Citation
Developments towards sustainable microbial electrochemical systems for simultaneous water deionization, electricity production, and wastewater treatment

Material Information

Title:
Developments towards sustainable microbial electrochemical systems for simultaneous water deionization, electricity production, and wastewater treatment
Creator:
Forrestal, Casey Bosch ( author )
Language:
English
Physical Description:
1 electronic file (126 pages). : ;

Thesis/Dissertation Information

Degree:
Doctorate ( Doctor of philosophy)
Degree Grantor:
University of Colorado Denver
Degree Divisions:
Department of Civil Engineering, CU Denver
Degree Disciplines:
Civil engineering

Subjects

Subjects / Keywords:
Sustainable engineering ( lcsh )
Microbial fuel cells ( lcsh )
Sewage disposal plants ( lcsh )
Saline water conversion ( lcsh )
Microbial fuel cells ( fast )
Saline water conversion ( fast )
Sewage disposal plants ( fast )
Sustainable engineering ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
Microbial Electrochemical Systems MESs , also known as MXCs, have the potential to dramatically affect the social, economic, and environmental systems around the world. This enormous potential stems from mxc systems simultaneously performing multiple beneficial functions such as waste treatment, energy production, chemical production, and water desalination. The treatment of waste and the production of value-added products such as electricity or fresh water, would benefit the environment and provide additional revenue for the economy. This dissertation focuses on developing novel microbial electrochemical systems while reducing costs and applying the technologies to benefit people, the planet, and the economy. ( ,,,, )
Review:
1. New methods for simultaneous deionization of salt water and energy production were developed through the incorporation of electrochemical adsorption with capacitive deionization in novel microbial electrochemical systems. With this fundamental principle established, several reactor systems were developed to remove salt and charged chemical pollutants.
Review:
2. New reactor configurations and materials were investigated. A modular reactor configuration called a spiral wound microbial electrochemical system was developed to improve MXC performance. This enhanced energy production and wastewater treatment system incorporates the maximum surface area to volume ration while reducing the overall footprint of the system. The modular component was developed for multiple applications, such as hydrogen production, desalination, and wastewater treatment. A new low cost membrane barrier was also developed to reduce the material cost for the commercialization of MXC technology. The agarose salt bridge barrier reduces the cost of the MXC system by more than 50 per cent without compromising in performance.
Review:
3. The treatment of produced water from the oil and gas industry is one of the most difficult types of water to treat. A new MXC system was developed which is capable of concurrently removing organic matter and total dissolved solids, while also producing electricity from the system.
Review:
With the development of new methods of desalination, new reactor configurations, and new reduced cost materials, the microbial electrochemical system will become more sustainable and commercially viable than ever before.
Thesis:
Thesis (Ph.D.)--University of Colorado Denver.
Bibliography:
Includes bibliographic references.
System Details:
System requirements: Adobe Reader.
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Casey Bosch Forrestal.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
930378432 ( OCLC )
ocn930378432

Downloads

This item has the following downloads:


Full Text
DEVELOPMENTS TOWARDS SUSTAINABLE MICROBIAL
ELECTROCHEMICAL SYSTEMS FOR SIMULTANEOUS WATER
DEIONIZATION, ELECTRICITY PRODUCTION, AND WASTEWATER
TREATMENT
by
Casey Bosch Forrestal
BS, Colorado State University, 2005
MS, Colorado School of Mines, 2010
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Civil Engineering
2013


2013
CASEY FORRESTAL
ALL RIGHTS RESERVED


This thesis for the Doctor of Philosophy degree by
Casey Bosch Forrestal
has been approved for the
Civil Engineering Program
by
Peter Jenkins, Chair
Zhiyong Jason Ren, Advisor
Jae Do Park
Angela Bielefeldt
Joann Silverstein


Forrestal, Casey, Bosch (Ph.D., Civil Engineering)
Developments Towards Sustainable Microbial Electrochemical Systems for
Simultaneous Water Deionization, Electricity Production and Wastewater Treatment
Thesis directed by Associate Professor Zhiyong Ren
ABSTRACT
Microbial Electrochemical Systems (MESs), also known as MXCs, have the
potential to dramatically affect the social, economic, and environmental systems around
the world. This enormous potential stems from MXC systems simultaneously performing
multiple beneficial functions such as waste treatment, energy production, chemical
production, and water desalination. The treatment of waste and the production of value-
added products such as electricity or fresh water, would benefit the environment and
provide additional revenue for the economy. This dissertation focuses on developing
novel microbial electrochemical systems while reducing costs and applying the
technologies to benefit people, the planet, and the economy.
1. New methods for simultaneous deionization of salt water and energy
production were developed through the incorporation of electrochemical adsorption with
capacitive deionization in novel microbial electrochemical systems. With this
fundamental principle established, several reactor systems were developed to remove salt
and charged chemical pollutants.
2. New reactor configurations and materials were investigated. A modular reactor
configuration called a spiral wound microbial electrochemical system was developed to
improve MXC performance. This enhanced energy production and wastewater treatment
system incorporates the maximum surface area to volume ratio while reducing the overall
iii


footprint of the system. The modular component was developed for multiple
applications, such as hydrogen production, desalination, and wastewater treatment. A
new low cost membrane barrier was also developed to reduce the material cost for the
commercialization of MXC technology. The agarose salt bridge barrier reduces the cost
of the MXC system by more than 50% without compromising on performance.
3. The treatment of produced water from the oil and gas industry is one of the
most difficult types of wastewater to treat. A new MXC system was developed which is
capable of concurrently removing organic matter and total dissolved solids, while also
producing electricity from the system.
With the development of new methods of desalination, new reactor
configurations, and new reduced cost materials, the microbial electrochemical system
will become more sustainable and commercially viable than ever before.
The form and content of this abstract are approved. I recommend its publication.
Approved: ZhiyongRen
IV


DEDICATION
I dedicate this work to my loving family and friends.
v


ACKNOWLEDGMENTS
I would like to thank my family and friends for all of your support. I would also
like to thank my friend and advisor Zhiyong (Jason) Ren for believing in me and guiding
me through the many obstacles faced in research. I would also like to thank all of my
advisers for your support and advice. Thank you to my financial sponsors, the Office of
Naval Research, National Science Foundation, University of Colorado Cleantech, and
University of Colorado Technology Transfer Office. I would also like to thank all of the
members of the Ren lab for your support and advice over the last few years.
Finally, I would like to make special acknowledgement to my wife Jessica
Forrestal for all of her love and support and for whom without I could not have been able
to pursue this degree.
vi


TABLE OF CONTENTS
Chapter
1. Introduction....................................................................1
1.1 Topics Covered in Dissertation and the Need for Sustainable Systems..........1
1.2 Background Information........................................................3
1.2.1 Microbial Desalination Cells and Other Desalination Devices.................3
1.2.2 Capacitive Deionization......................................................7
1.2.3 Microbial Electrochemical System Designs.....................................8
1.2.4 Membranes and Salt Bridge Barriers...........................................9
2. Capacitive Microbial Desalination Cell Integrated to Better Control Ion Migration1.. 12
2.1 Abstract.....................................................................12
2.2 Introduction.................................................................13
2.3 Materials and Methods........................................................15
2.3.1 cMDC Reactor Design.........................................................15
2.3.2 Reactor Innoculum and Growth Media..........................................16
2.3.3 Reactor Operation, Analysis and Calculations................................17
2.4 Results and Discussion.......................................................18
2.4.1 Removal and Adsoprtion of Desalination Chamber Salts........................18
2.4.2 Change In pH Over The Course of Desalination................................20
2.4.3 The Potential And Challenge Of The cMDC Configuration.......................21
2.5 Conclusion...................................................................22
3. Ion Migration and Control of pH Improved Through the Use of a New Reactor Design
Called a Microbial Capacitive Desalination Cell ..................................23
3.1 Abstract.....................................................................23
3.2 Introduction.................................................................24
vii


3.3 Materials and Methods
26
3.3.1 MCDC Reactor Design...........................................................26
3.3.2 MCDC Operating Conditions....................................................28
3.3.3 Analysis and Calculations....................................................31
3.4 Results and Discussion.........................................................32
3.4.1 Reactor Desalination Performance..............................................32
3.4.2 Assembly Regeneration and Salt Recovery......................................38
3.4.3 Reduced pH Fluctuation.......................................................40
3.5 Conclusion.....................................................................42
4. Investigations of Different Methods for the Construction and Operation of Spirally
Wound Microbial Electrochemical Systems3............................................44
4.1 Abstract.......................................................................44
4.2 Introduction...................................................................45
4.3 Materials and Methods..........................................................48
4.3.1 Spiral Wound Configurations...................................................48
4.3.1.1 Spiral Wound I Design.......................................................48
4.3.1.2 Spiral wound II and III Design.............................................50
4.3.2 Reactor Start-up and Operation..............................................53
4.3.2.1 Spiral Wound I Operation....................................................53
4.3.2.2 Spiral wound II and Spiral wound III........................................54
4.3.3 Analysis and Calculations.....................................................55
4.4 Results and Discussion.........................................................57
4.4.1 Electricity Production and Efficiency of Spiral Wound 1.......................57
4.4.2 Electricity Production and Efficiency of Spiral Wound II and III............60
4.5 Conclusion.....................................................................63
viii


5. An Efficient Method for Treating and Desalinating Produced Water Using a Microbial
Capacitive Desalination Cell4...................................................66
5.1 Abstract...................................................................66
5.2 Introduction...............................................................68
5.3 Material and Methods.......................................................71
5.3.1 Design of Microbial Capacitive Desalination Cell..........................71
5.3.2 MDC Design................................................................73
5.3.3 Operation of the MCDC and MDC Reactors for Proof of Concept Study........73
5.3.4 Operation of the MCDC for Maximum Desalination...........................76
5.3.5 Analysis..................................................................78
5.4 Results and Discussion.....................................................79
5.4.1 Proof of Concept MCDC Vs MDC Desalination and COD Removal Capacity.......79
5.4.2 Regeneration and Energy Harvesting Using the MCDC System.................86
5.4.3 Maximum Desalination Capacity and Rate of the MCDC System.................87
5.5 Conclusion.................................................................92
6. The Development of an Efficient Low Cost Barrier for the Use in a Microbial
Electrochemical System Called an Agarose Salt Bridge Membrane5..................94
6.1 Abstract...................................................................94
6.2 Introduction...............................................................95
6.3 Materials and Methods......................................................97
6.3.1 Agarose Salt Bridge Membrane Design.......................................97
6.3.2 Reactor Configuration.....................................................97
6.3.3 Analysis..................................................................99
6.4 Results and Discussion.....................................................99
6.4.1 Agarose Salt Bridge Membrane Power Performance from Produced Water........99
6.4.2 Agarose Salt Bridge Membrane Change in pH................................101
ix


6.4.3 Internal and Polarization Resistance for the ASBM Compared to the Astom CEM
102
6.4.4 Cost Comparison and Sustainability..........................................104
6.4.5 ASBM Challenges.............................................................105
6.5 Conclusion..................................................................105
References.........................................................................107
x


LIST OF TABLES
Table
3.1 MCDC Salt Removal .........................................................33
3.2 Comparison of Different Capacitive Deionization Methods..................34
4.1 Comparison of Spiral Wound Reactors........................................47
5.1 Produced Water Characteristics.............................................78
5.2 Energy Harvested From Single Batch Cycle of MCDC Using Produced Water....87
5.3 Percent TDS Removal and Change in Conductivity for the Three Desalination Cycles
89
5.4 Total Ions Removed and Recovered with the MCDC System....................90
6.1 Thicknesses of the ASBM....................................................97
xi


LIST OF FIGURES
Figure
2.1 Diagram of Capacitive Microbial Desalination Cell (cMDC) reactor with physical
and electrical adsorption. Picture of Spiral Wound II...........................16
2.2 cMDC Desalination And Energy Production Picture of Spiral Wound II..........19
2.3 cMDC Change in Conductivity 4.3 Picture of Spiral Wound II..................20
2.4 Change in pH for cMDC System 4.3 Picture of Spiral Wound II.................21
3.1 Operation of MCDC System.....................................................28
3.2 Diagram of MCDC System 4.3 Picture of Spiral Wound II.......................29
3.3 Correlation of Applied Potential and Deionization...........................35
3.4 MCDC Ion Migration..........................................................36
3.5 Electrochemical Impedance Spectroscopy for the MCDC ........................37
3.6 MCDC Regeneration...........................................................38
3.7 MCDC Ion Recovery...........................................................40
3.8 Change In pH for MCDC.......................................................41
4.1 Diagram for Spiral Wound I.................................................48
4.2 Picture of Spiral Wound I...................................................50
4.3 Picture of Spiral Wound III.................................................51
4.4 Computer Aided Drawing of Spiral Wound II and III...........................52
4.5 Internal Diagram of Spiral Wound II and III.................................53
4.6 Operation of Spiral Wound III...............................................54
4.7 Acclimation Times For Different reactor Configurations .....................56
4.8 Effect of Retention Time On Power Density SWI..............................57
4.9 Coulombic Efficiency Correlated to Retention Time and Maximum Power Point... 58
xii


4.10 Change In Polarization Resistance Over Time...........................59
4.11 Spiral Wound II and III Initial Overshoot Potential...................60
4.12 Spiral Wound III Loss of Overshoot Potential..........................61
4.13 Comparison of EIS for the Three Spiral Wound Reactors..................62
5.1 Operation of MCDC for Produced Water....................................72
5.2 MDC for Produced Water...................................................73
5.3 Voltage Profile..........................................................79
5.4 Charged Formed on CDI....................................................82
5.5 Change in Conductivity for MDC and MCDC..................................83
5.6 Change in pH for the MDC and MCDC........................................84
5.7 Change in COD per Hour for the MDC and MCDC..............................86
5.8 Change in TDS for the MCDC Operating Under Maximum Conditions............88
5.9 Change in COD for the MCDC for the three Chambers Plus the Recovered COD 92
6.1 Power Density of the ASBM...............................................101
6.2 Change in pH for the Anode and Cathode Chamber with the ASBM using Produced
Water in the Anolyte........................................................102
6.3 Nyquist Plot for the ASBM, ASBM-2 and the Astom CEM.....................103
6.4 EIS Comparison of the ASBM and ASBM-2...................................104
xiii


1. Introduction
1.1 Topics Covered in Dissertation and the Need for Sustainable Systems
Since the earliest recorded history of Thales of Miletus in 650 BC, humans have
strived to understand and record the mysteries of nature. The development of scientific
knowledge over the last 3000 years has led to amazing advancements in the quality of
human life. This progress, however, has also brought us to the precipice of annihilation
with the production of nuclear weapons and the destruction of the environment. Only in
the last 100 years has man realized the need to develop technology that will sustain life
now and in the future. In 1987 the Brundtland Commission published the most commonly
accepted definition of sustainable development, sustainable development is development
that meets the needs of the present without compromising the ability of future generations
to meet their own needs. 1 The following topics in this dissertation were developed with
this goal in mind.
Microbial electrochemical technology (MET), is a potential solution to many
needs in sustainable development. Microbial electrochemical systems (MES), also
known as MXC, use microorganisms to catalyze the oxidation of organic or inorganic
matter to generate electrical current in a fuel cell system. The ability of microorganisms
to transfer electrons and generate electrical current has been credited to MC Potter in
1911. However, only in the last 20 years has this phenomenon been extensively studied.
The explosion of research in this field is a result of the potential to dramatically affect the
social, economic, and environmental systems around the world. This enormous potential
stems from MXC systems simultaneously performing multiple beneficial functions, such
as wastewater treatment, electrical energy production, chemical production, and water
1


desalination. MXCs have the potential to revolutionize humanity by solving two of the
greatest problems to ever to face mankind: the availability of energy and water.
However, to date, few large scale MXCs have been developed and implemented because
of the inherent inefficiency and high cost of a newly developed technology. Presented,
here are some of the most recent advancements in METs ability to desalinate water,
efficiently treat wastewater, and generate electricity with a minimal cost and
environmental footprint.
Each chapter focuses on a single method to improve MXC for sustainability.
Chapter 2 and 3 focuses on developing a sustainable method for deionization of salt water
and energy production, through incorporating electrochemical adsorption by capacitive
deionization, in novel microbial electrochemical systems. With this fundamental
technology, a more sustainable method for desalinating salt water and removing charged
chemical pollutants can be achieved. Two articles have been published on this topic. The
article in chapter 2 demonstrates the advantages and challenges of a capacitive Microbial
Desalination Cell (cMDC). The second article, in chapter 3, demonstrates the newly
developed MXC called the microbial capacitive desalination cell (MCDC) which
demonstrates its advantages and challenges in desalination of salt water.
Chapter 4 focuses on the development of a modular spiral wound microbial
electrochemical system for enhanced energy production and wastewater treatment. The
system incorporates the maximum surface area to volume ratio while reducing the overall
footprint of the system. The spiral wound reactor is a modular system; therefore it can be
scaled to meet the need of large wastewater treatment operations as well small onsite
wastewater treatment. The modular component also helps develop the system for
2


multiple applications, such as hydrogen production, desalination, and wastewater
treatment. The manuscript Investigation of Different Methods for the Construction and
Operation of Spirally Wound Microbial Fuel Cells is in preparation.
Chapter 5 focuses on the development and characterization of a MXC for the
treatment of produced water from oil and natural gas mining. The technology uses the
newly developed MCDC, outlined in chapter 3, to remove both organic and dissolved
salts within the produced water. Additionally demonstrated is the MCDCs capability to
harvest electricity. The manuscript An Efficient and Sustainable Method of Treating and
Desalinating Natural Gas Produced Water Using a Microbial Capacitive Desalination
Cell is in preparation.
Lastly, chapter 6 focuses on the development of a low cost membrane barrier
which could be used in MXC reactors for the production of electricity, treatment of
wastewater, and potentially desalinate salt water. An agarose salt bridge membrane could
dramatically improve the economic viability of MXCs by reducing the capital cost of
manufacturing MXC systems, without compromising on performance. A thorough
investigation into the manufacturing and chemical composition of the agarose salt bridge
membrane yielded an efficient low cost MXC system. However, many practical
challenges still exist with this material. Further research is needed before this material
would be viable for commercial use and publication.
1.2 Background Information
1.2.1 Microbial Desalination Cells and Other Desalination Devices
While 97% of the planet is covered by water, only approximately 0.4% of that
water is available for consumption. Currently the most popular methods to increase the
3


supply of fresh water are through either evaporative processes such as multistage flash
evaporation (MSF) or filtration processes such as reverse osmosis (RO). Evaporative
desalination processes were the first to be developed and widely commercially used. The
technology works by heating the salt water to convert the liquid water into the gaseous
phase separating the water from the salt. The gaseous phase water then needs to be
condensed back into the liquid phase for use. This process, while highly effective, is also
energy intensive requiring up 650 kWh/m of fresh water produced. RO technology uses
a different method to separate water from salt. RO uses specially designed membranes
that allow water molecules to pass across the membrane but prevent larger molecules. In
order to facilitate the movement of water across the membrane, RO systems need to
operate at a high pressure using pumps which require electrical energy. The energy
needed for filtration desalination systems is much smaller through the evaporative
process, however RO still require between 1-6 kWh/m water produced. One other
currently commercialized process for desalination is called electrodialysis (ED).
Electrodialysis works by moving ions in a potential field across anion and cation
exchange membranes. The applied potential causes anions to migrate across the anion
exchange membrane towards the positive charge, and cations to migrate across the cation
exchange membrane towards the negative charge. Electrodialysis usually incorporates
multiple membranes so that ions migrate from desalination chambers into concentration
chambers. ED requires energy at a rate of 0.3-0.8kWh/m Microbial desalination cells
(MDC) are one of the many MXCs previously developed. MDCs desalinate salt water in
the exact same method as electrodialysis except the electrical potential is generated
internally by microorganisms rather than supplied externally. The oxidation of a
4


substrate in the anode chamber causes an electrochemically positive charge to be
generated. Electrons flow from the anode chamber to a cathode where the electrons are
combined with the reduction of oxygen to water. The electrochemical reduction of a
chemical (O2, or K3Fe(CN)6) causes a negative potential charge to be formed. By adding
a positively charged anion exchange membrane (AEM) adjacent to the anode chamber,
and a negatively charged cation exchange membrane (CEM) adjacent to the cathode
chamber, a MDC is formed. Desalination occurs between the AEM and the CEM called
the desalination chamber. When sodium chloride is dissociated, sodium ions pass
through the CEM to the cathode chamber while the negatively charged chloride ions pass
through the AEM. This technology was first demonstrated by Cao et al. 2009, and has
been improved upon by many researchers. The MDC system was improved upon by
incorporating multiple chambers for the production of desalinate and concentrate
solutions. Additionally work done by Haiping Luo at the University of Colorado Denver,
demonstrated that hydrogen gas could be generated in a MDC system by incorporating an
additional electrical supply. The MDC in 2009 was touted as a major breakthrough in
sustainably desalinating salt water because it could generate electricity, treat a
wastewater, and supply desalinated water. The proof of concept article was selected as
the Top Technology Paper of the Year by Environmental Science & Technology. While
the MDC was a major breakthrough it still contained many challenges. Generally there
are three problems associated with the MDC reactor configuration. As the ions are
removed from the desalination chamber the conductivity drops causing an increase in
internal resistance. In the Cao et al. 2009 paper, during the start of desalination the
internal resistance was low and so the voltage output was high. Over the course of
5


desalination the ohmic resistance, determined by electrochemical impedance
spectroscopy (EIS), increased to over 970 Ohms; thus causing the voltage and the rate of
desalination to decrease. The increase in ohmic resistance from the operation of the MDC
occurs as quickly as the salt solution is desalinated. This problem is compounded by
scaling and biofouling on the membranes also identified by Haiping Luo.4 Biofouling
generally occurs slowly over a period of days or months on the AEM due to the
microorganisms in the anode chamber. Chemical scaling was found to be most prevalent
on the CEM next to the cathode chamber. The affect of biofouling and chemical scaling
greatly reduced the MDCs reactors performance. Additionally, because of the use of
anion exchange membrane, which prevents the migration of positively charge ions such
as protons, the pH decrease in the anode and increase in cathode. The decrease in pH in
the anode chamber inhibits the microorganisms ability to transfer electrons. Lastly, if
wastewater is used in the anode chamber, ions that migrate from the desalination chamber
to the anode chamber may inhibit the treated wastewater reusability.
Because of the MXC ability to generate electricity it would be prudent to ponder
whether it would be more beneficial to generate electricity for external desalination or to
use the MDC system for desalination. Currently the energy produced and harvested in
MXCs, is small.5 The MDC does not actually consume any energy for desalination but
rather uses the electrical potential naturally generated. Therefore at this time it would not
be beneficial to use a MXC to generate electricity for desalination in a different
desalination reactor such as reverse osmosis.
6


1.2.2 Capacitive Deionization
Capacitive deionization (CDI) is an additional method for desalinating salt water
which uses many of the same principles as ED. CDI is an electrochemical process,
sometimes referred to as electrosorption6. CDI operates by adsorbing ions onto a high
surface area electrode by applying an electrical potential as in ED. When an electrical
potential is applied to an electrode a charge is formed on the surface of the electrode
causing electrochemical attraction of oppositely charged ions. As ions are adsorbed onto
the electrode a double layer capacitor is formed. This was first identified by Hermann
von Helmholtz in 1883, and is sometimes referred to as the Helmholtz double layer
capacitor. The theory of the electrical double layer was later modeled by Guoy-Chapman
in 1913 and has been widely accepted as true. The theory treats ions as small point
charges and the capacity of adsorptions is dependent upon potential at the surface of the
electrode, area of electrode, and distance between the electrodes. The region directly
adjacent to the electrode has been termed the Helmholtz layer and can be modeled using
Langmuirs adsorption isotherm. The outer layer, or the diffusive layer, can be modeled
using the Guoy-Chapman method. Molecules or ions in a solution form hydration shells
which can also be modeled using the double layer principles. A potential difference is
then formed between the surface of the ion and the bulk solution. This parameter is called
the electrokinetic (zeta) potential, or just the zeta potential. The zeta potential is
important in understanding the formation of colloids and is related to pH and
conductivity. Ideally CDI works by removing charged ions without involving redox
reactions which would make the process fully reversible; thus providing a means of
removing ions (desalination) and subsequently concentrating those ions (regeneration).
7


While the first CDI systems were developed in the 1960s, CDIs have not been
considered feasible until recently, when high surface area electrodes were developed,
including the carbon aerogel, activated carbon cloth, and graphene. The advantage of the
CDI system is that it requires a relatively small amount of energy, operating at potentials
between 0.6V and 1.2V consuming as low as 0.1 kwh/m water produced. Also, because
the applied energy is stored in a capacitor, some of that energy can be recovered. Work
on supercapacitors has demonstrated a round trip energy efficiency of greater than 95%.
A CDI system operating at a round trip energy efficiency similar to that of a
supercapacitor would approach the thermodynamic limitation for the removal of salt from
water. The CDI system is extremely complicated with many challenges, but the
technology has the greatest potential to reach the thermodynamic limitation for
desalination.
1.2.3 Microbial Electrochemical System Designs
MXC have been developing over the last century with most of the research done
in the last decade. The goal in designing a MXC system is to maximize energy output,
treatment of wastewater, and operate at a reduce cost compared to traditional wastewater
treatment. Paramount to achieving these three goals is the system design. The system
design consists of two facets, the external operating design and the materials used. Since
2000 the power output in a MXC system has be steadily increasing, from 13 W/m to
2.87 kW/m Researchers have increased volumetric power density generally by
increasing the total surface area to reactor volume, and by decreasing the total volume of
the reactor. In 2007, Fan et al. used a plate and frame reactor design with a 2.5mL
double cloth electrode assembly and a cathode surface area of 230m /m to produce
8


1.55kW/m3.9 Additionally, Nevin et al. 2008 with a volume of 0.335mL, using
3 10
ferricyanide as the electron acceptor produced a power density of 2.15 kW/m It
appeared that high power density in MXC systems could be achieved only through the
reduction of volume until recently, when Fan et al. 2012 demonstrated that at a volume of
30mL a power density of 2.87kW/m could be achieved. These advancements in system
design have been a long time coming. MXC systems have been developed using a single
chamber or two chambers with a membrane barrier. These two general reactor designs
have been modified on numerous occasions. Some of the designs have incorporated
serpentine or upflow pathways for fluid flow, and there have been designs for multiple
anode and cathodes. Some reactors have been designed in a tubular fashion, where either
the anode or cathode electrode is wrapped around the other. In addition to external
system designs, most of the research on MXC systems has been to identify materials
which improved performance and reduced cost. To date, the most widely studied
material for use as an anode or cathode electrode are inorganic carbons such as graphite,
carbon cloth, or graphite granules. This material has a relatively low cost at $0.24-10/kg
but has a resistance ranging from 1.47x10 -1.9x10 micro Ohms/cm. With greater
improvements to the system design a commercially viable MXC system will be
developed.
1.2.4 Membranes and Salt Bridge Barriers
The purpose of membrane in a MXC system is to separate different chambers
while still allowing ions to transfer. Membranes are not always required for MXC
systems; however, for the production of desalinated salt water or chemical production the
use of a membrane, is a must. Membranes can be beneficial and harmful for the operation
9


of a MXC system. They were first designed to separate the anode and cathode chamber
to prevent oxygen diffusion from the cathode chamber into the anode chamber, reducing
the electron transfer efficiency. They can also prevent the formation of biofilm on the
cathode which increases electron transfer. However, there are many challenges
associated with using membranes in MXC systems. The use of a membrane causes the
internal resistance of the reactor to increase, reducing power output. The use of a
membrane may inhibit proton transfer causing the pH in the anode chamber to become
more acidic and basic in the cathode chamber. Also, as previously mentioned, microbes
in the anode chamber can become attached to the membranes causing an increase an
internal resistance called biofouling. Membranes can become scaled with a chemical,
which causes an increase an internal resistance. These problems are common in MXC
research especially when real instead of synthetic wastewater is used.
The two most extensively studied membrane materials are cation exchange
membranes (CEM), also referred to as proton exchange membranes, and anion exchange
membranes (AEM). To date the use of an AEM in a two chamber MXC system has
demonstrated the highest energy efficiency. This is likely due to the concentration
gradient diffusion of anions across AEM and the reduced scalability of AEMs over
CEMs. In addition to standard ion exchange membranes, salt bridges, bipolar
membranes, ultrafiltration membranes, microfiltration membranes, glass fibers, porous
fabric and other porous filters have been investigated. Salt bridges were the first ion
barrier used in a fuel cell which was developed by using a glass tube filled with
electrolyte solution between two chambers. Salt bridge separators have been used in
MXC systems for many years but have been widely discredited due to its higher internal
10


resistance and thus lower power density. Most salt bridge separators use KC1 or
phosphate buffer as ion conductors with added agar to prevent fluids from mixing. Min
et al. 2005 found that a phosphate buffered salt bridge increased the internal resistance to
19,920 Ohms over a Nafion CEM membrane which had an internal resistance of 1286
Ohms. However, there were two major advantages of salt bridge membranes
previously investigated: it limited the diffusion of oxygen, and cost a faction of other
membrane separators. Traditional ion exchange membranes can cost between $100/m to
$1400/m In order to scale MXC systems, a sustainable separator will need to be
developed.
11


2. Capacitive Microbial Desalination Cell Integrated to Better Control Ion
Migration1
2.1 Abstract
A new microbial desalination cell with capacitive adsorption capability (cMDC)
was developed to solve the ion migration problem facing current MDC systems.
Traditional MDCs remove salts by transferring ions to the anode and cathode chambers,
which may prohibit wastewater beneficial reuse due to increased salinity. The cMDC
uses adsorptive activated carbon cloth (ACC) as the electrodes and utilizes the formed
capacitive double layers for electrochemical ion adsorption. The cMDC removed an
average of 69.4% of the salt from the desalination chamber through electrode adsorption
during one batch cycle, and it did not add salts to the anode or cathode chamber. It was
estimated that 61-82.2 mg of total dissolved solids (TDS) was adsorbed to 1 g of ACC
electrode. The cMDC provides a new approach for salt management, organic removal,
and energy production. Further studies will be conducted to optimize reactor
configuration and achieve in situ electrode regeneration.
1 The work presented in this chapter has been published by Casey Forrestal, Pei Xu, Peter
Jenkins, and Zhiyong Ren. Bioresource Tech. 2012, 120, 332-336
12


2.2 Introduction
The sustainable supply of fresh water through saltwater desalination has been
developed significantly in the past century, but one remaining challenge is high energy
use during the desalination process. Popular desalination methods include reverse
osmosis (RO) and multistage flash evaporation (MSF) which are considered energy
intensive, because for treating 1 m of seawater, RO typically uses 3-7 kwh/m of
electricity and MSF may require up to 68 kwh/m31415. Recently, a new desalination
technology called microbial desalination cell (MDC) was developed and demonstrated
that salt water can be desalinated without using external energy. Moreover, this process
can also simultaneously achieve wastewater treatment and energy production in the
format of electricity or hydrogen gas. 16-20 MDC reactor uses exoelectrogenic bacteria to
oxidize biodegradable substrate (i.e. wastewater) in an anode chamber and transfer the
electrons to the anode. The electron flows through an external circuit to a cathode, where
external electron acceptors (i.e. O2) are reduced. When a middle chamber is inserted in
between the anode and cathode chamber using a pair of ion exchange membranes,
desalination can be achieved. The potential difference between the anode and cathode
electrodes drives the migration of ions out of the desalination chamber, with anions (CF)
migrating to the anode chamber across an anion exchange membrane and cations (Na+)
migrating to the cathode chamber across a cation exchange membrane. The process can
remove more than 99% of the salt water and potentially produce more energy than the
external energy required for the system, making it a promising desalination process with
net energy gaining. 1721
13


One main challenge with the MDC technology is that while the salts are removed
from the middle chamber, they become concentrated in the anode and cathode chambers,
which results in salinity increases in the anolyte and catholyte. While this ion addition is
generally acceptable for wastewater treatment and helps with conductivity conditioning,
it may cause concerns for water beneficial reuse, where the total dissolved solids (TDS)
is regulated. Additionally, as ions are desalinated from the desalination chamber, the
internal resistance immediately increases causing the rate of desalination to also decrease
immediately.
One solution to manage salt removal from MDC is to incorporating capacitive
deionization (CDI) concept into MDC systems. When a saline solution flows
between two charged electrodes, the ions can be adsorbed by the double layer capacitor
formed on the high surface electrodes. When the potential is removed the ions can be
released back into the liquid to form a concentrate for salt recovery. Using this integrated
deionization approach in combination with the traditional MDC, salt water can be
deionized through electrochemical salt adsorption on the electrodes, so no ions will
migrate into the electrolyte solutions. In this study, the integration of an MDC and
electrochemical adsorption was developed into a single reactor for capacitive microbial
desalination (cMDC). This system demonstrates the feasibility of a new process for
concurrent power production and saltwater desalination without contaminating the anode
or cathode chambers.
14


2.3 Materials and Methods
2.3.1 cMDC Reactor Design
Each cMDC reactor consisted of three polycarbonate cube-shaped blocks. Each
block has a 3-cm diameter hole, and three blocks were clamped together to form one
anode chamber, one cathode chamber, and one desalination chamber, with the volume of
23 mL, 27mL, and lOmL, respectively (Figure 2.1).4 Zorflex Activated Carbon Cloth
(ACC) (Chemviron Carbon, UK) was used as the electrode material and was pretreated
by washing in acetone overnight and heating to 350C for 30 minutes. An anode
assembly consisted of one layer of AEM (AMX, Astom Corporation, Japan), one Ni/Cu
mesh current collector (Grade 400, McMaster Carr, IL), and an ACC anode was formed
to separate the anode chamber and the desalination chamber. Similarly, a cathode
assembly was formed by pressing one piece of CEM (CMX, Astom Corporation, Japan),
one layer of Ni/Cu mesh, and an ACC cathode together to separate the cathode chamber
2
and the desalination chamber. The total surface area of the ACC electrodes were 18 cm
(9 cm each) with the weight of one gram. The specific surface area of the ACC is 1020
m2/g, determined by the Brunauer-Emmet-Teller (BET) method (ASAP 2020,
Micromeritics, Norcross, GA). 26 Prior to use, the membranes were pretreated in 10 g/L
NaCl for 24 hours to remove impurities and maximize ion exchange capacity. The Ni/Cu
current collectors were connected to a titanium wire that connected the anode and
cathode across a one Q resistance (Figure 2.1).
15


