. 406-415.
17.
Qiu, S., L. Wu, X. Pan, L. Zhang, H. Chen, and C. Gao, Preparation and properties of
functionalized carbon nanotube/PSF blend ultrafiltration membranes. Journal of Membrane
Science, 2009. 342(1-2): p. 165-172.
18.
Kim, E., G. Hwang, M.G. El-Din, and Y. Liu, Development of nanosilver and multi-walled
carbon nanotubes thin-film nanocomposite membrane for enhanced water treatment. Journal of
19.
Rahimpour, A., M. Jahanshahi, S. Khalili, A. Mollahosseini, A. Zirepour, and B. Rajaeian, Novel
functionalized carbon nanotubes for improving the surface properties and performance of
polyethersulfone (PES) membrane. Desalination, 2012. 286: p. 99-107.
20.
Vatanpour, V., S.S. Madaeni, R. Moradian, S. Zinadini, and B. Astinchap, Fabrication and
characterization of novel antifouling nanofiltration membrane prepared from oxidized
multiwalled carbon nanotube/polyethersulfone nanocomposite. Journal of Membrane Science,
2011. 375(1-2): p. 284-294.
21.
22.
Singh, G. and L.F. Song, Quantifying the effect of ionic strength on colloidal fouling potential in
membrane filtration. Journal of Colloid and Interface Science, 2005. 284(2): p. 630-638.
23.
Naidu, G., S. Jeong, and S. Vigneswaran, Interaction of humic substances on fouling in
membrane distillation for seawater desalination. Chemical Engineering Journal, 2015. 262(0): p.
946-957.
145
24.
Jeong, S., S.J. Kim, L.H. Kim, M.S. Shin, S. Vigneswaran, T.V. Nguyen, and I.S. Kim, Foulant
analysis of a reverse osmosis membrane used pretreated seawater. Journal of Membrane Science,
2013. 428: p. 434-444.
25.
W. Stumm, J.J.M., Aquatic Chemistry-Chemical Equilibria and Rates in Natural Waters (3rd
ed.)Wiley & Sons, Inc. 3rd edition ed. 1996: Wiley & Sons, Inc.
26.
Jermann, D., W. Pronk, S. Meylan, and M. Boller, Interplay of different NOM fouling
mechanisms during ultrafiltration for drinking water production. Water Research, 2007. 41(8): p.
1713-1722.
27.
Prihasto, N., Q.-F. Liu, and S.-H. Kim, Pre-treatment strategies for seawater desalination by
reverse osmosis system. Desalination, 2009. 249(1): p. 308-316.
28.
Baalousha, M., M. Motelica-Heino, and P.L. Coustumer, Conformation and size of humic
substances: Effects of major cation concentration and type, pH, salinity, and residence time.
Colloids and Surfaces A: Physicochemical and Engineering Aspects, 2006. 272(1–2): p. 48-55.
146
Chapter 6. Protein fouling behaviour under different solution chemistry (pH and ionic strength)
Chapter 6. Protein fouling in carbon nanotubes enhanced
ultrafiltration membrane: fouling mechanism as a function
of pH and ionic strength
6.1.
Introduction
Ultrafiltration (UF)/nanofiltration (NF) membrane technologies have been increasingly used in the
secondary effluent from wastewater treatment plant as well as been examined for whey protein
fractionation on a laboratory scale [1-3], bioseparation [4-6] and cation macromolecule recovery in
bioprocess [7] due to its superior separation performance. However, such applications appear to have
several drawbacks in treating the feed waters containing effluent organic matters (EfOMs) such as
protein, polysaccharides, humic and fulvic acids [8]. Protein fouling in UF membrane is attributed to
proteins adsorption/deposition on the membrane surface or inside the pores via pore blocking and/or
cake layer formation, causing significant flux decline, increased chemical cleaning, and membrane
replacement [9].
The separation performance of UF/NF membrane can be significantly influenced by the solution
chemistry such as ionic strength, concentration and solution pH [10-12]. It is mainly due to the fact
that ionic strength and solution pH strongly influence membrane surface and protein interaction,
protein, and protein interaction by shifting isoelectric point of protein and charge of the protein [13].
Thus, ionic compounds contained in wastewater lead to accelerating membrane fouling when the
foulants are oppositely charged with the membrane and intensive chemical cleaning [14]. There have
been numerous studies examining the effects of solution chemistry (ionic strength and solution pH)
and model foulant in separation performance of UF/NF membrane [15-17]. Conventional polymeric
membranes experience severe fouling by negatively charged protein at near to its isoelectric point (IEP)
due to the decreased electrostatic repulsion between foulant-foulant, and foulant-membrane surface
[18, 19]. Further, high ionic strength can induce membrane fouling due to electrical double layer
147
compression effect on the foulants [20, 21]. Such effects of solution chemistry on protein adsorption
on the membrane surface and pores are governed by electrostatic interaction/repulsion [22], indicating
that the protein fouling in the UF membrane could be reduced by controlling the electrostatic
interaction [23, 24]. Like negatively charged polymeric membranes, positively charged conventional
membrane showed reduced fouling to the same charged proteins (Lys) mostly due to the less
adsorption on the membrane surface via electrostatic repulsion [25].
Recent progress on carbon nanotubes (CNTs) membrane fouling reveals protein-fouling resistance due
to its modified surface properties [26-28]. Celik et al. [29] also studied the protein fouling behavior in
CNTs composite membrane at different solution pHs and membrane properties (such as
hydrophilicity). The challenge in protein fouling stimulated further development on positively charged
nanohybrid UF membranes having antifouling properties due to electrostatic interaction with foulants.
For example, poly (arylene ether sulfone) (PAES) block copolymers with carboxyl (COOH)
functionalized CNTs incorporated UF hybrid membrane exhibited relatively low protein adsorption
and irreversible protein fouling by increasing hydration of membrane surface and reducing the net
charge density [30]. Further, currently fabricated positively charged graphene oxide (GO) nanosheet
UF membrane presented enhanced antifouling properties due to the combined effect of hydrophilicity,
surface charge and morphology of the membrane at different pH [31]. Recently fabricated CNTs/PANI
composite membrane shows superior performance in natural organic matter (NOM) removal [32]. This
membrane has unique physicochemical characteristics by incorporating the MWCNTs/PANI complex
to UF membrane. It is positively charged below pH 9.2 (IEP) and removes NOM via electrostatic
interaction with negatively charged NOM.
As such there have been numerous studies on protein fouling behavior in the positively charged
membranes [33-35]. However, in-depth study on the influence of solution chemistry on protein fouling
mechanism in a positively charged nanocomposite membrane has not reported in the literature.
Therefore, this study is focused on the fouling behavior caused by positively/negatively charged
148
protein on the positively charged membrane (CNT/PANI membrane) under different ionic strengths
and solution pHs. It would give useful information to the wide application of CNT engineered
membrane for wastewater reclamation and bioseparation.
This study aims to i) investigate the influence of feed solution chemistry on protein fouling behavior
in the positively charged membrane: flux pattern under different pHs and ionic strengths, and ii)
evaluate protein fouling mechanism using filtration models.
6.2.
Materials and methods
6.2.1
Materials
6.2.1.1 Model protein foulants and chemicals
Bovine serum albumin (BSA) and Lysozyme (Lys) were used as model protein foulants. BSA (powder,
Sigma-Aldrich) was used as the negatively charged model protein foulant. Lys (powder, SigmaAldrich) was selected as the positively charged protein foulant. The molecular weight (MW) of BSA
and Lys is reported to be 67 kDa and14.3 kDa, respectively [12, 19]. The IEP of BSA and Lys 4.7 and
10.4, respectively. Each stock solution was prepared at a concentration of 1 g/L and stored in a glass
bottle at 4 °C prior to use. The pH was adjusted using 1M-HCl and 1M-NaOH. The ionic strength of
solution was adjusted by adding different concentrations of NaCl (VWR, ACS reagent). All stock
solutions were prepared by ultrapure water (conductivity = 0.055 µS/cm).
6.2.1.2 Membranes
In this study, a multiwall CNT (MWCNT) enhanced membrane was tested and compared with a
commercial polyethersulfone (PES) membrane (Nanostone Flat Sheet Membrane, PES UF OPE10HR).
Fabrication procedure of 0.5 wt% MWCNTs enhanced membrane was adopted from Chapter 3. Prior
to use, all membranes were soaked in deionized (DI) water for 1-2 d to remove any impurities.
Characteristics of membranes used in this study are summarized in Table 6.1.
149
Table 6.1 Information of the membranes used in this chapter.
UPW* permeability
Pore size by
MWCO***
Contact angle
(LMH**/bar)
BET (nm)
(kDa)
(°)
MWCNT
490.8 ± 46.3
4.4-4.6
12
45.4 ± 0.1
9.2
PES
85.5 ± 8.5
-
10
57.6 ± 0.4
2.9
Isoelectric pH
*
UPW: Ultrapure water, **LMH: L/m2h; and ***MWCO: Molecular weight cut-off.
6.2.2
Fouling experiments
The protein fouling experiments were carried out in a dead-end filtration mode under 2 bar for 12 h.
The effective membrane area was 0.00146 m2. Constant pressure was applied during the filtration test
using a compressed nitrogen gas cylinder. The flux data were continuously acquired using an electronic
balance connected to a computer. Each feed solution was prepared under the following conditions:
-
Concentration: BSA = 10 mg/L, Lys = 10 mg/L, and BSA-Lys mixture = 20 mg/L (10mg/L BSA
+ 10mg/L Lys)
-
pH: 4.7, 7.0, and 10.4
-
Ionic strength: NaCl in the concentration of 0.001, 0.01, and 0.1 M
6.2.3
Membrane characterization
The zeta potential of membrane surface was measured by using the streaming potential technique
(Anton Paar electrokinetic analyzer). The zeta potential measurements were carried out as a function
of pH (4.0-11.0) and ionic strength (with 0.001, 0.005, and 0.010 M KCl at pH 7.0).
6.2.4
Membrane fouling model
Protein fouling mechanism of the MWCNT membrane by protein was examined. Three types of
membrane fouling model were used to explain the flux decline with protein deposition during
150
membrane filtration (standard blocking, and cake layer formation) [36, 37]. The equations of each
fouling models are summarized in Table 6.2.
Table 6.2 Fouling mechanism equations.
Fouling mechanisms
Equations
𝑉
𝑑( )
𝐴 = 𝐽 − 𝑘 (𝑉 )
0
𝑏
𝑑𝑡
𝐴
𝑑𝑡
1
= + 𝑘𝑖 𝑡
𝑉
𝐽
0
𝑑( )
