Hamid's Dissertation

BIOCONVERSION OF CELLULOSE INTO ELECTRICAL ENERGY IN MICROBIAL FUEL CELLS

DISSERTATION

Presented in Partial Fulfillment of the Requirements for The Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Hamid Rismani-Yazdi, M.S. * * * * *

The Ohio State University 2008

Dissertation Committee: Approved by Dr. Ann D. Christy, Advisor

Dr. Burk A. Dehority

Dr. Olli H. Tuovinen

Dr. Alfred B. Soboyejo Advisor Food, Agricultural and Biological Dr. Zhongtang Yu Engineering Graduate Program

ABSTRACT

In microbial fuel cells (MFCs), bacteria generate electricity by mediating the

oxidation of organic compounds and transferring the resulting electrons to an anode

electrode. The first objective of this study was to test the possibility of generating

electricity with rumen microorganisms as biocatalysts and cellulose as the electron donor

in two-compartment MFCs. Maximum power density reached 55 mW/m2 (1.5 mA, 313

mV) with cellulose as the electron donor. Cellulose hydrolysis and electrode reduction

were shown to support the production of current. The electrical current was sustained for over two months with periodic cellulose addition. Clarified rumen fluid and a soluble

carbohydrate mixture, serving as the electron donors, could also sustain power output.

The second objective was to analyze the composition of the bacterial communities

enriched in the cellulose-fed MFCs. Denaturing gradient gel electrophoresis of PCR

amplified 16S rRNA genes revealed that the microbial communities differed when

different substrates were used in the MFCs. The anode-attached and the suspended

consortia were shown to be different within the same MFC. Cloning and analysis of 16S

rRNA gene sequences indicated that the most predominant bacteria in the anode-attached

consortia were related to Clostridium spp., while Comamonas spp. was abundant in the

suspended consortia.

ii The external resistance affects the characteristic outputs of MFCs by controlling

the flow of electrons from the anode to the cathode. The third objective of this study was

to determine the effect of various external resistances on power output and coulombic

efficiency of cellulose-fed MFCs. Four external resistances (20, 249, 480, and 1000 ohms)

were tested with a systematic approach of operating parallel MFCs independently at

constant circuit loads for three months. A maximum power density of 66 mWm-2 was achieved by MFCs with 20 ohms circuit load, while MFCs with 249, 480 and1000 ohms external resistances produced 57.5, 53 and 47 mWm-2, respectively. The anode potential

varied under the different circuit loads employed. Higher coulombic efficiencies were

achieved in MFCs with lower external resistance.

The effect of different external resistances on the bacterial diversity and

metabolism in cellulose-fed MFCs was investigated as the fourth objective. DGGE

analysis of partial 16S rRNA genes showed clear differences between the planktonic and

the anode-attached populations at various external resistances. Cellulose degradation was

complete (< 0.1% residual), and there were no discernible differences among the MFCs.

HPLC analysis of short chain fatty acids (SCFA) revealed that anaerobic degradation of cellulose was accompanied by production of acetic, propionic, butyric, isobutyric, valeric, isovaleric, and lactic acids, with acetic acid being predominant. The profile of metabolites was different among the MFCs. The concentrations of SCFA were higher in

MFCs with larger external resistance. High levels of SCFA indicated that fermentative metabolism dominated over anaerobic respiration, resulting in relatively low coulombic efficiencies. The accumulation of SCFA at higher circuit resistances corresponded to lower power outputs.

iii Methanogenesis shifts the flow of electrons available from the substrate away

from electricity generation in MFCs. The fifth objective of this research was to assess the

influence of methane formation on the performance of cellulose-fed MFCs under long-

term operation. Two-compartment MFCs were inoculated with a ruminal microbial

consortium and fed colloidal cellulose (0.5 g l-1 d-1) as the sole substrate. Replicate MFCs

were operated under two different external resistances of 20 and 100 ohms, designated as

R20Ω and R100Ω. During the first week of operation 0.31 ± 0.004 (± SD) and 0.44 ± 0.004

mmol of methane was produced in R20Ω and R100Ω MFCs, respectively. Methanogenesis was suppressed, however, to below the detection limit (< 0.5 ×10-3 mmol) after 90 days

of operation. The decrease in methane production was accompanied with an increase in

the performance of MFCs. The current output of the MFCs increased from 0.1 mA during

the first week to 3.3 and 2.2 mA on day 90, resulting in 29 and 25% coulombic efficiency

for the R20Ω and R100Ω MFCs, respectively. A maximum volumetric power density of 3.5

-3 W m was achieved in R20Ω MFCs, which was three times greater than that obtained with

-3 R100Ω MFCs (1.03 W m ).