R


Ferricyanide
Ferrocyanide
Ni/Cu AEM
NaCl
Feed
CEM
ACC
Ni/Cu
Figure 2.1 Diagram of Capacitive Microbial Desalination Cell (cMDC) reactor with
physical and electrical adsorption.
2.3.2 Reactor Innoculum and Growth Media
The reactor was initially inoculated with anaerobic sludge and operated in fed-
batch microbial fuel cell mode by using a CEM to separate the anode and cathode
chambers. 19 When the repeatable voltage profile was obtained in three consecutive
batches from the MFC, the middle chamber was inserted to the reactor to form a cMDC
as described previously. The anolyte contained per liter: 1.6g NaCEECOO, 0.62 g
NH4CI, 4.9 g NaEEPCE -EEO, 9.2 g Na2HP04, 0.3g KCL, and 10ml trace metals and
10ml vitamin solution.28 The catholyte contained per liter: lOOmM KFe(CN)6, 5mM
KH2PO4, and 5mM K2HPO4. Ferricyanide solution was used as the catholyte to
minimize cathode mass transfer effects. The salt solution for desalination contained 10
g/L NaCl. This concentration of salt water would be representative of a brackish water
16


found around the world. A treatment efficiency of 70- 90% would be safe to dispose to
most environmental conditions.
2.3.3 Reactor Operation, Analysis and Calculations
The active cMDC reactor was operated in fed-batch mode. A single fed-batch was
defined as the time required for a single desalination cycle. As soon as the voltage across
the reactor reached below 0.025mA, all electrolyte solution was replaced. Electrolyte
conductivity was determined by a conductivity meter (Sension 156, HACH Co., CO), and
pH was determined with a pH meter (Sension 4, HACH Co., CO). Before the reactor was
connected for each batch, the anode and cathode chambers were allowed to reach the
maximum open circuit potential (OCP), which was determined using a potentiostat with a
saturated Ag/AgCl reference electrode (G 300, Gamry Instruments Inc. NJ). Using a data
acquisition system (model 2700, Keithley Instruments, inc. OH), the voltage across the
external resistor was recorded every one minute. Conductivity and pH measurements for
all three chambers were taken at the beginning and the end of each desalination cycle.
Conductivity was converted to TDS (mg/L) using the HACH Co. general calculation
equation presented below.
mS/cm 500 = mg/L
The ACC assemblies were manually regenerated by removing the ACC from the reactor
and rinsed in 1 L of deionized (DI) water for 30 minutes. The DI rinse was repeated a
total of three times till no salt residue was left before adding the electrodes back to the
reactor for additional experiments. Negative control experiments were conducted by
removing the acclimated anode and replacing it with an unacclimated piece of ACC and
performed as described previously for desalination experiments.
17


2.4 Results and Discussion
2.4.1 Removal and Adsoprtion of Desalination Chamber Salts
During cMDC operation an electrical potential is formed on the ACC across the
anode and the cathode due to microbial oxidation of substrate and electron transfer to the
cathode. This potential forms a double layer capacitor on the ACC electrodes, which
adsorbs ions to achieve water deionization. At the start of the desalination cycle the
cMDC reactor had an OCP of ~712mV. When a 1-Ohm resistor was used in the circuit, a
maximum current of 2.5mA was generated. The reactor was operated for three months,
and Figure 2.2 shows the production of electric current across the resistor is proportional
to the percent removal of NaCl in the desalination chamber for three consecutive cycles.
The substrate consumption and ion loss in the desalination chamber caused an increase of
internal resistance, which resulted in a decrease of current along with one batch cycle.
Such phenomenon is consistent with other MDC studies using similar reactor
configurations.1619 This traditional 3-chamber cube MDC configuration was used in the
study to demonstrate the feasibility of the process, and system optimization such as using
stack configurations with narrow chambers and/or operating the reactor in a continuous
mode have been shown effective to address the increased resistance problem.
18


2.5
80
70
60
50
40
30
20
10
0
u
o
>
o
E
&
3
Current
(mA)
NaCI
Remold
0
50 100 150
Time (hr)
Figure 2.2 cMDC Desalination And Energy Production
Three typical cycles of cMDC operation and current generation with NaCI removal profile. Arrows
indicate the manual washing and regenerating the ACC electrodes.
Figure 2.3 shows the salt removal from three chambers as represented by
conductivity changes. The average salt removal in the desalination chamber during one
batch cycle was 69.4%, which correlates to the removal of 69.4 mg (1.19mM) of NaCI
from the lOmL desalination chamber. The 1.19mM of salt removed from the desalination
chamber correlates to 42.2 mg of CT migrated to the anode and 27.2 mg of Na+ migrated
to the cathode. The anode chamber conductivity in average increased slightly by 0.7% or
l.lmg presumably due to the limited diffusive ion release from the anode. Such ion
balance suggests that 41.1 mg of chloride was electrochemically adsorbed on the anode,
which represents 97.3% of the desalinated chloride from the desalination chamber. In the
cathode chamber the conductivity decreased in average by 1.4% or by 3.2 mg. Therefore
the cathode chamber with the ACC adsorbed 30.4 mg of salt or 100% of the desalinated
salt plus an additional 12% from the cathode chamber. These results indicate that the
electrical adsorption capacity of the ACC assembly was between 61-82.2 mg TDS/g
19


ACC. Such results are comparable to the findings from a previous study, which showed
27
the physical and electrical adsorption was 72.7mg/g ACC. This salt removal profile
from the desalination chamber is different from traditional MDC systems, because the
removed salts didnt transfer from the desalination chamber to the anode and cathode
chamber but rather got adsorbed onto the ACC electrodes. The conductivity of the
anolyte and catholyte were kept quite stable, which prevented significant salinity changes
that might affect effluent reuse. The abiotic control reactor without microbial activities
showed no current generation or desalination performance, as shown in Figure 2.3. This
finding confirms that the potential generated by the microbial exoelectrogenic activities
was the driving force of desalination.
Figure 2.3 cMDC Change in Conductivity
The initial and final conductivity for the anode desalination and cathode chamber for the active cMDC
system as well as a negative control reactor.
2.4.2 Change In pH Over The Course of Desalination
While the initial pH of all three chambers was measured at 6.8, the final pH after
a batch cycle varied between chambers. As shown in Figure 2.4, the desalination chamber
pH slightly decreased to 6.4 after a batch cycle. The anolyte pH dropped to 5.9, while the
20


catholyte pH increased to an average of 7.9. Such pH variations have been reported by
previous MDC studies, because the accumulation of protons in the anode chamber caused
pH drop, while the loss of protons in the cathode chamber due to water formation led to
pH increase. To alleviate pH fluctuation, electrolyte recirculation can be implemented to
neutralize the anolyte and catholyte, and further studies are underway to address such
problems.4,30 The negative control without microbial activities showed no pH changes,
confirming no electrons or protons were transferred in the system.
Initial
Final
Negative
Control
Figure 2.4 Change in pH for cMDC System
The initial and final pH for the anode, cathode and desalination chambers of the cMDC system as well
as a negative control reactor.
2.4.3 The Potential And Challenge Of The cMDC Configuration
Compared with traditional MDC systems, where the salt removal from the middle
chamber is accompanied by the salinity increase in the anode and cathode chambers, this
cMDC configuration is able to incorporate capacitive deionization with microbial
desalination and captures salts on electrodes without releasing the salts to the electrolytes.
This integrated process addressed the concerns of increased salinity on cMDC effluent
reuse and provides a new approach for more complete salt management. Moreover, it can
21


simultaneously achieve organic removal and energy production together with
desalination, demonstrating a great potential for the water industry.
By directly using high surface activated carbon electrodes, the cMDC was able to
adsorb ions on the double layer capacitor formed on the electrode surface. However, it is
not clear how the increased ion concentration might affect anode biofilm activity and
community on the electrode. During this study, manual cleaning and regeneration of
electrodes was performed after reactor operation, which prevented a mass balance or salt
recovery calculation. It also impacted the microbial community evident in the last cycle
of Figure 2.2, which showed lower current production. Further studies will be conducted
to develop in situ electrode regeneration methods. One possible method could be to
develop a reactor that can switch the ACC electrodes in situ once they have become fully
adsorbed. Switching the electrodes would cause the adsorbed salts to desorb due to the
reverse potential, which would solve the problem of having to manually regenerate the
27
electrodes.
2.5 Conclusion
This study presents a step forward in sustainably desalinating salt water with a
capacitive microbial desalination cell. The cMDC reactor was capable of removing an
average of 69.4% of the salt from the desalination chamber through electrochemical ion
adsorption on the electrodes without adding salinity to the anode or cathode chamber.
The physical and electrical adsorption capacity of the ACC electrodes was between 61
and 82.2 mg/g ACC. Further studies are needed to improve system efficiency and
develop in situ ACC regeneration process. By combining this process with optimized
reactor configurations a sustainable method of desalination can be obtained.
22


3. Ion Migration and Control of pH Improved Through the Use of a New Reactor
Design Called a Microbial Capacitive Desalination Cell2
3.1 Abstract
Microbial desalination cells (MDCs) use the electrical current generated by
microbes to simultaneously treat wastewater, desalinate water, and produce bioenergy.
However, current MDC systems transfer salts to the treated wastewater and affect
wastewater beneficial use. A microbial capacitive desalination cell (MCDC) was
developed to address the salt migration and pH fluctuation problems facing current
MDCs and improve the efficiency of capacitive deionization. The anode and cathode
chambers of the MCDC were separated from the middle desalination chamber by two
specially designed membrane assemblies, which consisted of cation exchange membranes
and layers of activated carbon cloth (ACC). Taking advantage of the potential generated
across the microbial anode and the air-cathode, the MCDC was capable of removing 72.7
mg total dissolved solids (TDS) per gram of ACC without using any external energy. The
MCDC desalination efficiency was 7 to 25 times higher than traditional capacitive
deionization processes. Compared to MDC systems, where the volume of concentrate can
be substantial, all of the removed ions in the MCDC were adsorbed in the ACC assembly
double layer capacitors without migrating to the anolyte or catholyte, and the electrically
adsorbed ions could be recovered during assembly regeneration. The two cation exchange
membrane based assemblies allow the free transfer of protons across the system and thus
prevented significant pH changes observed in traditional MDCs.
The work presented in this chapter has been published by Casey Forrestal, Pei Xu, and
ZhiyongRen. Energy Environ. Sci. 2012, 5, 7161-7167
23


3.2 Introduction
The increasing awareness of the water-energy nexus is compelling the
development of technologies that reduce energy requirements during water treatment as
well as water demands for energy production. Microbial desalination cells (MDCs)
recently emerged as a promising technology to simultaneously treat wastewater,
desalinate saline water, and produce renewable energy such as electricity or hydrogen
gas.3, 14, 18, 33-40 MDCs share the same principle of bioelectrochemical reactions with
microbial fuel cells (MFCs): electrochemically active bacteria in the anode chamber
oxidize biodegradable substrates and generate electron flow (i.e. current) to reduce the
electron acceptors in the cathode chamber. The additional desalination function can be
achieved in an MDC by adding a middle chamber containing saline water and utilizing
the anode-cathode potential difference to drive the migration of anions (e.g., CT) to the
anode chamber and cations (e.g., Na ) to the cathode chamber for charge neutrality. The
MDC process carries great potential in desalination systems, because it can either be used
as a stand-alone process or serve as a pretreatment for conventional desalination
processes such as reverse osmosis (RO) to reduce salt concentration of RO feed, and
minimize energy consumption and the membrane fouling potential. Current desalination
technologies, such as RO and electrodialysis (ED) are energy and capital intensive. Even
the most advanced large scale seawater RO units require 3-7 kWh/m for water
desalination, while conventional multi-stage flash evaporation requires 68 kWh/m3.40 In
contrast, the MDC system is considered to be an energy gaining process, because it
converts the biochemical energy stored in wastewater to electricity or hydrogen gas. Lab
scale MDC studies showed that 180-231% more energy can be recovered as H2 than the
24


reactor energy input when desalinating 5-20 g/L NaCl solutions,3536 and a recent study
calculated that a liter-scale MDC can produce up to 58% of the electrical energy required
by downstream RO systems.34
Current MDC systems use an anion exchange membrane (AEM) to separate the
anode and middle chamber, and a cation exchange membrane (CEM) to separate the
cathode and middle chamber. Similar to electrodialysis, desalination in MDC is achieved
by direct transport of salts from the middle chamber to the anode and cathode chamber.
Such system faces two main problems. While salts get removed from the middle
chamber, they become concentrated in the anode and cathode chambers, resulting in an
increase of the volume of saline water. This concern becomes more imperative when
wastewater is treated as the anolyte. Although the addition of ions (or total dissolved
solids, TDS) increases wastewater conductivity and benefits electricity generation, the
increased salinity may affect effluent water quality and prevent subsequent beneficial use
of treated wastewater.39, 41 The high salinity may also affect wastewater treatment
efficiency in MDCs because studies showed that high chloride concentration is inhibitory
to biological treatment, especially for nutrient removal.41 In addition, the AEM between
the anode and middle chamber inhibits the free transfer of H+ accumulated in the anolyte
to other chambers, which causes a significant pH drop in the anode chamber and pH
increase in the cathode chamber.30, 42 A previous study showed that the pH of the
wastewater anolyte dropped to 4.2 in one batch cycle if no buffer was added to the
anolyte. Additionally, the catholyte pH could increase to 11-13 due to the loss of H
42
Such pH fluctuation significantly inhibits bioelectrochemical reaction efficiency and
reduces system performance.
25


In order to modulate the movement of salts to the anode and cathode chambers,
the concept of capacitive deionization (CDI) was incorporated in this study to develop a
sustainable desalination system called a microbial capacitive desalination cell (MCDC).
In the proof-of-concept MCDC, salt water can be deionized through electrochemical ion
adsorption driven by the electrical field generated by microorganisms. Two activated
carbon cloth (ACC) membrane assemblies were designed to connect with the anode and
cathode and adsorb ions from water. During desalination, the ions are stored in the
electrical double layer capacitors between the solution and the ACC assembly interfaces,
thus preventing the salinity increase in treated wastewater. After the ACC is saturated
with adsorbed ions, the assembly can be regenerated by removing the electrical potential
and the retained salts can be fully recovered in situ for disposal or further salt recovery.
Another innovative aspect of the MCDC as compared to conventional MDC, is the use of
a second CEM in lieu of AEM between the anode and desalination chamber (Figure 3.1).
Such configuration allows for cations and protons to move freely from the anode
chamber throughout the reactor and therefore maintains electrochemical neutrality and
prevents pH fluctuation. In this study, the proof-of-concept MCDC development and
operation were demonstrated, and its advantages over current systems and application
potentials were discussed.
3.3 Materials and Methods
3.3.1 MCDC Reactor Design
The MCDC reactors consisted of three polycarbonate cube-shaped blocks with 3-
cm diameter holes forming an internal anode, cathode, and desalination chamber volume
of 23 mL, 27 mL, and 10 mL respectively. The anode and cathode chambers had a length
26


of 4 cm, while the desalination chamber had a length of 1.5 cm. The anode electrode was
a graphite brush (Golden brush, CA) and pretreated by washing in acetone and heating to
350C for 30 minutes.43 Traditional air-cathodes were made by applying 10% Pt/C (0.5
mg/cm2) and four PTFE diffusion layers on 30% wet-proofed carbon cloth as previously
described.44 The desalination chamber was separated from the anode and cathode
chamber by two assemblies. Each assembly was constructed by placing a cation
exchange membrane (CMX-SB, Astom Corporation, Japan), a Ni/Cu Mesh current
collector (McMaster Carr, IL), and 3 layers of Zorflex activated carbon cloth (ACC,
Chemviron Carbon, UK) together. Additionally, the CEM faced the anode/cathode
chamber to prevent microbial growth on the assembly. The total weight of the ACC was
1 gram with the specific surface area of 1019.8 m /g, determined by the Brunauer-
Emmet-Teller (BET) method (ASAP 2020, Micromeritics, Norcross, GA).45 The ACC
assemblies were connected to the anode/cathode by titanium wires (Figure 3.2).
27


n

Anode Desalination Cathode
Chamber Chamber Chamber
Cationr-^ Cation^* * Cation
Anionch ^ Anion g Anion
o
Treated Waste Water Desalinated Salt Water (DESALINATION)
:B3=za
Anode
Chamber
desalination
Chamber
Cation-*
Anion
TF
Cathode
Chamber
Collected Salts
(REGENERATION)
Figure 3.1 Operation of MCDC System
The operation of the MCDC system consists of wastewater entering the anode chamber, and salt water
enter the desalination and cathode chambers. When electrical potential is applied to the membrane
assemblies ions inside the desalination chamber become physically and electrically adsorbed. Flowing out
of the reactor is treated wastewater, and desalinated water. When the system is disconnected from the
anode and cathode chambers the previously adsorbed salts, desorb and can be collected.
3.3.2 MCDC Operating Conditions
Two reactors were inoculated with anaerobic sludge from the Englewood-
Littleton Wastewater Treatment Plant (Englewood, CO) and operated in fed-batch MFC
mode. When a repeatable voltage profile was obtained for consecutive batch cycles, the
reactors were shifted to fed-batch MCDC mode by inserting a pair of assemblies and
adding one middle chamber as described previously. The anolyte growth media contained
per liter: 1.6g NaCH3COO, 0.62g NH4C1, 4.9g NaH2P04 -H20, 9.2g Na2HP04, 0.26g
KC1, and lOmL trace metals and lOmL vitamin solution.46 The catholyte contained per
liter: lOg KC1, 0.68g KH2P04, 0.87g K2HP04. Potassium chloride was used in the
cathode chamber to differentiate with sodium transport and monitor the movement of
cations from the cathode to the desalination chamber. The salt solution in the desalination
28


chamber contained per liter: lOg NaCl, 0.49g NaH2P04 -H2O, 0.92g Na2HP04. A small
amount of buffer was added to the salt solution to some extent mimic the 300 to 700
pmole/kg natural buffering capacity of seawater and prevent potential scaling at high pH
values.47
Anode
Chamber
Desalination
Chamber
Cathode
Chamber

CEM/ACC/NiCu NiCu/ACC/CEM
Figure 3.2 Diagram of MCDC System
Diagram of MCDC reactor configuration and operation. Two CEM-ACC assemblies were used to
separate the three chambers and capture the removed salts, as well as allow for the free transfer of protons.
Two experimental procedures and two controls were performed to investigate the
desalination performance of the MCDC system. The first experiment investigated
simultaneous physical and electrical adsorption capacity by directly adding salt solution
into the desalination chamber equipped with ACC assemblies free of adsorbed ion.
When the anode and cathode electrodes were connected to the ACC assemblies, physical
and electrical adsorption on the ACC assemblies could occur concurrently. The second
experimental procedure investigated only electrical adsorption capacity. Electrical
29


adsorption capacity of the ACC assemblies was determined by first adding salt solution
to the desalination chamber to allow complete physical adsorption. Electrical adsorption
was then characterized by replacing the desalted solution with fresh solution, and
connecting the two assemblies to the anode and cathode, respectively. Any residual water
from previous experimental washing would have been removed when the salt solution
was replaced.
Abiotic control experiments were performed by using new brush anodes without
bacterial acclimation. The first control experiment measured the physical adsorption
capacity by short circuiting the assemblies to ensure no charge was formed across the
electrodes. The adsorption capacity of the assemblies was defined as the change in initial
and final salt concentration. The second control investigated the electrical adsorption
capacity by first allowing complete physical adsorption to occur then by connecting the
assemblies to an external power supply at a voltage of 0.53V to simulate the voltage
generated by a microbial fuel cell. The MDC control experiment used an anion exchange
membrane next to the anode chamber (Astom Corporation, Japan) and a CEM next to the
cathode chamber without ACC assemblies in the desalination chamber. An external
resistor of 1000 Ohms was used between the anode and cathode electrodes, and all other
experimental procedures were identical to the MCDC experiments.
To regenerate the ACC assemblies in situ for all experiments, the assemblies were
either allowed to naturally regenerate or were regenerated by applying an external voltage
to increase the rate of regeneration. The natural regeneration was performed by
disconnecting the anode and cathode from the assemblies and creating a short circuit
between the assemblies with an external wire. Alternatively, an external voltage of IV in
30


reverse polarity was applied to the assemblies by a programmable power source. The
external voltage was applied for 5-10 minutes and followed by short circuiting the ACC
assemblies, as mentioned above, for 20-30 minutes. When the potential difference
reached 0.5mV, the ACCs were assumed to be regenerated, meaning that any
electrically adsorbed ions should have been removed from the electrodes. After
regeneration all electrolytes were emptied and washed with deionized (DI) water to
remove any residual salt remaining in the chambers before starting a new batch cycle.
3.3.3 Analysis and Calculations
Conductivity and pH were measured for all three chambers using a conductivity
meter and pH meter (HACH Co., CO). The change in the reactors internal resistance was
determined through electrochemical impedance spectroscopy (EIS) tests using a
Potentiostat. EIS measurements were performed using the anode as the working
electrode, the cathode as the counter electrode, and a saturated Ag/AgCl reference
electrode placed in the anode chamber. Results were fitted into equivalent circuit models
developed in our previous EIS studies and plotted using Nyquist plots where the ohmic
resistance is defined as the intercept of the Zreal axis. Samples of all three chambers
were collected before and after desalination, and after regeneration. Ion concentrations
were measured using the Optima 3000 Inductive Coupled Plasma (ICP) Spectrometer
(Perkin Elmer, CT) and Dionex DC80 ion chromatography system (IC) (Dionex, CA).
Using the data from the IC and ICP a mass balance of the major ions were determined by
summing the concentrations of the ions in each chamber initially, after desalination, and
after recovery of the salts. Internal power used was calculated using the following
equations:
31


Where P is power in terms of watt hours, V is the voltage, R is the resistance, p is
resistivity, L is length of the resistance, and A is the cross sectional area. Comparisons
between the MCDC and CDI were made based off either presented data, or estimated
from figures in published papers. Comparison to membrane capacitive deionization
(MCDI) was not conducted to do the incompatibility in methodology to the MCDC.
3.4 Results and Discussion
3.4.1 Reactor Desalination Performance
During MCDC operation an electrical potential was generated across the microbial
anode and air-cathode and applied to the two ACC assemblies to form a double layer
capacitor 48-55 (Figure 3.2). The formation of the double layer capacitor has been fully
modeled using the Gouy Chapman-Stem theory.54 The potential drives the ions to move
from the salt solution and adsorb on the activated carbon cloths. The ion adsorption can
be observed proportional to the charge formed between the ACC assemblies (Figure 3.3).
Figure 3.3 shows that in repeated batch cycles that when the potential on the assemblies
increases from 0 to more than 530 mV in each cycle, the solution conductivity in the
desalination chamber decreased by 12-18%, from 18 mS/cm to below 16 mS/cm. The
desalination rate was the greatest at the beginning of each cycle and then decreased
gradually, suggesting the adsorption capacity of the ACC assemblies decreased along
with the increased amount of salt been adsorbed in the assemblies. Salt removal was
characterized by both conductivity, measured using a conductivity meter, and total
32


dissolved solids (TDS) concentration, measured by IC and ICP (Table 3.1). Through
simultaneous physical and electrical adsorption, the MCDC removed 26.9% of the
conductivity or 25.5% of TDS from the desalination chamber in one batch cycle. In
addition, a small percentage of salt was removed from the anolyte (4.4%) and catholyte
(10.4%) as well. This is likely due to the ions being driven across the membranes by the
electrical potential of the ACC assemblies from the anode and cathode chamber then
adsorbed onto the ACC. Further experiments showed that electrical adsorption alone
removed 22.3% TDS from the desalination chamber, which contributed up to 88% of the
TDS removal compared to the combined physical and electrical adsorption experiments.
Table 3.1 MCDC Salt Removal
Both physical and electrical adsorption for the MCDC system. Table results indicate the total removal
capacity of the system in TDS per gram ACC electrode.
Physical/Electrical Adsorption Electrical Adsorption
Desalination Anode Cathode Desalination Anode Cathode
Chamber Chamber Chamber Chamber Chamber Chamber
% removal in conductivity 26.95.1 13.13.8 5.64.4 10.00.2 10.63.5 -2.02.7
% removal in TDS 25.23.6 4.43.6 10.43.6 22.33.6 7.63.6 23.6
Total TDS adsorption (mg TDS/g ACC) 72.7 50.7
Table 3.2 compares the normalized TDS removal between the MCDC and CDI
studies. The results showed that for the same amount of adsorptive material (ACC), the
MCDC improved TDS adsorption by 7-25 times. Both MCDC and CDI use an electric
field between two electrodes that electrochemically adsorb ions, but the high adsorption
from the MCDC may attributes to the unique feature that the MCDC uses the internal
potential generated by microorganisms. Such in situ approach avoided the use of external
33


power supply and circuit and reduced transportation energy loss, so it demonstrated
higher efficiency than traditional CDI processes. The salt adsorption rate in MCDC,
however, is lower than published CDI studies, and that is mainly due to the low kinetics
of the fed-batch operation and the limited amount of ACC available for ion adsorption. In
this study, the MCDC configuration was modified from traditional cubic type MDCs,
which only allowed for a total of 1 g activated carbon cloth being used in the assembly.
This may explain why the amount of salt removed in the desalination chamber was
relatively small. It was calculated that the amount of salt added in the desalination
chamber (114 mg TDS) was drastically beyond the control electrical adsorption capacity
of the ACC (8.5 mg TDS for the lg ACC applied).
Table 3.2 Comparison of Different Capacitive Deionization Methods
The table illustrates the advantage of the MCDC system for capacitive deionization.
Method Electrode Materials
MCDC Activated Carbon Cloth
CDI Carbon Aerogel
CDI Activated Carbon Powder
CDI Activated Carbon Powder
CDI Activated Carbon Powder Activated Carbon Powder with
CDI Microporous Carbon Black MnCVnanoporous Carbon
CDI Composite
CDI Activated Carbon Cloth Activated Carbon Cloth with
CDI Titania
Electrode Net Wh/g mg TDS/g
Distance TDS adsorptive
(mm) removed material Reference #
15 -2.18 50.74 This paper
2.3 +0.21 7.00 17
NA +1 1.95 24
0.1 +1.78 2.88 25
0.1 +1.68 3.11 26
0.22 NA 3.82 27
NA NA 0.10 28
NA +0.52 NA 30
NA NA 4.38 34
Moreover, compared to CDI systems that consume 0.21-1.78 Watt hour external
energy to generate the potential to remove lg TDS, the MCDC system does not use any
external energy but instead utilized the in situ potential difference between the ACC
34


assemblies generated during microbial activities. It was calculated that the MCDC reactor
saved 2.18 Watt hour for lg of TDS removed. That is why in Table 3.2 the net energy
used for the MCDC is negative, indicating that 2.18 Wh/g TDS removed was not
required, while for the CDI systems an external energy of 0.2-1.78 Wh is required for
removal of lg TDS. While the MCDC reactor directly uses generated current for
desalination, it is possible for electricity to be generated by applying an external load
across the ACC assemblies during regeneration. Reactor configuration optimization is
underway to increase the ACC loading and further improve desalination efficiency.
>
E
ro
IP
c
a
0
Q.
>.
Q
E
01


<
O
O
<
Figure 3.3 Correlation of Applied Potential and Deionization
Figure 3.3 demonstrates as the charge across the membrane assemblies increases the amount of
deionization also increase. Results demonstrated for three typical cycles.
Sodium, chloride, potassium, and phosphate accounted for greater than 85% of
the TDS, and their specific concentration changes in the three chambers are shown in
Figure 3.4. In addition to direct capacitive electrical adsorption that caused concentration
decreases in the desalination chamber, a small amount of charged ions migrated from the
35


anode and cathode chamber to the desalination chamber due to the electrical potential or
concentration gradient. However the desalination efficiency for the anode and cathode
chambers is low compared to the salt removal in the desalination chamber due to the lack
of electrical double layer adsorption and the inhibited anion transfer across cation
exchange membranes. Results in Table 3.1 showed that saline water can also be used as
the catholyte and partially desalinated. Further desalination can be achieved by feeding
the treated catholyte to subsequent reactors desalination chamber.
Figure 3.4 MCDC Ion Migration
The results from IC/ICP analysis demonstrate the ion migration in the MCDC system with ions
moving from the anode, desalination, and cathode chambers.
36


Figure 3.5 Electrochemical Impedance Spectroscopy for the MCDC
Results from electrochemical impedance spectroscopy illustrate that the change in internal resistance
illustrates that over the course of desalination the change in internal resistance only changes a small
amount.
The reactor internal resistance measured by EIS at the beginning of the desalination
cycle was on average 8.5 Ohms. After desalination, the internal resistance increased to an
average of 13 Ohms. The change in conductivity in the desalination chamber correlated
closely with the change in internal resistance for the reactor over the course of
desalination. The MCDC reactor ability to transfer electrons was not inhibited as occurs
over the course of desalination in standard MDCs. It is theorized that this is due to the
MCDC ability to maintain charge neutrality better than in MDC reactors. Because in
standard MDC reactors charge neutrality is reached by ion migrating out of the
desalination chamber, while in the MCDC reactor charge neutrality is performed by ion
migrating through the entire reactor.
37


600
Time (min)
Figure 3.6 MCDC Regeneration
The regeneration or collection of salts in the MCDC system was investigated using two methods. The
red line illustrates regeneration using a IV external power supply. The blue line show the regeneration
when the ACC assemblies are comiected in short circuit.
3.4.2 Assembly Regeneration and Salt Recovery
The ion saturated ACC assemblies were regenerated using two approaches. The
natural regeneration was accomplished by directly connecting the two assemblies in short
circuit. The electrical potential across the assemblies was dissipated with the adsorbed
salts being released back into solution. When the potential difference across the ACC
assemblies reached 0.5mV, it was assumed that the ACC assemblies were regenerated
with complete electrical salt desorption. The regeneration rate can be significantly
increased by connecting the assemblies to an external power supply of IV with reverse
polarity to facilitate ion desorption (Supporting Information). Figure 3.5 shows that
among the four major ion species, almost all of the electrical adsorbed salts were
38


recovered during assembly regeneration, shown as a direct correlation between the initial
and recovered salt concentrations. The capability of in situ regeneration of the ACC
assemblies is another advantage of the MCDC, because the assemblies can be reused
many times without investing significantly in materials. Almost all of the adsorbed salts
can be recovered in concentrates during regeneration, and the recovered salts can be
dewatered or extracted for beneficial uses. Furthermore, MCDC stacks can be developed
and integrated with reverse electrodialysis (RED) to capture the energy generated due to
the salinity gradient across the concentrate and freshwater.56 57 The current MCDC is
operated in batch mode, and the desalination and regenerated processes were conducted
sequentially. More efficient operation can be achieved by connecting multiple reactors in
series or in parallel and operating them in complementary sequential batch reactor (SBR)
modes. While some of the units perform desalination, others conduct assembly
regeneration at the same time. This operation not only provides continuous flow of
produced freshwater but also allows for the direct usage of the electricity produced from
regeneration units for desalination units.
39


Figure 3.7 MCDC Ion Recovery
The amount of recover ions from regeneration are indicated in Figure 3.7 for the four major ions
investigated. It can be observed that almost 100% of the desalinated salts can be recovered.
3.4.3 Reduced pH Fluctuation
Figure 3.8 shows the change in pH units among the three chambers over one
typical batch cycle for both the MCDC and the control MDC. The initial pH values in the
chambers were all within 7.0 + 0.2. The change in pH for the anode chamber in both the
MCDC and the MDC was relatively small with a drop in pH of between 0.2 and 0.5 pH
units, which was presumably attributed to the high buffering capacity of the anolyte.
However, the catholyte had drastically different results between the MCDC and MDC,
with the MCDC increasing in pH on average 1.5 pH units and the MDC increasing 4.4
pH units. Interestingly the change in pH for the desalination chamber for the MCDC is
greater than for the MDC control. Previous capacitive deionization studies showed that
water electrolysis may cause slight pH variation at low voltages, which may explain the
40


pH increase in the MCDC desalination chamber.54 It is difficult to compare the MCDC
results directly with CDI studies, because no known CDI experiments have been
conducted at a set potential lower than 0.6 V.47 49,5052-55 Further investigations should
explore the cause of this phenomenon.
MCDC
MDC
Figure 3.8 Change In pH for MCDC
Figure 3.8 show the change in pH values for the MCDC compared to the MDC system. The MCDC
has reduced pH fluctuations in the anode and cathode chambers. There was a slight increase in the pH of
the desalination chamber.
Because the average percent change between the cathode and desalination
chamber were essentially the same, it is assumed that the proton transfer capability of the
reactor was not inhibited. The MCDC employs a CEM to separate the anode and
desalination chamber. This is different from the AEM used in current MDCs and releases
the pH fluctuations in the reactor. In traditional MDCs, anions (CT) migrate from the
desalination chamber to the anode chamber to compensate for the accumulation of H+,
41


but because the AEM prevents the transfer of H+ out of the anode chamber a decrease in
pH is observed. By using a CEM, the accumulated H+ not only can transfer to the
desalination chamber but also can transfer further to the cathode chamber and therefore
solves the pH change problem in the entire MCDC reactor. Previous studies show that
other ions such as Na+ and K+ also play important roles in maintaining charge balances
across different chambers in microbial fuel cells, but the majority of such ions are
adsorbed in the ACC assemblies so electrolyte charge balance due to ion transfer is not a
concern in the MCDC.
3.5 Conclusion
The integration of capacitive deionization with microbial desalination provides a
sustainable solution that not only addresses the salt migration and pH fluctuation
problems facing current MDC systems, but also improves salt removal and energy
efficiency compared to CDI systems. Traditional MDCs remove salts from the
desalination chamber, but they also add TDS to the anode and cathode chambers and may
increase the volume of saline water significantly, depending on different operation
configurations. The MCDC reactor demonstrated that desalination can be accomplished
in the middle chamber without adding salts to the anolyte and catholyte, and therefore
released the concerns on the viability of wastewater treatment and reuse due to increased
TDS concentration. This proof-of-concept system also demonstrates a microbial
desalination reactor to reduce salinity in all three chambers of the reactor. The MCDC
system offers a sustainable desalination, renewable energy production, and wastewater
treatment. To maximize the benefits and prevent negative effects of salinity changes on
the wastewater anolyte, salt migration from the desalination chamber could be modulated
42


by constructing modular plate-shaped ACC-membrane assemblies. If the added salt is
desired in wastewater to improve the anolyte conductivity, regular MDC operation could
be performed. If the salt should be prevented from migrating into the anode chamber, the
modular ACC assembly plate can be inserted into the reactor to perform salt adsorption.
Such system integration and operation will provide microbial desalination systems great
flexibility in salt migration management as well as better pH fluctuation control.
Despite the potential benefits offered by the MCDC system, many challenges
remain to be addressed based on the information collected from this proof-of-concept
study. In addition to low-cost material development that is required for all
bioelectrochemical systems, the adsorptive material can be improved such as with silica
or titanium modification.59 60 Reactor configuration needs to be optimized to provide
more ACC loading and improve diffusion rate and adsorption capability. Modular stack
reactors and flexible operation strategies need to be developed to maximize the
integration of desalination and assembly regeneration in multiple units, optimize water
recovery as well as enhance salt migration management. The cost of the MCDC system is
proportional to the desired level of desalination. The MCDC system is cost competitive
for low to high range brackish water (<20g TDS/L), above that level of salt concentration
other more developed desalination technology would be more cost beneficial.
Improvements in MCDCs will also benefit from the continued advances of other
bioelectrochemical systems such as microbial fuel cells, and capacitive deionization, with
the eventual goal of developing a full scale sustainable system directed toward the
integration of multiple functions, such as extracting energy from wastewater and water
desalination.
43


4. Investigations of Different Methods for the Construction and Operation of
Spirally Wound Microbial Electrochemical Systems3
4.1 Abstract
Microbial electrochemical systems (MES) are a platform technology designed to
generate electricity, treat wastewater, desalinate salt water, and produce chemicals. In
order for MES technology to be scaled for practical application, further research needs to
be conducted to ensure sustainability. The MES reactor design and configuration greatly
affects its sustainability. Demonstrated here for the first time are three spiral wound
(SW) reactor configurations that incorporated a fully wound anode cathode and
membrane assembly. Spiral wound I (SWI) is a single chamber air cathode reactor,
which achieved a maximum power point (MPP) of 27 W/m Spiral wound II and III
(SWII, SWIII) are two chamber reactors using potassium ferricyande as the catholyte.
SWII used two cation exchange membranes to enclose the anode chamber, while SWIII
used two anion exchange membranes. SWII and SWIII had the same MPP of 29 W/m .
After a month of operation SWI power density decreased as the polarization resistance
increased. This phenomenon was not observed in SWII and SWIII. Additionally, it was
found that the most favorable operating condition for SWI was at the longest retention
time of 5.37 hours. Results presented here indicate that a two chamber spiral wound
reactor using anion exchange membranes would yield the most sustainable spiral wound
configuration.
The work presented in this chapter is in preparation for publication by Casey Forrestal,
Pei Xu, Peter Jenkins, and Zhiyong Ren. Water Res.
44