𝐴
Completed blocking
Intermediate blocking
𝑡
1
𝑘𝑐 𝑉
= +
( )
𝑉/𝐴 𝐽0
2 𝐴
Cake layer formation
In these equations, t is the filtration time, V is the cumulative permeate volume, A is the effective
filtration area (0.00146 m2), J0 is the initial flux, kb (1/s), ks (1/m), ki (1/m) and kc (s/m2) are the
coefficients of complete blocking, standard blocking, intermediate blocking, and cake filtration models,
respectively. From these equations, three data plots have been proposed where the linearity of the
filtration data in the plot of d(V/A)/dt vs V/A, t/(V/A) vs t, dt/d(V/A) vs t and t/(V/A) vs V/A offers
proof of the complete blocking, intermediate blocking, and cake filtration model, respectively. Two
pore blocking models (complete and intermediate) were applied at the initial stage of filtration (~2 h)
because those two blocking models are expected to be dominant in both the MWCNT and PES-UF
membranes with bigger pore size than the two model proteins. The initial stage of filtration was
determined by the time when the initial permeate flux was declined by more than half and showed the
highest value in all permeate flux profiles. At the later stage (2 – 12 h), cake layer filtration was applied.
The dominant filtration model was determined by comparing correlation coefficient (R2). The fouling
potential of both membranes at different solution chemistry was examined by comparing the fouling
coefficients (Kb, Ki and Kc), which were derived from those equations.
151
6.3. Results and discussion
6.3.1 Effects of solution chemistry on surface charge of the membranes
It is essential to study the surface charge of the membrane for an understanding of protein fouling
mechanisms on membranes. Zeta potentials of both MWCNT and PES membranes were analyzed by
streaming potential measurement to evaluate the effect of solution chemistry on the membrane surface
charge. Fig 6.1(a) shows the zeta potential of the MWCNT membrane, which varies from positive to
negative value depending on the pH. The zeta potential of the MWCNTs membrane was shown to be
highly positive (17.12 mV) under acidic condition (pH 4.7). Then, it slightly decreased to 10.9 mV
when pH increased to 7.0. However, the membrane shifted to be a negative charge (-25 mV) at pH
10.4. Thus, it is noted that the isoelectric point (IEP) of the MWCNT membrane was 9.2 as shown in
Table 6.1. On the other hand, the zeta potential of the PES-UF membrane remained in a negative value
over the entire pH range, although its negative value slightly decreased (from -35 to -20 mV) as the
pH decreased (pH 10.4 to 4.7). In addition to pH, ionic strength reduced the zeta potential of the
MWCNT membrane as shown in Fig 6.1(b). The zeta potential of the MWCNT membrane
dramatically dropped to the nearly zero (0) mV as the electrolyte (KCl) concentration increased to 0.01
M, indicating that the surface charge of the MWCNT membrane was strongly affected by ionic strength.
Electrolyte ions could reduce the membrane surface charge density via charge neutralization. In
contrast, there was no significant change in the negatively charged PES-UF membrane with the
increase in electrolyte concentration. Meanwhile, the zeta potential at 0.1 M KCl could not be
measured due to the limit of the equipment used. It is assumed that the zeta potential at 0.1 M would
be almost zero (0) or near to negative.
Overall, solution chemistry, such as pH and ionic strength of feed solution, significantly influenced
the surface charge of the MWCNT membrane while PES-UF membrane maintained being negatively
charged regardless of solution chemistry.
152
(a)
(b)
Fig 6.1 Zeta potential of MWCNT membrane and PES membrane (a) at different pHs (4.7, 7.0 and
10.4) and (b) different ionic strengths by adding different concentrations of KCl (background
electrolyte) (0.001, 0.005 and 0.010 M) at pH 7.0.
153
6.3.2 Effects of pH on permeate flux with single proteins
6.3.2.1 Permeate flux behavior in positively charged protein (Lys)
6.3.2.1.1
MWCNT membrane
Fig 6.2 shows the flux decline of MWCNT and PES-UF membrane filtration with Lys under different
pH levels. Overall, the MWCNT membrane was found to be sensitive to the feed properties both at
the beginning of filtration and the corresponding filtration process. A rapid permeate flux decline was
observed at the initial stage of the MWCNT membrane filtration with Lys, which is associated with
the membrane pore size. In addition, flux pattern of the MWCNT membrane was considerably affected
by pH. It is noted that the IEP of Lys is pH 10.4, which indicates that the charge of Lys is neutralized
at pH 10.4. As can be seen in Fig 6.2(a), permeate flux was shown in the following order: pH 4.7 >
pH 10.4 > pH 7.0. At pH 4.7, the initial permeate flux of MWCNT membrane was higher, and it was
more than six times (from 50 to 340 LMH/bar). The high permeate flux at pH 4.7 in MWCNT
membrane filtration may be due to less Lys adsorption on the membrane surface by the increased
electrostatic repulsion between more positively charged membrane surface and positively charged Lys.
Since MW of Lys (14.3kDa) was slightly larger than the pore size (12kDa) of the MWCNT membrane,
feed solution with Lys had a possibility to block the pore of the membrane affecting the water transport
through the membrane. In particular, at pH 4.7, the MWCNT membrane becomes more positively
charged (17.12 mV) as shown in Fig 6.1(a), resulting in an increased electrostatic repulsion between
membrane and Lys.
154
(a) Lys filtration with MWCNT membrane
(b) Lys filtration with PES-UF membrane
(c) BSA filtration with MWCNT membrane
(d) BSA filtration with PES-UF membrane
Fig 6.2 Comparison of permeate flux pattern of MWCNT and PES-UF membrane filtration with
positively charged protein (Lys) and negatively charged protein (BSA) under different pHs (4.7, 7.0
and 10.4).
The higher flux at pH 10.4 than pH 7.0 can be explained by the foulant intermolecular interaction. As
mentioned in the earlier section, the IEP is the pH at which a particular molecule carries no
net electrical charge. The net charge on the molecule is affected by pH of its surrounding environment
as the solution and can become more positively or negatively charged due to the gain or loss
of protons (H+). The IEP value can affect the solubility of a molecule at a given pH. Such molecules
have a minimum solubility in water or salt solutions at the pH that corresponds to their IEP and often
155
precipitates out of solution. Hence, the weakened electrostatic repulsion between the membrane and
Lys at pH 10.4 led to a higher flux of MWCNT membrane compared to at pH 7.0. It is probably due
to the more aggregation between Lys molecules. It indicates that intermolecular interaction (Lys-Lys)
could play a major role in permeate flux pattern in addition to the interaction between membrane and
foulant.
Table 6.3 Comparison of the correlation coefficient (R2) in the MWCNT membrane filtration with
proteins under different solution chemistry at (a) the initial stage of filtration (2h) and (b) final stage
of filtration (2 to 12h).
pH
Lys (10
Ion strength (NaCl) (M)
4.7
7.0
10.4
0.001
0.01
0.1
*
0.8694
0.9162
0.9910
0.8836
0.8554
0.9202
**
0.9727
0.9644
0.9457
0.9129
0.9634
0.9219
***
0.9983
0.9849
0.9978
0.8844
0.9954
0.9973
*
0.9369
0.9205
0.9024
0.9162
0.9134
0.9569
**
0.9664
0.9604
0.9349
0.9620
0.9520
0.9874
***
0.9927
0.9830
0.9471
0.9457
0.9760
0.9990
*
0.9213
0.9268
0.9273
0.9406
0.9462
0.9631
**
0.9409
0.9571
0.9498
0.9896
0.9926
0.9871
***
0.9990
0.9565
0.9930
0.9981
0.9847
0.9951
(a)
mg/L)
(b)
BSA
(a)
(10
mg/L)
(b)
Lys +
BSA
(a)
(20
mg/L)
(b)
* R2 of complete pore blocking model, ** R2 of intermediate pore blocking model, *** R2 of cake layer
filtration model.
156
Overall, complete pore blocking and intermediate pore blocking at the initial stage were dominant
except for the pH 4.7 (The initial stage of filtration was determined to be 2 hours because severe
permeate flux decline in most cases occurred in 2 hours, and R2 did not vary in 2 ± 0.5 hours of
filtration). Table 6.3 shows that the correlation coefficients (R2) of complete and intermediate pore
blockings are above 0.9 at pH 7.0 and 10.4. At the later stage, cake layer filtration was dominant, as
seen in Table 6.3. The fouling potentials at different pH are presented in Table 6.4. For 10 mg/L Lys
filtration, complete blocking and intermediate blocking appeared simultaneously at the initial filtration
stage (2 h). Then, cake layer formation developed at the later filtration stage (2 h – 12 h). As seen in
Table 6.4, Kb, Ki and Kc were high at pH 7.0 while they were low at pH 4.7. It indicates that severe
pore blocking at the initial stage and cake formation at the later stage developed on MWCNT
membrane at pH 7.0. Severe permeate flux decline with MWCNT membrane at pH 7.0 appeared
possibly due to the pore blocking at the initial filtration stage (2h) and developing cake layer form at
the later stage. However, decreases in Kb, Ki and Kc at pH 4.7 indicate that pore blocking and protein
adsorption was alleviated by increased electrostatic repulsion. Further, a decrease of Kc at the final
stage of filtration shows that protein deposition rate slowed down as filtration proceeds. It is noted that
at pH 4.7, both membrane and Lys are strongly positively charged; thus, there is relatively strong
electrostatic repulsion between the membrane and foulant. For this reason, at pH 4.7, the highest flux
was observed due to less Lys adsorption by strong electrostatic repulsion (Fig 6.2). It is good
agreement with the effect of electrostatic interaction on the fouling behavior. Fouling was also
alleviated at the IEP of Lys (pH 10.4), but it was higher than that at pH 4.7 (Table 6.4). At pH 10.4,
the membrane became negatively charged while Lys was neutralized. It may induce adsorption of Lys
on the membrane surface by removing electrostatic repulsion, corresponds to slight increases of Kb,i
and Kc compared to that at pH 10.4. Previous studies on protein fouling reported that severe fouling
was observed at IEP of model protein due to decreased electrostatic interaction [38]. However, the
MWCNT membrane showed low fouling potential at IEP of Lys. Such a discrepancy may be mainly
157
due to the change in zeta potential of the MWCNT membrane at IEP of the model protein (pH 10.4).
More interestingly, a discrepancy occurs between the fouling coefficient and flux patterns in Lys