The diversity of methanogens in cellulose-fed MFCs was also characterized. It

was shown that the suppression of methanogenesis was accompanied by a decrease in the

diversity of methanogens and changes in the concentration of SCFA, as revealed by

DGGE analysis of PCR-amplified 16S rRNA genes and HPLC analysis, respectively.

Analysis of partial 16S rRNA gene Sequences indicated that the most predominant

methanogens were related to the family Methanobacteriaceae.

The results demonstrate that electricity can be generated from cellulose by

exploiting rumen microorganisms as biocatalysts. Results suggest that oxidation of

iv metabolites with the anode as an electron sink was a rate limiting step in the conversion

of cellulose to electricity in MFCs. This study also demonstrates that the size of external resistance significantly affects the bacterial diversity and characteristic output of MFCs.

Thus the external resistance may be a useful tool to control microbial communities and consequently enhance performance of MFCs. Furthermore, this study shows that methanogenesis competes with electricity generation at the early stages of MFC operation but operating conditions suppress methanogenic activity over time. An improved understanding of the microbial communities, interspecies interactions and processes involved in electricity generation is essential to effectively design and control cellulose-fed MFCs for enhanced performance. In addition, technical and biological optimization is needed to maximize power output of these systems.

Dedicated to my family

vi ACKNOWLEDGMENTS

I would like to express my gratitude and deep appreciation to Dr. Ann D. Christy, advisor of my dissertation research, for her infinite support, encouragement, and patience as well as the magnificent mentorship she provided me with throughout the time it took me to complete this research and write the dissertation. It was a tremendous honor working with her.

The members of my dissertation committee, Dr. Olli Tuovinen, Dr. Zhongtang Yu,

Dr. Burk Dehority and Dr. Alfred Soboyejo, have generously given their time and expertise to improve my work. I thank them for their contribution and their good-natured support. Specifically, I wish to extend my sincere thanks and appreciation to Dr. Olli

Tuovinen whose knowledge, wisdom, and advice have supported me over the course of my academic training at the Ohio State University. His technical and editorial guidance have been an inspiration to me and a critical key to the completion of my dissertation. He was always available, even during his sabbatical leave, to work on my writings and discuss new topics, ideas, and results. I am grateful to Dr. Zhongtang Yu for his valuable insights and ideas as well as countless advice that supported and expanded my work. His wide expertise in the field of molecular biology has played as essential role in this research project. I would like to sincerely thank Dr. Burk Dehority who triggered my

vii academic journey in the US, and provided me with an incredible technical training and innumerable guidance before and throughout my dissertation research. I am also thankful to Dr. Alfred Soboyejo, for his continued encouragement and inspiring advice that supported and enlightened my research. I am grateful too for the support and advice from my faculty colleagues in Department of Food, Agricultural and Biological Engineering.

I would like to thank many persons who helped me with my experiments, especially Chris Gecik for helping with electrical instrumentation, Carl Cooper and Kevin

Duemmel for their assistance with fabrication of MFCs, and Don Irvine for providing computer support. Carol Moody, Kay Elliot, and Kevin Davison are also thanked for their administrative support.

I must acknowledge as well the many friends, colleagues and students who assisted, advised, and supported my research and writing efforts over the years.

Especially, I need to express my gratitude and deep appreciation to Sarah Carver, Mike

Nelson, James Douglass, Kerry Hughes Zwierschke, Eun Kyoung Kim, Jill Stephen,

Nikki Skrinak, Bethany Frew Corcoran, Brian Henslee, Peter Gehres, and Clayton Bettin.

The Iranian Students and Scholars at the Ohio State University and their families made these years in Ohio, away from home, very special, very lively and very enriching.

The enjoyable times spent with friends provided a well-needed balance to the work; thank you for your support and friendship.

I am deeply indebted to my family for their care, support and encouragement. I am grateful to my parents without whom I would never have been able to achieve so much. I am also thankful to my mother and father in-law for their support, and encouragement. My brothers Saied and Ehsan, I am thankful for having you in my life. I

viii have been blessed with an absolutely superb grandfather and honor the memory of the three of my grandparents whom I have lost, peace be upon their souls.