4.2 Introduction
The world is currently facing complicated challenges associated with meeting the
need for water and energy. Over one-third of the world has an inadequate supply of safe
drinking water, and rising energy consumption around the world is depleting our natural
resources. The water energy nexus is a relationship between these two seemingly separate
topics. Energy is required for the production and treatment of water, and water is required
for the production of energy. One clear indicator of this relationship is in a newly
emerging technology known as Microbial Electrochemical Systems (MES). In the last
two decades bioelectrochemical systems, and related technologies, have been extensively
researched because of their unique ability to address the water energy nexus with a single
technology.61 MES technology works by utilizing microorganisms to generate an
electrical current by breaking down an organic or inorganic source of an electron donor.
Numerous electron donors have been investigated including municipal wastewater,
benthic sediments, and industrial wastewater. Using wastewater as the source of
electricity production in a MES would convert the current nexus of requiring energy to
treat to an energy gaining system. Furthermore, because MES generate an electrical
potential, many additional value added benefits can be obtained such as the production of
hydrogen gas or the removal of ions through desalination.62
In order to achieve the greatest benefit from MES technology, electricity
produced and wastewater treated must be low cost and highly efficient. To accomplish
this goal researchers have focused on optimizing reactor materials and reactor design.
The basic MES consists of three parts: a container, electrodes, and separators. The
container is the external reactor shape or housing for the reactor. The most commonly
45


researched containers consist of a polycarbonate plastic in a cube shape.63 The main
purpose of the electrode is for generating or accepting electrons; however, MES
electrodes have been used recently for other purposes such as desalination of salt water or
the production of chemical.63,64 The need for a separator in a MES is specific to the
purpose of the MES. Generally the name separators in MES research refers to
membranes that prevent fluid or oxygen transfer throughout the reactor.65 In addition to
the materials used for the construction of the reactor the overall design is crucial for
increasing the reactor performance. By improving the system design the overall cost of
the reactor can be reduced, increasing the wastewater treatment efficiency and the
amount of energy generated. With these improvements, MES technology will be able to
move from the lab scale to commercial scale. There have been numerous MES designs
that have been developed and tested, such as the cube shape, plate and frame, tubular,
serpentine, upflow, and helical reactors.66'71 Recently, a reactor design was published
called a spiral wound by Boyang Jia et al. in 2012. Jia et al. 2012 was the first to
demonstrate the spiral wound design had potential advantages in wastewater treatment
and energy production. The spiral wound design has been developed in many
applications related to energy and water production such as in lithium batteries and
reverse osmosis membrane systems. Due to its compact nature the spiral wound design
has several benefits, including a high surface area to volume ratio, reduced internal
resistance, and a smaller area footprint. This study investigates three different spiral
wound reactors, and the potential challenges and advantages of each configuration. The
spiral wound configuration published by Jia et al. 2012 consisted of two spirally wound
membranes with an anode electrode within the membranes to form the anode chamber.
46


The cathode electrode was not spirally wound, rather consisted of three pieces of
platinum coated carbon paper placed next to the spirally wound anode membrane
assembly. Air was pumped to supply dissolved oxygen to the cathode electrode. This
paper presents for the first time a fully wound single chamber spiral wound air cathode
reactor (SWI) where all reactor materials were spirally wound. Two additional reactor
designs were investigated, consisting of a two chamber membrane spiral wound reactor
using potassium ferricyanide as the catholyte (SWII and SWIII). Investigated here for
first time are the advantages of using a cation exchange membrane (SWII) versus an
anion exchange membrane (SWIII). The spiral wound configuration provides an
efficient, compact, and scalable design for further development in commercial
applications. Additionally, because of the modularity of the reactor design it has the
potential to be integrated for desalination and hydrogen gas production. Comparison of
the three SW reactors and performance characteristics can bee seen in Table 4.1
Table 4.1 Comparison of Spiral Wound Reactors
The table outlines the three spiral wound reactor configurations and performance characteristics.
Reactor Anode Surface Area (cm2) Cathode Surface Area (cm2) Anode Volume (mL) Cathode Volume (mL) Membrane Type Ohmic Resistance (Q) Polarization Resistance (Q) Maximum Power Point (W/m3) Coulombic Efficiency (CE%)
SWI 245 245 32 NA NA 15 45 27 12
SWII 310 310 46 490 CEM 25 30 29 6.5
SWIII 310 310 46 490 AEM 27 50 29 6.4
47


4.3 Materials and Methods
4.3.1 Spiral Wound Configurations
4.3.1.1 Spiral Wound I Design
The MES spiral wound reactor was designed for use as a single chamber air
cathode. (Figure 4.1) The one exterior reactor wall was constructed by cutting a SKC gas
sampling bag coated with Tedlar PVF (Dupont). This prevented oxygen from penetrating
the anode chamber from the one exterior chamber wall. The anode electrode was made
from activated carbon cloth (ACC) (FM70, Chemviron Carbon) and was pretreated with
acetone to remove residual contaminants. The projected surface area of the SWI anode
was 245 cm2 (12.7 cm x 11.4 cm) (Figure 4.1). A U channel was formed in the SWI
for anolytle fluid flow using Plummers Amazing Goop (3.2 cm x 7 cm). A small piece of
titanium wire was sealed to the edge of the ACC and placed through the outer wall.
Anolyte
Effluent
Figure 4.1 Diagram for Spiral Wound I
A) Anolyte influent fluid flow into the reactor through a center tube across the anode electrode in a
U shape caused by a silicone barrier and exits the reactor through the center tube. B) The
expanded layers of the spiral wound I reactor.
Non
permeable
plastic
§3 cathode
48


The anode chamber was created by placing two pieces of plastic Mesh (0.63 mm
x 15 cm x 15 cm) with diamond shaped holes (0.32 cm ) on both sides of the anode
electrode.(Figure 4.1) A traditional air-cathode was made by applying 10% Pt/C (0.5
mg/cm2) and four PTFE diffusion layers on 30% wet-proofed carbon cloth.44 The
cathode electode had the same internal dimensions as the anode with an additional 2.54
cm x 11.4 cm overhang to connect the electrode. Anolyte entered and exited the single
chamber from a center chlorinated polyvinyl chloride (CPVC) manifold pipe which was
30.5 cm in length by 1.27 cm inner diameter. The manifold was fabricated to allow
influent and effluent fluid flow by cutting the pipe in half and sealing one end of each
pipe with epoxy. Then the two pieces were joined together with additional epoxy.
(Figrue 4.2b) Four 0.64 cm holes were drilled into each side of the center tubing at 0.64
cm apart. The four holes were designed to be used as an inlet and exit to the reactor. The
working volume of the reactor was 32 mL. The reactor was rolled to a width of 7 cm. A
reference electrode port was added on the edge of the rolled reactor by inserting a
polyvinyl chloride (PVC) tube (O.D 0.9cm, I.D. 0.6cm) between the two pieces of Mesh
and the outer wall. The PVC tube was plugged with an 11 mm septa (Restek) during
operation. Air was allowed to naturally pass across the wound air cathode. The air
cathode electrode and the oxygen-impermeable plastic layer were sealed together using
silicon. Before adhering the plastic, the surface was scratched using sand paper to
increase the surface area. Silicone was placed on the inside and outside of both layers to
ensure a tight seal. (Figure 4.2)
49


Figure 4.2 Pictures of Spiral Wound I
A) The folly sealed air cathode spiral wound showing the influent and effluent tubing. B) The center
tubing with holes drilled for anolyte influent and effluent.
4.3.1.2 Spiral wound II and in Design
Spiral wound reactor version II and III were identical except for the use of the a
Neosepta cation exchange membrane (CEM) from Astom incorporated for SWII and an
anion exchange membrane from Membranes International for SWIII.(Figure 4.3) SWII
and SWIII were two chamber MESs, where the anode chamber was formed by two
membranes. All components of reactors SWII and SWIII were fully wound including the
anode electrode, membranes and cathode electrode. The reactors consisted of two 1.27
cm inner diameter CPVC pipe, 35.5 cm in length. In both pipes at 12.7 cm from the end
three 0.47 cm holes were drilled into one side of the pipe at 2.54 cm margins. Both ends
of both pipes were tapped with a 0.63 cm #18 NPT tap. One pipe was used as the
influent for the anolyte while the second pipe was used for effluent flow. Two
membranes were sealed in a spiral wound fashion with the center pipe used as the
influent pipe and the outer pipe as the effluent pipe. The anode and cathode electrode
dimensions were 30.5 cm x 10.2 cm for a total surface area of 3. lx 10' m Both anode
50


and cathode electrode material were activated carbon cloth (ACC) (Chemviron Carbon,
UK).
Figure 4.3 Picture of Spiral Wound II
Spiral Wound II fully wound and sealed. A space between the two sealed membranes provides a
channel for catholyte fluid to flow.
A total of 4 spacers were used for the reactor; one on the outside and inside of the
anode electrode, and two spacers were placed on the outside and inside of the cathode
electrode. The spacers were made of a nylon plastic mesh as described for SWI. The
dimensions for the membranes were 33 cm x 15.2 cm on the inside, and 35.5 cm x 15.2
cm on the outside. (Figure 4.4) Each of the four spacers were 2.5 cm different in length
starting from the inside at 24 cm, 26.5 cm, 29 cm, and 31,5cm. The anode chamber was
sealed by using a combination of 3M wet surface epoxy and marine epoxy. Adhesive
was placed on the inside and the outside of the anode chamber membranes. On the inside
of the membranes, a 1.2 cm width adhesive seal was used on the top and bottom to seal
the anode chamber. Additionally, a 1.2 cm width adhesive seal was applied to the outside
of the membranes on both the top and bottom to ensure a good seal. The surface of the
membranes were scratched using a piece of sand paper to increase surface adhesion. The
total volume of the anode chamber was 45.8mL.
51


Figure 4.4 Computer Aided Drawing of Spiral Wound II and III
The computer aided drawing shows what the ideal reactor should look like where the center tube is
used for influent and the tube to the left of the center tube is used for the effluent. The space between the
wound membranes allows for catholyte fluid to flow.
Electrical connections were made by crimping a piece of titanium wire to the
anode and cathode electrode and threading the wire to the outside of the reactor. The
fully formed anode and cathode assembly was tested for leaks before startup by using
combination air and water pressure tests. The wound anode and cathode assemblies of
SWII and SWIII were designed to be used with a liquid catholyte such as ferricyanide.
Catholyte fluid passively flowed on both sides of the anode chamber. (Figure 4.5) In
order for this passive catholyte fluid flow to occur, the spirally wound anode chamber
and cathode electrode assembly was placed inside a 7.6 cm diameter x 20.3 cm PVC pipe
forming the cathode chamber. The cathode chamber was sealed at the top and bottom
with 7.6 cm diameter PVC caps. Catholyte entered the cathode chamber from one end of
the 7.6 cm diameter PVC cap through a 0.635 cm tapped hole and exited in the same
manner on the opposite side of the reactor. The influent and effluent pipes of the spirally
wound assembly were threaded through 5 cm holes drilled in the top and bottom caps,
52


and sealed with epoxy.(Figure 4.6) The total catholyte volume was 490mL. Both reactors
were horizontally fixed to a piece of plywood for operation.
Figure 4.5 Internal Diagram of Spiral Wound II and III
The drawing shows how the spiral wound II and III operates with influent and effluent. The layers of
the wound material can also be seen, which consists of the anode, spacers, and membranes.
4.3.2 Reactor Start-up and Operation
4.3.2.1 Spiral Wound I Operation
The reactor was acclimated with microbes from a previously acclimated MES as
well as by adding activated sludge from the city of Broomfield wastewater treatment
facility. At startup, the anolyte solution was stationary inside the reactor. After the
voltage reached greater than -lOOmV at 1000 Ohm external resistance, the reactor was
switched to batch with continuous fluid flow. Anolyte was pumped initially at startup
with a with a peristaltic variable speed pump, of 0. lmL/min. After the reactor became
fully acclimated with a voltage of greater than 500 mV at 1000 Ohm resistance, anolyte
fluid flow was varied from 0. lmL/min to 1.5mL/min (Fisher Scientific Co., Fairlawn,
o
'o
o
o
Spacers
53


NJ). The anolyte contained per liter: 20mM sodium acetate (1.6g/L), 11.6mM (0.62 g)
NH4C1, 35.5mM (4.904 g) NaH2P04 -H20, 64.5mM (9.15 g) Na2HP04100mM, 3.4mM
(0.26g) KC1, and lOmL trace metals, and lOmL vitamin solution. An external resistor of
1000Q connected the reactor and was used for all experimental conditions. A saturated
Ag/AgCl reference electrode was used for electrochemical analysis. The anolyte volume
used for batch tests was 250mL.
Figure 4.6 Operation of Spiral Wound III
Figure 4.6 shows the computer, the pump, electrolyte solution one of the fully sealed spiral wound
reactors in operation. Electrical and fluid connections are made at the ends of the reactors.
4.3.2.2 Spiral wound II and Spiral wound III
SWII and SWIII anolyte media were the same as above for SWI. The catholyte
contained per liter: 16.5g/L (50mM) KFe(CN)6, 4.9g/L (35.5mM) NaH2P04, and 9.1g/L
(64.5mM) Na2HP04 A peristaltic pump, (Masterflex, Cole Palmer) was used for
anolyte and catholyte fluid flow at a flow rate of lmL/min. An external resistor of
1000Q connected the reactor during startup. To start the SWII, activated sludge was
added to the anolyte solution that was pumped into the reactor. Sludge and anolyte
solution, as well as the catholyte solution, was allowed to sit stationary inside the reactor
until the voltage was greater than lOOmV at 1000 Ohm resistance. SWII was further
54


acclimated at a lmL/min flow rate under batch conditions. SWIII was acclimated without
the use of activated sludge. Effluent from the fully acclimated SWII was fed into SWIII
at a flow rate of lmL/min.
4.3.3 Analysis and Calculations
Voltage data was recorded using a data acquisition system (model 2700, Keithley
Instruments, inc. OH) where the voltage drop across the external resistor (Re) was
recorded at a 1 minute interval. Power density was determined by linear sweep
voltametry (LSV) using a potentiostat (Gamry Instruments). The reactors were
disconnected from the external resistor for greater than one hour prior to running LSV.
Then reactors were connected to the potentiostat with the anode as the working electrode,
the cathode as the counter, and reference electrode. The scan rate was set at O.lmV/s,
with a step size of lmV and a maximum current of 10mA. Coulombic efficiency (ecb)
was calculated using the change in Chemical Oxygen Demand (COD)(Hach, Loveland
CO) under steady conditions for batch conditions.
Tb
M \ldt
scb =--------------
Fb VanACOD
Where M is the molecular weight of oxygen (32g/mol), I is current (A), integrated
between the interval between 0 and time Tb, F is Faradays constant (96,485 C/mol), b is
the number of electrons transferred for every mole of oxygen, Van is the volume of the
73
anode chamber ACOD is the change of COD between the influent and the effluent.
The internal resistance of the reactor and the polarization resistance was
determined by potentiostatic electrochemical impedance spectroscopy (EIS) at a
frequency range of 10 to 10' Hz, with a 10 mV sinusoidal perturbation. The anode
55


electrode was used as the working electrode, the cathode electrode as the counter, and
reference electrode. The data was plotted using a Nyquist plot, which indicates the
internal resistance (Rin) defined as the intercept with the Zreal axis, and the polarization
resistance, which is defined as the diameter of the semicircle by the graph. Open circuit
voltage (OCV) was calculated by disconnecting the reactors anode and cathode
electrodes from the external resistance. The OCV is the reactors electron motive force,
or the difference in potential energy of the stabilized anode and cathode electrodes. The
OCV was achieved in about 1 hour under open circuit conditions.
Figure 4.7 Acclimation Times For Different reactor Configurations
The current produced in the spiral wound configurations increased dependant upon the reactor
acclimation process. SW III acclimated the fasted because it used previously acclimated anolyte from SW
II. The maximum current produced was in SW II and SWIII.
56


4.4 Results and Discussion
4.4.1 Electricity Production and Efficiency of Spiral Wound I
Upon start up of the reactor, it reached a maximum current of 540mA. The reactor
was operated under batch continuous flow for greater than 30 days with only a slight
variation in current produced.(Figure 4.7) The voltage would slightly drop because of a
decrease in available COD concentration and the standard deviation for the reactor
energy production varied by 0.035mA. Because the volume was relatively small
compared to the overall size of the reactor, any interruptions in the flow of anolyte
resulted in a quick drop in produced current.
HRT 0.36 hr
1.5mL/min
HRT 1.58 hr
0.34mL/min
HRT 2.15 hr
0.25ml/min
HRT 5.37 hr
0.10mL/min
0
50
100
150
200
A/nr
250
Figure 4.8 Effect of Retention Time On Power Density SWI
Power density curves for the various retention times indicate little change in maximum power point.
The highest power point was achieved at the lowest flow rate 0.1 mL/min.
The maximum open circuit potential was 632mV which is similar other MES
reactors operating with an air cathode. The maximum current produced, however, was
57


only 540mA which indicates that improvements to the reactors construction and
operations could improve power production and reduce ohmic and polarization losses.
Figure 4.9 Coulombic Efficiency Correlated to Retention Time and Maximum
Power Point
Power density curves for the various retention times indicate little change in maximum power point.
CE varied greatly over changing retention times.
The polarization and power density curves for the reactor were determined at varying
hydraulic retention time (HRT). (Figure 4.8) The maximum power density was formed at
a HRT of 5.37 hours or 0. lmL/min, which generated a maximum power point of 27.4
W/m anode volume. (Figure 4.8) However, the maximum power point for the reactor
operating at increasing flow rates had only a negligible effect on the power produced. At
a flow rate of 1.5 mL/min, the maximum power point decreased to 20 W/m Because
the maximum power point for SWI did not decrease linearly, with a linear increase in
flow rate, this indicates that potentially the reactor configuration could allow for a higher
flow rate across the anode electrode without inhibiting the microbes ability to consume
the organic waste and generate an electrical current. The highest CE for SWI was not
58


achieved at the longest HRT of 5.37 hours but rather at the retention time of 0.83 hours.
(Figure 4.9). CE for SWI did not appear to be directly correlated to the maximum power
point. Further research will help to better understand this phenomenon.
Figure 4.10 Change In Polarization Resistance Over Time
From the Nyquist plot, the internal resistance and the polarization resistance can be seen. The internal
resistance of the reactor did not change over time but the polarization resistance changed significantly.
The maximum voltage, and maximum power density for SWI, was produced just
after the reactor was fully acclimated. After the initial period of operation, the OCV and
the voltage produced dropped from 540mV to approximately 510mV over a 50 day
period. This drop in OCV and voltage was correlated to the increase in polarization
resistance as seen in Figure 4.10. The fully acclimated reactor had a starting internal
resistance of 15 Ohm and a polarization resistance of 45 Ohms. After 26 days of
operation the internal resistance remained at 15 Ohms, but the polarization resistance
increased to 65 Ohms. After 49 days of operation the internal resistance remained almost
the same as the starting point at 16 Ohm, but the polarization resistance increased to 85
Ohms. The increase in polarization resistance of 40 Ohms over 50 days of operation can
be directly correlated to the drop in OCV. Because SWI was an air cathode electrode
59


with minimized spacing between the anode electrode and cathode electrode, it is believed
that the increase in polarization resistance was due to the increase in biofouling of the
75
cathode electrode common with traditional air cathode reactors.
Figure 4.11 Spiral Wound II and III Initial Overshoot Potential
The initial overshoot potential was high for the SWII and SWIII indicating that the microbes had not
fully acclimated.
4.4.2 Electricity Production and Efficiency of Spiral Wound II and III
The results for the SWII and III reactors indicate that through improved reactor
construction and with the incorporation of the higher potential electron acceptor
ferricyanide, the maximum current produced for SWII and SWIII was 680mA for both
reactors. SWII was started one week before SWIII. SWII required 6 days to become
fully acclimated using activated sludge. SWIII was acclimated much faster taking only
1.5 days. The faster acclimation was achieved by using the effluent from the fully
acclimated SWII.(Figure 4.7) The maximum current produced in SWIII was 26% higher
than SWI. The maximum power point for the reactors SWII and SWIII was not
immediately achieved, even after operating the reactor at a stable high current for over
60


three weeks. This was indicated in LSV test which showed both reactors having a large
overshoot potential which prevented the observation of the maximum power point.
(Figure 4.11) The maximum power point with the overshoot potential was 20 W/m To
resolve the overshoot potential for both SWII and SWIII, both reactors were operated
under a low external resistance of 2 Ohms for 3 weeks. (Figure 4.12)
Figure 4.12 Spiral Wound III Loss of Overshoot Potential
Overshoot potential was overcome by acclimating the reactor at a low external resistance. The change
in power density can be seen proportional to time.
The maximum power point of the acclimated reactors SWII and SWIII was 29
W/m anode volume after resolving the overshoot potential. By week 4 of operation,
SWII which used a CEM membrane from the Astom Corporation developed a small leak
inside the reactor. This was first identified by a color change in the anolyte effluent. The
normal effluent color was clear-slightly pink, as potassium ferricyanide leaked into the
anode chamber the color slowly changed to a bluish green. The reactor was taken offline
to identify and seal the leak. After an exhaustive effort, the location of the leak could not
61


be identified. Identifying leaks for the spiral wound configuration and sealing the
membranes are two of the most difficult aspects of the spiral wound configuration. The
leak for SWII was later identified by disassembling the reactor. The leak was found
towards the center of the anolyte influent tubing. No further experiments were conducted
with SWII. Unlike SWI, there was no noticeable change in polarization and internal
resistance of SWII or SWIII over 4 weeks of operation. This is likely due to the use of
potassium ferricyanide as the catholyte and completely sealing the anode chamber from
the cathode chamber. However, the internal resistance for SWII and SWIII were larger
than SWI. This is because SWII and SWIII used two membranes to enclose the anode
chamber, thus increasing the internal resistance. The internal resistance for SWII was
approximately 25 Ohms and SWIII was slightly higher at 27 Ohms. (Figure 4.13) The
difference in internal resistance for SWII and SWIII indicates that the Neosepta CEM
membrane from Astom Incorporated led to a smaller resistance than the Membranes
International anion exchange membrane. Additionally, the polarization resistance for the
Astom membrane was smaller than the polarization resistance for the membranes internal
membrane. SWII had a polarization resistance of approximately 30 Ohms while SWIII
had a polarization resistance of approximately 50 Ohms. Even though the reactor
resistance profiles were significantly different, the maximum OCV and the maximum
power point for both reactors were almost identical. This result indicates that using an
anion exchange membrane with a higher polarization resistance and a slightly higher
internal resistance yields almost the same results as using a cation exchange membrane
with a lower internal resistance and polarization resistance. It was theorized that using
the higher priced Astom corporation membrane would yield the highest performance.
62


However, because of the improved performance generally indentified with an anion
exchange membrane, the higher resistance of the Membranes International membrane can
be overcome. SWII and SWIII had a higher internal resistance over SWI because of the
use of two membranes the increase in maximum power point for SWII and SWIII was
still 31% higher than SWI.
Figure 4.13 Comparison of EIS for the Three Spiral Wound Reactors
The Nyquist plot for the three spiral wound reactors indicates a high internal resistance for the SWII
and SWIII.
4.5 Conclusion
The results from the construction and operation of three different spiral wound
reactors indicates that while the use of a two membrane system will increase the internal
resistance of the reactor, it may not necessarily reduce the performance. The maximum
power achieved for the spiral wound reactor is more dependent upon the operating
conditions rather than the configuration. By using the higher strength electron acceptor
potassium ferricyanide, the highest power point was achieved despite the larger internal
resistance. For SWI, the difficulty in constructing the serpentine flow path may be
63


prohibitory for the further development of the spiral wound system. However, because
the SWI serpentine flow path uses the same manifold for influent and effluent, the total
area used can be reduced by up to 25%. For further development of the spiral wound
system this reduction in total area would reduce the cost for installing the system at a
wastewater treatment facility.
SWII and SWIII, in addition to being compared to the single chamber SWI, were
also used to compare the performance of an AEM verses a CEM membrane. The
performance characteristics between SWII and SWIII indicated that while a slight
fluctuation was observed, the results between SWII and SWIII were similar. This
indicates that the use of the more expensive Astom membrane would not be beneficial to
the commercialization of the technology. Additionally, the small leak that formed in
SWII indicates that further research needs to be conducted on membrane sealing
techniques to ensure sustained performance. Ideally, a new reactor design should be
developed where the reactor could be easily sealed and unsealed if necessary allowing
access to replace materials or resolve problems as they arose. Furthermore, all of the SW
reactors used a significant quantity of adhesives to seal the reactors. This increases the
cost of the reactor and may prohibit it from being further commercialized. Further
research needs to be conducted towards developing an adhesive-less spiral wound
reactor, which will reduce the cost of manufacturing and potentially simplify the
manufacturing process.
Currently, the cost of the spiral wound reactor would not make it a competitor
against large scale municipal wastewater treatment facilities. Until the spiral wound
design makes a major breakthrough in energy production, the advantage of the design
64


will be for site specific conditions such as on a boat or a small island. The operation of
the SW MEC for wastewater treatment would need to be performed similarly to
traditional municipal wastewater treatment facilities. The only major difference would be
replacing the aeration basin, or trickling filter, for the SW MES. The biggest potential
advantage for operating a MES system on municipal wastewater would be the reduction
of produced sludge. At the current stage of the SW MES development, energy
production would not be its main business attractant. The compact nature of the
technology and its minimal requirement for operation and maintenance are the biggest
assets.
As indicated, the use of an anion exchange membrane verses a cation exchange
membrane does not significantly alter the performance of the reactor. However, the
configuration of the reactor whether single or two chambers, does drastically affect the
performance of the system. The single chamber reactor had a problem of biofouling on
the cathode which decreased its long term performance. The ideal reactor configuration
appears to be an inexpensive anion exchange membrane in a two chamber system. With
improvements in reactor design and operation, and further reduction of internal and
polarization resistance, the SW MES could achieve higher power densities and increased
coulombic efficiencies. While there are still many challenges to address with the SW
configuration, the small environmental footprint and high surface area to volume ratio,
makes the SW a potential commercially viable product.
65


5. An Efficient Method for Treating and Desalinating Produced Water Using a
Microbial Capacitive Desalination Cell4
5.1 Abstract
A microbial capacitive desalination cell (MCDC) is a simultaneous wastewater
treatment, electricity production, and desalination device. Demonstrated here is a MCDC
system that can efficiently treat produced water without requiring an external energy
source. Produced water (PW) is a term generally associated with the production of water
from oil and natural gas mining. Produced water consists of mainly organic and
inorganic salts. The produced water MCDC system is a three chamber desalination
system which uses capacitive deionization (CDI) to in situ remove dissolved salts and
organics. The MCDC ability to remove organics, salts, and generate electricity is
demonstrated for the first time. Additionally, a maximum desalination capacity was
identified. A proof of concept study using a small cube MCDC system, indicated that the
desalination chamber on average removed 450 mg TDS/L/g ACC/hr, and the anode
chamber removed 40 mg TDS/L/g ACC/hr. The cumulative ions removed in the MCDC
system was 390 mg TDS/L/g ACC/hr accounting for a small fraction of ions that
migrated from the desalination chamber into the cathode chamber. The MCDC removed
COD from the anode chamber at an average rate of 10 mg COD/L/hr. The desalination
chamber had a high COD removal rate of 93 mg COD/L/hr. Additionally with the use of
a charge pump connected to a 2.5V, 12F capacitor, a cumulative sum of 0.054 mJ of
external electricity was recovered from the in situ MCDC capacitor from a single
desalination cycle. Experiments investigating the maximum desalination capacity of the
66


MCDC system indicated that, in three successive cycles, greater than 65% of the TDS in
the desalination chamber could be removed. Additionally, the CDI assembly was able to
remove 83% of the COD in the PW in the desalination chamber. The MCDC system
operating under physical and electrical adsorption had an adsorption capacity of 65 mg
TDS/g ACC. Due to the ability of the MCDC system to simultaneously treat the
organic and dissolved solids content, as well as produce direct electrical current, the
MCDC system is a potential solution for site specific produced water treatment.
4 The work presented in this chapter is in preparation for publication by Casey Forrestal,
Pei Xu, Peter Jenkins, and Zhiyong Ren. Environ. Sci. Technol.
67


5.2 Introduction
Water and energy are directly related in energy production and consumption and
the need for sustainable solutions to generate clean water and energy have never been
more important. Water and energy are also closely related to global conflict, which has
caused the United States to invest heavily in domestic sources of energy.76 Within the
next ten years many countries important to the United States will experience water
problems which will risk instability and increase regional tensions, according to a 2012
intelligence report by the US State Department. One rapidly emerging domestic energy
source is the production of natural gas from hydraulic fracturing. Hydraulic fracturing
works by pumping large volumes of water and chemicals into deep underground shale
rock formations. This process releases the trapped methane which is extracted from the
flowback water. After all of the valuable products have been removed from the flowback
water, the remaining water is referred to as produced water (PW). In the United States it
77
was estimated that up to 2.3 billion gallons of produced water are generated everyday.
More recent estimates indicate that the average daily volume of generated produced water
is around 80 million barrels(bbl) (42 gallons/bbl). The produced water contains a wide
range of hydrocarbons, salts, hazardous chemicals, and naturally occurring radioactive
material, which is highly dependant upon the geologic conditions of the natural gas well.
Salt concentrations range from a few mg/L to over 300 g/L, and hydrocarbons can range
up to 2000 mg/L Total Organic Carbon (TOC). Because produced water contains both
high salinity and high organic content, many traditional wastewater treatment options are
not viable. Some produced water does not require treatment and can be used directly for
beneficial purposes such as crop irrigation, water for livestock, or upstream
68


augmentation. If the PW is too highly concentrated in salts or hydrocarbons, the water
needs to be treated to regulator standards prior to discharge. Discharge requirements are
also highly variable depending upon the location and method of disposal. Because of the
need to treat both organics and inorganics the cost of treating PW can be expensive. The
highest costs observed for the treatment of produced water is related to the transportation
of the PW to a treatment facility. If onsite treatment is viable for the producing well,
multiple stages are required to treat the PW. The first step is to remove the suspended oil
through an air flocculation separator; 100% of natural gas well have this system
installed. Following the removal of the residual oils, large particles are removed though
a centrifuge, hydrocyclone, or filtration process. The remaining organic content is
removed with traditional biological wastewater treatment methods such as continuous
activated sludge, solid retention basin, trickling filter or through the construction of a
wetland. Removal of inorganic salts is most commonly achieved through the use of a
membrane filtration system such as reverse osmosis, microfiltration and ultrafiltration.
The least expensive options for disposal of the PW are reinjection into the subsurface or
surface discharge to an evaporation pond. The cost for this method of disposal ranges
from $0.30-$10/bbl.79 A thermal treatment facility can cost up to $105/bbl. The average
cost of produced water treatment is around $3/bbl but is highly dependent PW conditions.
The main reasons for the high cost to treat PW are energy requirements and waste
77
disposal costs. Sludge disposal can account for 40% of the total cost of PW treatment.
Recently, researchers around the world have been heavily investigating methods
of using wastewater to generate energy. One of these methods is through the use of a
28
Microbial Electrochemical Systems (MES) The MES work by using microorganisms
69


to breakdown organic or inorganic sources of electrons and transferring those electrons to
a terminal electron acceptor such as oxygen. Over the last 2 decades, MESs has been
developed to perform multiple functions such as the treatment of municipal wastewater,
production of hydrogen gas, and desalination of salt water. The microbial
desalination cell (MDC) was a major breakthrough in simultaneously treating
wastewater, generating an energy source, and desalinating salt water. The MDC works
by generating an electrical potential which is used to desalinate salt water contained in a
separate chamber by electrodialysis. Therefore, the salts in a traditional MDC only
migrate from the center desalination chamber to the anode and cathode chambers. To
alleviate this problem, the Microbial Capacitive Desalination Cell (MCDC) was
developed.63 The Microbial Capacitive Desalination Cell (MCDC) used microorganisms
to generate an electrical potential which is applied onto high surface area electrodes
within the reactor in order to capacitively desalinate salt water. The desalination process
of CDI, in the MCDC, has been modeled using the Gouy-Chapman-Stern model, which
was adapted from the double layer model by Herman von Helmholtz in 1883.6 When the
electrical potential is removed, the capacitively desalinated salts are removed from the
high surface area electrodes and are collected. The MCDC system provides a method of
wastewater treatment, energy generation, and desalination without contaminating the
anode and cathode chamber with the desalinated salts. Presented in this paper for the first
time is a proof of concept study in which produced water was used to generate the
electrical potential needed to desalinate the organics and dissolved salts in produced
water, and generate and external electrical current. The development of the PW MCDC
system can potentially transform the treatment of PW from a multistage process system
70


requiring a lot of energy, to an energy gaining process. The PW MCDC system was first
evaluated to remove the organics and salt from the PW. After demonstrating the
capability of the MCDC to treat and desalinate PW, the MCDC experimental conditions
were modified to achieve a maximum desalination capacity in the shortest amount of
time. Additionally, the MCDC system was compared to the traditional MDC system to
evaluate advantages and challenges of both systems.
5.3 Material and Methods
5.3.1 Design of Microbial Capacitive Desalination Cell
The MCDC reactor consisted of three small cubic polycarbonate chambers with
3cm diameter hole forming an internal anode, desalination, and cathode chamber volume
of 23, 12, and 27mL respectively. (Figure 5.1) For the proof of concept study, the anode
chamber was connected to a 200mL reservoir containing raw produced water from a
natural gas well in Colorado. The PW was recirculated into the anode compartment at a
rate of lmL/min (Master flex, Cole Palmer). The total volume for the anode chamber
was 223 mL. The anode and cathode chambers had a width of 4 cm while the
desalination chamber had width of 1.5cm. A carbon brush electrode (Golden Brush, CA)
was used as the anode electrode. The anode electrode was pretreated by washing in
acetone and preheating to 350C for 30 min. The cathode electrode was a 9cm
traditional air cathode, coated with 0.5mg/cm Pt/C (10%) and four PTFE diffusion layers
on 30% Teflon coated carbon cloth.44 Inside the desalination chamber, a capacitive
deionization (CDI) system was used to desalinate the PW. The CDI consisted of two
electrodes each split into three parts and place overlapping to reduce the spacing between
the electrode. Each of the three parts of the electrode contained a central current
71


2
collector (2.5cm ) made of Ni/Cu mesh (McMaster Carr, IL), a total of two pieced of
activated carbon cloth (ACC) were placed (each 2.5cm ) (Chemviron Carbon, UK)on
both sides of the current collector. On the outside of the Ni/Cu/ACC assembly, a fine
plastic mesh separated the CDI electrodes (3cm ). In total, the CDI consisted of six
pieces of Ni/Cu mesh, twelve pieces of ACC, (0.72g) and twelve pieces of fine plastic
mesh. Three pieces of the fully assembled Ni/Cu were welded together to form a single
electrode. Fluid flow between the electrodes was achieved by the design of the plastic
mesh. The space between two electrodes was between 1.2-2mm. The anode chamber,
desalination, and cathode chambers were separated by pieces of CEM (CMX-SB, Astom
Corporation, Japan). No external resistance was used during the operation of the MCDC.
Conductive electrodes conductive
Plate Mesh
Figure 5.1 Operation of MCDC for Produced Water
Diagram of the MCDC shows an expanded view of the capacitive deionization component inside the
desalination chamber.
72