filtration. Severe fouling appeared at pH 7.0 at which the membrane surface and Lys are still oppositely
charged. Due to an electrostatic repulsion, its fouling should be reduced compared to that at pH 10.4
at which the electrostatic repulsion disappears due to neutrally charged Lys. It indicates that the fouling
behavior of Lys filtration in the MWCNT membrane is also controlled by intermolecular interaction
(Lys-Lys). Severe foulant aggregation occurs at its IEP (pH 10.4), resulting in the enhanced permeate
flux due to its insolubility compared to at pH 7.0.
Table 6.4 Comparison of fouling potential in the MWCNT membrane filtration with proteins under
different solution chemistry at (a) the initial stage of filtration (2h) and (b) final stage of filtration (2
to 12h)
pH
Lys (10
4.7
7.0
10.4
0.001
0.01
0.1
Kb
1.00×10-4
5.00×10-5
5.00×10-5
1.00×10-4
1.00×10-3
8.00×10-5
Ki
1.63
10.35
4.02
0.66
0.99
3.83
Kc
1.30×104
3.70×105
9.43×104
3.6×103
7.18×104
5.68×104
Kb
1.00×10-4
4.00×10-5
1.00×10-4
1.00×10-4
1.00×10-4
1.00×10-4
Ki
0.53
4.10
0.58
0.93
1.40
1.75
Kc
5.13×103
7.89×104
1.60×103
3.63×103
7.69×103
2.08×104
Kb
1.00×10-5
6.00×10-5
1.00×10-5
3.00×10-5
4.00×10-5
8.00×10-5
Ki
19.50
1.29
20.59
14.64
11.80
2.15
Kc
2.15×106
9.29×103
1.95×106
9.42×105
6.56×105
3.76×104
(a)
mg/L)
(b)
BSA
(a)
(10
mg/L)
(b)
Lys +
BSA
(a)
(20
mg/L)
(b)
158
Based on the synergetic effect of membrane-foulant interaction and foulant-foulant interaction (LysLys), it is obvious that the permeate flux, at pH 7.0 at which both two interactions are dominant, was
the lowest, and the fouling was greater.
6.3.2.1.2
PES-UF membrane
Permeate flux in the PES-UF membrane filtration was almost similar even at different pHs. However,
there was no flux difference by pH level from 4.7 to 10.4 in the PES-UF membrane (the initial flux
was around 70 LMH/bar). The dominant filtration mechanisms at the initial stage were found to
complete and intermediate pore blockings (Table 6.5). There was no noticeable change in Kb and Ki,
indicating that adsorption rate of Lys particle was not affected by different pH. It may be interpreted
that the electrostatic interaction between the membrane and Lys particle may not be the critical factor.
Meanwhile, the relatively higher cake layer formation was observed at pH 7.0 while it was slightly
reduced at IEP of Lys (pH 10.4). It is mostly due to the increase in the deposition of the neutrally
charged Lys at the later stage of filtration. Consequently, the foulant-foulant interaction may have an
important role in fouling in the PES-UF membrane filtration.
159
Table 6.5 Comparison of the correlation coefficient (R2) in the PES membrane filtration with proteins
under different solution chemistry at (a) the initial stage of filtration (2h) and (b) final stage of filtration
(2 to 12h).
pH
Lys (10
4.7
7.0
10.4
0.001
0.01
0.1
*
0.9763
0.9871
0.9681
0.9888
0.9673
0.9981
**
0.9885
0.9954
0.9845
0.9855
0.9789
0.9890
***
0.9995
0.9919
0.9993
0.9563
0.9946
0.9849
*
0.9488
0.9759
0.9536
0.9706
0.9684
0.9524
**
0.9928
0.9962
0.9881
0.9965
0.9937
0.9924
***
0.9947
0.9950
0.9809
0.9925
0.9935
0.9947
*
0.9183
09681
0.9782
0.9159
0.9290
0.9318
**
0.9859
0.9858
0.9969
0.9817
0.9913
0.9844
***
0.9990
0.9980
0.9950
0.9991
0.9943
0.9993
(a)
mg/L)
(b)
BSA
(a)
(10
mg/L)
(b)
Lys +
BSA
(a)
(20
mg/L)
(b)
* R2 of complete pore blocking model, ** R2 of intermediate pore blocking model, *** R2 of cake layer
filtration model.
6.3.2.2 Permeate flux behavior in negatively charged protein (BSA)
6.3.2.2.1
MWCNT membrane
In the case of negatively charged protein (BSA), the BSA adsorption on the surface of membranes
with a much smaller pore size (MW of BSA = 67kDa vs. membrane pore size = 10 kDa) caused a rapid
decline in permeate flux of the MWCNT membrane at the initial stage of filtration (Fig 6.2(c)). Based
on the flux decline, fouling by BSA on MWCNT membrane was less severe compared to that by Lys.
BSA adsorption caused severe flux decline at around pH 7.0 where both the membrane and BSA are
160
oppositely charged. However, higher fluxes were observed at pH 10.4 where the membrane is
negatively charged and at the IEP of BSA (pH 4.7). As shown in Fig 6.2(c), the permeate flux in
MWCNT membrane was different depending on pH, and it had the following order: pH 10.4 > pH 4.7
> pH 7.0. Such a result could be interpreted due to the shift of electrostatic interaction between BSA
and membrane. At below pH 10.4, the surface charge of the MWCNTs membrane shifted to negative
charge while BSA was still negatively charged. Thus, BSA was less adsorbed on the membrane surface
due to the electrostatic repulsion between BSA and membrane, indicating that BSA in feed solution at
pH 10.4 did not seriously hinder the water transport. The comparable flux was observed at pH 4.7. It
may be due to the zero charged BSA at its IEP. Similar as Lys, weakened interactions between BSA
molecules by neutralized charge resulted in less adsorption of BSA-BSA foulant on the membrane
surface. BSA molecules were aggregated each other, rather than adsorption of BSA onto the membrane.
Both complete and intermediate blockings were simultaneously dominant at the initial stage (Table
6.3). As shown in Table 6.4, fouling potentials were relatively weak, compared to those in Lys
filtration. The severe fouling was observed at pH 7.0 while it was lessened at pH 4.7 and 10.4,
corresponding to the flux pattern (Fig 6.2). It is probably due to the effect of pH on the electrostatic
interaction between both charged membrane surface and BSA. At pH 7.0, as both membrane surface
and BSA are oppositely charged, and it resulted in an increase in electrostatic interaction between the
membrane and foulants. It induces the foulant adsorption on the membrane surface, leading to
membrane fouling. However, at IEP of BSA (pH 4.7), due to the weakened electrostatic interaction,
neutrally charged BSA became to be less adsorbed on the positively charged membrane surface
compared at pH 7.0, leading to a slower fouling rate. Further, at pH 10.4, both membrane and BSA
have the same charge (in negative), hindering BSA adsorption on the membrane surface via
electrostatic repulsion.
161
6.3.2.2.2
PES-UF membrane
Overall, Severer flux decline was observed with BSA than that with Lys. Although permeate flux in
the PES-UF membrane filtration seemed to be affected by pH, its influence was much smaller than the
MWCNT membrane. Permeate flux in the PES-UF membrane filtration at pH 10.4 was slightly higher
than at other pH values probably due to the strong electrostatic repulsion between membrane and BSA
(Fig 6.2(d)). Permeate flux at IEP (pH 4.7) of BSA was lower than that at pH 7.0 and 10.4 most
probably due to deposition of BSA-BSA aggregates formed by surface charge neutralization. It showed
that negatively charged PES-UF membrane was dominated by complete and intermediate pore
blocking at the initial stage (Table 6.5). As shown in Table 6.6, Ki and Kc at pH 4.7 was the highest
(6.25 and 3.09×105, respectively), indicating that rapid flux decline at the initial stage and building up
the cake layer at the later stage occurred at the IEP of BSA. It is mostly due to the low electrostatic
repulsion between BSA and the membrane. It is in agreement with the previous studies [19, 21].
162
Table 6.6 Comparison of fouling potential in the PES-UF membrane filtration with proteins under the
different solution chemistry (a) at the initial stage (2h) and (b) later stage of filtration
pH
Lys (10
4.7
7.0
10.4
0.001
0.01
0.1
Kb
3.00×10-5
3.00×10-5
3.00×10-5
3.00×10-5
4.00×10-5
8.00×10-5
Ki
2.03
2.02
2.11
2.53
1.03
3.52
Kc
5.43×104
8.99×104
5.07×104
1.51×105
1.32×104
2.19×105
Kb
3.00×10-5
3.00×10-5
4.00×10-5
4.00×10-5
4.00×10-5
4.00×10-5
Ki
6.25
4.18
4.00
3.62
3.90
4.80
Kc
3.09×105
1.34×105
8.05×104
8.85×104
1.05×105
1.43×105
Kb
3.00×10-5
2.00×10-5
2.00×10-5
2.00×10-5
3.00×10-5
3.00×10-5
Ki
2.15
13.02
5.83
14.47
9.64
6.79
Kc
9.82×105
5.19×105
4.18×105
1.08×106
5.64×105
2.89×105
(a)
mg/L)
(b)
BSA
(a)
(10
mg/L)
(b)
Lys +
BSA
(a)
(20
mg/L)
(b)
Based on the flux decline and fouling potential, the PES-UF membrane fouling is more affected by
electrostatic repulsion with negatively charged protein (BSA).
163
6.3.3
Effect of ionic strength on permeate flux with single protein
6.3.3.1 Permeate flux pattern with positively charged protein (Lys)
6.3.3.1.1
MWCNT membrane
With positively charged protein (Lys), the permeate flux in the MWCNT membrane was also
significantly influenced by ionic strength (in terms of NaCl) at pH 7.0. It is noted that at pH 7.0,
MWCNT membrane charged positively and the charge of membrane gradually decrease to nearly zero
as ionic strength decreases to 0.01 M. As can be seen in Fig 6.3(a), the permeate flux of MWCNT
membrane when 0.001 and 0.01 M of NaCl were added to the solution, was much higher than zero
ionic strength (without NaCl addition), and it was rather slightly increased in the presence of 0.1M of
NaCl. Slightly lower permeate flux at higher ion strength can be explained by the concentration
polarization near to membrane. Overall, the low flux of the MWCNT membrane at pH 7.0 (Fig 6.2)
was overcome in the presence of ionic strength (0.001–0.1 M). Such a flux enhancement may appear
due to the charge shielding effect of chloride ion (Cl-) to the positive charged Lys (Lys+). It probably
prevented deposition of Lys on the pores and surface of MWCNT membrane.
At 0.001 and 0.01 M NaCl, intermediate pore blocking was dominant, and both complete and
intermediate pore blockings became simultaneously dominant at 0.1 M NaCl at the initial stage as
presented in Table 6.3. Severe Lys fouling at pH 7.0 was reduced by the addition of ionic strength. As
shown in Table 6.4, the coefficients (Kb, Ki and Kc) seemed to increase when NaCl increased from
0.001 M to 0.1 M, but overall, much lower than at pH 7.0 (without NaCl). Such a trend is consistent
with the permeate flux pattern (Fig 6.3(a)). The flux had the following order: 0.001 M >0.01 M>0.1
M >0 M.
It may appear probably due to the effect of ionic strength on Lys particles. Due to the presence of ionic
strength, Lys-Lys interaction decreased via increasing repulsion hydrating forces between Lys-Lys
[10]. Further, Cl- may interfere the interaction between the membrane and foulant by weakening
164