Last and most importantly, I would like to thank the love of my life, my wife

Najmeh, for her unyielding devotion and love, endless support, patience, and encouragement and all the helps she provided me with during the course of completing this research and writing the dissertation. She is the only one who knows the real price of this dissertation as we worked hard together to make this accomplishment happen.

ix VITA

February 8, 1979 ……………………… Born – Isfahan, Iran

1997 – 2001 …………………………… B.S. Animal Sciences Isfahan University of Technology, Isfahan, Iran

2001 – 2003 ……………………………M.S. Animal Sciences Isfahan University of Technology, Isfahan, Iran

2004 – Present ………………………… Graduate Research and Teaching Associate, Food, Agricultural and Biological Engineering, Ohio State University

PUBLICATIONS

H. Rismani-Yazdi, A.D. Christy, B.A. Dehority, M. Morrison, Z. Yu and O.H. Tuovinen. 2007. Electricity generation from cellulose by rumen microorganisms in microbial fuel cells. Biotechnol. Bioeng. 97:1398-1407.

H. Rismani-Yazdi, S. Carver, A.D. Christy, and O.H. Tuovinen. 2008. Cathodic limitations and optimization in microbial fuel cells: An overview. J. Power Sources. 180:683-694.

x FIELDS OF STUDY

Major Field:

Food, Agricultural and Biological Engineering

Study in:

Biological Engineering

x TABLE OF CONTENTS

Page Abstract …………………………………………………………………………… ii

Dedication …..…………………………………………………………………….. vi

Acknowledgments ……..………………………………………………………….. vii

Vita ….……………………………………………………………………………. x

List of Tables …………………….……………………………………………….. xv

List of Figures …………………………………………………………………….. xvi

Chapters:

1. Introduction …..…………………………………………….…………………... 1

1.1 Bio-generation of electricity in microbial fuel cell ……………………….. 1

1.2 Cellulosic biomass: A sustainable source of renewable energy …………… 8

1.3 Research objectives …………………………………………………………. 10

1.4 Outline of dissertation ……………………………………………………... 10

1.5 Contribution of dissertation …………………………………………………. 11

2. Cathodic limitations in microbial fuel cells: An overview ……………………... 12

2.1 Introduction …………………………………………………………………. 12

2.2 Cathodic limitations ……………………………………………………….. 17

2.2.1 Activation losses ……………………………………………………... 17

2.2.2 Ohmic Losses …………………………………………………………. 18

2.2.3 Mass transport losses …………………………………………………. 19

2.3 Reducing cathodic activation losses ……………………………………….. 20

xi 2.3.1 Mediators ……………………………………………………………... 21

2.3.2 Catalysts …………………………………………………….……….. 24

2.3.2.1 Metal-based catalysts …………………………………………… 24

2.3.2.2 Biocatalysts ……………………………………………………... 28

2.3.3 Cathode surface area …………………………………………………… 30

2.3.4 Operational conditions …………………………………………………. 31

2.4 Reducing cathodic ohmic losses …………………………………………… 33

2.4.1 Catholyte ……………………………………………………………….. 33

2.4.2 Membrane ……………………………………………………………….. 34

2.4.3 Electrode-spacing ……………………………………………………….. 39

2.5 Reducing cathodic mass transport losses …………………………………… 45

2.5.1 Oxidant concentration …………………………………………………. 45

2.5.2 Cathode electrode design ……………………………………………….. 49

2.5.3 Cathode compartment design …………………………………………… 49

2.6 Other losses …………………………………………………………………. 51

2.7 Concluding remarks ………………………………………………………. 52

3. Electricity generation from cellulose by rumen microorganisms in microbial fuel cells …………………………………………………………… 55

3.1 Introduction …………………………………………………………………. 55

3.2 Materials and methods ……………………………………………………... 58

3.2.1 Microorganism …………………………………………………………. 58

3.2.2 Medium …………………………………………………………………. 58

3.2.3 Microbial fuel cells ……………………………………………………... 59

xii 3.2.4 DNA extraction and quantification ……………………………………... 61

3.2.5 DGGE …………………………………………………………………… 62

3.2.6 Cloning, sequencing, and DNA sequence analysis ……….…………….. 62

3.2.7 Nucleic acid accession numbers ……………………………………….. 63

3.3 Results ……………………………………………………………………... 70

3.3.1 Current production ……………………………………………………... 70

3.3.2 Polarization characteristics …………….……………………………….. 72

3.3.3 Substrate comparison …………………………………………………… 74

3.3.4 DGGE analysis of microbial community ……………………………….. 76

3.3.5 Phylogenetic diversity revealed by cloning and sequencing …………… 76

3.4 Discussion …………………………………………………………………. 78

3.5 Conclusion …………………………………………………………………. 84

4. External resistance affects the power output and bacterial diversity of cellulose-fed microbial fuel cells ……………………………………………….. 85