5.3.2 MDC Design
The design of the MDC was identical to the MCDC except the desalination
chamber volume was increased to 15 mL because it did not contain the CDI component.
An AEM (AMX-SB, Astom Corporation, Japan) was placed next to the anode chamber
and a CEM (CMX-SB, Astom Corporation, Japan) was placed next to the cathode
chamber. An external resistance of 1000 Ohms was used during the operation of the
MDC. (Figure 5.2)
R

1 An odi Desalination Cathode

AEM CEM
Figure 5.2 MDC for Produced Water
Diagram of the MDC shows the operation of the system for desalination. No additional components
added to desalination chamber.
5.3.3 Operation of the MCDC and MDC Reactors for Proof of Concept Study
Both the MDC and MCDC reactors were fully acclimated in MFC mode before
converting the reactors to either the MCDC or MDC for desalination. Prior to converting
the reactors, the anode electrodes were acclimated by adding 13 mL of activated sludge
from the Broomfield Municipal Wastewater Treatment Facility along with 10 mL of
anolyte growth media. The anolyte growth media contained per liter: 1.6g NaCFECOO,


0.62g NH4CI, 4.9g NaH2P04 -H2O, 9.2g Na2HP04, 0.26g KC1, and lOmL trace metals
and lOmL vitamin solution.46 After the anode electrode became fully acclimated over a
one month period, the anolyte solution was slowly transitioned to 100% raw PW by
adding 10 mL of raw PW to the anode with the anolyte media described above. This was
done over a two week period until the anode chamber completely used raw PW. After
converting the reactors for desalination, raw PW was also immediately used in the
desalination chamber. For all proof of concept experiments, the catholyte contained per
liter: 9.8 g/L (71mM) NaH2P04-H20 and 18.3g/L (129mM) Na2HP04. The electrolyte
solution was replaced with fresh media for the MCDC prior to starting a new desalination
experiment. Electrolyte solution for the MDC was replaced when the voltage dropped
below 50mV. The MDC reactor was connected to a data acquisition system (model 2700,
Keithley Instruments, inc. OH) the voltage across the external resistor was recorded every
1.1 minutes. The MCDC system was not connected to the data acquisition system
because previous experiments indicated that current interruption affected the MCDC
desalination capacity. Electrical potential of the anode, cathode and CDI assembly was
determined with a portable multimeter. Desalination cycles for the MCDC PW proof of
concept study, were achieved by connecting the anode chamber to the pump and
recirculating 200mL of PW for 25 minutes. During this time PW as added to the
desalination chamber and phosphate buffer to the cathode chamber. The PW in the
desalination chamber was replaced a total of three times prior to starting the experiment
in order to insure that the CDI assembly was fully physically adsorbed with salts. The
MCDC proof of concept study only investigated the capacity of the reactor to remove
salts using electrical adsorption. For the MDC experiments, the PW was not replaced
74


prior to start because the conductivity would not change prior to starting experiment.
Before connecting the MDC and MCDC reactors conductivity and pH measurements for
all three chambers were recorded. At the end of the desalination cycle, all electrolyte
solutions (anode, desalination, cathode) were removed and recorded for conductivity and
pH. The end of the desalination cycle for the MDC experiments was caped at 24 hours.
Desalination for the MDC would have continued past 24 hours of operation. A 24 hour
operation period of the MDC was chosen based off results from the initial desalination
runs. The end of a desalination cycle for the MCDC was determined when the voltage
across the CDI became stable, approximately two hours. The CDI assembly was
regenerated using one of two methods. The CDI electrodes were either short circuited and
regenerated with in one hour, or the CDI electrodes were connected to an external charge
pump (S-882Z24, Seiko Instruments) connected to a 2.5V, 12F capacitor for energy
harvesting from the charged CDI. Energy harvested from CDI was determined based off
the following equation:
E = '/2 CV2
Where E is energy in Joules, C is capacitance in Farads, and V is voltage in Volts.
A control experiment was conducted for the MCDC to determine the ion migration if the
CDI assemblies capacity was removed from operation. MCDC control experiment was
conducted as outlined above, except the CDI assembly was connected in short circuit for
the entire desalination cycle. The anode and cathode electrodes were connected to
generate an electron motive force similar to the MDC.
75


5.3.4 Operation of the MCDC for Maximum Desalination
The MCDC system operating conditions were optimized based off results from
the proof of concept study to investigate the maximum desalination rate of the MCDC
system. The MCDC reactor design was identical to what was described previously. The
anode, desalination, and cathode chamber were all converted into a recirculating fluid
flow system using a Vi pipe to 1/16 tube adapters to attach to the desalination, and
cathode chamber. The anode and cathode recirculating volume were reduced to lOOmL.
The anode and desalination chambers were recirculated at a flow rate of 2mL/min. For
this experiment all three chambers used raw produced water. The time for capacitive
desalination was reduced from 2 hours to 1 hour. Regeneration was achieved by
connecting the CDI in short circuit and pumping deionized water through the desalination
chamber at a rate of lOmL/min. The total volume of regenerated solution was lOOmL.
The MCDC reactor was started by replacing the lOOmL anolyte and catholyte
solution with fresh raw PW, under open circuit conditions. The same anode and cathode
electrodes were used from the proof of concept study and a new CDI system was
constructed following the procedure previously outlined. The anolyte and catholyte
solutions were allowed to equilibrate for 35 minutes prior to connecting the anode and
cathode electrodes to the CDI assembly. Before starting the experiment, the lOOmL of
deionized water (DI) was recirculated through the desalination chamber at 2mL/min and
the CDI electrodes were connected in short circuit. After the anode and cathode chamber
were equilibrated, the lOOmL DI was completely removed and the desalination chamber
was purged of all residual liquid. Following the removal of the DI water from the
desalination chamber, 12 mL of raw produced water was pumped into the desalination
76


chamber and capped for operation. Samples for conductivity, pH, salinity, and organic
content were collected. Open circuit voltage as well as the starting voltage on the CDI
assembly (~0 mV) were measured with the multimeter prior to starting,. Next, the anode
anode/cathode electrodes were connected to the CDI assembly for 1 hour. Following the
one hour of operation, the CDI assembly voltage was measured and 10 mL samples were
collected from the anolyte and catholyte reservoirs. The anode and cathode electrodes
were still connected to the CDI when the 12 mL desalination PW was completely purged
from the desalination chamber. Following the desalination procedure, the lOOmL DI
regeneration solution was connected to the desalination chamber and the CDI assembly
was connected in short circuit. The regeneration solution was recirculated at a rate of
lOmL/min for 20min. During this 20 minute regeneration period, the anode and cathode
chamber were disconnected from the pump in order to not disturb the acclimated anode
microorganisms and the reactor was operated under open circuit conditions. At the end of
the 20 minute the 100 mL regeneration solution was completely removed from the
desalination chamber which signaled the end of the first desalination cycle. The second
desalination cycle begin immediately after the end of the regeneration. lOmL samples
were collected from the anode and cathode chamber, for analysis and voltage potential
were measured. The same 12 mL desalination solution from cycle one was reintroduced
to the desalination as previously described. The experimented concluded at the end of
three cycles. The full 12 ml desalination solution was collected for analysis. The
experimental procedure for maximum salt removal was designed to have the smallest
operating anode and cathode volumes to allow for sample collection while maintaining a
sufficient volume for operation.
77


5.3.5 Analysis
Conductivity and pH was measured for all three chambers for all experiments.
The removal of oxidizable substrates and inorganic chemicals in all three chambers were
determined by testing chemical oxygen demand (COD) (Hach Co, Loveland CO). Total
alkalinity was measured using a HACH alkalinity test kit (Hach Co, Loveland CO). The
dissolved organic content (DOC) was measured for the maximum desalination
experiment using a Sievers 53 IOC Series TOC analyzer. The change in total dissolved
solids (TDS) concentration was determined by measuring cation concentration using the
Optima 3000 Inductive Coupled Plasma (ICP) Spectrometer (Perkin Elmer, CT) and
anion concentration using the Dionex DC80 ion chromatography system (IC) (Dionex,
CA). Prior to TOC, IC, and ICP analysis, the collected samples were filtered through a
0.45 pM filter. Average produced water characteristics are identified in Table 5.1.
Table 5.1 Produced Water Characteristics
Table 5.1 illustrates the characteristics of the untreated, raw produced waters that were used in all
experiments.
Raw Produced Water Characteristics
Analysis Cone.
TDS (mg/L) 15870 290
pH 7.8 0.2
Conductivity (mS/cm) 25 0.15
Alkalinity (mg/L as CaC03) 700 8
Ba (mg/L) 44.4 18.7
Ca (mg/L) 236.6 64.2
K (mg/L) 49.5 1.9
Mg 285 (mg/L) 30.0 2.3
Na (mg/L) 5992.0 82.2
Si (mg/L) 30.1 0.5
Sr (mg/L) 27.4 4.2
Cl (mg/L) 9290.2 241.1
Br (mg/L) 67.1 2.9
P04 (mg/L) 54.4 5.43
DOC (mg/L) 33.5 4.3
78


Figure 5.3 Voltage Profile
Fluctuations in voltage over three typical cycles indicates that variation in operating a microbial
electrochemical system with raw produced water. Arrows indicate when the electrolyte solutions for the
anode/desalination/cathode chambers were replaced.
5.4 Results and Discussion
5.4.1 Proof of Concept MCDC Vs MDC Desalination and COD Removal Capacity
The produced water used in all experiments would be considered in the low range of
salinity and organic content compared to other PW wells. However, greater than 99% of
all unconventional PW wells in the United States contain a TDS value of less than 20 g
TDS/L. Therefore, the results presented here for the MCDC system would be applicable
to most unconventional PW wells. A full analysis of the untreated produced water can be
observed in Table 5.1.
The acclimation of the anode electrode to utilize produced water as a substrate
required that the microbes be slowly transitioned. Without this slow transition, or else
they could not tolerate the high salinity and did not contain the necessary enzymes to
break down the organic content of the produced water. Even after converting the reactors
to MCDC and MDC, the voltage outputs from the reactors were variable as can be seen
79


in Figure 5.3. Data recorded from the data logger for the MDC clearly shows how the
voltage after replacing the electrolyte solutions took time to reach the maximum voltage.
Figure 5.3 also clearly illustrates how energy can be generated over successive batch
cycles. However, unlike traditional MDC experiments, the drop in voltage over time was
not directly correlated to the removal of ions in the middle chamber. Figure 5.5 shows the
rate of change in conductivity for the anode, desalination, and cathode chambers for the
MDC. Over a 24 hour period the change in conductivity was extremely slow. The
conductivity in the desalination chamber decreased on average 1.9 mS/cm, while the
anode and cathode chambers increased 0.8mS/cm and 1.9mS/cm respectively. At the end
of the 24 hour desalination cycle the conductivity in the desalination chamber was still
high. Therefore the decrease in voltage over time illustrated in Figure 5.3 shows that ion
transfer capability was not the limiting factor, rather that desalination would stop due to
the loss of available oxidizable material. Additionally, the rate of desalination in the
MDC was believed to be slow because of the poor kinetics of the MDC reactor. All of the
desalination in the MDC using PW in the anode and desalination chambers was achieved
through electrodialysis. In almost all other MDC studies the anode and cathode chambers
contained salt concentrations that were significantly lower than the desalination chamber,
allowing for concentration gradients to aid in electrodialysis.
The ability of produced water to be used as a substrate for microbial exoelectrogenic
electron transfer is exciting. The maximum current produced in the MDC reactor was
450mA. For a small desalination reactor operating on a relatively unstudied substrate,
450mA is an acceptable current output. With improvements to the reactor design and
reduction of the external resistance, a higher current could have been produced which
80


would have resulted in a higher desalination capacity. However, it would not have likely
made a dramatic affect on the performance of the reactor due to the concentration
gradients across the membrane and the rapid consumption of available oxidizable
material.
For the MCDC reactor, voltage and current profiles cannot be recorded using the data
acquisition system. The voltage drop measured by the datalogger over a short period of
time prevents the CDI from developing a sustained potential. This was discovered over
many failed attempts to use the datalogger. The open circuit potential for the MCDC was
>780mV, but the maximum voltage accieved on the CDI assembly was only 530mV in a
two hour period. Interestingly, the achieved CDI voltage potential like the MDC reactor
fluctuated from batch to batch depending upon the microorganism ability to transfer
electrons at any given point.(Figure 5.4) The greatest variability in applied voltage was
observed at the end of the desalination cycle. This fluctuation could be partially due to
the loss of available organic content but may also be associated with a change in pH and
ion concentration in the anode and desalination chamber. However, Figure 5.4 clearly
shows that despite minor flucations in voltage at any given time, the results for the
applied voltage potential on the CDI assembly were reproducable.
81


Figure 5.4 Charged Formed on CDI
The standard deviation for the charge formed on the CDI assembly of the MCDC system over a single
desalination cycle. The maximum voltage achieved was 0.53 V, but the average maximum voltage
achieved over multiple cycles was 0.5V.
The removal of salts for the desalination chamber and the anode chamber for the
MCDC reactor was drastically faster than the MDC system. The conductivity in the
desalination chamber using only electrical adsorption dropped on average 3.0mS/cm.
The anode chamber also decreased slightly: O.lmS/cm in two hours with the cathode
chamber slightly increasing by 0. lmS/cm. It can be observed from Figure 5.5 that the
fastest change in conductivity was achieved in the desalination chamber using the MCDC
system with 0.8 mS/cm/hr desalinated in desalination chamber and O.lmS/cm/hr
desalinated in the anode chamber. When the MCDC CDI assembly was connected in
short circuit, as was in the MCDC control experiment, the performance of the
desalination chamber dropped to 0. lmS/cm/hr, the same as the desalination rate for the
MDC. This indicates that it was the electrical adsorption that led to the faster
desalination rate. Because the MCDC uses two cation exchange membranes, ions migrate
from the anode chamber to the desalination chamber and to the cathode chamber to
82


balance the transfer of electrons from the anode to the cathode. This causes the cathode
chamber for the MDC and the MCDC to increase in conductivity at a rate proportional to
the migration of salts from the desalination chamber to the cathode chamber. The MCDC
desalination chamber ion removal rate is the fastest because it is actually employing two
methods of desalination: capacitive deionization and electrodialysis. Accounting for all
of the ions removed in the MCDC system accounting for volume the reactor removed an
average of 20 mg TDS/L.
1.6
c 1.4
£
i 1.2
o 1
(0
E, 0.8
o
+4 2 0.6
ra 0.4
>
o 0.2
£
£ 0
c o -0.2
-0.4
Anode
Desal
MDC
MCDC
MCDC Negative Control
1
Cathode
Figure 5.5 Change in Conductivity for MDC and MCDC
Change in conductivity for the anode, desalination, and cathode chamber; positive values indicate salt
removal and negative values indicate an increase salt concentration.
The pH fluctuation between the three chambers also showed stark differences
between the MDC and MCDC reactor shown in Figure 5.6. The pH in the MCDC
decreased in the anode chamber by 0.01 pH units/hr while the MDC increased by 0.08
pH units/hr. However, the MDC had little change in pH in the desalination chamber,
while the MCDC decreased 0.08 pH units/hr. Also, in the cathode chamber the MDC
only increased 0.03 pH units/hr while the MCDC increased at rate of 0.04 pH units/hr.
83


These results indicate that by using the MCDC system, the pH of the anode and cathode
chambers can be potentially further stabilized. The biggest change in pH is in the
desalination chamber of the MCDC. What exactly is happening is still a mystery, as
mentioned in a previous publication, fluctuations in pH in CDI units have been observed
and are potentially linked to water hydrolysis. Additionally, because the MCDC is
capable of adsorbing buffers such as phosphates and carbonates, migration of hydrogen
ions may be observed more easily in the MCDC system than the MDC system.
0.2
0.15
- 0.05
o
o>
O
-0.05
-0.1
--
~i I

Anode
MDC
MCDC
MCDC Negative Control
Desal Cathode
Figure 5.6 Change in pH for the MDC and MCDC
Change in pH for the MDC and MCDC shows how the pH changed for the entire reactor. Positive
values indicate a decrease in pH and negative values indicate an increase in pH.
One of the most interesting findings presented here is illustrated in Figure 5.7.
As mentioned previously, the advantage of using a MES for the treatment of PW is in its
ability to simultaneously remove the organic content as well as its salinity. It was
believed originally that the anode chamber would remove organics and the desalination
chamber would serve as the salt remover. However, the results presented in Figure 5.7
illustrate a different picture. For both the MDC and MCDC, COD was removed by the
84


microbes in the anode chamber at a rate of 40 mg COD/L/hr. However, COD was
actually removed at a much faster rate in the desalination chamber of the MCDC than any
other chamber investigated. Almost no COD was removed from the MDC experiment in
the desalination chamber or the cathode chamber. The MCDC desalination chamber was
capable of removing 100-160 mg COD/L/hr. The capacity of the CDI assembly in the
MCDC has been demonstrated in other CDI study using produced water. However,
unlike the anode chamber which oxidizes the organic faction of the COD to CO2, the
COD removed in the desalination chamber is electrochemically removed through
adsorption. Potentially any charged organic molecule could be removed using capacitive
deionization. When the CDI assembly is regenerated, the adsorbed molecules are
recovered back into solution. Ideally, COD should be fully oxidized, but using the CDI is
a quick method of COD removal and concentration. The only potential problem is in the
disposal of the concentrated COD waste. While it is important to understand the full
ramifications of any new technology, the disposal of the regenerated salts and COD is
beyond the scope of the present study.
85


Figure 5.7 Change in COD per Hour for the MDC and MCDC
Figure 5.7 indicates the change in COD for the anode, desalination and cathode chamber per hour.
Positive values demonstrate a removal in COD and negative values indicate a gain in COD.
5.4.2 Regeneration and Energy Harvesting Using the MCDC System
Regeneration of the MCDC was achieved by either connecting the assemblies in
short circuit for one hour or by connecting the fully charged assemblies to an external
circuit in order to harvest the energy store in the in situ capacitors. Table 5.2 shows the
increase in harvested energy from the stored energy of the MCDC CDI assembly. In a
single batch desalination cycle, the maximum voltage reached across the CDI assembly
was 530mV. When connected to the charge pump, electrons can only flow in a single
direction. Electrical energy stored on the MCDC was pumped across the charge pump
and was stored on the 12F 2.5V external capacitor. Capacitors are designed to prevent
jumps in voltage therefore, as can be seen in Table 5.2, the extracted energy was not
achieved quickly. Over 48 hours almost all of the charge on the MCDC CDI assembles
was removed and stored on the on the external capacitor. The external capacitor had a
much higher capacitance and the charge pump efficiency has been shown to be low,
86


Full Text

PAGE 1

DEVELOPMENTS TOWARDS SUSTAINABLE MICROBIAL ELECTROCHEMICAL SYSTEMS FOR SIMULTANEOUS WATER DEIONIZATION, ELECTRICITY PRODUCTION, AND WASTEWATER TREATMENT by Casey Bosch Forrestal BS, Colorado State University, 2005 MS, Colorado School of Mines, 2010 A thesis submitted to the Faculty of the Graduate School of the University of Colorado in partial fulfillment of the requirements for the degree of Doctor of Philosophy Civil Engineering 2013

PAGE 2

2013 CASEY FORRESTAL ALL RIGHTS RESERVED

PAGE 3

ii This thesis for the Doctor of Philosophy degree by Casey Bosch Forrestal has been approved for the Civil Engineering Program by Peter Jenkins Chair Zhiyong Ren, Advisor Jae Do Park Angela Bielefeldt Joann Silverstein April 18, 2013

PAGE 4

iii Forrestal, C asey, Bosch (Ph.D., Civil Engineering ) Developments Towards Sustainable Microbial Electrochemical Systems for Simultaneous Water Deionization, Electricity Production and Wastewater Treatment Thesis directed by Associate Professor Zhiyong Ren ABSTRACT Microbial Electrochemical Systems (MESs), also known as MXCs, have the potential to dramatically affect the socia l, economic, and environmental systems around the world. This enormous potential stems from MXC systems simultaneously performing multiple beneficial functions such as waste treatment, energy production, chemical production, and water desalination. The treatment of waste and the production of value added products such as electricity or fresh water, would benefit the environment and provide additional revenue for the economy. This dissertation focuses on developing novel microbial electrochemical systems while reducing costs and applying the technologies to benefit people, the planet, and the economy. 1. New methods for simultaneous deionization of salt water and energy production were developed through the incorporation of electrochemical adsorption wit h capacitive deionization in novel microbial electrochemical systems. With this fundamental principle established, several reactor systems were developed to remove salt and charged chemical pollutants. 2. New reactor configurations and materials were in vestigated. A modular reactor configuration called a spiral wound microbial electrochemical system was developed to improve MXC performance. This enhanced energy production and wastewater treatment system incorporates the maximum surface area to volume rat io while reducing the overall

PAGE 5

iv footprint of the system. The modular component was developed for multiple applications, such as hydrogen production, desalination, and wastewater treatment. A new low cost membrane barrier was also developed to reduce the ma terial cost for the commercialization of MXC technology. The agarose salt bridge barrier reduces the cost of the MXC system by more than 50% without compromising on performance. 3. The treatment of produced water from the oil and gas industry is one of t he most difficult types of wastewater to treat. A new MXC system was developed which is capable of concurrently removing organic matter and total dissolved solids, while also producing electricity from the system. With the development of new methods of de salination, new reactor configurations, and new reduced cost materials, the microbial electrochemical system will become more sustainable and commercially viable than ever before. The form and content of this abstract are approved. I recommend its public ation. Approved: Zhiyong Ren

PAGE 6

v DEDICATION I dedicate this work to my loving family and friends.

PAGE 7

vi ACKNOWLEDGMENTS I would like to thank my family and friends for all of your support. I would also like to thank my friend and advisor Zhiyong (Jason) Ren for believing in me and guiding me through the many obstacles faced in research I would also like to thank all of my advisers for your support and advice. Thank you to my financial sponsors, the Offi ce of Naval Research, National Science Foundation University of Colorado Cleantech, and University of Colorado Technology Transfer Office. I would also like to thank all of the members of the Ren lab for your support and advice over the last few years. Finally, I would like to make special acknowledgement to my wife Jessica Forrestal for all of her love and support and for whom without I could not have been able to pursue this degree.

PAGE 8

vii TABLE OF CONTENTS Chapter 1. Introduction ................................ ................................ ................................ ..................... 1 1.1 Topics Covered in Dissertation and the Need for Sustainable Systems ....................... 1 1.2 Background Information ................................ ................................ ............................... 3 1.2.1 Microbial Desalination Cells and Other Desalination Devices ................................ 3 1.2.2 Capacitive Deionization ................................ ................................ ............................. 7 1.2.3 Microbial Electrochemical System Designs ................................ .............................. 8 1.2.4 Membranes and Salt Bridge Barriers ................................ ................................ ......... 9 2. Capacitive Microbial Desalination Cell Integrated to Better Control Ion Migration 1 .. 12 2.1 Abstract ................................ ................................ ................................ ....................... 12 2.2 Introduction ................................ ................................ ................................ ................. 13 2.3 Materials and Methods ................................ ................................ ................................ 15 2.3.1 cMDC Reactor Design ................................ ................................ ............................. 15 2.3.2 Reactor Innoculum and Growth Media ................................ ................................ .... 16 2.3.3 Reactor Operation, Analysis and Calculations ................................ ........................ 17 2.4 Results and Discussion ................................ ................................ ............................... 18 2.4.1 Removal and Adsoprtion of Desalination Chamber Salts ................................ ....... 18 2.4.2 Change In pH Over The Course of Desalination ................................ ..................... 20 2.4.3 The Potential And Challenge Of The cMDC Configuration ................................ ... 21 2.5 Conclusi on ................................ ................................ ................................ .................. 22 3. Ion Migration and Control of pH Improved Through the Use of a New Reactor Design Called a Microbial Capacitive Desalinati on Cell 2 ................................ ............................ 23 3.1 Abstract ................................ ................................ ................................ ....................... 23 3.2 Introduction ................................ ................................ ................................ ................. 24

PAGE 9

viii 3.3 Materials and Methods ................................ ................................ ................................ 26 3.3.1 MCDC Reactor Design ................................ ................................ ............................ 26 3.3.2 MCDC Operating Conditions ................................ ................................ .................. 28 3.3.3 Analysis and Calculations ................................ ................................ ........................ 31 3.4 Results and Discussion ................................ ................................ ............................... 32 3. 4.1 Reactor Desalination Performance ................................ ................................ .......... 32 3.4.2 Assembly Regeneration and Salt Recovery ................................ ............................. 38 3.4.3 Reduced pH Fluctuation ................................ ................................ .......................... 40 3.5 Conclusion ................................ ................................ ................................ .................. 42 4. Investigations of Different Methods for the Construction and Operation of Spirally Wound Microbial Electroche mical Systems 3 ................................ ................................ ... 44 4.1 Abstract ................................ ................................ ................................ ....................... 44 4.2 Introduction ................................ ................................ ................................ ................. 45 4.3 Material s and Methods ................................ ................................ ................................ 48 4.3.1 Spiral Wound Configurations ................................ ................................ .................. 48 4.3.1.1 Spiral Wound I Design ................................ ................................ .......................... 48 4.3.1.2 Spiral wound II and III Design ................................ ................................ ............. 50 4.3.2 Reactor Start up and Operation ................................ ................................ ............... 53 4.3.2.1 Spiral Wound I Operation ................................ ................................ ..................... 53 4.3.2.2 Spiral wound II and Spiral wound III ................................ ................................ ... 54 4.3.3 Analysis and Calculations ................................ ................................ ........................ 55 4.4 Results and Discussion ................................ ................................ ............................... 57 4.4.1 Electricity Production and Efficiency of Spiral Wound I ................................ ........ 57 4.4.2 Electricity Production and Efficiency of Spiral Wound II and III ........................... 60 4.5 Conclusion ................................ ................................ ................................ .................. 63

PAGE 10

ix 5. An Efficient Method for Treating and Desalinating Produced Water Using a Microbial Capacitive Desalination Cell 4 ................................ ................................ ........................... 66 5.1 Abstract ................................ ................................ ................................ ....................... 66 5.2 Introduction ................................ ................................ ................................ ................. 68 5.3 Material and M ethods ................................ ................................ ................................ 71 5.3.1 Design of Microbial Capacitive Desalination Cell ................................ .................. 71 5.3.2 MDC Design ................................ ................................ ................................ ............ 73 5.3.3 Operation of the MCDC and MDC Reactors for Proof of Concept Study .............. 73 5.3.4 Operation of the MCDC for Maximum Desalination ................................ .............. 76 5.3.5 Analysis ................................ ................................ ................................ .................... 78 5.4 Results and Discussion ................................ ................................ ............................... 79 5.4.1 Proof of Concept MCDC Vs MDC Desalination and COD Removal Capacity ...... 79 5.4.2 Regeneration and Energy Harvesting Using the MCDC System ............................ 86 5.4.3 Maximum Desali nation Capacity and Rate of the MCDC System .......................... 87 5.5 Conclusion ................................ ................................ ................................ .................. 92 6. The Development of an Efficient Low Cost Barrier for the Use in a Microbial Electrochemical System Called an Agarose Salt Bridge Membrane 5 .............................. 94 6.1 Abstract ................................ ................................ ................................ ....................... 94 6.2 Introduction ................................ ................................ ................................ ................. 95 6.3 Materials and Methods ................................ ................................ ................................ 97 6.3.1 Agarose Salt Bridge Membrane Design ................................ ................................ .. 97 6.3.2 Reactor Configuration ................................ ................................ .............................. 97 6.3.3 Analysis ................................ ................................ ................................ .................... 99 6.4 Resu lts and Discussion ................................ ................................ ............................... 99 6.4.1 Agarose Salt Bridge Membrane Power Performance from Produced Water .......... 99 6.4.2 Agarose Salt Bridge Membrane Change in pH ................................ ...................... 101

PAGE 11

x 6.4.3 Internal and Polarization Resistance for the ASBM Compared to the Astom CEM ................................ ................................ ................................ ................................ ......... 102 6.4.4 Cost Comparison and Sustainability ................................ ................................ ...... 104 6.4.5 ASBM Challenges ................................ ................................ ................................ 105 6.5 Conclusion ................................ ................................ ................................ ................ 105 References ................................ ................................ ................................ ....................... 107

PAGE 12

xi LIST OF TABLES Table 3.1 MCDC Salt Removal ................................ ................................ ................................ 33 3.2 Comparison of Different Capacitive Deionization Methods ................................ ..... 34 4.1 Comparison of Spiral Wound Reactors ................................ ................................ ...... 47 5 .1 Produced Water Characteristics ................................ ................................ ................ 78 5.2 Energy Harvested From Single Batch Cycle of MCDC Using Produced Water ..... 87 5.3 Percent TDS Removal and Change in Conductivity for the Three Desalination Cycles ................................ ................................ ................................ ................................ .......... 89 5.4 Total Ions Removed and Recovered with the MCDC System ................................ 90 6.1 Thicknesses of the ASBM ................................ ................................ ....................... 9 7

PAGE 13

xii LIST OF FIGURES Figure 2.1 Diagram of Capacitive Microbial Desalination Cell (cMDC) reactor with physical and electrical adsorption. Picture of Spiral Wound II. ................................ ..................... 1 6 2.2 cMDC Desalination And Energy Production Picture of Spiral Wound II. ................ 19 2.3 cMDC Change in Conductivity 4.3 Picture of Spiral Wound II. .............................. 20 2.4 Change in pH for cMDC System 4.3 Picture of Spiral Wound II. ........................... 21 3.1 Operation of MCDC System ................................ ................................ ..................... 28 3.2 Diagram of MCDC System 4.3 Picture of Spiral Wound II. ................................ .... 29 3.3 Correlation of App lied Potential and Deionization ................................ .................. 35 3.4 MCDC Ion Migration ................................ ................................ ............................... 36 3.5 Electrochemical Impedance Spectroscopy for the MCDC ................................ ....... 37 3.6 MC DC Regeneration ................................ ................................ ............................... 38 3.7 MCDC Ion Recovery ................................ ................................ ............................... 40 3.8 Change In pH for MCDC ................................ ................................ .......................... 41 4.1 Diagram for Spiral Wound I ................................ ................................ ..................... 48 4 .2 Pic ture of Spiral Wound I ................................ ................................ ........................ 50 4.3 Picture of Spiral Wound II I ................................ ................................ ...................... 51 4.4 Computer Aided Drawing of Spiral Wound II and III ................................ ............. 52 4.5 Internal Diagram of Spiral Wound II and III ................................ ........................... 53 4.6 Operation of Spiral Wound III ................................ ................................ ................. 54 4.7 Acclimation Times For Different reactor Configurations ................................ ........ 56 4.8 Effect of Retention Time On Power Density SWI ................................ .................... 57 4.9 Coulombic Efficiency Correlated to Rete ntion Time and Maximum Power Point .. 58

PAGE 14

xiii 4.10 Change In Polarization Resistance Over Time ................................ ........................ 59 4.11 Spiral Wound II and III Initial Overshoot Potential ................................ .............. 60 4.12 Spiral Wound III Loss of Overshoot Potential ................................ ...................... 61 4.13 Comparison of EIS for the Three Spiral Wound Reactors ................................ .... 62 5.1 Operation of MCDC for Produced Water ................................ ................................ 72 5.2 MDC for Produced Water ................................ ................................ ........................ 73 5.3 Voltage Profile ................................ ................................ ................................ ......... 79 5.4 Charged Formed on CDI ................................ ................................ ........................... 82 5.5 Change in Conductivity for MDC and MCDC ................................ ......................... 83 5.6 Change in pH for the MDC and MCDC ................................ ................................ ... 84 5.7 Change in COD per Hour for the MDC and MCDC ................................ ................ 86 5.8 Change in TDS for the MCDC Operating Under Maximum Conditions .................. 88 5.9 Change in COD for the MCDC for the three Chambers Plus the Recovered COD 92 6.1 Power Density of the ASBM ................................ ................................ .................. 101 6. 2 Change in pH for the Anode and Cathode Chamber with the ASBM using Produced Water in the Anolyte ................................ ................................ ................................ ..... 102 6.3 Nyquist Plot for the ASBM, ASBM 2 and the Astom CEM ................................ .. 103 6.4 EIS Co mparison of the ASBM and ASBM 2 ................................ ........................... 10 4

PAGE 15

1 1. Introduction 1.1 Topics Covered in Dissertation and the Need for Sustainable Systems Since the earliest recorded history of Thales of Miletus in 650 BC, humans have strived to understand and record the mysteries of nature. The development of scientific knowledge over the last 3000 years has led to amazing advancements in the quality of hu man life. This progress, however, has also brought us to the precipice of annihilation with the production of nuclear weapons and the destruction of the environment. Only in the last 100 years has man realized the need to develop technology that will sust ain life now and in the future. In 1987 the Brundtland Commission published the most commonly that meets the needs of the present without compromising the ability of fu ture generations 1 The following topics in this dissertat ion were developed with this goal in mind. Microbial electrochemical technology (MET), is a potential solution to many needs in sustainable development. Microbial electrochemical systems (MES), also known as MXC, use microorganisms to catalyze the oxida tion of organic or inorganic matter to generate electrical current in a fuel cell system. The ability of microorganisms to transfer electrons and generate electrical current has been credited to MC Potter in 1911. However, only in the last 20 years has thi s phenomenon been extensively studied. The explosion of research in this field is a result of the potential to dramatically affect the social, economic, and environmental systems around the world. This enormous potential stems from MXC systems simultaneo usly performing multiple beneficial functions, such as wastewater treatment, electrical energy production, chemical production, and water

PAGE 16

2 desalination. MXCs have the potential to revolutionize humanity by solving two of the greatest problems to ever to fac e mankind: the availability of energy and water. However, to date, few large scale MXCs have been developed and implemented because of the inherent inefficiency and high cost of a newly developed technology. Presented, here are some of the most recent ad vancements in METs ability to desalinate water, efficiently treat wastewater, and generate electricity with a minimal cost and environmental footprint. Each chapter focuses on a single method to improve MXC for sustainability. Chapter 2 and 3 focuses on de veloping a sustainable method for deionization of salt water and energy production, through incorporating electrochemical adsorption by capacitive deionization, in novel microbial electrochemical systems. With this fundamental technology, a more sustainab le method for desalinating salt water and removing charged chemical pollutants can be achieved. Two articles have been published on this topic. The article in chapter 2 demonstrates the advantages and challenges of a capacitive Microbial Desalination Cell (cMDC). The second article, in chapter 3, demonstrates the newly developed MXC called the microbial capacitive desalination cell (MCDC) which demonstrates its advantages and challenges in desalination of salt water. Chapter 4 focuses on the development of a modular spiral wound microbial electrochemical system for enhanced energy production and wastewater treatment. The system incorporates the maximum surface area to volume ratio while reducing the overall footprint of the system. The spiral wound reac tor is a modular system; therefore it can be scaled to meet the need of large wastewater treatment operations as well small onsite wastewater treatment. The modular component also helps develop the system for