surface charge of both membrane and foulant, leading to fouling mitigation. It demonstrates that ionic
strength mainly lessened fouling potential via charge shielding effect on the model proteins. The
increased fouling potential at the increasing ionic strength can be explained according to the interaction
between both neutralized foulants and the membrane surface. Under increasing ionic strength, both the
membrane and Lys particle became neutralized, leading to the increase in Lys deposition on the
membrane surface.
6.3.3.2 PES-UF membrane
In contrast to the MWCNT membrane, a different trend was observed in flux behavior of the PES-UF
membrane by Lys fouling. The flux without ionic strength at pH 7.0 increased when 0.01 M of NaCl
was added, but it showed a decreased flux with 0.001 and 0.1 M of NaCl. It indicates that there
presented the critical point to neutralize the charge of Lys and PES-UF membrane. Ionic strength (NaCl)
affected the charge of both Lys and membrane, especially negatively charged PES-UF membrane.
However, at relatively high or low ionic strength, residual ions (Na+ and Cl-) after their reaction with
Lys and membrane affected again to them. Thus, it is assumed that flux decline behavior in the PESUF membrane is governed by complex mechanism due to the altered surface charge of both membrane
and protein at different ionic strengths, which affects the interaction between protein and the membrane.
Like as the MWCNT membrane, both complete and intermediate pore blockings were simultaneously
dominant at the initial stage (Table 6.5). As shown in Table 6.6, Kc increased as ionic strength
increased, indicating that cake layer formation developed during the final stage of filtration (2h-24h).
For this reason, ionic strength may induce the building up of deposition layer of Lys at the final stage
of filtration.
165
6.3.3.2 Permeate flux behavior with negatively charged protein (BSA)
6.3.3.2.1
MWCNT membrane
Overall, ionic strength seemed to contribute considerably to permeate flux enhancement in the filtering
of BSA with MWCNT membrane, but no significant difference was observed at the initial stage of
filtration when the concentration of NaCl increased from 0.001 to 0.1 M. Then, there were minimal
changes as increasing ionic strength at the final stage of filtration. As shown in Fig 6.3(c), permeate
flux increased three times at the initial stage (2 h), continuing to have a stable flux for the rest filtration
period (100-130 LMH/bar) when 0.001 – 0.1 M NaCl was added.
At the initial stage of filtration, both complete and intermediate pore blockings were simultaneously
dominant (Table 6.3). The presence of ionic strength seemed to alleviate pore blocking and cake layer
formation on the MWCNT membrane in the filtration of BSA, but the fouling potential induced at the
increase in ionic strength. As shown in Table 6.4, Ki at 0 M NaCl (4.10) decreased to 0.93 at 0.001 M
NaCl, and then gradually increased (to 1.75) as the ionic strength increased to 0.1 M. Similarly, at the
later stage, cake formation seemed to reduce in the presence of ionic strength. However, increase in
ionic strength led to the building up the layer of BSA via protein adsorption, corresponding to the
increase in Kc as NaCl increased to 0.1 M. Overall, fouling was reduced mainly due to the charge
shielding effects on the negatively charged BSA, resulting in a decrease of electrostatic interaction
with positively charged MWCNT membrane. However, an increase in ionic strength lessened the
charge of both membrane surface and BSA particle. Thus, it induced the neutralized BSA deposition
on the less positively membrane surface at increasing ionic strength. Such trend corresponds to the
permeate order: 0 M NaCl <<0.1 M < 0.01 M< 0.001M.
166
Fig 6.3(d) shows that the permeate flux increased by nearly 30 % at the initial stage (2 h) as ionic
strength increased, but it remained nearly same over the whole filtration period (12 h). It indicates that
ionic strength did not significantly influence BSA fouling on the PES-UF membrane surface at the
final stage of filtration. It corresponds to the fouling potential as shown in Table 6.6. There was no
noticeable change in Kc as ionic strength increased. Both complete and intermediate pore blockings
were simultaneously dominant at the initial stage (Table 6.5). However, the intermediate pore blocking
was promoted as the ionic strength increased, corresponding to the slightly increasing Ki. It may be
due to the reduced electrostatic repulsion via charge shielding effect as reported in the literature [21].
The result is in agreement with previous research presenting that increase in ionic strength can enhance
fouling due to a decrease in electrostatic repulsion via hydration repulsion forces between BSA-BSA
and membrane-BSA [10].
167
(a) Lys filtration with MWCNT membrane
(b) Lys filtration with PES-UF membrane
(c) BSA filtration with MWCNT membrane
(d) BSA filtration with PES-UF membrane
Fig 6.3 Comparison of permeability pattern with positively charged protein (Lys) and negatively
charged protein (BSA) fouling under different ionic strength (concentration, monovalent/divalent ionic
strength) (0.001, 0.01, 0.1 M NaCl).
168
6.3.4 Effect of solution chemistry on permeate flux with binary mixture
6.3.4.1 Permeate flux pattern as a function of pH
6.3.4.1.1
MWCNT membrane
Fig 6.4 shows the permeability MWCNT membrane in filtering mixture of two positively/negatively
charged proteins (Lys + BSA). Overall, extremely low flux pattern was observed in the filtration with
MWCNT membrane compared to that with a single protein. Interestingly, an opposite trend - the
lowest flux in the vicinity of nearly IEP of each model protein was observed in permeate flux under
protein mixture filtration test, compared to those in the single protein filtration. As seen in Fig 6.4(a),
the permeate flux exhibited a considerable decrease in flux pattern (10 LMH/bar) at pH 4.7 and 10.4
compared to that (80 LMH/bar) at pH 7.0.
At pH 7.0, complete and intermediate pore blockings appeared to be simultaneously dominant at the
initial stage while intermediate blocking played the main role in fouling at pH 4.7 and 10.4 (Table
6.3). As shown in Table 6.4, the fouling potential was low at pH 7.0 while severe intermediate pore
blocking and cake formation at the initial stage occurred at IEP of both Lys (pH 10.4) and BSA (pH
4.7). It corresponds to the permeate flux pattern (Fig 6.4). It is suggested that IEP of both charged
proteins may contribute to the permeate flux pattern, but it led to the opposite trend probably due to
the intramolecular interactions (Lys-BSA) and that with the membrane surfaces. Apart from
contributing to membrane transport via electrostatic interaction (repulsion) between single proteins
and membrane surface, mixing of oppositely charged proteins seemed to generate more complicated
intermolecular interactions between macro solutes (Lys-BSA, Lys-Lys, BSA-BSA) and charged
membrane surface. Based on the fouling and permeate flux behavior at different pH, protein mixture
fouling mechanism tends to be mainly governed by the intermolecular interaction (Lys-BSA). At pH
7.0, Lys-BSA mixture may agglomerate by intermolecular interaction in the feed solution. It decreased
the deposition of the binary mixture on the membrane surface, leading to fouling alleviation, which is
169
comparable to the BSA filtration. However, at pH 4.7 and 10.4 where each of proteins is neutrally
charged, the proteins deposited on the membrane, then induced the deposition of the counter charged
proteins by bridge effect.
6.3.4.1.2
PES-UF membrane
In contrast, negatively charged PES-UF membrane showed the lowest permeate flux (20 LMH/bar) at
pH 7.0 while the permeate flux was improved at pH 10.4 and 4.7 (to 24 and 43 LMH/bar, respectively)
(Fig 6.4(b)). Complete and intermediate pore blockings were simultaneously dominant at the initial
stage (Table 6.5). Like as severe flux decline at the initial stage at pH 7.0, high potential intermediate
pore blocking appeared at the initial stage. It indicates that oppositely charged foulants were easily
adsorbed on the membrane surface, leading to the severe pore blocking at the initial stage. The fouling
at the initial stage was significantly alleviated at pH 4.7 and 10.4, at which each protein is neutralized.
It means that the interaction between the neutralized protein, and charge membrane and counter protein
may reduce pore blocking at the initial stage, but it promoted cake layer formation at the later stage
via bridge effect, particularly at pH 4.7 where the membrane and model protein are oppositely charged.
6.3.4.2 Permeate flux pattern as a function of ionic strength
6.3.4.2.1
MWCNT membrane
Similar to the effect of pH, ionic strength also caused extremely low flux pattern in the filtering of
protein mixture with MWCNT membrane. As seen in Fig 6.4(c), permeate flux of the MWCNT
membrane remained at 80 LMH/bar for 12 h in the absence of NaCl. However, at the low concentration
(0.001-0.01 M), flux decreased significantly to 20-27 LMH/bar, and then was recovered to that without
NaCl. Such a trend in the presence of binary mixture is opposite to that of a single protein. It may be
due to the charge shielding effect of NaCl ion, which probably had a negative impact on flux decline
behavior. The intermolecular interaction (Lys-BSA, Lys-Lys and BSA-BSA) may contribute to protein
170
aggregation on the membrane surface, leading to an extremely low permeate flux behavior in the
MWCNT membrane.
In mixed protein (binary) filtration at different ionic strength, complete and intermediate pore
blockings appeared simultaneously dominant at the initial stage (Table 6.3). The addition of 0.001 M
NaCl led to severe pore blocking and cake layer formation by reducing Lys-BSA interaction. However,
the membrane fouling was reduced gradually when ionic strength increased to 0.1 M. It is probably
due to the charge shielding effect of NaCl ion on the two oppositely charged proteins and the
membrane surface, and it reduced the membrane-foulant interaction. Such effects may induce protein
adsorption at a low concentration of NaCl and then tends to prevent deposition of binary protein
particles on the membrane surface via charge shielding effect.
6.3.4.2.2
PES-UF membrane
An opposite trend was observed in the filtration of the PES-UF membrane with protein mixture