PAGE 17

3 multiple applications, such as hydrogen produc tion, desalination, and wastewater Investigation of Different Methods for the Construction and Operation of Spirally Wound Microbial Fuel Cells Chapter 5 focuses on the development and characterization of a MX C for the treatment of produced water from oil and natural gas mining. The technology uses the newly developed MCDC, outlined in chapter 3, to remove both organic and dissolved ility to An Efficient and Sustainable Method of Treating and Desalinati ng Natural Gas Produced Water Using a Microbial Capacitive Desalination Cell Lastly, chapter 6 focuses on the development of a l ow cost membrane barrier which could be used in MXC reactors for the production of electricity, treatment of wastewater, and potentially desalinate salt water. An agarose salt bridge membrane could dramatically improve the economic viability of MXCs by re ducing the capital cost of manufacturing MXC systems, without compromising on performance. A thorough investigation into the manufacturing and chemical composition of the agarose salt bridge membrane yielded an efficient low cost MXC system. However, many practical challenges still exist with this material. Further research is needed before this material would be viable for commercial use and publication. 1.2 Background Information 1.2.1 Microbial Desalination Cells and Other Desalination Devices While 97% of the planet is covered by water, only approximately 0.4% of that water is available for consumption 2 Currently the most popular methods to increase the

PAGE 18

4 supply of fresh water are through either evaporative processes such as multistage flash evaporation (MSF) or filtration processes such as reverse osmosis (RO). Evaporative desalination processes were the first to be developed and widely commercially used. The technology works by heating the salt water to convert the liquid water into the gaseous phase separating the water from the salt. The gaseous phase water then needs to be condensed back into the liqu id phase for use. This process, while highly effective, is also energy intensive requiring up 650 kWh/m 3 of fresh water produced. RO technology uses a different method to separate water from salt. RO uses specially designed membranes that allow water mol ecules to pass across the membrane but prevent larger molecules. In order to facilitate the movement of water across the membrane, RO systems need to operate at a high pressure using pumps which require electrical energy. The energy needed for filtration desalination systems is much smaller through the evaporative process, however RO still require between 1 6 kWh/m 3 water produced. One other currently commercialized process for desalination is called electrodialysis (ED). Electrodialysis works by moving ions in a potential field across anion and cation exchange membranes. The applied potential causes anions to migrate across the anion exchange membrane towards the positive charge, and cations to migrate across the cation exchange membrane towards the neg ative charge. Electrodialysis usually incorporates multiple membranes so that ions migrate from desalination chambers into concentration chambers. ED requires energy at a rate of 0.3 0.8kWh/m 3 Microbial desalination cells (MDC) are one of the many MXCs previously developed. MDCs desalinate salt water in the exact same method as electrodialysis except the electrical potential is generated internally by microorganisms rather than supplied externally. The oxidation of a

PAGE 19

5 substrate in the anode chamber caus es an electrochemically positive charge to be generated. Electrons flow from the anode chamber to a cathode where the electrons are combined with the reduction of oxygen to water. The electrochemical reduction of a chemical (O 2 or K 3 Fe(CN) 6 ) causes a neg ative potential charge to be formed. By adding a positively charged anion exchange membrane (AEM) adjacent to the anode chamber, and a negatively charged cation exchange membrane (CEM) adjacent to the cathode chamber, a MDC is formed. Desalination occurs between the AEM and the CEM called the desalination chamber. When sodium chloride is dissociated, sodium ions pass through the CEM to the cathode chamber while the negatively charged chloride ions pass through the AEM. This technology was first demonstr ated by Cao et al. 2009, 3 and has been improved upon by many researchers. The MDC system was improved up on by incorporating multiple chambers for the production of desalinate and concentrate solutions. Additionally work done by Haiping Luo at the University of Colorado Denver, demonstrated that hydrogen gas could be generated in a MDC system by incorporatin g an additional electrical supply. The MDC in 2009 was touted as a major breakthrough in sustainably desalinating salt water because it could generate electricity, treat a wastewater, and supply desalinated water. The proof of concept article was selecte d as the Top Technology Paper of the Year by Environmental Science & Technology While the MDC was a major breakthrough it still contained many challenges. Generally there are three problems associated with the MDC reactor configuration. As the ions are removed from the desalination chamber the conductivity drops causing an increase in internal resistance. In the Cao et al. 2009 paper, during the start of desalination the internal resistance was low and so the voltage output was high. Over the course of

PAGE 20

6 desalination the ohmic resistance, determined by electrochemical impedance spectroscopy (EIS), increased to over 970 Ohms; thus causing the voltage and the rate of desalination to decrease. The increase in ohmic resistance from the operation of the MDC oc curs as quickly as the salt solution is desalinated. This problem is compounded by scaling and biofouling on the membranes also identified by Haiping Luo. 4 Biofouling generally occurs slowly over a period of days or months on the AEM due to the microorganisms in the anode chamber. Chemical scaling was found to be most prevalent on the CEM next to the cathode chamber. The affect of biofouling and chemical scaling greatly reduced the MDCs reactors performance. Additionally, because o f the use of anion exchange membrane, which prevents the migration of positively charge ions such as protons, the pH decrease in the anode and increase in cathode. The decrease in pH in ectrons. Lastly, if wastewater is used in the anode chamber, ions that migrate from the desalination chamber to the anode chamber may inhibit the treated wastewater reusability. Because of the MXC ability to generate electricity it would be prudent to po nder whether it would be more beneficial to generate electricity for external desalination or to use the MDC system for desalination. Currently the energy produced and harvested in MXCs, is small. 5 The MDC does not actually consume any energy for desalination but rather uses the electrical potential naturally generated. Therefore at this time it would not be beneficial to use a MXC to generate electricity for desalination in a different desalination reactor such as reverse osmosis.

PAGE 21

7 1.2.2 Capacitive Deionization Capacitive deionization (CDI) is an additional method for desalinating salt water which uses ma ny of the same principles as ED. CDI is an electrochemical process, sometimes referred to as electrosorption 6 CDI operates by adsorbing ions onto a high surface area electrode by applying an electrical potential as in ED. When an electrical potential is applied to an electrode a ch arge is formed on the surface of the electrode causing electrochemical attraction of oppositely charged ions. As ions are adsorbed onto the electrode a double layer capacitor is formed. This was first identified by Hermann von Helmholtz in 1883, and is s ometimes referred to as the Helmholtz double layer capacitor. The theory of the electrical double layer was later modeled by Guoy Chapman in 1913 and has been widely accepted as true. The theory treats ions as small point charges and the capacity of ads orptions is dependent upon potential at the surface of the electrode, area of electrode, and distance between the electrodes. The region directly adjacent to the electrode has been termed the Helmholtz layer and can be modeled using isotherm. The outer layer, or the diffusive layer, can be modeled using the Guoy Chapman method. Molecules or ions in a solution form hydration shells which can also be modeled using the double layer principles. A potential difference is then formed betw een the surface of the ion and the bulk solution. This parameter is called the electrokinetic (zeta) potential, or just the zeta potential. The zeta potential is important in understanding the formation of colloids and is related to pH and conductivity. Ideally CDI works by removing charged ions without involving redox reactions which would make the process fully reversible; thus providing a means of removing ions (desalination) and subsequently concentrating those ions (regeneration).

PAGE 22

8 While the first CD considered feasible until recently, when high surface area electrodes were developed, including the carbon aerogel, activated carbon cloth, and graphene. The advantage of the CDI system is that i t requires a relatively small amount of energy, operating at potentials between 0.6V and 1.2V consuming as low as 0.1 kwh/m 3 water produced. Also, because the applied energy is stored in a capacitor, some of that energy can be recovered. Work on supercapa citors has demonstrated a round trip energy efficiency of greater than 95%. A CDI system operating at a round trip energy efficiency similar to that of a supercapacitor would approach the thermodynamic limitation for the removal of salt from water. The C DI system is extremely complicated with many challenges, but the technology has the greatest potential to reach the thermodynamic limitation for desalination. 1.2.3 Microbial Electrochemical System Designs MXC have been developing over the last century wit h most of the research done in the last decade. The goal in designing a MXC system is to maximize energy output, treatment of wastewater, and operate at a reduce cost compared to traditional wastewater treatment. Paramount to achieving these three goals is the system design. The system design consists of two facets, the external operating design and the materials used. Since 2000 the power output in a MXC system has be steadily increasing, from 13 W/m 3 7 to 2.87 kW/m 3 8 Researchers have increased volumetric power density generally by increasing the total surface area to reactor volume, and by decreasing the total volume of the reactor. In 2007, Fan et al. used a plate and frame reactor design with a 2.5mL double cloth electrode assembly and a cathode surface area of 230m 2 /m 3 to produce

PAGE 23

9 1.55kW/m 3 9 Additionally, Nevin et al. 2008 with a volume of 0.335mL, using ferricyanide as the electron acceptor produced a power density of 2.15 kW/m 3 10 It appeared that high power density in MXC systems could be achieved only through the reduction of volume until recentl y, when Fan et al. 2012 demonstrated that at a volume of 30mL a power density of 2.87kW/m 3 could be achieved. These advancements in system design have been a long time coming. MXC systems have been developed using a single chamber or two chambers with a membrane barrier. These two general reactor designs have been modified on numerous occasions. Some of the designs have incorporated serpentine or upflow pathways for fluid flow, and there have been designs for multiple anode and cathodes. Some reactors have been designed in a tubular fashion, where either the anode or cathode electrode is wrapped around the other. In addition to external system designs, most of the research on MXC systems has been to identify materials which improved performance and red uced cost. To date, the most widely studied material for use as an anode or cathode electrode are inorganic carbons such as graphite, carbon cloth, or graphite granules. This material has a relatively low cost at $0.24 10/kg but has a resistance ranging from 1.47x10 3 1.9x10 3 micro Ohms/cm. 11 With greater improvements to the system design a commercially viable MXC system will be developed. 1.2.4 Membranes and Salt Bridge Barriers The purpose of membrane in a MXC system is to separate d ifferent chambers while still allowing ions to transfer. M embranes are not always required for MXC systems; h owever for the production of desalinated salt water or chemical production the use of a membrane is a must. Membranes can be beneficial and harmful for the operation

PAGE 24

10 of a MXC system. They were first des igned to separate the anode and cathode chamber to prevent oxygen diffusion from the cathode chamber into the anode chamber reducing the electron transfer efficiency. They can also prevent the formation of biofilm on the cathode which increases electron transfer However there are many challenges associated with using membranes in MXC systems. T he use of a membrane causes the internal resistance of the reactor to increase reducing power output T he use of a membrane may inhibit proton transfer causin g the pH in the anode chamber to become more acidic and basic in the cathode chamber. Also, as previously mentioned, microbes in the anode chamber can become attached to the membranes causing an increase an internal resistance called bio foul ing. Membrane s can become scale d with a chemical, which causes an increase an internal resistance. These problems are common in MXC research especially when real instead of synthetic wastewater is used. The two most extensively studied membrane materials are cation e xchange membranes (CEM), also referred to as proton exchange membranes, and anion exchange membranes (AEM). To date the use of an AEM in a two chamber MXC system has demonstrated the highest energy efficiency. This is likely due to the concentration grad ient diffusion of anions across AEM and the reduced scalability of AEMs over CEMs. In addition to standard ion exchange membranes, salt bridges bipolar membranes, ultrafiltration membranes, microfiltration membranes, glass fibers, porous fabric and other porous filters have been investigated 12 Salt bridges were the fi rst ion barrier used in a fuel cell which was developed by using a glass tube filled with electrolyte solution between two chambers. Salt bridge separators have been used in MXC systems for many years but have been widely discredited due to its higher int ernal

PAGE 25

11 resistance and thus lower power density. Most salt bridge separators use KCl or phosphate buffer as ion conductors with added agar to prevent fluids from mixing. Min et al. 2005 found that a phosphate buffered salt bridge increase d the internal res istance to 19,920 Ohms over a Nafion CEM membrane which had an internal resistance of 1286 Ohms. 13 However there were two major adva ntages of salt bridg e membranes previously investigated: it limited the diffusion of oxygen, and cost a faction of other membrane separators. T radition al ion exchange membranes can cost between $100/m 2 to $1400/m 2 In order to scale MXC systems a sustainable separator will need to be developed.

PAGE 26

12 2. Capacitive Microbial Desalination Cell Integrated to Better Control Ion Migration 1 2.1 Abstract A new microbial desalination cell with capacitive adsorption capability (cMDC) was developed to solve the ion migration problem facing current MDC systems. Traditional MDCs remove salts by transferring ions to the anode and cathode chambers, which may prohibit wastewater beneficial reuse due to increased salinity. The cMDC uses adsorptive activated carbon cloth (ACC) as the electrodes and utilizes the formed capacitive double layers for electrochemical ion adsorption. The cMDC removed an average of 69.4% of the salt from the desalination chamber through electrode adsorption during one batch cycle, and it did not add salts to the anode or cathode chamber. It was estimated that 61 82.2 mg of total dissolved solids (TDS) was adsorbed to 1 g of ACC electrode. The cMDC provide s a new approach for salt management, organic removal, and energy production. Further studies will be conducted to optimize reactor configuration and achieve in situ electrode regeneration. 1 The work presented in this chapter has been published by Casey Forrestal, Pei Xu, Peter Jenkins, and Zhiyong Ren. Bioresource Tech 2012, 120, 332 336

PAGE 27

13 2.2 Introduction The sustainable supply of fresh water through saltwater desalination has been developed significantly in the past century, but one remaining challenge is high energy use during the desalination process. Popular desalination methods include reverse osmosis (RO) and multistage flash evaporation (MSF) which are considered energy intensive, because for treating 1 m 3 of seawater, RO typically uses 3 7 kwh/m 3 of electricity and MSF may require up to 68 kwh/m 3 14, 15 Recently, a new des alination technology called microbial desalination cell (MDC) was developed and demonstrated that salt water can be desalinated without using external energy. Moreover, this process can also simultaneously achieve wastewater treatment and energy production in the format of electricity or hydrogen gas. 16 20 MDC reactor use s exoelectrogenic bacteria to oxidize biodegradable substrate (i.e. wastewater) in an anode chamber and transfer the electrons to the anode. The electron flows through an external circuit to a cathode, where external electron acceptors (i.e. O 2 ) are reduce d. When a middle chamber is inserted in between the anode and cathode chamber using a pair of ion exchange membranes, desalination can be achieved. The potential difference between the anode and cathode electrodes drives the migration of ions out of the de salination chamber, with anions (Cl ) migrating to the anode chamber across an anion exchange membrane and cations (Na + ) migrating to the cathode chamber across a cation exchange membrane. The process can remove more than 99% of the salt water and potent ially produce more energy than the external energy required for the system, making it a promising desalination process with net energy gaining. 17, 21

PAGE 28

14 One main challenge with the MDC technology is that while the salts are removed from the middle chamber, they become concentrated in the anode and cathode chambers, which re sults in salinity increases in the anolyte and catholyte. While this ion addition is generally acceptable for wastewater treatment and helps with conductivity conditioning, it may cause concerns for water beneficial reuse, where the total dissolved solids (TDS) is regulated. 19, 22 Additionally, as ions are desalinated from the desalination chamber, the internal resistance immediately increases causing the rate of desalination to also decrease immediately. One solution to manage salt removal from MDC is to incorporating capacitive deionization (CDI) concept into MDC systems. 23, 24 When a saline solution flows between two charged electrodes, the ions can be adsorbed by the double layer capacitor formed on the high surface electrodes. When the potential is removed the ions can be released back into the liquid to form a concentrate for salt recovery. Using this integrated deionization approach in co mbination with the traditional MDC, salt water can be deionized through electrochemical salt adsorption on the electrodes, so no ions will migrate into the electrolyte solutions. In this study, the integration of an MDC and electrochemical adsorption was d eveloped into a single reactor for capacitive microbial desalination (cMDC). This system demonstrates the feasibility of a new process for concurrent power production and saltwater desalination without contaminating the anode or cathode chambers.

PAGE 29

15 2.3 Materials and Methods 2.3.1 cMDC R eactor Design Each cMDC reactor consisted of three polycarbonate cube shaped blocks. Each block has a 3 cm diameter hole, and three blocks were clamped together to form one anode chamber, one cathode chamber, and one d esalination chamber, with the volume of 23 mL, 27mL, and 10mL, respectively (Figure 2.1). 4 Zorflex Activated Carbon Cloth (ACC) (Chemviron Carbon, UK) was used as the electrode material and was pretreated by washing in acetone overnight and heating to 350C for 30 minutes. 25 An anode assembly consisted of one layer of AEM (AMX, Astom Corporation, Japan), one Ni/Cu mesh current collector (Grade 400, McMaster Carr, IL), and an ACC anode was formed to separate the anode chamber and the desalination chamber. Similarly, a cathode assembly was formed by pressing one piece of CEM (CMX, Astom Corporation, Japan), one layer of Ni/Cu mesh, and an ACC cathode together to separate the cathode chamber and the desalination chamber. The total surface area of the ACC electrodes were 18 cm 2 ( 9 cm 2 each) with the weight of one gram. The specific surface area of the ACC is 1020 m 2 /g, determined by the Brunauer Emmet Teller (BET) method (ASAP 2020, Micromeritics, Norcross, GA). 26 Prior to use, the membranes were pretreated in 10 g/L NaCl for 24 hours to remove impurities and maximize ion ex change capacity. The Ni/Cu current collectors were connected to a titanium wire that connected the anode and

PAGE 30

16 Figure 2.1 Diagram of Capacitive Microbial Desalination Cell ( cMDC) r eactor with physical and electrical adsorption. 2.3.2 Reactor Innoculum and Growth Media The reactor was initially inoculated with anaerobic sludge a nd operated in fed batch microbial fuel cell mode by using a CEM to separate the anode and cathode chambers. 19 When the repeatable voltage profile was obtained in three consecutive batches from the MFC, the middle chamber was inserted to the reactor to form a cMDC as described previously. 27 The anolyte contained per liter: 1.6g NaCH 3 COO, 0.62 g NH 4 Cl, 4.9 g NaH 2 PO 4 H 2 O, 9.2 g Na 2 HPO 4 0.3g KCL, and 10ml trace metals and 10ml vitamin solution. 28 The catholyte contained per liter: 100mM KFe(CN) 6 5mM KH 2 PO 4 and 5mM K 2 HPO 4 Ferricyanide solution was used as th e catholyte to minimize cathode mass transfer effects. The salt solution for desalination contained 10 g/L NaCl. This concentration of salt water would be representative of a brackish water AEM CEM R Cl Na + NaCl Feed Cl Cl Na + Na + Ferricyanide Acetate ACC ACC Ni/Cu Ni/Cu CO 2 Microbes F erro cyanide

PAGE 31

17 found around the world. A treatment efficiency of 70 90% would be safe to dispose to most environmental conditions. 2.3.3 Reactor Operation, Analysis and Calculations The active cMDC reactor was operated in fed batch mode. A single fed batch was defined as the time required for a single desalination cycle. As soon as the voltage across the reactor reached below 0.025mA, all electrolyte solution was replaced. Electrolyte conductivity was determined by a conductivity meter (Sension 156, HACH Co., CO), and pH was determined with a pH meter (Sension 4, HACH Co., CO). Befo re the reactor was connected for each batch, the anode and cathode chambers were allowed to reach the maximum open circuit potential (OCP), which was determined using a potentiostat with a saturated Ag/AgCl reference electrode (G 300, Gamry Instruments Inc NJ). Using a data acquisition system (model 2700, Keithley Instruments, inc. OH), the voltage across the external resistor was recorded every one minute. Conductivity and pH measurements for all three chambers were taken at the beginning and the end of each desalination cycle. Conductivity was converted to TDS (mg/L) using the HACH Co. general calculation equation presented below. mS/cm 500 = mg/L The ACC assemblies were manually regenerated by removing the ACC from the reactor and rinsed in 1 L of d eionized (DI) water for 30 minutes. The DI rinse was repeated a total of three times till no salt residue was left before adding the electrodes back to the reactor for additional experiments. Negative control experiments were conducted by removing the acc limated anode and replacing it with an unacclimated piece of ACC and performed as described previously for desalination experiments.

PAGE 32

18 2.4 Results and Discussion 2.4.1 Removal and Adsoprtion of Desalination Chamber Salts During cMDC operation an electrical potential is formed on the ACC across the anode and the cathode due to microbial oxidation of substrate and electron transfer to the cathode. This potential forms a double layer capacitor on the ACC electrodes, which adsorbs ions to achieve water deioniza tion. At the start of the desalination cycle the cMDC reactor had an OCP of ~712mV. When a 1 Ohm resistor was used in the circuit, a maximum current of 2.5mA was generated. The reactor was operated for three months, and Figure 2.2 shows the production of electric current across the resistor is proportional to the percent removal of NaCl in the desalination chamber for three consecutive cycles. The substrate consumption and ion loss in the desalination chamber caused an increase of internal resistance, whic h resulted in a decrease of current along with one batch cycle. Such phenomenon is consistent with other MDC studies using similar reactor configurations. 16, 19 This traditional 3 chamber cube MDC conf iguration was used in the study to demonstrate the feasibility of the process, and system optimization such as using stack configurations with narrow chambers and/or operating the reactor in a continuous mode have been shown effective to address the increa sed resistance problem. 18, 29, 30

PAGE 33

19 Figure 2.2 cMDC Desalination And Energy Production Three typical cycles of cMDC operation and current generation with NaCl removal profile. Arrows indicate the manual washing and regenerating the ACC electrodes. Figure 2.3 shows the salt removal from three chambers as represented by conductivity changes. The average salt removal in the desalination chamber during one batch cycle was 69.4%, which correlates to the removal of 69.4 mg (1.19mM) of NaCl from the 10mL desalination cham ber. The 1.19mM of salt removed from the desalination chamber correlates to 42.2 mg of Cl migrated to the anode and 27.2 mg of Na + migrated to the cathode. The anode chamber conductivity in average increased slightly by 0.7% or 1.1mg presumably due to the limited diffusive ion release from the anode. Such ion balance suggests that 41.1 mg of chloride was electrochemically adsorbed on the anode, which represents 97.3% of the desalinated chloride from the desalination chamber. In the cathode chamber the cond uctivity decreased in average by 1.4% or by 3.2 mg. Therefore the cathode chamber with the ACC adsorbed 30.4 mg of salt or 100% of the desalinated salt plus an additional 12% from the cathode chamber. These results indicate that the electrical adsorption capacity of the ACC assembly was between 61 82.2 mg TDS/g

PAGE 34

20 ACC. Such results are comparable to the findings from a previous study, which showed the physical and electrical adsorption was 72.7mg/g ACC. 27 This salt removal profile from the desalination chamber is different from traditional MDC systems, because the remove chamber but rather got adsorbed onto the ACC electrodes. The conductivity of the anolyte and catholyte were kept quite stable, which prevented significant salinity changes that might affect effluent reuse. The abiotic control reactor without microbial activities showed no current generation or desalination performance, as shown in Figure 2.3. This finding confirms that the potential generated by the microbial exoelectrogenic acti vities was the driving force of desalination. Figure 2.3 cMDC Change in Conductivity The initial and final conductivity for the anode desalination and cathode chamber for the active cMDC system as well as a negative control reactor 2.4.2 Change In p H Over The Course of Desalination While the initial pH of all three chambers was measured at 6.8, the final pH after a batch cycle varied between chambers. As shown in Figure 2.4, the desalination chamber pH slightly decreased to 6.4 after a batch cycle. T he anolyte pH dropped to 5.9, while the

PAGE 35

21 catholyte pH increased to an average of 7.9. Such pH variations have been reported by previous MDC studies, because the accumulation of protons in the anode chamber caused pH drop, while the loss of protons in the ca thode chamber due to water formation led to pH increase. To alleviate pH fluctuation, electrolyte recirculation can be implemented to neutralize the anolyte and catholyte, and further studies are underway to address such problems 4, 30 The negative control without microbial activities showed no pH changes, confirming no electrons or protons were transferred in the system. Figure 2.4 Change in pH for cMDC System The initial and final pH for the anode, catho de and desalination chambers of the cMDC system as well as a negative control reactor. 2.4.3 The Potential And Challenge Of The cMDC Configuration Compared with traditional MDC systems, where the salt removal from the middle chamber is accompanied by the salinity increase in the anode and cathode chambers, this cMDC configuration is able to incorporate capacitive deionization with microbial desalination and captures salts on electrodes without releasing the salts to the electrolytes. This integrated proces s addressed the concerns of increased salinity on cMDC effluent reuse and provides a new approach for more complete salt management. Moreover, it can

PAGE 36

22 simultaneously achieve organic removal and energy production together with desalination, demonstrating a g reat potential for the water industry. By directly using high surface activated carbon electrodes, the cMDC was able to adsorb ions on the double layer capacitor formed on the electrode surface. However, it is not clear how the increased ion concentratio n might affect anode biofilm activity and community on the electrode. During this study, manual cleaning and regeneration of electrodes was performed after reactor operation, which prevented a mass balance or salt recovery calculation. It also impacted the microbial community evident in the last cycle of Figure 2.2, which showed lower current production. Further studies will be conducted to develop in situ electrode regeneration methods. One possible method could be to develop a reactor that can switch the ACC electrodes in situ once they have become fully adsorbed. Switching the electrodes would cause the adsorbed salts to desorb due to the reverse potential, which would solve the problem of having to manually regenerate the electrodes. 27 2.5 Conclusion This study presents a step forward in sustainably desalinating salt water with a capacitive microbial desalination cell. The cMDC reactor was capable of removing an average of 69.4% of the salt from the desalination chamber through electrochemical ion ads orption on the electrodes without adding salinity to the anode or cathode chamber. The physical and electrical adsorption capacity of the ACC electrodes was between 61 and 82.2 mg/g ACC. Further studies are needed to improve system efficiency and develop in situ ACC regeneration process. By combining this process with optimized reactor configurations a sustainable method of desalination can be obtained.

PAGE 37

23 3. Ion Migration and Control of pH Improved Through the Use of a New Reactor Design Called a Microbial Capacitive Desalination Cell 2 3.1 Abstract Microbial desalination cells (MDCs) use the electrical current generated by microbes to simultaneously treat wastewater, desalinate water, and produce bioenergy. However, current MDC systems transfer salts to th e treated wastewater and affect wastewater beneficial use. A microbial capacitive desalination cell ( MCDC ) was developed to address the salt migration and pH fluctuation problems facing current MDCs and improve the efficiency of capacitive deionization. Th e anode and cathode chambers of the MCDC were separated from the middle desalination chamber by two specially designed membrane assemblies, which consisted of cation exchange membranes and layers of activated carbon cloth (ACC). Taking advantage of the pot ential generated across the microbial anode and the air cathode, the MCDC was capable of removing 72.7 mg total dissolved solids (TDS) per gram of ACC without using any external energy. The MCDC desalination efficiency was 7 to 25 times higher than traditi onal capacitive deionization processes. Compared to MDC systems, where the volume of concentrate can be substantial, all of the removed ions in the MCDC were adsorbed in the ACC assembly double layer capacitors without migrating to the anolyte or catholyte and the electrically adsorbed ions could be recovered during assembly regeneration. The two cation exchange membrane based assemblies allow the free transfer of protons across the system and thus prevented significant pH changes observed in traditional M DCs. 2 The work presented in this chapter has been published by Casey Forrestal, Pei Xu, and Zhiyong Ren. Energy Environ. Sci 2012, 5, 7161 7167

PAGE 38

24 3.2 Introduction The increasing awareness of the water energy nexus is compel ling the development of technologies that reduce energy requirements during water treatment as well as water demands for energy production. 31, 32 Microbial desalination cells (MDCs) recently emerged as a promising technology to simultaneously treat wastewater, desalinate saline water, and produce renewable energy such as electricity or hydrogen gas. 3, 14, 18, 33 40 MDCs share the same principle of bioelectrochemical reactions with microbial fuel cells (MFCs): electrochemically active bacteria in the anode chamber oxidize biodegradable substrates and generate electron flow (i.e. current) to reduce the electron acceptors in the cathode chamber. The additional desalination function can be achieved in an MDC by adding a middle chamber containing saline wa ter and utilizing the anode cathode potential difference to drive the migration of anions (e.g., Cl ) to the anode chamber and cations (e.g., Na + ) to the cathode chamber for charge neutrality. 3 The MDC process carries great potential in desalination systems, because it can either be used as a stand alone process or serve as a pretreatment for conventional des alination processes such as reverse osmosis (RO) to reduce salt concentration of RO feed, and minimize energy consumption and the membrane fouling potential. Current desalination technologies, such as RO and electrodialysis (ED) are energy and capital inte nsive. Even the most advanced large scale seawater RO units require 3 7 kWh/m 3 for water desalination, while conventional multi stage flash evaporation requires 68 kWh/m 3 40 In contrast, the MDC system is considered to be an energy gaining process, because it converts the biochemical energy stored in wastewater to electricity or hydrogen gas. Lab scale MDC studies showed that 180 231% more en ergy can be recovered as H 2 than the

PAGE 39

25 reactor energy input when desalinating 5 20 g/L NaCl solutions, 35, 36 and a recent study calculated that a liter scale MDC can produce up to 58% of the electrical energy required by downstream RO systems 34 Current MDC systems use an anion exchange membrane (AEM) to separate the anode and middle chamber, and a cation exchange membrane (CEM) to separate the cathode and middle chamber. Similar to electrodialysis, desalination in MDC is achieved by direct transport of salts from the middle chamber to the anode and cathode chamber. Such system faces two main p roblems. While salts get removed from the middle chamber, they become concentrated in the anode and cathode chambers resulting in an increase of the volume of saline water. This concern becomes more imp erative when wa stewater is treated as the anolyte. Al t hough the addition of ions (or total dissolved solids, TDS) increases wastewater conductivity and b enefits electricity generation, the increased salinity may affect efflue nt water quality and prevent subsequent beneficial use of treated wastewater. 39, 41 The high salinity may also affect wastewater treatment efficiency in MDCs because studies showed that high chloride concentration is inhibitory to biological treatment, especially for nutrient removal. 41 In addition, the AEM between the anode and middle chamber inhibits the free transfer of H + accumulated in the anolyte to other chambers, which cause s a significa nt pH drop in the anode chamber and pH increase in the cathode chamber 30, 42 A previous study show ed that the pH of the wastewater ano lyte dropped to 4.2 in one batch cycle if no buffer was added to the anolyte 35 Additionally, the catholy te pH could increase to 11 13 due to the loss of H + 20, 42 Such pH fluctuation significantly inhibits bioelectrochemical reaction efficiency and reduces system performance.

PAGE 40

26 In order to modulate the movement of salts to the anode and cathode chambers, the concept of capacitive deionization (CDI) was incorporated in this s tudy to develop a sustainable desalination system called a microbial capacitive desalination cell (MCDC). In the proof of concept MCDC, s alt water can be deionized through electrochemical ion adsorption driven by the electric al field generated by microorganisms. Two activated carbon cloth (ACC) membrane assemblies were designed to connect with the anode and cathode and adsorb ions from water. During desalination, the ions are stored in the electrical double layer capacitors between t he solution and the ACC assembly interface s, thus preventing the salinity increase in treated wastewater. After the ACC is saturated with adsorbed ions the assembly can be regenerated by removing the electrical potential and the retained salts can be full y recovered in situ for disposal or further salt recovery. Another innovative aspect of the MCDC as compared to conventional MDC, is the use of a second CEM in lieu of AEM between the anode and desalination chamber (Figure 3.1). Such configuration allows f or cations and protons to move freely from the anode chamber throughout the reactor and therefore maintains electrochemical neutrality and prevents pH fluctuation. In this study, the proof of concept MCDC development and operation were demonstrated, and it s advantages over current systems and application potentials were discussed. 3.3 Materials and Methods 3.3.1 MCDC Reactor Design The MCDC reactors consisted of three polycarbonate cube shaped blocks with 3 cm diameter holes forming an internal anode, cath ode, and desalination chamber volume of 23 mL, 27 mL, and 10 mL respectively. The anode and cathode chambers had a length

PAGE 41

27 of 4 cm, while the desalination chamber ha d a length of 1.5 cm. The anode electrode was a graphite brush (Golden brush, CA) and pretre ated by washing in acetone and heating to 350C for 30 minutes 43 Traditional air cathode s were made by applying 10% Pt/C (0.5 mg/cm 2 ) and four PTFE diffusion layers on 30% wet proofed carbon cloth as previously described 44 The desalination chamber was separated from the anode and cathode chamber by two assemblies. Each assembly wa s constructed by p lacing a cation exchange membrane ( CMX SB, Astom Corporation, Japan), a Ni/Cu Mesh current collector (McMaster Carr, IL), and 3 layers of Zorflex activated carbon cloth (ACC, Chemviron Carbon, UK) together. Additionally, the CEM faced the anode/catho de chamber to prevent microbial growth on the assembly. The total weight of the ACC was 1 gram with the specific surface area of 1019.8 m 2 /g, determined by the Brunauer Emmet Teller (BET) method (ASAP 2020, Micromeritics, Norcross, GA). 45 The ACC assemblies were connected to the a node/cathode by titanium wires ( Figure 3.2).

PAGE 42

28 Figure 3 1 Operation of MCDC System The operation of the MCDC system consists of wastewater entering the anode chamber, and salt water enter the desalination and cathode chambers. When electrical potential is applied to the membrane assemblies ions inside the desalination chamber become physically and electrically adsorbed. Flowing out of the reactor is treated wastewater, and desalinated water. When the system is disconnected from the anode and cathode chambers the previously adsorbed salts, desorb and can be collected. 3.3 .2 MCDC Operating Conditions Two reactors were inoculated with anaerobic sludge from the Englewood Littleton Wastewater Treatment Plant (Englewood, CO) and operated in fed bat ch MFC mode. When a repeatable voltage profile was obtained for consecutive batch cycles, the reactors were shifted to fed batch MCDC mode by inserting a pair of assemblies and adding one middle chamber as described previously. The anolyte growth media con tained per liter: 1.6g NaCH 3 COO, 0.62g NH 4 Cl, 4.9g NaH 2 PO 4 H 2 O, 9.2g Na 2 HPO 4 0.26g KCl, and 10mL trace metals and 10mL vitamin solution. 46 The catholyte contained per liter: 10g KCl, 0.68g KH 2 PO 4 0.87g K 2 HPO 4 Potassium chloride was used in the cathode chamber to differentiate with sodium transport and monitor the movement of cations from the cathode to the desalination chamber. The salt solution in the desalination

PAGE 43

29 chamber contained per liter: 10g NaCl, 0.49g NaH 2 PO 4 H 2 O, 0.92g Na 2 HPO 4 A small amount of buffer was added to the salt solution to some extent mimic the 300 to 700 mole/kg natural buffering capacity of seawater and prevent potential scaling at high pH values. 47 Figure 3.2 Diagram of MCDC System Diagram of MCDC reactor configuration and operation. Two CEM ACC assemblies were used to separate the three chambers and capture the removed salts, as well as allow for the free trans fer of protons. Two experimental procedures and two controls were performed to investigate the desalination performance of the MCDC system. The first experiment investigated simultaneous physical and electrical adsorption capacity by directly adding salt solution into the desalination chamber equipped with ACC assemblies free of adsorbed ion. When the anode and cathode electrodes were connected to the ACC assemblies, physical and electrical adsorption on the ACC assemblies could occur concurrently. The second experimental procedure investigated only electrical adsorption capacity. Electrical Anode Chamber Cathode Chamber Ai r Cathode Desalination Chamber e H + H + H + Cation Cation Cation Anion Anion Anion CEM/ACC/NiCu NiCu/ACC/CEM Anode With Microbe s e

PAGE 44

30 adsorption capacity of the ACC assemblies was determined by first adding salt solution to the desalination chamber to allow complete physical adsorption. Electrical adsorption was then characterized by replacing the desalted solution with fresh solution, and connecting the two assemblies to the anode and cathode, respectively. Any residual water from previous experimental washing would have been removed when the salt solution was replac ed. Abiotic control experiments were performed by using new brush anodes without bacterial acclimation. The first control experiment measured the physical adsorption capacity by short circuiting the assemblies to ensure no charge was formed across the ele ctrodes. The adsorption capacity of the assemblies was defined as the change in initial and final salt concentration. The second control investigated the electrical adsorption capacity by first allowing complete physical adsorption to occur then by connect ing the assemblies to an external power supply at a voltage of 0.53V to simulate the voltage generated by a microbial fuel cell. The MDC control experiment used an anion exchange membrane next to the anode chamber (Astom Corporation, Japan) and a CEM next to the cathode chamber without ACC assemblies in the desalination chamber. An external resistor of 1000 Ohms was used between the anode and cathode electrodes, and all other experimental procedures were identical to the MCDC experiments. To regenerate the ACC assemblies in situ for all experiments, the assemblies were either allowed to naturally regenerate or were regenerated by applying an external voltage to increase the rate of regeneration. The natural regeneration was performed by disconnecting the an ode and cathode from the assemblies and creating a short circuit between the assemblies with an external wire. Alternatively, an external voltage of 1V in

PAGE 45

31 reverse polarity was applied to the assemblies by a programmable power source. The external voltage w as applied for 5 10 minutes and followed by short circuiting the ACC assemblies, as mentioned above, for 20 30 minutes. When the potential difference reached 0.5mV, the ACCs were assumed to be regenerated, meaning that any electrically adsorbed ions shoul d have been removed from the electrodes. After regeneration all electrolytes were emptied and washed with deionized (DI) water to remove any residual salt remaining in the chambers before starting a new batch cycle. 3.3.3 Analysis and Calculations Conduct ivity and pH were measured for all three chambers using a conductivity meter and pH meter (HACH Co., CO). The change in the reactors internal resistance was determined through electrochemical impedance spectroscopy (EIS) tests using a Potentiostat. EIS mea surements were performed using the anode as the working electrode, the cathode as the counter electrode, and a saturated Ag/AgCl reference electrode placed in the anode chamber. Results were fitted into equivalent circuit models developed in our previous E IS studies and plotted using Nyquist plots where the ohmic resistance is defined as the intercept of the Zreal axis. 21 Samples of all three chambers were collected before and after desalination, and after regeneration. Ion concentrations were measured usin g the Optima 3000 Inductive Coupled Plasma (ICP) Spectrometer (Perkin Elmer, CT) and Dionex DC80 ion chromatography system (IC) (Dionex, CA). Using the data from the IC and ICP a mass balance of the major ions were determined by summing the concentrations of the ions in each chamber initially, after desalination, and after recovery of the salts. Internal power used was calculated using the following equations:

PAGE 46

32 Where P is power in terms of watt resistivity, L is length of the resistance, and A is the cross sectional area. Comparisons between the MCDC and CDI were made based off either presented data, or estimated from figures in published papers. Comparison to membrane capacitive deionization (MCDI) was not conducted to do the incompatibility in methodology to the MCDC. 3.4 Results and Discussion 3.4.1 Reactor Desalination Performance During MCDC operation an electrical potential was generated across the microbial anode and air cathode and applied to the two ACC assemblies to form a double layer capacitor 48 55 ( Figure 3.2 ). The formation of the double layer capacitor has been fully modeled using the Gouy Chapman Stern theory. 54 The potential drives the ions to move from the salt solution and adsorb on the activated carbon cloths. The ion adsorption can be observed proportional to the charge formed between the ACC assemblies (Figure 3.3 ). Figure 3.3 shows that i n repeated batch cycles that when the potential on the assemblies increases from 0 to more than 530 mV in each cycle, the solution conductivity in the desalination chamber decreased by 12 18%, from 18 mS/cm to below 16 mS/cm. The desalination rate was the greatest at the beginning of each cycle and then decreased gradually, suggesting the adsorption capacity of the ACC assemblies decreased along with the increased amount of salt been adsorbed in the assemblies. Salt removal was characterized by both conduct ivity, measured using a conductivity meter, and total

PAGE 47

33 dissolved solids (TDS) concentration, measured by IC and ICP ( Table 3. 1 ). Through simultaneous physical and electrical adsorption, the MCDC removed 26.9% of the conductivity or 25.5% of TDS from the des alination chamber in one batch cycle. In addition, a small percentage of salt was removed from the anolyte (4.4%) and catholyte (10.4%) as well. This is likely due to the ions being driven across the membranes by the electrical potential of the ACC assembl ies from the anode and cathode chamber then adsorbed onto the ACC. Further experiments showed that electrical adsorption alone removed 22.3% TDS from the desalination chamber, which contribute d up to 88 % of the TDS removal compared to the combined physical and electrical adsorption experiments. Table 3. 1 MCDC Salt Removal Both physical and electrical adsorption for the MCDC system. Table results indicate the total removal capacity of the system in TDS per gram ACC electrode. Physical/Electrical Adsorption Electrical Adsorption Desalination Chamber Anode Chamber Cathode Chamber Desalination Chamber Anode Chamber Cathode Chamber % removal in conductivity 26.95.1 13.13.8 5.64.4 10.00.2 10.63.5 2.02.7 % removal in TDS 25.23.6 4.43.6 10.43.6 22.33.6 7.63.6 23.6 Total TDS adsorption (mg TDS/g ACC) 72.7 50.7 Table 3. 2 compares the normalized TDS removal between the MCDC and CDI studies. The results showed that for the same amount of adsorptive material (ACC), the MCDC improved TDS adsorption by 7 25 times. Both MCDC and CDI use an electric field between two electrodes that electrochemically adsorb ions, but the high adsorption from the MCDC may attributes to the unique feature that the MCDC uses the internal potential generated by microorganisms. Such in situ approach avoided the use of external

PAGE 48

34 power supply and circuit and reduced transportation energy loss, so it demonstrated higher efficiency than traditional CDI processes. The salt adsorption rate in MCDC, however, is lower than published CDI studies, and that is mainly due to the low kinetics of the fed batch operat ion and the limited amount of ACC available for ion adsorption. In this study, t he MCDC configuration was modified from traditional cubic type MDCs, which only allowed for a total of 1 g activated carbon cloth being used in the assembly. This may explain w hy the amount of salt removed in the desalination chamber was relatively small. I t was calculated that the amount of salt added in the desalination chamber (114 mg TDS) was drastically beyond the control electrical adsorption capacity of the ACC (8.5 mg TD S for the 1g ACC applied). Table 3.2 Comparison of Different Capacitive Deionization Methods The table illustrates the advantage of the MCDC system for capacitive deionization. Method Electrode Materials Electrode Distance (mm) Net Wh/g TDS removed mg TDS/g adsorptive material Reference # MCDC Activated Carbon Cloth 15 2.18 50.74 This paper CDI Carbon Aerogel 2.3 +0.21 7.00 17 CDI Activated Carbon Powder NA +1 1.95 24 CDI Activated Carbon Powder 0.1 +1.78 2.88 25 CDI Activated Carbon Powder 0.1 +1.68 3.11 26 CDI Activated Carbon Powder with Micr oporous Carbon Black 0.22 NA 3.82 27 CDI MnO 2 /nanoporous Carbon Composite NA NA 0.10 28 CDI Activated Carbon Cloth NA +0.52 NA 30 CDI Activated Carbon Cloth with Titania NA NA 4.38 34 Moreover, compared to CDI systems that consume 0.21 1.78 Watt hour external energy to generate the potential to remove 1g TDS, the MCDC system does not use any external energy but instead utilized the in situ potential difference between the ACC

PAGE 49

35 assemblies generated during microbial activities. It was calculated that the MCDC reactor saved 2.18 Watt hour for 1g of TDS removed. That is why in Table 3.2 the net energy used for the MCDC is negative, indicating th at 2.18 Wh/g TDS removed was not required, while for the CDI systems an external energy of 0.2 1.78 Wh is required for removal of 1g TDS. While the MCDC reactor directly uses generated current for desalination, it is possible for electricity to be generate d by applying an external load across the ACC assemblies during regeneration. Reactor configuration optimization is underway to increase the ACC loading and further improve desalination efficiency. Figure 3 3 Correlation of Applied Potential and Deioniz ation Figure 3.3 demonstrates as the charge across the membrane assemblies increases the amount of deionization also increase. Results demonstrated for three typical cycles. Sodium, chloride, potassium, and phosphate accounted for greater than 85% of the TDS, and their specific concentration changes in the three chambers are shown in Figure 3.4 In addition to direct capacitive electrical adsorption that caused concentration decreases in the desalination chamber, a small amount of charged ions migrated fr om the

PAGE 50

36 anode and cathode chamber to the desalination chamber due to the electrical potential or concentration gradient. However the desalination efficiency for the anode and cathode chambers is low compared to the salt removal in the desalination chamber d ue to the lack of electrical double layer adsorption and the inhibited anion transfer across cation exchange membranes. Results in Table 3.1 showed that saline water can also be used as the catholyte and partially desalinated. Further desalination can be a chieved by feeding Figure 3.4 MCDC Ion Migration The results from IC/ICP analysis demonstrate the ion migration in the MCDC system with ions moving from the anode, desalination, and c athode chambers.

PAGE 51

37 Figure 3.5 Electrochemical Impedance Spectroscopy for the MCDC Results from electrochemical impedance spectroscopy illustrate that the change in internal resistance illustrates that over the course of desalination the change in internal resistance only changes a small amount. The reactor internal resistance measured by EIS at the beginning of the desalination cycle was on average 8.5 Ohms. After desalination, the internal resistance increased to an average of 13 Ohms. The change in conductivity in the desalination chamber correlated closely with the change in internal resistance for the reactor over the course of desalination. The MCDC reactor ability to transfer electrons was not inhibited as occurs over the course of desalinati on in standard MDCs. It is theorized that this is due to the MCDC ability to maintain charge neutrality better than in MDC reactors. Because in standard MDC reactors charge neutrality is reached by ion migrating out of the desalination chamber, while in th e MCDC reactor charge neutrality is performed by ion migrating through the entire reactor.

PAGE 52

38 Figure 3.6 MCDC Regeneration The regeneration or collection of salts in the MCDC system was investigated using two methods. The red line illustrates regeneratio n using a 1V external power supply. The blue line show the regeneration when the ACC assemblies are connected in short circuit. 3.4.2 Assembly Regeneration and Salt Recovery The ion saturated ACC assemblies were regenerated using two approaches. The natural regeneration was accomplished by directly connecting the two assemblies in short circuit. The electrical potential across the assemblies was dissipated with the adsorbed sa lts being released back into solution. When the potential difference across the ACC assemblies reached 0.5mV, it was assumed that the ACC assemblies were regenerated with complete electrical salt desorption. The regeneration rate can be significantly incr eased by connecting the assemblies to an external power supply of 1V with reverse polarity to facilitate ion desorption (Supporting Information). Figure 3.5 shows that among the four major ion species, almost all of the electrical adsorbed salts were

PAGE 53

39 recov ered during assembly regeneration, shown as a direct correlation between the initial and recovered salt concentrations. The capability of in situ regeneration of the ACC assemblies is another advantage of the MCDC, because the assemblies can be reused many times without investing significantly in materials. Almost all of the adsorbed salts can be recovered in concentrates during regeneration, and the recovered salts can be dewatered or extracted for beneficial uses. Furthermore, MCDC stacks can be developed and integrated with reverse electrodialysis (RED) to capture the energy generated due to the salinity gradient across the concentrate and freshwater. 56, 57 The current MCDC is operated in batch mode, and the desalination and regenerated processes were conducted sequentially. More efficient operation can be a chieved by connecting multiple reactors in series or in parallel and operating them in complementary sequential batch reactor (SBR) modes. While some of the units perform desalination, others conduct assembly regeneration at the same time. This operation n ot only provides continuous flow of produced freshwater but also allows for the direct usage of the electricity produced from regeneration units for desalination units.

PAGE 54

40 Figure 3.7 MCDC Ion Recovery The amount of recover ions from regeneration are indi cated in Figure 3.7 for the four major ions investigated. It can be observed that almost 100% of the desalinated salts can be recovered. 3.4.3 Reduced pH Fluctuation F igure 3.8 shows the change in pH units among the three chambers over one typical batch cycle for both the MCDC and the control MDC. The initial pH values in the chambers were all within 7.0 0.2. The change in pH for the anode chamber in both the MCDC and the MDC was relatively small with a drop in pH of between 0.2 and 0.5 pH units, which wa s presumably attributed to the high buffering capacity of the anolyte. However, the catholyte had drastically different results between the MCDC and MDC, with the MCDC increasing in pH on average 1.5 pH units and the MDC increasing 4.4 pH units. Interestin gly the change in pH for the desalination chamber for the MCDC is greater than for the MDC control. Previous capacitive deionization studies showed that water electrolysis may cause slight pH variation at low voltages, which may explain the

PAGE 55

41 pH increase in the MCDC desalination chamber. 54 It is difficult to compare the MCDC results directly with CDI studies, because no known CDI experiments have been conducted at a set potential lower than 0.6 V. 47, 49, 50, 52 55 Further investigations should explore the cause of this phenomenon. Figure 3.8 Change In pH for MCDC Figure 3.8 show the change in pH values for the MCDC compared to the MDC system. The MCDC has reduced pH fluctuations in the anode and cathode chambers. There was a slight increase in the pH of the desalination chamber. Because the average percent change between the cathode and desalina tion chamber were essentially the same, it is assumed that the proton transfer capability of the reactor was not inhibited. The MCDC employs a CEM to separate the anode and desalination chamber. This is different from the AEM used in current MDCs and relea ses the pH fluctuations in the reactor. In traditional MDCs, anions (Cl ) migrate from the desalination chamber to the anode chamber to compensate for the accumulation of H +

PAGE 56

42 but because the AEM prevents the transfer of H + out of the anode chamber a decrea se in pH is observed. By using a CEM, the accumulated H + not only can transfer to the desalination chamber but also can transfer further to the cathode chamber and therefore solves the pH change problem in the entire MCDC reactor. Previous studies show tha t other ions such as Na + and K + also play important roles in maintaining charge balances across different chambers in microbial fuel cells, 58 but the majority of such ions are adsorbed in the ACC assemblies so electrolyte charge balance due to ion transfer is not a concern in the MCDC. 3.5 Conclusion The integration of capacitive deionization with microbial desalination pr ovides a sustainable solution that not only addresses the salt migration and pH fluctuation problems facing current MDC systems, but also improves salt removal and energy efficiency compared to CDI systems. Traditional MDCs remove salts from the desalinati on chamber, but they also add TDS to the anode and cathode chambers and may increase the volume of saline water significantly, depending on different operation configurations. The MCDC reactor demonstrated that desalination can be accomplished in the middl e chamber without adding salts to the anolyte and catholyte, and therefore released the concerns on the viability of wastewater treatment and reuse due to increased TDS concentration. This proof of concept system also demonstrates a microbial desalination reactor to reduce salinity in all three chambers of the reactor. The MCDC system offers a sustainable desalination, renewable energy production, and wastewater treatment. To maximize the benefits and prevent negative effects of salinity changes on the wast ewater anolyte, salt migration from the desalination chamber could be modulated

PAGE 57

43 by constructing modular plate shaped ACC membrane assemblies. If the added salt is desired in wastewater to improve the anolyte conductivity, regular MDC operation could be per formed. If the salt should be prevented from migrating into the anode chamber, the modular ACC assembly plate can be inserted into the reactor to perform salt adsorption. Such system integration and operation will provide microbial desalination systems gre at flexibility in salt migration management as well as better pH fluctuation control. Despite the potential benefits offered by the MCDC system, many challenges remain to be addressed based on the information collected from this proof of concept study. In addition to low cost material development that is required for all bioelectrochemical systems, the adsorptive material can be improved such as with silica or titanium modification. 59, 60 Reactor configuration needs to be optimized to provide more ACC loading and improve diffusion rate and adsorption capability. Modular stack reactors and fl exible operation strategies need to be developed to maximize the integration of desalination and assembly regeneration in multiple units, optimize water recovery as well as enhance salt migration management. The cost of the MCDC system is proportional to t he desired level of desalination. The MCDC system is cost competitive for low to high range brackish water (<20g TDS/L), above that level of salt concentration other more developed desalination technology would be more cost beneficial. Improvements in MC DCs will also benefit from the continued advances of other bioelectrochemical systems such as microbial fuel cells, and capacitive deionization, with the eventual goal of developing a full scale sustainable system directed toward the integration of multipl e functions, such as extracting energy from wastewater and water desalination.

PAGE 58

44 4. Investigations of Different Methods for the Construction and Operation of Spirally Wound Microbial Electrochemical Systems 3 4.1 Abstract Microbial electrochemical systems (M ES) are a platform technology designed to generate electricity, treat wastewater, desalinate salt water, and produce chemicals. In order for MES technology to be scaled for practical application, further research needs to be conducted to ensure sustainabi lity. The MES reactor design and configuration greatly affects its sustainability. Demonstrated here for the first time are three spiral wound (SW) reactor configurations that incorporated a fully wound anode cathode and membrane assembly. Spiral wound I (SWI) is a single chamber air cathode reactor, which achieved a maximum power point (MPP) of 27 W/m 3 Spiral wound II and III (SWII, SWIII) are two chamber reactors using potassium ferricyande as the catholyte. SWII used two cation exchange membranes t o enclose the anode chamber, while SWIII used two anion exchange membranes. SWII and SWIII had the same MPP of 29 W/m 3 After a month of operation SWI power density decreased as the polarization resistance increased. This phenomenon was not observed in SWII and SWIII. Additionally, it was found that the most favorable operating condition for SWI was at the longest retention time of 5.37 hours. Results presented here indicate that a two chamber spiral wound reactor using anion exchange membranes would yi eld the most sustainable spiral wound configuration. 3 The work presented in this chapter is in preparation for publication by Casey Forrestal, Pei Xu, Peter Jenkins, and Zhiyong Ren. Water Res.

PAGE 59

45 4.2 Introduction The world is currently facing complicated challenges associated with meeting the need for water and energy. Over one third of the world has an inadequate supply of safe drinking water, and rising energy consumption around the world is depleting our natural resources. The water energy nexus is a r elationship between these two seemingly separate topics. Energy is required for the production and treatment of water, and water is required for the production of energy. 32 One clear indicator of this relationship is in a newly emerging technology known as Microbial Electrochemical Systems (MES). In the last two decades bioelectrochemical systems, and related technologies, have been extensively researched because of their unique ability to address the water energy nexus with a single technology. 61 MES technology works by utilizing microorganisms to generate an electrical current by breaking down an organic or inorganic source of an electron donor. Numerous elec tron donors have been investigated including municipal wastewater, benthic sediments, and industrial wastewater. Using wastewater as the source of electricity production in a MES would convert the current nexus of requiring energy to treat to an energy ga ining system. Furthermore, because MES generate an electrical potential, many additional value added benefits can be obtained such as the production of hydrogen gas or the removal of ions through desalination. 62 In order to achieve the greatest benefit from MES technology, electricity produced and wastewater treated must be low cost and highly efficient. To accomplish this goal researchers have focused on optimizing reactor materials and reactor design. The basic MES cons ists of three parts: a container, electrodes, and separators. The container is the external reactor shape or housing for the reactor. The most commonly

PAGE 60

46 researched containers consist of a polycarbonate plastic in a cube shape. 63 The main purpose of the electrode is for generating or accepting electrons; however, MES electrodes hav e been used recently for other purposes such as desalination of salt water or the production of chemical. 63, 64 The need for a separator in a MES is specific to the purpose of the MES. Generally the name separators in MES research refers to membranes that prevent fluid or oxygen transfer throughout the reactor. 65 In addition to the materials used for the construction of the reactor the overall design is crucial for increasing the reactor performance. By improving the system design the overall cost of the reactor can be reduced, increasing the wastewater treatment efficiency and the amount of energy generated. With these improvements, MES technology will be able to move from the lab scale to commercial scale. Ther e have been numerous MES designs that have been developed and tested, such as the cube shape, plate and frame, tubular, serpentine, upflow, and helical reactors. 66 71 Recently, a reactor design was published called a spiral wound by Boyang Jia et al. in 2012. 72 Jia et al. 2012 was the first to demonstrate the spiral wound design had potential advantages in wastewater treatment and energy production. The spiral wound design has been developed in many applicatio ns related to energy and water production such as in lithium batteries and reverse osmosis membrane systems. Due to its compact nature the spiral wound design has several benefits, including a high surface area to volume ratio, reduced internal resistance and a smaller area footprint. This study investigates three different spiral wound reactors, and the potential challenges and advantages of each configuration. The spiral wound configuration published by Jia et al. 2012 consisted of two spirally wound membranes with an anode electrode within the membranes to form the anode chamber.

PAGE 61

47 The cathode electrode was not spirally wound, rather consisted of three pieces of platinum coated carbon paper placed next to the spirally wound anode membrane assembly. Air was pumped to supply dissolved oxygen to the cathode electrode. This paper presents for the first time a fully wound single chamber spiral wound air cathode reactor (SWI) where all reactor materials were spirally wound. Two additional reactor designs w ere investigated, consisting of a two chamber membrane spiral wound reactor using potassium ferricyanide as the catholyte (SWII and SWIII). Investigated here for first time are the advantages of using a cation exchange membrane (SWII) versus an anion exc hange membrane (SWIII). The spiral wound configuration provides an efficient, compact, and scalable design for further development in commercial applications. Additionally, because of the modularity of the reactor design it has the potential to be integ rated for desalination and hydrogen gas production. Comparison of the three SW reactors and performance characteristics can bee seen in Table 4.1 Table 4.1 Comparison of Spiral Wound Reactors The table outlines the three spiral wound reactor configurations and performance characteristics

PAGE 62

48 4.3 Materials and Methods 4.3 .1 Spiral Wound Configurations 4.3.1.1 Spiral Wound I Design The MES spiral wound reactor was designed for use as a single chamber air cathode. (Figure 4.1) The one exterior reactor wall was constructed by cutting a SKC gas sampling bag coated with Tedlar PVF (Dupont). This prevented oxygen from penetrating the anode chamber from the one exterior chamber wall. The anode electrode was made from activated carbon cloth ( ACC) (FM70, Chemviron Carbon) and was pretreated with acetone to remove residual contaminants. The projected surface area of the SWI anode was 245 cm 2 for anolytle fluid flow using Plum mers Amazing Goop (3.2 cm x 7 cm). A small piece of titanium wire was sealed to the edge of the ACC and placed through the outer wall. Figure 4.1 Diagram for Spiral Wo u nd I A) Anolyte influent fluid flow into the reactor through a center tube across the anode electrode in a be. B) The expanded layers of the spiral wound I reactor. Anolyte In fluent Anolyte Effluent Anode Cathode Channel Former Spacers Spacer Air c athode Anode Non permeable plastic B A

PAGE 63

49 The anode chamber was created by placing two pieces of plastic Mesh (0.63 mm x 15 cm x 15 cm) with diamond shaped holes (0.32 cm 2 ) on both sides of the anode electrode.(Figure 4.1) A traditional a ir cathode was made by applying 10% Pt/C (0.5 mg/cm 2 ) and four PTFE diffusion layers on 30% wet proofed carb on cloth. 44 The cathode electode had the same internal dimensions as the anode with an additional 2.54 cm x 11.4 cm overhang to connect t he electrode. Anolyte entered and exited the single chamber from a center chlorinated polyvinyl chloride (CPVC) manifold pipe which was 30.5 cm in length by 1.27 cm inner diameter. The manifold was fabricated to allow influent and effluent fluid flow by cutting the pipe in half and sealing one end of each pipe with epoxy. Then the two pieces were joined together with additional epoxy. (Figrue 4.2b) Four 0.64 cm holes were drilled into each side of the center tubing at 0.64 cm apart. The four holes were d esigned to be used as an inlet and exit to the reactor. The working volume of the reactor was 32 mL. The reactor was rolled to a width of 7 cm. A reference electrode port was added on the edge of the rolled reactor by inserting a polyvinyl chloride (PVC ) tube (O.D 0.9cm, I.D. 0.6cm) between the two pieces of Mesh and the outer wall. The PVC tube was plugged with an 11 mm septa (Restek) during operation. Air was allowed to naturally pass across the wound air cathode. The air cathode electrode and the oxy gen impermeable plastic layer were sealed together using silicon. Before adhering the plastic, the surface was scratched using sand paper to increase the surface area. Silicone was placed on the inside and outside of both layers to ensure a tight seal. ( Figure 4.2)

PAGE 64

50 Figure 4.2 Pictures of Spiral Wound I A) The fully sealed air cathode spiral wound showing the influent and effluent tubing. B) The center tubing with holes drilled for anolyte influent and effluent. 4.3.1.2 Spiral wound II and III Design Spiral wound reactor version II and III were identical except for the use of the a Neosepta cation exchange membrane (CEM) from Astom incorporated for SWII and an anion exchange membrane from Membranes International for SWIII.(Figure 4.3) SWII and SWIII were two chamber MESs, where the anode chamber was formed by two membranes. All components of reactors SWII and SWIII were fully wound including the anode electrode, membranes and cathode electrode. The reactors consisted of two 1.27 cm inner di ameter CPVC pipe, 35.5 cm in length. In both pipes at 12.7 cm from the end three 0.47 cm holes were drilled into one side of the pipe at 2.54 cm margins. Both ends of both pipes were tapped with a 0.63 cm #18 NPT tap. One pipe was used as the influent f or the anolyte while the second pipe was used for effluent flow. Two membranes were sealed in a spiral wound fashion with the center pipe used as the influent pipe and the outer pipe as the effluent pipe. The anode and cathode electrode dimensions were 3 0.5 cm x 10.2 cm for a total surface area of 3.1x 10 2 m 2 Both anode A B

PAGE 65

51 and cathode electrode material were activated carbon cloth (ACC) (Chemviron Carbon, UK). Figure 4 3 Picture of Spiral Wound II Spiral Wound II fully wound and sealed. A space betwe en the two sealed membranes provides a channel for catholyte fluid to flow. A total of 4 spacers were used for the reactor; one on the outside and inside of the anode electrode, and two spacers were placed on the outside and inside of the cathode electrod e. The spacers were made of a nylon plastic mesh as described for SWI. The dimensions for the membranes were 33 cm x 15.2 cm on the inside, and 35.5 cm x 15.2 cm on the outside. (Figure 4.4) Each of the four spacers were 2.5 cm different in length star ting from the inside at 24 cm, 26.5 cm, 29 cm, and 31.5cm. The anode chamber was sealed by using a combination of 3M wet surface epoxy and marine epoxy. Adhesive was placed on the inside and the outside of the anode chamber membranes. On the inside of t he membranes, a 1.2 cm width adhesive seal was used on the top and bottom to seal the anode chamber. Additionally, a 1.2 cm width adhesive seal was applied to the outside of the membranes on both the top and bottom to ensure a good seal. The surface of the membranes were scratched using a piece of sand paper to increase surface adhesion. The total volume of the anode chamber was 45.8mL.

PAGE 66

52 Figure 4.4 Computer Aided Drawing of Spiral Wound II and III The computer aided drawing shows what the ideal reactor should look like where the center tube is used for influent and the tube to the left of the center tube is used for the effluent. The space between the wound membranes allows for catholyte fluid to flow. Electrical connections were made by crimping a pie ce of titanium wire to the anode and cathode electrode and threading the wire to the outside of the reactor. The fully formed anode and cathode assembly was tested for leaks before startup by using combination air and water pressure tests. The wound anode and cathode assemblies of SWII and SWIII were designed to be used with a liquid catholyte such as ferricyanide. Catholyte fluid passively flowed on both sides of the anode chamber. (Figure 4.5) In order for this passive catholyte fluid flow to occur, th e spirally wound anode chamber and cathode electrode assembly was placed inside a 7.6 cm diameter x 20.3 cm PVC pipe forming the cathode chamber. The cathode chamber was sealed at the top and bottom with 7.6 cm diameter PVC caps. Catholyte entered the ca thode chamber from one end of the 7.6 cm diameter PVC cap through a 0.635 cm tapped hole and exited in the same manner on the opposite side of the reactor. The influent and effluent pipes of the spirally wound assembly were threaded through 5 cm holes dril led in the top and bottom caps,

PAGE 67

53 and sealed with epoxy.(Figure 4.6) The total catholyte volume was 490mL. Both reactors were horizontally fixed to a piece of plywood for operation. Figure 4.5 Internal Diagram of Spiral Wound II and III The drawing shows how the spiral wound II and III operates with influent and effluent. The layers of the wound material can also be seen, which consists of the an ode, spacers, and membranes. 4.3 .2 Reactor Start up a nd Operation 4.3.2.1 Spiral Wound I Operation The reactor was acclimated with microbes from a previously acclimated MES as well as by adding activated sludge from the city of Broomfield wastewater treatment facility. At startup, the anolyte solution was stationary inside the reactor. After the voltage reached greater than ~100mV at 1000 Ohm external resistance, the reactor was switched to batch with continuous fluid flow. Anolyte was pumped initi ally at startup with a with a peristaltic variable speed pump, of 0.1mL/min. After the reactor became fully acclimated with a voltage of greater than 500 mV at 1000 Ohm resistance, anolyte fluid flow was varied from 0.1mL/min to 1.5mL/min (Fisher Scientifi c Co., Fairlawn, Anode Influent Membrane/ Spacers Cathode/ Spacer s Membrane/ Spacers Wound Layers / Catholyte Influent Anode/ Spacers Anolyte Distribution

PAGE 68

54 NJ). The anolyte contained per liter: 20mM sodium acetate (1.6g/L), 11.6mM (0.62 g) NH 4 Cl, 35.5mM (4.904 g) NaH 2 PO 4 H 2 O, 64.5mM (9.15 g) Na 2 HPO 4 100mM, 3.4mM (0.26g) KCl, and 10mL trace metals, and 10mL vitamin solution. An external resistor of Ag/AgCl reference electrode was used for electrochemical analysis. The anolyte volume used for batch tests was 250mL. Figure 4.6 Operation of Spiral Wound III Figure 4.6 shows the computer, the pump, electrolyte solution one of the fully sealed spiral wound reactors in operation. Electrical and fluid connections are made at the ends of the reactors. 4.3.2.2 Spiral wound II and Spiral wound III SWII an d SWIII anolyte media were the same as above for SWI. The catholyte contained per liter: 16.5g/L (50mM) KFe(CN) 6 4.9g/L (35.5mM) NaH 2 PO 4 and 9.1g/L (64.5mM) Na 2 HPO 4 A peristaltic pump, (Masterflex, Cole Palmer) was used for anolyte and catholyte flu id flow at a flow rate of 1mL/min. An external resistor of added to the anolyte solution that was pumped into the reactor. Sludge and anolyte solution, as well as the cat holyte solution, was allowed to sit stationary inside the reactor until the voltage was greater than 100mV at 1000 Ohm resistance. SWII was further

PAGE 69

55 acclimated at a 1mL/min flow rate under batch conditions. SWIII was acclimated without the use of activated sludge. Effluent from the fully acclimated SWII was fed into SWIII at a flow rate of 1mL/min. 4.3 .3 Analysis and Calculations Voltage data was recorded using a data acquisition system (model 2700, Keithley Instruments, inc. OH) where the voltage drop ac ross the external resistor (Re) was recorded at a 1 minute interval. Power density was determined by linear sweep voltametry (LSV) using a potentiostat (Gamry Instruments). The reactors were disconnected from the external resistor for greater than one hou r prior to running LSV. Then reactors were connected to the potentiostat with the anode as the working electrode, the cathode as the counter, and reference electrode. The scan rate was set at 0.1mV/s, with a step size of 1mV and a maximum current of 10mA Cb ) was calculated using the change in Chemical Oxygen Demand (COD)(Hach, Loveland CO) under steady conditions for batch conditions. Where M is the molecular weight of oxygen (32g/mol), I is current (A), int egrated the number of electrons transferred for every mole of oxygen, V AN is the volume of the 73 The internal resistance of the reactor and the polarization resistance was determined by potentiostatic electrochemical impedance spectroscopy (EIS) at a fre quency range of 10 5 to 10 2 Hz, with a 10 mV sinusoidal perturbation. The anode

PAGE 70

56 electrode was used as the working electrode, the cathode electrode as the counter, and reference electrode. The data was plotted using a Nyquist plot, which indicates the inter nal resistance (Rin) defined as the intercept with the Zreal axis, and the polarization resistance, which is defined as the diameter of the semicircle by the graph. 74 Open circuit voltage (OCV) was calculated by disconnecting the reactors anode and cathode or the difference in potential energy of the stabilized anode and cathode electrodes. The OCV was achieved in about 1 hour under open circuit conditions. Figure 4.7 Acclimation Times For Different reactor Configurations The current produced in the spiral wound configurations increased dependant upon the reactor acclimation process. SW III acclimated the fa sted because it used previously acclimated anolyte from SW II. The maximum current produced was in SW II and SWIII.