solution compared with that of the MWCNT membrane. Fig 6.4(d) shows that there seemed to have a
slight flux decline in the presence of 0.001 M NaCl at the initial filtration period. However, as ionic
strength increased up to 0.1 M, the permeate flux increased nearly 1.5-fold (from 17 to 27 LMH/bar)
as a function of ionic strength. The permeate flux with protein mixture seemed to decrease in the
presence of 0.001 M NaCl, but it gradually increased as the ionic strength increased to 0.1 M. Further,
severe flux decline at the initial stage was alleviated by the increase in ionic strength, corresponding
to decrease in fouling potential (Ki) as a function of ionic strength. It indicates that combined effect of
membrane-foulant and foulant-foulant interaction which resulting in a serious fouling at pH 7.0 was
reduced by ionic strength via charge shielding effect on the foulant and the membranes. These results
are in agreement with a previous study that increase in ionic strength can enhance the permeate flux
by suppressing Lys-BSA interaction [11]. Further, complete and intermediate pore blockings were
simultaneously dominant at the initial stage (Table 6.5).
171
Overall, a linear trend in permeate flux was observed as a function of ionic strength, compared to the
irregularly increased flux pattern under single protein filtration test (Lys and BSA are individually
added) (Fig 6.3(b) and 3(d)). With protein mixture, the charge-shielding effect seemed to contribute
to the flux enhancement of the PES-UF membrane in the binary mixture filtration presumably due to
the different mechanisms of negatively charged membrane.
(a) MWCNT filtration at different pH
(b) PES-UF filtration at different pH
(c) MWCNT filtration at different ionic strength
(d) PES-UF filtration at different ionic strength
Fig 6.4 Comparison of permeability pattern with binary mixture fouling under different pH (4.7, 7.0,
10.4) and ionic strength (NaCl 0.001, 0.01 and 0.1 M)
172
6.3.5 Fouling mechanisms under solution chemistry
Filtration models were used to explaining MWCNT filtration with two proteins (Lys and BSA) both
at initial (~2h) and final (2-24h) stages of filtration. Overall, standard blocking and cake layer
formation were found to be simultaneously dominant during MWCNTs filtration both with individual
Lys and BSA solution, and protein mixture under different pHs and ionic strengths.
Based on the fouling potential estimated by these two fouling models and change of zeta potential at
different pHs and ionic strengths in the MWCNT membrane and protein, fouling mechanisms can be
explained by the electrostatic interaction between the proteins and the membrane surface and
intermolecular interaction (Lys-Lys, BSA-BSA and Lys-BSA). The dominant organic matter removal
mechanism was found to be increased electrostatic interaction between charged model compound and
positively charged membrane surface [32]. Current research also reports that membrane fouling can
be determined by the surface chemistry such as foulant-membrane interactions [39]. Thus, protein
fouling can be explained by the pH and ionic strength influence on the electrostatic interaction between
positively charged MWCNT membrane and both charged proteins (BSA and Lys).
The effects of charged proteins and solution chemistry such as pH and ionic strength on fouling
mechanisms were examined by comparing coefficients from filtration model and flux decline behavior.
6.3.5.1 MWCNT membrane
a. For a same charged protein (Lys) filtration in the MWCNT membrane, the membrane fouling was
found to be more influenced by intrafoulant interaction (Lys-Lys) based on the effect of increasing
ionic strength on the flux decline and fouling mechanisms. For the effect of pH, at pH 4.7 the fouling
alleviation was achieved due to the increased electrostatic interaction between the oppositely charged
membrane surface and the foulants. At pH 10.4, Lys aggregation due to the neutralized particle reduced
membrane fouling compared to at pH 7.0 thus higher flux was achieved at IEP of Lys (pH 10.4). Even
173
if electrostatic repulsion still exists at pH 7.0, highly hydrated Lys may decline permeate flux, resulting
in severe membrane fouling.
b. For an oppositely charged protein (BSA) filtration in the MWCNT membrane, the membrane
fouling was more governed by the electrostatic interaction between the membrane surface and the
foulants according to the effects of pH and ionic strength. The membrane fouling was significantly
alleviated at pH 4.7 and 10.4 at which the electrostatic interaction between the membrane and BSA
weakened. Increasing ionic strength also reduced membrane fouling via charge shielding effect, but
still higher than at different pH, demonstrating that the foulant-foulant attraction may have a less
significant contribution to the membrane fouling than the membrane-fouling attraction.
c. For binary mixture filtration test in the MWCNT membrane, the foulant-foulant interaction was
found to affect its fouling mechanisms dominantly. At pH 7.0 at which oppositely charged proteins
are neutralized by agglomeration, fouling was alleviated. However, at pH 4.7 and 10.4, the neutralized
protein deposited on the membrane surface, inducing deposition of counter charged proteins via
bridging effect. Thus, severe fouling was observed. It demonstrates that foulant-foulant interaction
played a major role under different pH condition. Meanwhile, NaCl addition seemed to promote binary
proteins deposition on the membrane surface by reducing intermolecular reaction (Lys-BSA).
However, as the ionic strength increased, the protein deposition was reduced by charge shielding effect
on the foulant-foulant and membrane-foulant interactions. The fouling mechanisms of the MWCNTs
membrane are illustrated in Fig 6.5.
Regarding the membrane performance recovery after fouling, it is assumed that the MWCNTs
enhanced membrane after protein fouling is highly likely to get recovered by chemical cleaning that
was introduced in Section 3.5, Chapter 3. The results of cleaning efficiency on HA fouling in Chapter
3 demonstrates that the 100 % of recovery on the permeate flux and rejection efficiency were achieved
after proper cleaning procedure. It indicates that the membrane fouled by HA can be recovered by
174
chemical cleaning (0.1 M HCl and 0.1 m NaOH). Compared to MW of HA (350 Da), adsorbed Lys
(14.3 kDa) and BSA (67 kDa) with much bigger MW on the membrane surface (12 kDa) could be
more easily removed by chemical cleaning (0.1 M HCl and 0.1 M NaOH). It corresponds with the
main fouling models in the MWCNTs membrane: protein deposition on the membrane surface by pore
blocking and cake layer formation.
175
Foulant
Schematics
Fouling mechanisms
-Membrane-foulant
interaction
-Foulant-foulant interaction
(Lys-Lys)
Increased electrostatic
repulsion at away from IEP
(a)
Positively
charge
protein
- Foulant-foulant interaction
Lys aggregation by charge
shielding effect
-Membrane -foulant
interaction
(b)
Negatively
charged
protein
Decreased electrostatic
interaction at near IEP of both
membrane and foulant
176
-Membrane -foulant
interaction
interaction by charge
shielding effect
Foulant-foulant interaction
Protein mixture
agglomeration at between IEP
(c)
Mixed
protein
Decreased deposition by the
charge shielding effect on
protein mixture
Fig 6.5 Fouling mechanisms in the MWCNT membrane under different solution chemistry (a) Lys
protein filtration, (b) BSA filtration and (c) Mixed protein filtration.
177
a. For an oppositely charged protein (Lys) filtration in the PES-UF membrane, combined effect of
membrane-foulant and foulant-foulant interactions was found to control the fouling mechanism at
different pH. Particularly, at different ionic strength, membrane-foulant interaction played a major role
in fouling via charge shielding effect on the protein and membrane surface. Different solution
chemistry mainly affected building up the cake layer at the later stage of filtration.
b. For an oppositely charged protein (BSA) filtration in the PES-UF membrane, fouling was mainly
governed by the membrane-fouling interaction at different solution chemistry. Severe pore blocking
at the initial stage and cake layer formation at the later stage at pH 4.7 was alleviated at pH 7.0 and
10.4 due to the increased electrostatic repulsion. However, increase in ionic strength induced fouling
due to the decreased electrostatic repulsion via charge shielding effect.
c. For binary mixture filtration test in the PES-UF membrane, combined effect of membrane-foulant
and foulant-foulant interaction played an important role in fouling. Different pH affected oppositely
charged foulants and membrane. It led to severe pore blocking at pH 7.0 via electrostatic interaction
between the membrane and foulants. At pH 4.7, at which neutralized protein and oppositely charged
protein promoted building up the cake layer at the later stage. Severe fouling was alleviated by an
increase in ionic strength due to charge shielding effect. The fouling mechanisms of the PES-UF
membrane are illustrated in Fig 6.6.
178
lant
Schematics
Fouling mechanisms
Membrane-foulant
interaction and
Deposition of neutralized
Lys on the PES membrane
(a)
Positively
charge
protein
Membrane-foulant
interaction
Decreased protein
adsorption by charge
shielding effect
Membrane-foulant
interaction
Increased electrostatic
repulsion between BSA(-)
and PES membrane (-)
(b)
Negatively
charged
protein
Membrane-foulant
interaction
repulsion by double layer
compression and charge
shielding effect
179
Combined effect of
membrane-foulant and
foulant-foulant interaction
Electrostatic interaction
between oppositely charged
foulants and the membrane
(c)
Combined effect of
Mixed
protein
membrane-foulant and
foulant-foulant interaction
Decreased adsorption by the
charge shielding effect on
protein mixture
Fig 6.6 Fouling mechanisms in the PES membrane under different solution chemistry (a) Lys protein
filtration, (b) BSA filtration and (c) Mixed protein filtration.
180
6.4. Conclusion
The fouling behavior of positively (Lys) and negatively charged (BSA) proteins was observed in the
positively charged MWCNT membrane with high sensitivity to the pH and ionic strength of the feed
solution. Overall, the MWCNT membrane with high sensitivity to the feed properties considerably
affected the initial flux and fouling behaviour.