PAGE 71

57 4.4 Results and Discussion 4.4 .1 Electricity Production and Efficiency of Spiral Wound I Upon start up of the reactor, it reached a maximum current of 540mA. The reactor was operated under batch continuous flow for greater than 30 days with only a slight variation in current produced.(Figure 4.7) The voltage would slightly drop because of a decrease in available COD concentration and the standard deviation for the reactor energy production varied by 0.035mA. Because the volume was relatively small compared to the overall size of the reactor, any interruptions in the flow of anolyte resulted in a quick drop in produced current. Figu re 4.8 Effect of Retention Time On Power Density SWI Power density curves for the various retention times indicate little change in maximum power point. The highest power point was achieved at the lowest flow rate 0.1 mL/min. The maximum open circuit p otential was 632mV which is similar other MES reactors operating with an air cathode. The maximum current produced, however, was

PAGE 72

58 operations could improve power production and r educe ohmic and polarization losses. Figure 4.9 Coulombic Efficiency Correlated to Retention Time and Maximum Power Point Power density curves for the various retention times indicate little change in maximum power point. CE varied greatly over changing retention times. The polarization and power density curves for the reactor were determined at varying hydraulic retention time (HRT). (Figure 4.8) The maximum power density was formed at a HRT of 5.37 hours or 0.1mL/min, which generated a maximum power point of 27.4 W/m 3 anode volume. (Figure 4.8) However, the maximum power point for the reactor operating at increasing flow rates had only a negligible effect on the power produced. At a flow rate of 1.5 mL/min, the maximum power point decreased to 20 W/m 3 Because the maximum power point for SWI did not decrease linearly, with a linear increase in flow rate, this indicates t hat potentially the reactor configuration could allow for a higher the organic waste and generate an electrical current. The highest CE for SWI was not

PAGE 73

59 achieved at th e longest HRT of 5.37 hours but rather at the retention time of 0.83 hours. (Figure 4.9). CE for SWI did not appear to be directly correlated to the maximum power point. Further research will help to better understand this phenomenon. Figure 4.10 Chan ge In Polarization Resistance Over Time From the Nyquist plot, the internal resistance and the polarization resistance can be seen. The internal resistance of the reactor did not change over time but the polarization resistance changed significantly. Th e maximum voltage, and maximum power density for SWI, was produced just after the reactor was fully acclimated. After the initial period of operation, the OCV and the voltage produced dropped from 540mV to approximately 510mV over a 50 day period. This dr op in OCV and voltage was correlated to the increase in polarization resistance as seen in Figure 4.10. The fully acclimated reactor had a starting internal resistance of 15 Ohm and a polarization resistance of 45 Ohms. After 26 days of operation the int ernal resistance remained at 15 Ohms, but the polarization resistance increased to 65 Ohms. After 49 days of operation the internal resistance remained almost the same as the starting point at 16 Ohm, but the polarization resistance increased to 85 Ohms. The increase in polarization resistance of 40 Ohms over 50 days of operation can be directly correlated to the drop in OCV. Because SWI was an air cathode electrode

PAGE 74

60 with minimized spacing between the anode electrode and cathode electrode, it is believed that the increase in polarization resistance was due to the increase in biofouling of the cathode electrode common with traditional air cathode reactors. 75 Figure 4.11 Spiral Wound II and III Initial Overshoot Potential The initial overshoot potential was high for the SWII and SWIII indicating that the microbes had not fully acclimated. 4.4 .2 Electricity Production and Efficiency of Spiral Woun d II and III The results for the SWII and III reactors indicate that through improved reactor construction and with the incorporation of the higher potential electron acceptor ferricyanide, the maximum current produced for SWII and SWIII was 680mA for both reactors. SWII was started one week before SWIII. SWII required 6 days to become fully acclimated using activated sludge. SWIII was acclimated much faster taking only 1.5 days. The faster acclimation was achieved by using the effluent from the fully acc limated SWII.(Figure 4.7) The maximum current produced in SWIII was 26% higher than SWI. The maximum power point for the reactors SWII and SWIII was not immediately achieved, even after operating the reactor at a stable high current for over

PAGE 75

61 three weeks. This was indicated in LSV test which showed both reactors having a large overshoot potential which prevented the observation of the maximum power point. (Figure 4.11) The maximum power point with the overshoot potential was 20 W/m 3 To resolve the overs hoot potential for both SWII and SWIII, both reactors were operated under a low external resistance of 2 Ohms for 3 weeks. (Figure 4.12) Figure 4 12 Spiral Wound III Loss of Overshoot Potential Overshoot potential was overcome by acclimating the reactor at a low external resistance. The change in power density can be seen proportional to time. The maximum power point of the acclimated reactors SWII and SWIII was 29 W/m 3 anode volume after resolving the overshoot potential. By week 4 of operation, SWII which used a CEM membrane from the Astom Corporation developed a small leak inside the reactor. This was first identified by a color change in the anolyte effluent. The nor mal effluent color was clear slightly pink, as potassium ferricyanide leaked into the anode chamber the color slowly changed to a bluish green. The reactor was taken offline to identify and seal the leak. After an exhaustive effort, the location of the le ak could not

PAGE 76

62 be identified. Identifying leaks for the spiral wound configuration and sealing the membranes are two of the most difficult aspects of the spiral wound configuration. The leak for SWII was later identified by disassembling the reactor. The le ak was found towards the center of the anolyte influent tubing. No further experiments were conducted with SWII. Unlike SWI, there was no noticeable change in polarization and internal resistance of SWII or SWIII over 4 weeks of operation. This is likely due to the use of potassium ferricyanide as the catholyte and completely sealing the anode chamber from the cathode chamber. However, the internal resistance for SWII and SWIII were larger than SWI. This is because SWII and SWIII used two membranes to e nclose the anode chamber, thus increasing the internal resistance. The internal resistance for SWII was approximately 25 Ohms and SWIII was slightly higher at 27 Ohms. (Figure 4.13) The difference in internal resistance for SWII and SWIII indicates that the Neosepta CEM membrane from Astom Incorporated led to a smaller resistance than the Membranes International anion exchange membrane. Additionally, the polarization resistance for the Astom membrane was smaller than the polarization resistance for the m embranes internal membrane. SWII had a polarization resistance of approximately 30 Ohms while SWIII had a polarization resistance of approximately 50 Ohms. Even though the reactor resistance profiles were significantly different, the maximum OCV and the maximum power point for both reactors were almost identical. This result indicates that using an anion exchange membrane with a higher polarization resistance and a slightly higher internal resistance yields almost the same results as using a cation excha nge membrane with a lower internal resistance and polarization resistance. It was theorized that using the higher priced Astom corporation membrane would yield the highest performance.

PAGE 77

63 However, because of the improved performance generally indentified wit h an anion exchange membrane, the higher resistance of the Membranes International membrane can be overcome. SWII and SWIII had a higher internal resistance over SWI because of the use of two membranes the increase in maximum power point for SWII and SWII I was still 31% higher than SWI. Figure 4.13 Comparison of EIS for the Three Spiral Wound Reactors The Nyquist plot for the three spiral wound reactors indicates a high internal resistance for the SWII and SW III. 4.5 Conclusion The results from the c onstruction and operation of three different spiral wound reactors indicates that while the use of a two membrane system will increase the internal resistance of the reactor, it may not necessarily reduce the performance. The maximum power achieved for the spiral wound reactor is more dependent upon the operating conditions rather than the configuration. By using the higher strength electron acceptor potassium ferricyanide, the highest power point was achieved despite the larger internal resistance. For S WI, the difficulty in constructing the serpentine flow path may be

PAGE 78

64 prohibitory for the further development of the spiral wound system. However, because the SWI serpentine flow path uses the same manifold for influent and effluent, the total area used can b e reduced by up to 25%. For further development of the spiral wound system this reduction in total area would reduce the cost for installing the system at a wastewater treatment facility. SWII and SWIII, in addition to being compared to the single cham ber SWI, were also used to compare the performance of an AEM verses a CEM membrane. The performance characteristics between SWII and SWIII indicated that while a slight fluctuation was observed, the results between SWII and SWIII were similar. This indica tes that the use of the more expensive Astom membrane would not be beneficial to the commercialization of the technology. Additionally, the small leak that formed in SWII indicates that further research needs to be conducted on membrane sealing techniques to ensure sustained performance. Ideally, a new reactor design should be developed where the reactor could be easily sealed and unsealed if necessary allowing access to replace materials or resolve problems as they arose. Furthermore, all of the SW reac tors used a significant quantity of adhesives to seal the reactors. This increases the cost of the reactor and may prohibit it from being further commercialized. Further research needs to be conducted towards developing an adhesive less spiral wound reac tor, which will reduce the cost of manufacturing and potentially simplify the manufacturing process. Currently, the cost of the spiral wound reactor would not make it a competitor against large scale municipal wastewater treatment facilities. Until the s piral wound design makes a major breakthrough in energy production, the advantage of the design

PAGE 79

65 will be for site specific conditions such as on a boat or a small island. The operation of the SW MEC for wastewater treatment would need to be performed simil arly to traditional municipal wastewater treatment facilities. The only major difference would be replacing the aeration basin, or trickling filter, for the SW MES. The biggest potential advantage for operating a MES system on municipal wastewater would b e the reduction of produced sludge. At the current stage of the SW MES development, energy production would not be its main business attractant. The compact nature of the technology and its minimal requirement for operation and maintenance are the bigges t assets. As indicated, the use of an anion exchange membrane verses a cation exchange membrane does not significantly alter the performance of the reactor. However, the configuration of the reactor whether single or two chambers, does drastically affec t the performance of the system. The single chamber reactor had a problem of biofouling on the cathode which decreased its long term performance. The ideal reactor configuration appears to be an inexpensive anion exchange membrane in a two chamber system With improvements in reactor design and operation, and further reduction of internal and polarization resistance, the SW MES could achieve higher power densities and increased coulombic efficiencies. While there are still many challenges to address wit h the SW configuration, the small environmental footprint and high surface area to volume ratio, makes the SW a potential commercially viable product.

PAGE 80

66 5. An Efficient Method for Treating and Desalinating Produced Water Using a Microbial Capacitive Desalin ation Cell 4 5.1 Abstract A microbial capacitive desalination cell (MCDC) is a simultaneous wastewater treatment, electricity production, and desalination device. Demonstrated here is a MCDC system that can efficiently treat produced water without requirin g an external energy source. Produced water (PW) is a term generally associated with the production of water from oil and natural gas mining. Produced water consists of mainly organic and inorganic salts. The produced water MCDC system is a three chambe r desalination system which uses capacitive deionization (CDI) to in situ remove dissolved salts and organics. The MCDC ability to remove organics, salts, and generate electricity is demonstrated for the first time. Additionally, a maximum desalination c apacity was identified. A proof of concept study using a small cube MCDC system, indicated that the desalination chamber on average removed 450 mg TDS/L/g ACC/hr, and the anode chamber removed 40 mg TDS/L/g ACC/hr. The cumulative ions removed in the MCDC system was 390 mg TDS/L/g ACC/hr accounting for a small fraction of ions that migrated from the desalination chamber into the cathode chamber. The MCDC removed COD from the anode chamber at an average rate of 10 mg COD/L/hr. The desalination chamber had a high COD removal rate of 93 mg COD/L/hr. Additionally with the use of a charge pump connected to a 2.5V, 12F capacitor, a cumulative sum of 0.054 mJ of external electricity was recovered from the in situ MCDC capacitor from a single desalination cycle. Experiments investigating the maximum desalination capacity of the

PAGE 81

67 MCDC system indicated that, in three successive cycles, greater than 65% of the TDS in the desalination chamber could be removed. Additionally, the CDI assembly was able to remove 83% of the COD in the PW in the desalination chamber. The MCDC system operating under physical and electrical adsorption had an adsorption capacity of 65 mg TDS/g ACC. Due to the ability of the MCDC system to simultaneously treat the organic and dissolved so lids content, as well as produce direct electrical current, the MCDC system is a potential solution for site specific produced water treatment. 4 The work presented in this chapter is in preparation for publication by Casey Forrestal, Pei Xu, Peter Jenkins, and Zhiyong Ren. Environ. Sci. Technol.

PAGE 82

68 5.2 Introduction Water and energy are directly related in energy production and consumption and the need for sustainable solutions to generate clean water and energy have never been more important. 32 Water and energy are also closely related to global conflict, which has caused the United States to invest heavily in domestic sources of energy. 76 Within the next ten years many countries important to the United States will experience water problems which will risk instability and incre ase regional tensions, according to a 2012 intelligence report by the US State Department. One rapidly emerging domestic energy source is the production of natural gas from hydraulic fracturing. Hydraulic fracturing works by pumping large volumes of water and chemicals into deep underground shale rock formations. This process releases the trapped methane which is extracted from the flowback water. After all of the valuable products have been removed from the flowback water, the remaining water is referred to as produced water (PW). In the United States it was estimated that up to 2.3 billion gallons of produced water are generated everyday. 77 More recent estimates indicate that the average daily volume of generated produced water is around 80 million barrels(bbl) (42 gallons/bbl). 78 The produced water contains a wide range of hydrocarbons, salts, hazardous chemicals, and naturally occurring radioactive material, which is highly dependant upon the geologic conditions of the natural gas well. Salt concentrations range from a few mg/L to over 300 g/L, and hydrocarbons can range up to 2000 mg/L Total Organic Carbon (TOC). 78 Because produced water contains both high salinity and high organic content, many traditional wastewater treatment options are not viable. Some produced water does not require treatment and can be used directly for beneficial purposes such a s crop irrigation, water for livestock, or upstream

PAGE 83

69 augmentation. If the PW is too highly concentrated in salts or hydrocarbons, the water needs to be treated to regulator standards prior to discharge. Discharge requirements are also highly variable depe nding upon the location and method of disposal. Because of the need to treat both organics and inorganics the cost of treating PW can be expensive. The highest costs observed for the treatment of produced water is related to the transportation of the PW to a treatment facility. 77 If onsite treatmen t is viable for the producing well, multiple stages are required to treat the PW. The first step is to remove the suspended oil through an air flocculation separator; 100% of natural gas well have this system installed. 77 Following the removal of the residual oils, large particles are removed though a centrifuge, hydrocyclone, or filtration process. The remaining organic content is removed with traditional biological wastewater treatment methods such as continuous activated sludge, solid retention basin, trickling filter or through the construction of a wetland. Removal of inorganic salts is most commonly achieved through the use of a membrane filtration system such as reverse osmosis, microfiltration and ultrafiltration. The least expensive options for disposal of the PW are reinjection into the s ubsurface or surface discharge to an evaporation pond. The cost for this method of disposal ranges from $0.30 $10/bbl. 79 A thermal treatment facility can co st up to $105/bbl. The average cost of produced water treatment is around $3/bbl but is highly dependent PW conditions. The main reasons for the high cost to treat PW are energy requirements and waste disposal costs. Sludge disposal can account for 40% o f the total cost of PW treatment. 77 Recently researchers around the world have been heavily investigating methods of using wastewater to generate energy. 80 One of these methods is through the use of a Microbial Electrochemical Systems (MES) 28 The MES work by using microorganisms

PAGE 84

70 to breakdown organic or inorganic sources of electrons and transferring those electrons to a terminal electron acceptor such as oxygen. Over the last 2 decades, MESs has been developed to perform mult iple functions such as the treatment of municipal wastewater, production of hydrogen gas, and desalination of salt water. 3, 39, 81 The microbial desalination cell (MDC) was a major breakthrough in simultaneously treating wastewater, generating an energy source, and desalinating salt water. The MDC works by generating an electrical potential which is used to desalinate salt water containe d in a separate chamber by electrodialysis. Therefore, the salts in a traditional MDC only migrate from the center desalination chamber to the anode and cathode chambers. To alleviate this problem, the Microbial Capacitive Desalination Cell (MCDC) was de veloped. 63 The Microbial Capacitive Desalination Cell (MCDC) used microorgan isms to generate an electrical potential which is applied onto high surface area electrodes within the reactor in order to capacitively desalinate salt water. The desalination process of CDI, in the MCDC, has been modeled using the Gouy Chapman Stern model which was adapted from the double layer model by Herman von Helmholtz in 1883. 6 When the electrical potential is removed, the capacitively desalinated salts are removed from the high surface area electrodes and are collected. The MCDC system provides a method of wastewater treatme nt, energy generation, and desalination without contaminating the anode and cathode chamber with the desalinated salts. Presented in this paper for the first time is a proof of concept study in which produced water was used to generate the electrical poten tial needed to desalinate the organics and dissolved salts in produced water, and generate and external electrical current. The development of the PW MCDC system can potentially transform the treatment of PW from a multistage process system

PAGE 85

71 requiring a lo t of energy, to an energy gaining process. The PW MCDC system was first evaluated to remove the organics and salt from the PW. After demonstrating the capability of the MCDC to treat and desalinate PW, the MCDC experimental conditions were modified to ach ieve a maximum desalination capacity in the shortest amount of time. Additionally, the MCDC system was compared to the traditional MDC system to evaluate advantages and challenges of both systems. 5.3 Material and Methods 5.3.1 Design of Microbial Capac itive Desalination Cell The MCDC reactor consisted of three small cubic polycarbonate chambers with 3cm diameter hole forming an internal anode, desalination, and cathode chamber volume of 23, 12, and 27mL respectively. (Figure 5.1) For the proof of conce pt study, the anode chamber was connected to a 200mL reservoir containing raw produced water from a natural gas well in Colorado. The PW was recirculated into the anode compartment at a rate of 1mL/min (Master flex, Cole Palmer). The total volume for the anode chamber was 223 mL. The anode and cathode chambers had a width of 4 cm while the desalination chamber had width of 1.5cm. A carbon brush electrode (Golden Brush, CA) was used as the anode electrode. The anode electrode was pretreated by washing in a cetone and preheating to 350C for 30 min. The cathode electrode was a 9cm 2 traditional air cathode, coated with 0.5mg/cm Pt/C (10%) and four PTFE diffusion layers on 30% Teflon coated carbon cloth. 44 Inside the desalination chamber, a capacitive deionization (CDI) system was used to desalinate the PW. The CDI consisted of two electrodes each split into three parts and p lace overlapping to reduce the spacing between the electrode. Each of the three parts of the electrode contained a central current

PAGE 86

72 collector (2.5cm 2 ) made of Ni/Cu mesh (McMaster Carr, IL), a total of two pieced of activated carbon cloth (ACC) were placed (each 2.5cm 2 ) ( Chemviron Carbon, UK) on both sides of the current collector. On the outside of the Ni/Cu/ACC assembly, a fine plastic mesh separated the CDI electrodes (3cm 2 ). In total, the CDI consisted of six pieces of Ni/Cu mesh, twelve pieces o f ACC, (0.72g) and twelve pieces of fine plastic mesh. Three pieces of the fully assembled Ni/Cu were welded together to form a single electrode. Fluid flow between the electrodes was achieved by the design of the plastic mesh. The space between two elect rodes was between 1.2 2mm. The anode chamber, desalination, and cathode chambers were separated by pieces of CEM ( CMX SB, Astom Corporation, Japan ). No external resistance was used during the operation of the MCDC. Figure 5.1 Oper ation of MCDC for Produced Water Diagram of the MCDC shows an expanded view of the capacitive deionization component inside the desalination chamber. C EM CEM Anode Cathode Desalination Ni/Cu Conductive Plate ACC electrodes Non conductive Mesh spacer Energy Harvester

PAGE 87

73 5.3.2 MDC D esign The design of the MDC was identical to the MCDC except the desalination chamber volume was increased to 15 mL because it did not contain the CDI component. An AEM ( AMX SB, Astom Corporation, Japan) was placed next to the anode chamber and a CEM ( CMX SB, Astom Corporation, Japan) was placed next to the cathode chamber. An external re sistance of 1000 Ohms was used during the operation of the MDC. (Figure 5.2) Figure 5.2 MDC for Produced Water Diagram of the MDC shows the operation of the system for desalination. No additional components added to desalination chamber. 5.3.3 Operation of the MCDC and MDC Reactors for Proof of Concept Study Both the MDC and MCDC reactors were fully acclimated in MFC mode before converting the reactors to either the MCDC or MDC for desalination. Prior to converting the reactors, the anode electrodes were acclimated by adding 13 mL of activated sludge from the Broomfield Municipal Wastewater Treatment Facility alon g with 10 mL of anolyte growth media. The anolyte growth media contained per liter: 1.6g NaCH 3 COO, R AEM CEM Anode Cathode Desalination

PAGE 88

74 0.62g NH 4 Cl, 4.9g NaH 2 PO 4 H 2 O, 9.2g Na 2 HPO 4 0.26g KCl, and 10mL trace metals and 10mL vitamin solution. 46 After the anode electrode became fully acclimated over a one month period, the anolyte so lution was slowly transitioned to 100% raw PW by adding 10 mL of raw PW to the anode with the anolyte media described above. This was done over a two week period until the anode chamber completely used raw PW. After converting the reactors for desalinati on, raw PW was also immediately used in the desalination chamber. For all proof of concept experiments, the catholyte contained per liter: 9.8 g/L (71mM) NaH 2 PO 4 H 2 O and 18.3g/L (129mM) Na 2 HPO 4 The electrolyte solution was replaced with fresh media for the MCDC prior to starting a new desalination experiment. Electrolyte solution for the MDC was replaced when the voltage dropped below 50mV. The MDC reactor was connected to a data acquisition system (model 2700, Keithley Instruments, inc. OH) the voltage across the external resistor was recorded every 1.1 minutes. The MCDC system was not connected to the data acquisition system because previous experiments indicated that current interruption affected the MCDC desalination capacity. Electrical potential of the anode, cathode and CDI assembly was determined with a portable multimeter. Desalination cycles for the MCDC PW proof of concept study, were achieved by connecting the anode chamber to the pump and recirculating 200mL of PW for 25 minutes. During this time PW as added to the desalination chamber and phosphate buffer to the cathode chamber. The PW in the desalination chamber was replaced a total of three times prior to starting the experiment in order to insure that the CDI assembly was fully physicall y adsorbed with salts. The MCDC proof of concept study only investigated the capacity of the reactor to remove salts using electrical adsorption. For the MDC experiments, the PW was not replaced

PAGE 89

75 prior to start because the conductivity would not change pr ior to starting experiment. Before connecting the MDC and MCDC reactors conductivity and pH measurements for all three chambers were recorded. At the end of the desalination cycle, all electrolyte solutions (anode, desalination, cathode) were removed and recorded for conductivity and pH. The end of the desalination cycle for the MDC experiments was caped at 24 hours. Desalination for the MDC would have continued past 24 hours of operation. A 24 hour operation period of the MDC was chosen based off result s from the initial desalination runs. The end of a desalination cycle for the MCDC was determined when the voltage across the CDI became stable, approximately two hours. The CDI assembly was regenerated using one of two methods. The CDI electrodes were e ither short circuited and regenerated with in one hour, or the CDI electrodes were connected to an external charge pump ( S 882Z24, Seiko Instruments ) connected to a 2.5V, 12F capacitor for energy harvesting from the charged CDI. Energy harvested from CDI was determined based off the following equation: E = CV 2 Where E is energy in Joules, C is capacitance in Farads, and V is voltage in Volts. A control experiment was conducted for the MCDC to determine the ion migration if the CDI assemblies capacity was removed from operation. MCDC control experiment was conducted as outlined above, except the CDI assembly was connected in short circuit for the entire desalination cycle. The anode and cathode electrodes were connected to generate an electron motive force similar to the MDC.

PAGE 90

76 5.3.4 Operation of the MCDC for Maximum Desalination The MCDC system operating conditions were optimized based off results from the proof of concept study to investigate the maximum desalination rate of the MCDC system. The MCDC reactor design was identical to what was described previously. The anode, desal ination, and cathode chamber were all converted into a recirculating fluid cathode chamber. The anode and cathode recirculating volume were reduced to 100mL. The anode and desalination chambers were recirculated at a flow rate of 2mL/min. For this experiment all three chambers used raw produced water. The time for capacitive desalination was reduced from 2 hours to 1 hour. Regeneration was achieved by connecting the C DI in short circuit and pumping deionized water through the desalination chamber at a rate of 10mL/min. The total volume of regenerated solution was 100mL. The MCDC reactor was started by replacing the 100mL anolyte and catholyte solution with fresh raw PW, under open circuit conditions. The same anode and cathode electrodes were used from the proof of concept study and a new CDI system was constructed following the procedure previously outlined. The anolyte and catholyte solutions were allowed to equili brate for 35 minutes prior to connecting the anode and cathode electrodes to the CDI assembly. Before starting the experiment, the 100mL of deionized water (DI) was recirculated through the desalination chamber at 2mL/min and the CDI electrodes were conne cted in short circuit. After the anode and cathode chamber were equilibrated, the 100mL DI was completely removed and the desalination chamber was purged of all residual liquid. Following the removal of the DI water from the desalination chamber, 12 mL of raw produced water was pumped into the desalination

PAGE 91

77 chamber and capped for operation. Samples for conductivity, pH, salinity, and organic content were collected. Open circuit voltage as well as the starting voltage on the CDI assembly (~0 mV) were meas ured with the multimeter prior to starting,. Next, the anode anode/cathode electrodes were connected to the CDI assembly for 1 hour. Following the one hour of operation, the CDI assembly voltage was measured and 10 mL samples were collected from the anol yte and catholyte reservoirs. The anode and cathode electrodes were still connected to the CDI when the 12 mL desalination PW was completely purged from the desalination chamber. Following the desalination procedure, the 100mL DI regeneration solution was connected to the desalination chamber and the CDI assembly was connected in short circuit. The regeneration solution was recirculated at a rate of 10mL/min for 20min. During this 20 minute regeneration period, the anode and cathode chamber were disconnec ted from the pump in order to not disturb the acclimated anode microorganisms and the reactor was operated under open circuit conditions. At the end of the 20 minute the 100 mL regeneration solution was completely removed from the desalination chamber whic h signaled the end of the first desalination cycle. The second desalination cycle begin immediately after the end of the regeneration. 10mL samples were collected from the anode and cathode chamber, for analysis and voltage potential were measured. The same 12 mL desalination solution from cycle one was reintroduced to the desalination as previously described. The experimented concluded at the end of three cycles. The full 12 ml desalination solution was collected for analysis. The experimental procedu re for maximum salt removal was designed to have the smallest operating anode and cathode volumes to allow for sample collection while maintaining a sufficient volume for operation.

PAGE 92

78 5.3.5 Analysis Conductivity and pH was measured for all three chambers fo r all experiments. The removal of oxidizable substrates and inorganic chemicals in all three chambers were determined by testing chemical oxygen demand (COD) (Hach Co, Loveland CO). Total alkalinity was measured using a HACH alkalinity test kit (Hach Co, L oveland CO). The dissolved organic content (DOC) was measured for the maximum desalination experiment using a Sievers 5310C Series TOC analyzer. The change in total dissolved solids (TDS) concentration was determined by measuring cation concentration usin g the Optima 3000 Inductive Coupled Plasma (ICP) Spectrometer (Perkin Elmer, CT) and anion concentration using the Dionex DC80 ion chromatography system (IC) (Dionex, CA). Prior to TOC, IC, and ICP analysis, the collected samples were filtered through a 0. 45 M filter. Average produced water characteristics are identified in Table 5.1. Table 5.1 Produced Water Characteristics Table 5.1 illustrates the characteristics of the untreated, raw produced waters that were used in all experiments. Raw Produced Water Characteristics Analysis Conc. TDS (mg/L) 15870 290 pH 7.8 0.2 Conductivity (mS/cm) 25 0.15 Alkalinity (mg/L as CaCO3) 700 8 Ba (mg/L) 44.4 18.7 Ca (mg/L) 236.6 64.2 K (mg/L) 49.5 1.9 Mg 285 (mg/L) 30.0 2.3 Na (mg/L) 5992.0 82.2 Si (mg/L) 30.1 0.5 Sr (mg/L) 27.4 4.2 Cl (mg/L) 9290.2 241.1 Br (mg/L) 67.1 2.9 PO4 (mg/L) 54.4 5.43 DOC (mg/L) 33.5 4.3

PAGE 93

79 Figure 5.3 Voltage Profile Fluctuations in voltage over three typical cycles indicates that variation in operating a microbial electrochemical system with raw produced water. Arrows indicate when the electrolyte solutions for the anode/desalination/cathode chambers were replaced. 5.4 Results and Discussion 5.4.1 Proof of Concept MCDC Vs MDC Desalination an d COD Removal Capacity The produced water used in all experiments would be considered in the low range of salinity and organic content compared to other PW wells. However, greater than 99% of all unconventional PW wells in the United States contain a TDS value of less than 20 g TDS/L. Therefore, the results presented here for the MCDC system would be applicable to most unconventional PW wells. A full analysis of the untreated produced water can be observed in Table 5.1. The acclimation of the anode electro de to utilize produced water as a substrate required that the microbes be slowly transitioned. Without this slow transition, or else they could not tolerate the high salinity and did not contain the necessary enzymes to break down the organic content of t he produced water. Even after converting the reactors to MCDC and MDC, the voltage outputs from the reactors were variable as can be seen

PAGE 94

80 in Figure 5.3. Data recorded from the data logger for the MDC clearly shows how the voltage after replacing the elect rolyte solutions took time to reach the maximum voltage. Figure 5.3 also clearly illustrates how energy can be generated over successive batch cycles. However, unlike traditional MDC experiments, the drop in voltage over time was not directly correlated to the removal of ions in the middle chamber. Figure 5.5 shows the rate of change in conductivity for the anode, desalination, and cathode chambers for the MDC. Over a 24 hour period the change in conductivity was extremely slow. The conductivity in the d esalination chamber decreased on average 1.9 mS/cm, while the anode and cathode chambers increased 0.8mS/cm and 1.9mS/cm respectively. At the end of the 24 hour desalination cycle the conductivity in the desalination chamber was still high. Therefore the d ecrease in voltage over time illustrated in Figure 5.3 shows that ion transfer capability was not the limiting factor, rather that desalination would stop due to the loss of available oxidizable material. Additionally, the rate of desalination in the MDC was believed to be slow because of the poor kinetics of the MDC reactor. All of the desalination in the MDC using PW in the anode and desalination chambers was achieved through electrodialysis. In almost all other MDC studies the anode and cathode chamber s contained salt concentrations that were significantly lower than the desalination chamber, allowing for concentration gradients to aid in electrodialysis. The ability of produced water to be used as a substrate for microbial exoelectrogenic electron t ransfer is exciting. The maximum current produced in the MDC reactor was 450mA. For a small desalination reactor operating on a relatively unstudied substrate, 450mA is an acceptable current output. With improvements to the reactor design and reduction o f the external resistance, a higher current could have been produced which

PAGE 95

81 would have resulted in a higher desalination capacity. However, it would not have likely made a dramatic affect on the performance of the reactor due to the concentration gradients across the membrane and the rapid consumption of available oxidizable material. For the MCDC reactor, voltage and current profiles cannot be recorded using the data acquisition system. The voltage drop measured by the datalogger over a short period of t ime prevents the CDI from developing a sustained potential. This was discovered over many failed attempts to use the datalogger. The open circuit potential for the MCDC was >780mV, but the maximum voltage accieved on the CDI assembly was only 530mV in a t wo hour period. Interestingly, the achieved CDI voltage potential like the MDC reactor fluctuated from batch to batch depending upon the microorganism ability to transfer electrons at any given point.(Figure 5.4) The greatest variablity in applied voltage was observed at the end of the desalination cycle. This fluctuation could be partially due to the loss of available organic content but may also be associated with a change in pH and ion concentration in the anode and desalination chamber. However, Figu re 5.4 clearly shows that despite minor flucations in voltage at any given time, the results for the applied voltage potential on the CDI assembly were reproducable.

PAGE 96

82 Figure 5.4 Charged Formed on CDI The standard deviation for the charge formed on the C DI assembly of the MCDC system over a single desalination cycle. The maximum voltage achieved was 0.53V, but the average maximum voltage achieved over multiple cycles was 0.5V. The removal of salts for the desalination chamber and the anode chamber for the MCDC reactor was drastically faster than the MDC system. The conductivity in the desalination chamber using only electrical adsorption dropped on average 3.0mS/cm. The anode chamber also decreased slightly: 0.1mS/cm in two hours with the cathode cha mber slightly increasing by 0.1mS/cm. It can be observed from Figure 5.5 that the fastest change in conductivity was achieved in the desalination chamber using the MCDC system with 0.8 mS/cm/hr desalinated in desalination chamber and 0.1mS/cm/hr desalin ated in the anode chamber. When the MCDC CDI assembly was connected in short circuit, as was in the MCDC control experiment, the performance of the desalination chamber dropped to 0.1mS/cm/hr, the same as the desalination rate for the MDC. This indicates that it was the electrical adsorption that led to the faster desalination rate. Because the MCDC uses two cation exchange membranes, ions migrate from the anode chamber to the desalination chamber and to the cathode chamber to

PAGE 97

83 balance the transfer of elec trons from the anode to the cathode. This causes the cathode chamber for the MDC and the MCDC to increase in conductivity at a rate proportional to the migration of salts from the desalination chamber to the cathode chamber. The MCDC desalination chamber ion removal rate is the fastest because it is actually employing two methods of desalination: capacitive deionization and electrodialysis. Accounting for all of the ions removed in the MCDC system accounting for volume the reactor removed an average of 20 mg TDS/L. Figure 5.5 Change in Conductivity for MDC and MCDC Change in conductivity for the anode, desalination, and cathode chamber; positive values indicate salt removal and negative values indicate an increase salt concentration. The pH fluctuation between the three chambers also showed stark differences between the MDC and MCDC reactor shown in Figure 5.6. The pH in the MCDC decreased in the anode chamber by 0.01 pH units/hr while the MDC increased by 0.08 pH units/hr. However, the MDC had little change in pH in the desalination chamber, while the MCDC decreased 0.08 pH units/hr. Also, in the cathode chamber the MDC only increased 0.03 pH units/hr while the MCDC increased at rate of 0.04 pH units/hr.

PAGE 98

84 These results indicate that by using the MCDC system, the pH of the anode and cathode chambers can be potentially further stabilized. The biggest change in pH is in the desalination chamber of the MCDC. What exactly is happening is still a mystery, as mentioned in a previous public ation, fluctuations in pH in CDI units have been observed and are potentially linked to water hydrolysis. Additionally, because the MCDC is capable of adsorbing buffers such as phosphates and carbonates, migration of hydrogen ions may be observed more eas ily in the MCDC system than the MDC system. Figure 5.6 Change in pH for the MDC and MCDC Change in pH for the MDC and MCDC shows how the pH changed for the entire reactor. Positive values indicate a decrease in pH and negative values indicate an incr ease in pH. One of the most interesting findings presented here is illustrated in Figure 5.7. As mentioned previously, the advantage of using a MES for the treatment of PW is in its ability to simultaneously remove the organic content as well as its sal inity. It was believed originally that the anode chamber would remove organics and the desalination chamber would serve as the salt remover. However, the results presented in Figure 5.7 illustrate a different picture. For both the MDC and MCDC, COD was r emoved by the

PAGE 99

85 microbes in the anode chamber at a rate of 40 mg COD/L/hr. However, COD was actually removed at a much faster rate in the desalination chamber of the MCDC than any other chamber investigated. Almost no COD was removed from the MDC experiment in the desalination chamber or the cathode chamber. The MCDC desalination chamber was capable of removing 100 160 mg COD/L/hr. The capacity of the CDI assembly in the MCDC has been demonstrated in other CDI study using produced water. 82 However, unlike the anode chamber which oxidizes the organic faction of the COD to CO 2 the COD removed in the desalination chamber is electrochemically removed through adsorption. Potentially any charged organic molecule could be removed using capacitive deionization. When the CDI assembly is regenerated, the adsorbed molecules are recovered back into solution. Ideally, COD should be fully oxidized, but using the CDI is a quick method of COD removal and concentration. The only potential problem is in the disposal of the concentrated COD waste. While it is important to understand the full ramifications of any new technology, the disposal of the regenerated sal ts and COD is beyond the scope of the present study.

PAGE 100

86 Figure 5.7 Change in COD per Hour for the MDC and MCDC Figure 5.7 indicates the change in COD for the anode, desalination and cathode chamber per hour. Positive values demonstrate a removal in C OD and negative values indicate a gain in COD. 5.4.2 Regeneration and Energy Harvesting Using the MCDC System Regeneration of the MCDC was achieved by either connecting the assemblies in short circuit for one hour or by connecting the fully charged assemb lies to an external circuit in order to harvest the energy store in the in situ capacitors. Table 5.2 shows the increase in harvested energy from the stored energy of the MCDC CDI assembly. In a single batch desalination cycle, the maximum voltage reache d across the CDI assembly was 530mV. When connected to the charge pump, electrons can only flow in a single direction. Electrical energy stored on the MCDC was pumped across the charge pump and was stored on the 12F 2.5V external capacitor. Capacitors ar e designed to prevent jumps in voltage therefore, as can be seen in Table 5.2, the extracted energy was not achieved quickly. Over 48 hours almost all of the charge on the MCDC CDI assembles was removed and stored on the on the external capacitor. The ext ernal capacitor had a much higher capacitance and the charge pump efficiency has been shown to be low,

PAGE 101

87 resulting in the energy harvested from a single 2 hour desalination cycle of 0.54 mJ. Many advances will need to be made in energy harvesting if the MCD C is ever expected to produce a usable energy. However, the purpose of this study was to demonstrate that PW can be removed of organic, content be partially desalinated, and generate external energy. Table 5.2 Energy Harvested From Single Batch Cycle of MCDC Using Produced Water The table illustrates the increase in harvested energy using a simple charge pump connected to a 12 F capacitor. Time (hr) Capacitor Voltage (mV) Harvested Energy (mJ) 0 0 0 2 0.2 0.00024 6 0.4 0.00096 8 0.6 0.00216 24 2 0.024 48 3 0.054 5.4.3 Maximum Desalination Capacity and Rate of the MCDC System Following the proof of concept study for the treatment of PW using the MDC and the MCDC system, a more thorough investigation into the capacity of the technology was performed. It was observed in the proof of concept and through previous studies, that up to 90% of the TDS is removed in the MCDC system within the first hour. Therefore, in order to increase the rate of salt and organic removal, the MCDC was operated for only one hour. The maximum voltage achieved on the CDI assembly was 280mV, essentially half of the achievable potential. The additional advantage of achieving a smaller potential over a shorter period of time is that regeneration of the CDI assembly was faster. The CDI assembly was regenerated in 20 min as opposed to one hour in the proof o f concept study. Part of the reason for the faster regeneration was due to the desalination chamber

PAGE 102

88 being flushed with low salinity water at a high flow rate. Using this method, multiple cycles of desalination could be achieved in a single day while still being able remove greater than 90% of the potential desalination capacity in a single cycle. Three consecutive desalination cycles were performed in less than 5 hours. Also, from the proof of concept study it was indicated that pH fluctuation in the anode and cathode chambers were minimal. Consequently, instead of using phosphate buffer in the cathode chamber, raw PW could be used. This greatly simplified the operation of the reactor requiring only a single fluid to be passed through three chambers desig ned for different purposes. The anode chamber with acclimated microorganisms operating under anaerobic conditions could oxidize the PW. The desalination chamber with the CDI assembly could physically and electrically adsorb dissolved organics and salts, and the cathode chamber with the air cathode could reduce oxygen. The change in TDS for the three chambers from the initial starting conditions to the end of the three desalination cycles can be seen in Figure 5.8. Figure 5.8 Change in TDS for the MCDC Operating Under Maximum Conditions The change in TDS for the MCDC from initial starting conditions to the final condition after three successive desalination cycles indicates the greatest salt removal was achieved in the desalination chamber.

PAGE 103

89 IC/ICP analysis in Figure 5.8 indicates that TDS was removed in all three chambers. The desalination chamber had the highest amount of removed TDS over the anode and cathode chambers with 65% of the TDS being removed in three cycles. In addition to the d Changes in conductivity for each cycle can be seen in Table 5.3. Table 5.3 Percent TDS Removal and Change in Conductivity for the Three Desalination Cycles The table illustrate s the initial and final conductivity for the three chambers. The percent conductivity removed is also indicated. Cycle 1 Cycle 2 Cycle 3 Anode Conductivity Initial(mS/cm) 24.47 24.40 24.27 Conductivity Final (ms/cm) 24.40 24.27 24.23 % Removed 0.27 0.82 0.95 Desalination Conductivity Initial (mS/cm) 25.07 17.24 12.04 Conductivity Final (mS/cm) 17.24 12.04 8.90 % Removed 31.21 51.97 64.48 Cathode Conductivity Initial (mS/cm) 24.93 24.93 24.73 Conductivity Final (mS/cm) 24.93 24.73 24.67 % Removed 0.00 0.80 1.07 IC/ICP analysis was unable to be performed on each desalination cycle due to the small volumes used in the experiment. The final TDS concentration for the desalination chambers PW was 5.3 g/L, down from 15.9 g/L. While the final TD S for this experiment would still be too high for most disposal and reuse conditions, the removal of greater than 65% of the TDS in four hours of operation indicates the viability of the technology to quickly remove salts from PW without directly consuming any external electricity. The only electricity used for the operation of the MCDC was in pumping the PW into the

PAGE 104

90 reactor and for sample analysis; this type of energy consumption would be required regardless of the operated system. However, because the t echnology does not require high pressures, the energy required for pumping is low. By improving system design and operation, fluid flow through the reactor could potentially be gravity fed. The total amount of salts removed for the MCDC system was calcul ated by multiplying the volume of each chamber by the change in TDS and summing the three chambers. To compare the results to similar CDI studies and to the previous MCDC publication, the total salt removal was normalized by the amount of activated carbon used. Table 5.4 Total Ions Removed and Recovered with the MCDC System Table 5.4 illustrates over three successive cycles the total ions removed and recovered for the anode, desalination and cathode chambers. Total Ions Removed (mg TDS/g ACC) (Anode+ Desal+ Cathode) Total Ions Recovered (mg TDS/g ACC) (Anode+ Desal+ Cathode) Cycle 1 65.6 11.8 66.4 5.3 Cycle 2 68.2 1 3.0 41.5 4.4 Cycle 3 32.6 16.0 27.4 4.2 Table 5.4 shows the total ions removed in terms of mg TDS/ g ACC as well as the total ions recovered in mg TDS/g ACC. For the first cycle 65.6 mg TDS/g ACC were physically and electrically removed from the three chambers. Upon regeneration, greater than 98% of the desalinated salt was recovered. Table 5.4 indicates the capacity of t he MCDC system to remove salt physically, electrically, and through electrodialysis from the anode and cathode chambers. Interestingly, the capacity of the MCDC to remove and recover salts decreases over successive cycles. After two cycles the capacity f or removal remained approximately the same, but the recovery of desalinated salts slightly

PAGE 105

91 decreased. After three cycles the capacity for salt removal was almost half of the first cycle. The loss in desalination capacity over successive cycles is one of the limitations of capacitive deionization. This is due to the inability to fully regenerate the CDI. As a result, over successive cycles, the capacity for desalination decreases. This loss in capacity has been observed in most CDI studies. 54 With improvements in reactor design and operation, the loss of capacitance can be minimized; if the electrical potential is reversed the electrodes can be fully recovered. 6 As mentioned previously for the proof of concept experiment, one of the more inte resting findings was in the MCDCs ability to remove COD from the desalination chamber. In the maximum desalination capacity experiment, the results were directly correlated with the proof of concept study. The anode chamber removed only a small fraction of COD in four hours of operation (20 mg COD/L), while the desalination chamber removed 83% of the COD. After the MCDC was regenerated, 75% of the COD that was removed during desalination was fully recovered. It is believed that the remaining 25% of the COD that was not recovered was lost to experimental error and through electrochemical oxidation on the CDI assembly. Figure 5.9 shows the initial and final COD values for the three chambers operated under the maximum desalination capacity procedure. Whi le the MCDC desalination chamber does not actually break down the organics in the PW, physically and electrically removing and recovering COD is a much faster method of COD removal from wastewater than through direct oxidation.

PAGE 106

92 Figure 5.9 Change in COD for the MCDC for the three Chambers Plus the Recovered COD Figure 5.9 shows the change in COD over the three successive desalination cycles, for the anode, desalination and cathode chambers. Additionally, the recovered COD (REG) is shown from the reg eneration process to indicate how COD can be removed and recovered. 5.5 Conclusion A microbial capacitive desalination cell (MCDC) is a simultaneous wastewater treatment, electricity production, and desalination device. Demonstrated for the first time is the MCDC ability to treat both the salt and organic content of produced water, and t o generate external electricity. The proof of concept study demonstrated how using only electrical adsorption, a small amount of COD and TDS could be removed. The results from the proof of concept study indicated that a much higher TDS and COD removal ef ficiency could be achieved. Using PW as the single electrolyte, the MCDC can be completely operated. This will allow the technology to be deployed onsite without requiring a large amount of additional chemicals or systems for PW treatment. When the MCDC system was compared to the traditional MDC system, the MCDC vastly outperformed the MDC in terms of TDS removal/hr and COD removal/hr. This is likely

PAGE 107

93 due to the slow kinetics of electrodialysis under a small concentration gradient. It is the goal of futur e research studies to implement a control system for monitoring the change in salinity and organic content for all three chambers. As well as to automate the MCDC process of desalination and regeneration and energy harvesting so that it could be operated c ontinuously. This is the first study to demonstrate the production of energy from PW in addition to treating both the organic and TDS content within PW. Due to the ability of the MCDC system to simultaneously treat the organic and dissolved solids conten t of produced water, as well as produce direct current, the MCDC system is a potential solution for site specific produced water treatment.

PAGE 108

94 6. The Development of an Efficient Low Cost Barrier for the Use in a Microbial Electrochemical System Called an Aga rose Salt Bridge Membrane 5 6.1 Abstract Microbial electrochemical systems (MES) are a rapidly developing platform technology to meet the most challenging social, economic, and environmental problems. To address theses challenges MES technology needs to be manufactured out of low cost sustainable and renewable materials. One major cost associated with MES technology are ion exchange membranes used to separate chambers inside the reactor. Presented here for the first time is the development of a low cost agarose salt bridge membrane (ASBM). The unique manufacturing process and structure of the ASBM allows it to operate with a high maximum power point and low internal resistance. Compared to a commercial available cation exchange membrane from the Astom C orporation, the ASBM had a 44% higher power density. The ASBM is fabricated out of sustainable materials with a cost of $6/m2. Further research needs to be conducted to ensure the longevity of the ASMB for practical applications. 5 The work presented in this chapter is in preparation for publication by Casey Forrestal, Pei Xu, Peter Jenkins, and Zhiyong Ren.

PAGE 109

95 6.2 Introduction Across the globe the need for sustainable water and waste water treatment is paramount. In the United States 3% of all energy used is for the treatment of wastewater. 80 Much of the energy used for wastewater treatment comes from aerating the wastewater to allow for the microbes to quickly oxidize the organic content into carbon dioxide. Countries without access to abundant energy sources cannot efficiently and quickly treat waste which leads to the development of disease and socioeconomic depression. The World Health Or ganization has estimated that up to one third of the problem in developing countries many researchers have been investigating low cost methods in which to sustainably p roduce and treat water. One method of sustainably treating wastewater is though the use of a Microbial Electrochemical System (MES). 28 The MES is a wastewater treatment option which sustainably generates electricity. Additionally, MESs have been modified to perform many tasks, one of which is desalination of salt water. MESs work by using microorganisms under anaerobic conditions to breakdown an organic waste product, such as municipal wastewater, and transfer electrons exoelectrogenically to a terminal electron acceptor. The MES provides a met hod by which wastewater can be treated without using the high energy required to aerate the wastewater. The MES also has the added benefit of generating a small amount of electricity which could augment electrical needs required for pumping or analytical monitoring. The most common method of increasing the amount of available water is through the desalination of salt water. 2 There are many ways salt water can be desalinated, but the most widely used system is reverse osmosis (RO). 83 Reverse osmosis

PAGE 110

96 is a pressure driven membrane filtration desalination process where fresh water and brine water are se parated by the RO membrane. RO is one of the more energy efficient desalination processes, but it still requires a 3 7kWh/m 3 of desalinated salt water. 84 For energy poor countries, the energy demand for RO is too large to me rit the development of a desalination plant. Using a modified MES called a Microbial Desalination Cell (MDC) or Microbial Capacitive Desalination Cell (MCDC), energy can be generated from the desalination of salt water. 3, 63 While these technologies are major advancements in s ustainably producing desalinated salt water, they still require the use of a membrane which can dramatically increase the cost of desalination. Typical membranes used in the MDC or MCDC cost around $100/m 2 RO membranes can cost up to $1800/g. The cost o f the membranes and energy used account for 87% of the cost of desalination using RO. 85 Mem branes are used in RO and energy producing systems such as MESs to create a barrier for separation. Historically the first separators were salt bridge barriers and have been used in fuel cells since the voltaic pile. In MESs, salt bridge membranes have b een widely discredited as a viable option due to their perceived high internal resistance and subsequently low power density. The first paper published using a salt bridge separator was Booki Min et al 2005. 13 It was discovered that the salt bridge separator was a good barrier for oxygen diffusion but, due to the high internal resistance, a low power density was achieved. The salt bridge membrane used for t he first publication was 30 cm in length which led to poor performance statistics publish. Investigated here is a new method for the manufacturing of a low cost agarose salt bridge membrane (ASBM) to treat wastewater, and produce electricity.

PAGE 111

97 6.3 Material s and Methods 6.3.1 Agarose Salt Bridge Membrane Design The agarose salt bridge membrane contained per liter: 0.1 g KCl, 0.098 g NaH 2 PO 4 H 2 O, 0.183 g Na 2 HPO 4 and 0.2 g agar. The agarose media was continuously stirred while heated to a boil. While still hot, the agarose media was slowly poured onto a 1ft 2 slowly poured to insure that no bubbles formed in the agar which would affect the performance. The ASBM was allowed to naturally desiccate on the table for 24 hours. This step is vital to the production of a low resistance ASBM. Upon desiccation, the fiberglass appears glassy and brittle. Following the 24 hour desiccation, the ASBM is placed in DI water and is rehydrated for one hour. An additional ASBM was fabricated and tested: The ASBM 2 was exactly the same as the ASBM but it was allowed to desiccate and rehydrate a seco nd time. (Table 6.1) Table 6.1 Thicknesses of the ASBM Table 6.1 illustrates how the thickness of the ASBM changed from the initial point to the final point. ASBM Construction Thickness (in) Initial 0.102 0.115 After 24hr desiccation 0.016 After 1 hr rehydration (ASBM) 0.065 After second 24 desiccation 0.016 After second 1hr rehydration (ASBM 2) 0.057 6.3.2 Reactor Configuration The reactors used to investigate the performance of the ASBM consisted of two polycarbonate cube shaped blocks with 3 cm diameter holes forming an internal anode

PAGE 112

98 and cathode chambers volume of 23 mL, and 27 mL respectively. The anode and cathode chambers h ad a length of 4 cm. The anode electrode was a graphite brush (Golden brush, CA) and was pretreated by washing it in acetone and heating it to 350C for 30 minutes 43 Traditional air cathode s were made by applying 10% Pt/C (0.5 mg/cm 2 ) and four PTFE diffusion layers on 30% wet proofed carbon cloth as previously described 44 The two chamber MES w as fully acclimated using Neosepta cation exchange membrane (CEM) electrodes were acclimated by adding 13mL of activated sludge from the City of Broomfield Municipal Wastew ater Treatment Facility to the anode chamber along with 10mL of anolyte growth media. The anolyte growth media contained per liter: 1.6g NaCH 3 COO, 0.62g NH 4 Cl, 4.9g NaH 2 PO 4 H 2 O, 9.2g Na 2 HPO 4 0.26g KCl, and 10 mL trace metals and 10 mL vitamin solution. 46 After the anode electrode became fully ac climated over a one month period, the anolyte solution was slowly transitioned to 100% raw produced wastewater (PW) from a well in Colorado. The anode electrode was converted to the PW by adding 10mL of raw PW to the anode with the anolyte media described above. This was done over a two week period until the anode chamber completely used raw PW. The catholyte contained per liter: 9.8 g/L (71mM) NaH 2 PO 4 H 2 O and 18.3g/L (129mM) Na 2 HPO 4 Electrolyte solutions were replaced for prior to starting experiments, and were replaced for all studies when the voltage dropped below 50mV. The ASBM reactor was connected to a data acquisition system (model 2700, Keithley Instruments, inc. OH) and the voltage across the external resistor was recorded every 1.1 minutes. A 1 000 Ohm external resistance was used for all experiments. An

PAGE 113

99 Astom Neosepta membrane was used in an identical reactor to the ASBM rectors for comparison. 6.3.3 Analysis pH was measured before and after a single batch operation. A single batch was defined as the time from adding new electrolyte solution until the voltage dropped below 50mV. Anode and cathode potentials were measured using a potentiostat (G 300, Gamry Instruments Inc. NJ) with a Ag/AgCl reference electrode place inside the chamber o f the working electrode. Power density was determined using linear sweep voltametry (LSV) with a potentiostat (G 300, Gamry Instruments Inc. NJ). The reactors were disconnected from the external resistor for greater than one hour, and then connected with t he anode as the working electrode, the cathode as the counter, and reference electrode. The scan rate was set at 0.1mV/s, with a step size of 1mV and a maximum current of 10mA. The internal resistance of the reactor and the polarization resistance was det ermined by potentiostatic electrochemical impedance spectroscopy (EIS) at a frequency range of 10 5 to 10 2 Hz, with a 10 mV sinusoidal perturbation. The anode was the working electrode, the cathode the counter and reference. The data was plotted on a Nyqui st plot, which indicates the internal resistance (Rin) and the polarization resistance. On the Nyquist plot the intercept with the Zreal axis is defined as the internal resistance, and the polarization resistance, is the diameter of the semicircle formed from the EIS test. 74 6.4 Results and Discussion 6.4 .1 Agarose Salt Bridge Membrane Power Performance from Produced Water The performance of the ASBM and the ASBM 2 which was desiccated for a seco nd day indicate that the ASBM could potentially be used as separator for wastewater.

PAGE 114

100 Table 6.1 shows that the ASBM manufacturing procedure reduced the thickness of the ASBM from 0.102 inches to 0.065 inches. The ASBM 2 using the second desiccation, furth er reduced the thickness to 0.057 inches. The Astom CEM membrane has a thickness of 0.006 inches. The minimum thickness for the ASBM was observed under the desiccation conditions and, the thickness was equal to that of the fiberglass mesh. The purpose o f the fiberglass mesh is to provide structural support to the weak ASBM while allowing for ions to freely pass across the reactor. The maximum open circuit potential (OCP) for the reactors operating with the ASBM, ASBM, and Astom CEM were similar, with OC P values of 570mV for the ASBM, 533mV for the ASBM 2, and 553 mV for the Astom CEM. The maximum power density normalized to the anode volume of 27mL, showed that the ASBM and the ASBM 2 had almost identical power density curves. The maximum power point ( MPP) for the ASBM was 8.6 W/m 3 and the MPP for the ASBM 2 was 8.9 W/m 3 The MPP for the Astom CEM was significantly lower at 4.1 W/m 3 It is believed that the difference in the MPP for the Astom verses the ASBM and ASBM 2 is related to chemical fouling of the Astom membrane from the produced water. This only occurred after operating the reactors for more than two months. Previous studies using the Astom membranes did not observe this scaling because the membranes were replaced more frequently. The higher power density of the ASBM 2 was likely due to the smaller thickness allowing for ions to more easily transfer. Results from the LSV test can be seen in Figure 6.1

PAGE 115

101 Figure 6.1 Power Density of the ASBM Power density for the ASMB and Astom membrane. Th e power density curve is normalized by anode volume in cubic meters. 6.4 .2 Agarose Salt Bridge Membrane Change in pH The operation of MES is closely correlated to the fluctuation of pH in the anode and cathode chambers. If the anode chamber becomes to ac idic, it affect the becomes too basic, the cathode electrode cannot easily reduce oxygen. The results presented in figure 6.2 show the change in pH for the anode and cat hode chamber for the ASBM, ASBM 2, and the Astom CEM. The anode chamber decreased from 6.8 to approximately 5.5 pH units for all three experimental conditions. However, in the cathode chamber, the pH of the reactors operating with the agarose salt bridge increased to a higher value than the Astom CEM. The pH in the ASBM and ASBM 2 reactors were 8.5 while the Astom CEM reactor only had an average pH value of 8. The minor variation in the fluctuation of pH from the initial to final point indicates that the ASBM

PAGE 116

102 phosphate buffer is used in the manufacturing of the agarose membrane that the slight variation in cathode pH between the Astom membrane and the ASBM membranes is d ue to hydrogen consumed inside the membrane. As a result, hydrogen ions produced in the anode chamber migrate to the membrane, and hydrogen ions inside the catholyte are consumed for reactor operation, while hydrogen ions are able to more easily pass from the anode chamber to the cathode chamber. Figure 6.1 Change in pH for the Anode and Cathode Chamber with the ASBM using Produced Water in the Anolyte The change in pH for the anode and cathode chambers over a single batch cycle. The ASBM was compar ed to the Astom CEM membrane. 6.4 .3 Internal and Polarization Resistance for the ASBM Compared to the Astom CEM EIS results for the three experimental conditions help to illustrate the advantage of the ASBM as well as indicate why the Astom membrane did not perform well in power production. Figure 6.3 and 6.4 show the internal and polarization resistance of the ASB M, ASBM 2, and the Astom CEM membrane. The internal resistances for all three

PAGE 117

103 membranes were similar, with the ASBM 2 having the lowest resistance at 40.9 Ohms. The ASBM had a slightly larger internal resistance at 41.9 Ohms, and the Astom CEM had the hig hest internal resistance of 51 Ohms. The biggest difference between the ASBM and the Astom CEM is in the polarization resistance clearly observable in Figure 6.3. The polarization resistance for the Astom CEM reactor is approximately 100 Ohms, while the p olarization resistance is not even visible in Figure 6.3 for the agarose membranes. The polarization resistance for the agarose membranes can be seen in Figure 6.4. The ASBM polarization resistance is 0.3 Ohms, while the ASBM 2 is 0.2 Ohms. Figure 6.3 Nyquist Plot for the ASBM, ASBM 2 and the Astom CEM The Nyquist plot clearly shows the resistances of a reactor. The internal resistance is the lowest value where the graph touches the x axis. The polarization resistance is the diameter of the semicir cle formed where the graph would touch the x axis. The lower MPP for the reactor using the Astom membrane was most likely due to the high polarization resistance. The ability to transfer ions is the best aspect of the agarose salt bridge membrane. The s mall thickness created from desiccating the agar for two days improved the internal resistance as well as the polarization resistance, which likely caused the ASBM 2 to have a higher MPP. Through additional evaluation of the

PAGE 118

104 manufacturing process the ASB M membrane, a potentially lower internal resistance could be achieved. One easy method would be to reduce the thickness of the fiberglass mesh used for structural support. This would allow the desiccated ASBM to form a thinner membrane and creating smalle r internal resistance. Modifying the ASBM formula for the concentration of buffer and salts contained may affect the polarization resistance. Because the polarization resistance is already small, increasing the amount of salts may not be cost effective. Figure 6.4 EIS Comparison of the ASBM and ASBM 2 The Nyquist plot clearly shows the resistances of a reactor. The internal resistance is the lowest value where the graph touches the x axis. The polarization resistance is the diameter of the semicircle formed where the graph would touch the x axis. 6.4 .4 Cost Comparison and Sustainability There are many membranes available on the market. Most of the membranes investigated in MES research cost between $100 1000/m 2 and specialized membranes can range up to $25,000/m 2 A big advantage of the ASBM is the relatively low cost. Including only the cost of raw materials, the ASBM would cost around $6/m 2 The highest cost of the ASBM is in the fiberglass mesh used for structural support. Further

PAGE 119

105 research coul d be conducted into a more sustainable structural support that would allow for the price of the ASBM to be further reduced. The ASBM also has the advantage because it can be manufactured onsite using limited natural resources. Disposing of the ASBM would have a minimal environmental impact, because all material used for the manufacturing of the ASBM is natural except the structural support. 6.4 .5 ASBM Challenges The ASBM does come with some difficulties. After about two months of operation the ASBM needs to be replaced. Microorganisms began to grow on the ASBM which produced tiny holes in the membrane causing the chambers to leak. Additionally, without the formation of holes by microorganisms tares in the ASBM can occur easily. The ASBM reactors were co nverted to continuous flow, and almost immediately after starting the reactor, the increased pressure and shear force from the fluid flow caused holes to form in the ASBM. Further research needs to be conducted to increase the ASBM strength and microbial r esistance. 6.5 Conclusion Salt bridge barriers have been used in fuel cell systems for many years, but have been largely disregarded in MES research due to a few publications with faulty methodologies indicating that an agarose salt bridge has a high internal resistance. The first salt bridge barrier had an anode to cathode distance of almost twelve inches. The greater the distance between the anode and cathode chambers the harder it is for ions to migrate. This led to a high internal and polarization resistance which reduces the maximum power point. The ASBM and ASBM 2 presented are the thinnest salt bridge membrane ever investigated in a MES system. The reduced thickness of ASBM 2 caused

PAGE 120

106 it to slightly outperform the ASBM in terms of MPP, but made no difference in controlling the pH fluctuations of the anode and cathode chambers. The low cost and superior performance of the ASBM membranes over the Astom CEM make it a compelling choice for produced water treatment. However, without solving the problem of hole formation in the ASBM, it would not be a viable commercial option for the treatment of produced water.

PAGE 121

107 REFERENCES 1. G. H. Brundtland and D. World Commission on Environment and, Our common future Oxford University Press Oxford, 1987. 2. R. Semiat, Environmental Science & Technology 2008, 42 3. X. X. Cao, X. Huang, P. Liang, K. Xiao, Y. J. Zhou, X. Y. Zhang and B. E. Logan, Environmental Science & Technology 2009, 43 7148 7152. 4. H. Luo, P. Xu, P. Jenkins and Z. Ren, J Membrane Sci 2012, DOI:10.1016/j.memsci.2012.02.059 5. H. Wang, J. D. Park and Z. Ren, Environmental Science & Technology 2012, 46 5247 5252. 6. M. A. Anderson, A. L. Cudero and J. Palma, Electrochimica Acta 2010, 55 3845 3856 7. H. Liu and B. Logan, Abstracts of Papers of the American Chemical Society 2004, 228 193 ENVR. 8. Y. Fan, S. K. Han and H. Liu, Energy & Environmental Science 2012, 5 8273 8280. 9. Y. Fan, H. Hu and H. Liu, Environmental Science & Technology 2007, 41 8154 8158. 10. K. P. Nevin, H. Richter, S. F. Covalla, J. P. Johnson, T. L. Woodard, A. L. Orloff, H. Jia, M. Zhang and D. R. Lovley, Environmental Microbiology 2008, 10 2505 2514. 11. K. Rabaey, S. Butzer, S. Brown, J. Keller and R. A. Rozendal, Environmental Science & Technology 2010, 44 4315 4321. 12. H. B. Li, Y. Gao, L. K. Pan, Y. P. Zhang, Y. W. Chen and Z. Sun, Water Research 2008, 42 4923 4928. 13. B. Min, J. R. Kim, S. E. Oh, J. M. Regan and B. E. Logan, Water Research 2005, 39 14. P. Xu, T. Cath, G. Wang, J. E. Drewes and S. Dolnicar, Water Research Foundation, Denver, 2009.

PAGE 122

108 15. S. A. Avlonitis, K. Kouroumbas and N. Vlachakis, Desalination 2003, 157 151 158. 16. X. X. Cao, X. Huang, P. Liang, K. Xiao, Y. J. Zhou, X. Y. Zhang and B. E. Logan, Environ. Sci. Technol. 2009, 43 7148 7152. 17. K. S. Jacobson, D. M. Drew and Z. He, Environ. Sci. Technol. 2011, 45 4652 4657. 18. Y. Kim and B. E. Logan, Environ. Sci. Te chnol. 2011, 45 5840 5845. 19. H. Luo, P. Xu, T. M. Roane, P. E. Jenkins and Z. Ren, Bioresour. Technol. 2012, 105 60 66. 20. H. Luo, P. Jenkins and Z. Ren, Environ. Sci. Technol. 2011 45 340 344. 21. M. Mehanna, T. Saito, J. L. Yan, M. Hickner, X. X. Cao, X. Huang and B. E. Logan, Energy Environ. Sci. 2010, 3 1114 1120. 22. P. Xu, J. E. Drewes and D. Heil, Water Res. 2008, 225 139 155. 23. L. Yuan, X. Yang, P. Liang, L. Wang, Z. H. Huang, J. Wei and X. Huang, Bioresource Technology 2012, 110 735 738. 24. C. Forrestal, P. Xu and Z. Ren, Energy & Environmental Science 2012, 5 7161 7167. 25. X. Wang, S. A. Cheng, Y. J. Feng, M. D. Merrill, T. Saito and B. E. Logan, Environ. Sci. Technol. 2009, 43 6870 6874. 26. H. M. Wang, M. Davidson, Y. Zuo and Z. Y. Ren, J. Power Sources 2011, 196 5863 5866. 27. C. Forrestal, P. Xu and Z. Ren, Energy Environ Sci 2012, DOI: 10.1039/C2EE21121A 28. Z. Ren, H. Yan, W. Wang, M. Mench and J. Regan Environ. Sci. Technol. 2011, 45 2435 2441. 29. X. Chen, X. Xia, P. Liang, X. X. Cao, H. T. Sun and X. Huang, Environ. Sci. Technol. 2011, 45 2465 2470. 30. Y. Qu, Y. Feng, X. Wang, J. Liu, J. Lv, W. He and B. E. Logan, Bioresour. Technol. 2012, 10 6 89 94.

PAGE 123

109 31. M. M. Pendergast and E. M. V. Hoek, Energy & Environmental Science 2011, 4 1946 1971. 32. J. L. Schnoor, Environmental Science & Technology 2011, 45 5065 5065. 33. M. Mehanna, P. D. Kiely, D. F. Call and B. E. Logan, Environmental Scie nce & Technology 2010, 44 9578 9583. 34. K. S. Jacobson, D. M. Drew and Z. He, Environmental Science & Technology 2011, 45 4652 4657. 35. H. Luo, P. E. Jenkins and Z. Ren, Environmental Science & Technology 2011, 45 340 344. 36. M. Mehanna, T. Saito, J. L. Yan, M. Hickner, X. X. Cao, X. Huang and B. E. Logan, Energy & Environmental Science 2010, 3 1114 1120. 37. K. S. Jacobson, D. M. Drew and Z. He, Bioresource Technology 2011, 102 376 380. 38. X. Chen, X. Xia, P. Liang, X. Cao, H. Sun and X. Huang, Environmental Science & Technology 2011, 45 2465 2470. 39. H. Luo, P. Xu, T. M. Roane, P. E. Jenkins and Z. Ren, Bioresource Technology 2012, 105 60 66. 40. P. Xu, J. E. Drewes and D. Heil, Desalination 2008, 225 139 155. 41. O. Lefebvre and R. Moletta, Water Research 2006, 40 3671 3682. 42. Y. J. Zhang, Y. Wang, S. C. Xu, J. X. Wang, Z. Wang and S. C. Wang, Synthetic Metals 2010, 160 1392 1396. 43. X. Wang, S. Cheng, Y. Feng, M. D. Merrill, T. Saito and B. E. Logan, Environmental Science & Technology 2009, 43 6870 6874. 44. S. Cheng, H. Liu and B. E. Logan, Electrochemistry Communications 2006, 8 489 494. 45. H. Y. Wang, A. Bernarda, C. Y. Huang, D. J. Lee and J. S. Chang, Bioresour Technol 2011, 102 235 243. 46. Z. Ren, H. Yan, W. Wang, M. M. Mench and J. M. Regan, Environmental Science & Technology 2011, 45 2435 2441. 47. J. Z. Zhang, Marine Chemistry 2000, 70 121 131.

PAGE 124

110 48. P. M. Biesheuvel and A. van der Wal, Journal of Membrane Science 2010, 346 25 6 262. 49. R. Broseus, J. Cigana, B. Barbeau, C. Daines Martinez and H. Suty, Desalination 2009, 249 217 223. 50. Y. J. Kim and J. H. Choi, Separation and Purification Technology 2010, 71 70 75. 51. Y. J. Kim and J. H. Choi, Water Research 2010, 44 990 996. 52. S. Nadakatti, M. Tendulkar and M. Kadam, Desalination 2011, 268 182 188. 53. J. Yang, L. Zou, H. Song and Z. Hao, Desalination 2011, 276 199 206. 54. Y. Oren, Desalination 2008, 228 10 29. 55. J. B. Lee, K. K. Park, H. M. Eum and C. W. Lee, Desalination 2006, 196 125 134. 56. Y. Kim and B. E. Logan, Environmental Science & Technology 2011, 45 5840 5845. 57. B. B. Sales, M. Saakes, J. W. Post, C. J. N. Buisman, P. M. Biesheuvel and H. V. M. Hamelers, Environmental Science & Technology 2010, 44 5661 5665. 58. R. A. Rozendal, H. V. M. Hamelers and C. J. N. Buisman, Environmental Science & Technology 2006, 40 5206 5211. 59. M. W. Ryoo, J. H. Kim and G. Seo, Journal of Colloid and Interfac e Science 2003, 264 414 419. 60. C. M. Yang, W. H. Choi, B. K. Na, B. W. Cho and W. I. Cho, Desalination 2005, 174 125 133. 61. O. Lefebvre, A. Uzabiaga, I. S. Chang, B. H. Kim and H. Y. Ng, Applied Microbiology and Biotechnology 2011, 89 62. S. C hen, G. Liu, R. Zhang, B. Qin and Y. Luo, Environmental Science & Technology 2012, 46 2467 2472. 63. C. Forrestal, P. Xu and Z. Ren, Energy & Environmental Science 2012. 64. C. Forrestal, P. Xu, P. E. Jenkins and Z. Ren, Bioresource Technology 2012, 120 332 336.

PAGE 125

111 65. W. W. Li, G. P. Sheng, X. W. Liu and H. Q. Yu, Bioresource Technology 2011, 102 244 252. 66. Z. He, N. Wagner, S. D. Minteer and L. T. Angenent, Environmental Science & Technology 2006, 40 5212 5217. 67. B. Wang and J. I. Han, Biote chnology Letters 2009, 31 387 393. 68. P. Ledezma, A. Stinchcombe, J. Greenman and I. Ieropoulos, Physical Chemistry Chemical Physics 2013, 15 2278 2281. 69. M. M. Mardanpour, M. N. Esfahany, T. Behzad and R. Sedaqatvand, Biosensors & Bioelectronics 2012, 38 264 269. 70. J. R. Kim, H. C. Boghani, N. Amini, K. F. Aguey Zinsou, I. Michie, R. M. Dinsdale, A. J. Guwy, Z. X. Guo and G. C. Premier, Journal of Power Sources 2012, 213 382 390. 71. L. Zhuang, Y. Yuan, Y. Wang and S. Zhou, Bioresource Te chnology 2012, 123 406 412. 72. B. Jia, D. Hu, B. Xie, K. Dong and H. Liu, Biosensors & Bioelectronics 2013, 41 894 897. 73. B. E. Logan, B. Hamelers, R. A. Rozendal, U. Schrorder, J. Keller, S. Freguia, P. Aelterman, W. Verstraete and K. Rabaey, Env ironmental Science & Technology 2006, 40 5181 5192. 74. Z. He and F. Mansfeld, Energy & Environmental Science 2009, 2 215 219. 75. B. Erable, D. Feron and A. Bergel, Chemsuschem 2012, 5 975 987. 76. V. Srinivasan, E. F. Lambin, S. M. Gorelick, B. H. Thompson and S. Rozelle, Water Resources Research 2012, 48 77. C. E. Clark and J. A. Veil, Argonne National Laboratory (ANL), 2009. 78. Razi, A. Pendashteh, L. C. Abdullah, D. R. A. Biak, S. S. Madaeni and Z. Z. Abidin, Journal of Hazardous Materials 2009, 170 530 551. 79. Technical Assement of Produced Water Treatment Technologies Colorado School of Mines, 2009. 80. P. L. McCarty, J. Bae and J. Kim, Environmental Science & Technology 2011, 45 7100 7106.

PAGE 126

112 81. Y. H. Jia, J. H. Ryu, C. H. Kim, W. K. Lee, T. Thi Van Trinh, H. L. Lee, R. H. Zhang and D. H. Ahn, Journal of Industrial and Engineering Chemistry 2012, 18 715 719. 82. P. Xu, J. E. Drewes, D. Heil and G. Wang, Water Research 2008, 42 2605 2617. 83. J. Zhou, V. W. C Chang and A. G. Fane, Energy & Environmental Science 2011, 4 2267 2278. 84. P. Xu, T. Cath, G. Wang, J. E. Drewes and S. Dolnicar, Critical Assessment of Implementing Desalination Technology Water Research Foundation, Denver, 2009. 85. K. P. Lee, T. C. Arnot and D. Mattia, Journal of Membrane Science 2011, 370 1 22.