The permeate flux of the MWCNT membrane was found to be strongly influenced by pH and
ionic strength under single protein filtration test. In comparison, solution chemistry did not
affect the permeate flux considerably in PES-UF membrane. Although severe flux decline
behavior appeared on the MWCNT membrane at the initial stage of filtration, overall permeate
flux outperformed the PES-UF membrane due to the extremely high flux of the MWCNT
membrane. Further, ionic strength resulted in flux enhancement in single protein filtration with
the MWCNT membrane. Meanwhile, protein mixture filtration had different permeate flux
behavior in both membranes.

For the MWCNT membrane filtration, fouling by Lys was controlled by combined effect of
membrane-foulant and foulant-foulant interaction at different solution chemistry. Severe pore
blocking was alleviated at pH 4.7 and increase in ionic strength, at which electrostatic
interaction increased, and Lys became aggregated due to charge shielding effect. Severe pore
blocking by BSA was alleviated at pH 4.7 and 10.4 where electrostatic interaction decreased,
and an increase in ionic strength due to charge shielding effect. Intermolecular interaction
(BSA-Lys) dominantly affected the MWCNT membrane fouling mechanism at different pHs,
Meanwhile, in the presence of ionic strength, intermolecular interaction mainly controlled the
membrane fouling, but as ionic strength increased, both the foulant-foulant and membranefoulant interaction by charge shielding effect became a dominant factor in the fouling
mechanism.
181

For the PES-UF membrane filtration, severe pore blocking by singe proteins was dominantly
controlled by the interaction with foulants. However, in a protein mixture filtration, severe cake
layer formation was developed at pH, and it was alleviated by combined effect of membranefoulant and foulant-foulant interaction.

Severe flux decline and fouling, despite very high initial flux in the MWCNT membrane could
be significantly alleviated by controlling solution pH and chemistry. It is suggested that the
MWCNT membrane sensitive to pH and ionic strength can be widely applied in the wastewater
reclamation.
182
References
1.
Bhattacharjee, S., C. Bhattacharjee, and S. Datta, Studies on the fractionation of β-lactoglobulin from
casein whey using ultrafiltration and ion-exchange membrane chromatography. Journal of Membrane
Science, 2006. 275(1-2): p. 141-150.
2.
Konrad, G. and T. Kleinschmidt, A new method for isolation of native α-lactalbumin from sweet whey.
International Dairy Journal, 2008. 18(1): p. 47-54.
3.
Cheang, B. and A.L. Zydney, A two-stage ultrafiltration process for fractionation of whey protein
isolate. Journal of Membrane Science, 2004. 231(1-2): p. 159-167.
4.
Yu, D., M.D. McLean, J.C. Hall, and R. Ghosh, Purification of monoclonal antibody from tobacco
extract using membrane-based bioseparation techniques. Journal of Membrane Science, 2008. 323(1):
p. 159-166.
5.
Zydney, A.L. and R. van Reis, 5.40 - Bioseparations: Membrane Processes, in Comprehensive
Biotechnology (Second Edition), M.-Y. Editor-in-Chief: Murray, Editor. 2011, Academic Press:
Burlington. p. 499-520.
6.
Wang, L., D.M. Kanani, and R. Ghosh, Purification of humanized monoclonal antibodies by membranebased hybrid bioseparation technique. Journal of Immunological Methods, 2006. 314(1-2): p. 1-8.
7.
Cheng, S.Y., D.L. Oatley, P.M. Williams, and C.J. Wright, Positively charged nanofiltration
membranes: Review of current fabrication methods and introduction of a novel approach. Advances in
Colloid and Interface Science, 2011. 164(1-2): p. 12-20.
8.
Rebhun, M. and J. Manka, Classification of organics in secondary effluents. Environmental Science &
Technology, 1971. 5(7): p. 606-&.
9.
Marshall, A.D., P.A. Munro, and G. Tragardh, The Effect of Protein Fouling in Microfiltration and
Ultrafiltratin on Permeate Flux, Protein Retention and Selectivity - A Literature-Review. Desalination,
1993. 91(1): p. 65-108.
10.
Miao, R., L. Wang, N. Mi, Z. Gao, T. Liu, Y. Lv, X. Wang, X. Meng, and Y. Yang, Enhancement and
Mitigation Mechanisms of Protein Fouling of Ultrafiltration Membranes under Different Ionic
Strengths. Environmental Science & Technology, 2015.
11.
Wang, Y.-N. and C.Y. Tang, Fouling of Nanofiltration, Reverse Osmosis, and Ultrafiltration
Membranes by Protein Mixtures: The Role of Inter-Foulant-Species Interaction. Environmental
12.
Palecek, S.P. and A.L. Zydney, Intermolecular electrostatic interactions and their effect on flux and
protein deposition during protein filtration. Biotechnology Progress, 1994. 10(2): p. 207-213.
13.
Singh, G. and L.F. Song, Quantifying the effect of ionic strength on colloidal fouling potential in
membrane filtration. Journal of Colloid and Interface Science, 2005. 284(2): p. 630-638.
14.
183
15.
Chan, R., V. Chen, and M.P. Bucknall, Quantitative analysis of membrane fouling by protein mixtures
using MALDI-MS. Biotechnology and Bioengineering, 2004. 85(2): p. 190-201.
16.
Nilsson, M., G. Tragardh, and K. Ostergren, The influence of pH, salt and temperature on nanofiltration
performance. Journal of Membrane Science, 2008. 312(1-2): p. 97-106.
17.
Palacio, L., C.C. Ho, P. Pradanos, A. Hernandez, and A.L. Zydney, Fouling with protein mixtures in
microfiltration: BSA-lysozyme and BSA-pepsin. Journal of Membrane Science, 2003. 222(1-2): p. 4151.
18.
Wang, Y.N. and C.Y.Y. Tang, Protein fouling of nanofiltration, reverse osmosis, and ultrafiltration
membranes-The role of hydrodynamic conditions, solution chemistry, and membrane properties.
Journal of Membrane Science, 2011. 376(1-2): p. 275-282.
19.
Ang, W.S. and M. Elimelech, Protein (BSA) fouling of reverse osmosis membranes: Implications for
wastewater reclamation. Journal of Membrane Science, 2007. 296(1-2): p. 83-92.
20.
She, Q., C.Y. Tang, Y.-N. Wang, and Z. Zhang, The role of hydrodynamic conditions and solution
chemistry on protein fouling during ultrafiltration. Desalination, 2009. 249(3): p. 1079-1087.
21.
Mo, H., K.G. Tay, and H.Y. Ng, Fouling of reverse osmosis membrane by protein (BSA): Effects of pH,
calcium, magnesium, ionic strength and temperature. Journal of Membrane Science, 2008. 315(1-2): p.
28-35.
22.
Jones, K.L. and C.R. O'Melia, Protein and humic acid adsorption onto hydrophilic membrane surfaces:
effects of pH and ionic strength. Journal of Membrane Science, 2000. 165(1): p. 31-46.
23.
Rohani, M.M. and A.L. Zydney, Role of electrostatic interactions during protein ultrafiltration.
Advances in Colloid and Interface Science, 2010. 160(1-2): p. 40-48.
24.
Chen, V., A.G. Fane, and C.J.D. Fell, The Use of Anionic Surfactants for Reducing of Ultrafiltration
Membranes - Their Effects and Optimization. Journal of Membrane Science, 1992. 67(2-3): p. 249-261.
25.
Xu, J., Z. Wang, J. Wang, and S. Wang, Positively charged aromatic polyamide reverse osmosis
membrane with high anti-fouling property prepared by polyethylenimine grafting. Desalination, 2015.
365: p. 398-406.
26.
Qiu, S., L. Wu, X. Pan, L. Zhang, H. Chen, and C. Gao, Preparation and properties of functionalized
carbon nanotube/PSF blend ultrafiltration membranes. Journal of Membrane Science, 2009. 342(1-2):
p. 165-172.
27.
Vatanpour, V., S.S. Madaeni, R. Moradian, S. Zinadini, and B. Astinchap, Fabrication and
characterization of novel antifouling nanofiltration membrane prepared from oxidized multiwalled
carbon nanotube/polyethersulfone nanocomposite. Journal of Membrane Science, 2011. 375(1-2): p.
284-294.
28.
Rahimpour, A., M. Jahanshahi, S. Khalili, A. Mollahosseini, A. Zirepour, and B. Rajaeian, Novel
functionalized carbon nanotubes for improving the surface properties and performance of
polyethersulfone (PES) membrane. Desalination, 2012. 286: p. 99-107.
184
29.
Celik, E., L. Liu, and H. Choi, Protein fouling behavior of carbon nanotube/polyethersulfone composite
membranes during water filtration. Water Research, 2011. 45(16): p. 5287-5294.
30.
Kumar, M. and M. Ulbricht, Novel antifouling positively charged hybrid ultrafiltration membranes for
protein separation based on blends of carboxylated carbon nanotubes and aminated poly(arylene ether
sulfone). Journal of Membrane Science, 2013. 448: p. 62-73.
31.
Kumar, M., D. McGlade, M. Ulbricht, and J. Lawler, Quaternized polysulfone and graphene oxide
nanosheet derived low fouling novel positively charged hybrid ultrafiltration membranes for protein
separation. Rsc Advances, 2015. 5(63): p. 51208-51219.
32.
Lee, J., Y. Ye, A.J. Ward, C. Zhou, V. Chen, A.I. Minett, S. Lee, Z. Liu, S.-R. Chae, and J. Shi, High
flux and high selectivity carbon nanotube composite membranes for natural organic matter removal.
Separation and Purification Technology, 2016. 163: p. 109-119.
33.
Nakao, S., H. Osada, H. Kurata, T. Tsuru, and S. Kimura, Separation of Proteins by Charged
Ultrafiltration Membranes. Desalination, 1988. 70(1-3): p. 191-205.
34.
Hadidi, M. and A.L. Zydney, Fouling behavior of zwitterionic membranes: Impact of electrostatic and
hydrophobic interactions. Journal of Membrane Science, 2014. 452: p. 97-103.
35.
Ouammou, M., N. Tijani, J.I. Calvo, C. Velasco, A. Martin, F. Martinez, F. Tejerina, and A. Hernandez,
Flux decay in protein microfiltration through charged membranes as a function of pH. Colloids and
Surfaces a-Physicochemical and Engineering Aspects, 2007. 298(3): p. 267-273.
36.
Zeman, L.J. and A.L. Zydney, Microfiltration and ultrafiltration : principles and applications. 1996:
Marcel Dekker, INC.
37.
Jeong, S., Y.J. Choi, N. Tien Vinh, S. Vigneswaran, and T.M. Hwang, Submerged membrane hybrid
systems as pretreatment in seawater reverse osmosis (SWRO): Optimisation and fouling mechanism
determination. Journal of Membrane Science, 2012. 411: p. 173-181.
38.
Kelly, S.T. and A.L. Zydney, Protein fouling during microfiltration: Comparative behavior of different
model proteins. Biotechnology and Bioengineering, 1997. 55(1): p. 91-100.
39.
Mustafa, G., K. Wyns, A. Buekenhoudt, and V. Meynen, New insights into the fouling mechanism of
dissolved organic matter applying nanofiltration membranes with a variety of surface chemistries.
Water Research, 2016. 93: p. 195-204.
185
Chapter 7 Conclusions
This research aimed to develop multi-walled carbon nanotube enhanced membrane with high
permeability and effective natural organic matter which cannot be achieved by conventional low
pressure driven membranes.
In the first part of the results (Chapter 4), optimization in membrane fabrication was conducted,
and the membrane performance in natural organic matter removal was systematically examined.
Further study built up finding critical factors in fast water flux and high NOM removal
mechanism. Lastly, efficient cleaning method was thoroughly investigated for the purpose of
overcoming flux decline due to the high removal efficiency.
- Multi-walled carbon nanotubes (MWCNTs)/polyaniline (PANI)/polyethersulfone (PES)
membranes were fabricated by incorporation of in-situ polymerized MWCNTs/PANI complex
for effective removal of natural organic matter (NOM) in water. The membranes exhibited up to
1400 LMH/bar of pure water permeability as MWCNTs increased to 2
wt%,
which is 30 times
greater than pure PES UF membrane. The NOM removal efficiency increased to 80 %, which is
4 times higher than the one of PES UF membrane.
- Based on analysis of the physical/chemical properties of the membranes, MWCNTs/PANI
complex increased porosity (78→85 %) and hydrophilicity (73.5→52.9 °) as MWCNTs
increased up to 2 %. Further, MWCNTs/PANI complex was revealed to tune the membrane
structure to narrow pore size on the top surface and relatively large macrovoid below a well
developed thinner skin layer, leading to the slippage effect. Such a superior performance in water
flux is attributed to the synergetic effect of increased porosity, hydrophilicity and slippage effect.
186
- MWCNTs/PANI complex was found to alter the positively charged membranes. In addition,
the complex formed the membrane with narrow pore size distribution. Thus, effective NOM
removal was achieved mainly due to the synergetic effect of increased electrostatic interaction
and enhanced size exclusion.
-The membrane demonstrated 100 % water flux recovery and 65 % total fouling ratio after
cleaning with 0.1 M HCl/0.1 M NaOH solution for 1 hr, indicating that rapid flux decline can be
easily overcome by the proper cleaning method.
Chapter 3 demonstrates that MWCNTs/PANI electron transfer complex successfully engineered
chemical/physical properties of the membrane to have high NOM adsorption capacity and fast
water flux.
In a second part of the results (Chapter 5), newly developed MWCNTs enhanced membrane was
applied to SWRO pretreatment. The MWCNTs enhanced membrane outperformed conventional
PES UF membrane when coupled with PAC adsorption system. Further, the impact of the ionic
strength on the membrane performance was covered in depth by examining the applicability of
MWCNTs membrane to alleviating RO membrane fouling in a desalination plant.
- MWCNTs membrane achieved superior permeate flux (300 LMH/bar) in seawater filtration,
which was up to 4 times as PES-UF membrane. MWCNTs membranes itself reduced dissolved
organic matter in seawater and required fewer amounts of adsorbent to achieve high permeate
flux. The performance of MWCNTs membrane did not decline when PAC adsorbent decreased
from 1.5 g/L to 0.5 g/L.
187
- Ionic strength such as NaCl was found to influence decreased organic removal efficiency (80
→ 30 %) and nearly 2-fold enhanced permeate flux in seawater filtration by MWCNTs
membrane. In the case of a divalent ion, greatly enhanced organic matter removal (90 %) was
observed in the presence of 0.01 M CaCl2, but the permeate flux rapidly declined.
In a third part of the results (chapter 6), protein fouling behavior was examined under different
solution chemistry- pH, ionic strength and charge of proteins. This study would give a
contribution to the extensive application of CNTs composite membrane for wastewater treatment
and bioseparation for which solution chemistry such as pH, ionic strength and charge of model
protein are critical factors of membrane fouling.
-PH and ionic strength affected permeate flux with positively and negatively charged proteins in
the MWCNTs membrane. In contrast, permeate flux of the negatively charged PES membrane
was not noticeably affected by different pH and ionic strength. MWCNTs membrane showed
high permeate flux when the membrane had the same charge with the model protein.
- According to the results on the fouling mechanism, MWCNTs membrane filtration with
proteins (Lys and BSA) dominated by standard blocking and cake layer formation. In
comparison to previous studies on protein fouling, the MWCNTs membrane showed low protein
fouling potential at IEP of the model compound due to the altering zeta potential of pH as a
function of pH and ionic strength. Lys filtration resulted in more severe pore blocking and cake
layer formation potential than BSA filtration due to its comparable size to the membrane pore
diameter. Due to the electrostatic repulsion between the membrane surface and the model protein
(Lys), the weakest fouling potential was observed at pH 4.7. The fouling potential was alleviated
by increasing ionic strength most probably due to the shielding effect of the ion on the model
188
protein. For oppositely charged proteins filtration, membrane fouling was alleviated at IEP of
proteins (pH 4.7) and pH 10.4 at which electrostatic interaction with the MWCNTs membrane
lessens. Increasing ionic strength seemed to decrease fouling potential, but was found to be
minor compared to the pH effect. Mixed protein filtration showed much more severe pore
blocking and cake layer formation at the IEP of both BSA and Lys (4.7 and 10.4, respectively). It
may be due to the bridging effect of neutralized proteins at their IEP with their counter proteins,
corresponding to the high fouling potential.
Recommendation for future research:
- For the continuing research work in Chapter 3 and 4, more uniformed single wall carbon
nanotube will contribute to improving the performance of mixed matrix membranes (MMMs).
- The synthesis of inorganic nanotubes is generally expensive, which will restrict their wide
application in MMMs for water treatment. Thus, it is recommended that future work can be a
synthesis of halloysite nanotubes (HNTs) which is naturally abundant and low cost as the filler
for MMMs fabrication, and its performance evaluation.
- Tubular supported (porous polymeric and ceramic supports) carbon nanomaterials such as
graphene is recommended to be a promising material for MMMs for water treatment.
189
ABBREVIATIONS
BSA: Bovine serum albumin
CA: Cellulose acetate
CNT: Carbon nanotube
CVD: Chemical vapor deposition
DBPs: Disinfection by-products
DCMD: Direct contact membrane distillation
DOC: Dissolved organic carbon
DOM: Dissolved organic matter
EPS: Extracellular polymeric substances
GO: Graphene oxide
HACNTs membrane: Horizontally aligned carbon nanotubes membrane
ICP: Internal concentration polarization
LC-OCD: Liquid chromatography–organic carbon detection
LMW: Low molecular weight
LMW-DOM: Low molecular weight dissolved organic matter
190
Lys: Lysozyme
MD: Membrane distillation
MF: Microfiltration
MMMs: Mixed matrix membrane
MPD: m-phenylenediamine
MWCNTs: Multi-walled carbon nanotubes
NOM: Natural organic matter
NMP: N-methyl-2-pyrrolidone;
NF: Nanofiltration
NOM: Natural organic matter
PA: Polyamid
PAC: Powder activated carbo
PANI: Polyaniline
PECVD: Plasma-enhanced chemical vapor deposition
PES: Polyethersulfone
PPCPs: Pharmaceutical and personal care products
PSf: Polysulfone
191
PTFE: Polytethrafluoroethylene
PVDF: Polyvinylidene fluoride
RO: Reverse osmosis
ROS: Reactive oxygen species
SA: Sodium alginate
SEC: Size exclusion chromatography
SWCNTs: Single-walled carbon nanotubes
SWOM: Seawater organic matter
SWRO: Seawater reverse osmosis
TFC: Thin film composite
UF: Ultrafiltration
UVD: Ultraviolet detector
VACNTs membrane: Vertically aligned carbon nanotube membrane
192

2016 Jieun Lee Final Thesis

Transcription

Similar documents

Diffusion and Osmosis Worksheet

Structure of Cell Membranes (Insane in the membrane)

Entertainment/Leisure Only

- Membrane vacuum press

POLYGLASS® U.S.A., Inc. Corporate Office 1111 West Newport

Nitto Group Company

Cell = City - Ms. Feffer 6th and 7th Grade Science

MusselBound Install

BIOCHEMISTRY IN PERSPECTIVE Organelles and Human Disease

OutDry technology achieves a fully waterproof, windproof and