4.1 Introduction …………………………………………………………………. 85

4.2 Materials and methods ……………………………………………………... 90

4.2.1 Microorganisms and medium …………………………………………… 90

4.2.2 MFCs construction and operating conditions …………………………… 91

4.2.3 Electrical measurements ……………………………………………….. 93

4.2.4 Coulombic efficiency …………………………………………………. 94

4.2.5 Analysis of cellulose degradation ……………………………………... 95

4.2.6 Analysis of bacterial diversity …………………………………………. 96

4.2.7 Analysis of fermentative metabolites …………………………………. 96

xiii 4.2.8 Analysis of bacteria cell protein ……………………………………….. 97

4.3 Results ……………………………………………………………………... 98

4.3.1 MFC performance as a function of external resistance ……………….. 98

4.3.2 Bacterial diversity as a function of external resistance ……………….. 100

4.3.3 Effect of different external resistance on microbial metabolism ……... 102

4.4 Discussion …………………………………………………………………. 104

5. Evaluation of methane formation in microbial fuel cells generating electricity from cellulose ……………………………………………………….. 108

5.1 Introduction …………………………………………………….…………. 108

5.2 Materials and methods ……………………………………………………... 110

5.2.1 Microbial fuel cell construction and operation ………….……………. 110

5.2.2 Electrical measurements ………………………………….……………. 112

5.2.3 Analytical techniques …………………………………………………. 113

5.2.4 DNA extraction and amplification ……………………………………... 114

5.2.5 Recovery and sequencing of DGGE bands …………………………… 114

5.2.6 Band sequencing ……………………………………………………….. 115

5.3 Results and discussion ……………………………………………………... 115

5.3.1 Methane production and MFC performance …………………………… 115

5.3.2 Effect of methanogens on metabolites …………………………………. 119

5.3.3 Bacterial and archaeal diversity ……………………………………….. 122

6. Conclusions ……………………………………………………………………... 128

6.1 Implications of this research ……………….………………………………. 128

6.2 Suggestions for future research ……………………………………………... 130

xiv References …………………………………………………………………………. 133

xv LIST OF TABLES

Tables Page

2.1 Summary list of recent (2001-2007) MFC review papers ……………….……… 16

3.1 Bacterial composition as inferred from the cloned 16S rRNA gene fragments recovered from anode-associated bacteria (AB) in the cellulose-fed MFC ………………………………………………………….…… 64

3.2 Bacterial composition as inferred from the cloned 16S rRNA gene fragments recovered from suspended bacteria (SB) in the anode chamber in the cellulose-fed MFC ………………………………………….…… 68

4.1 Maximum power out put in MFCs with different external resistances, designs and operational conditions ……………………………………….……… 88

4.2 Effect of external resistance on performance characteristics of MFCs ….……….. 100

5.1 Microbial fuel cell specifications and experimental conditions …….…………... 111

5.2 Phylogenetic identification of the bands excised and sequenced from DGGE gels of 16S rRNA genes PCR-amplified using cathode electrode in a membrane-less MFC …………………………….………. 124

xvi LIST OF FIGURES

Figures Page

1.1 Schematic of a microbial fuel cell containing a model bacterial cell ……….……. 3

2.1 Schematic potential losses for a cathodic reaction ………………………….…… 18

2.2 Different approaches to enhance cathodic kinetics in MFCs. samples …….…….. 23

2.3 Effect of phosphate buffer concentration (pH 3.3) used as catholyte on the galvanodynamic polarization properties of pyr-FePc modified cathode for oxygen reduction …………………………………………………………….… 26

2.4 Effect of catholyte pH (500 mM phosphate buffer) on the galvanodynamic polarization properties of CoTMPP modified cathode for oxygen reduction …………………………………………….…….. 27

2.5 Power output as a function of cathode surface area (2 to 22.5 cm2) in a two-compartment MFC with fixed surface areas for anode (22.5 cm2) and PEM (30.6 cm2) ……….…………………………………………….……… 31

2.6 Power output as a function of anode surface area (2 to 22.5 cm2) in a two-compartment MFC with different PEM surface areas ……………….... 36

2.7 Schematic of a single-chamber membrane-less microbial fuel cell and SEM image of bacterial biofilm on the interior-surface of the cathode electrode in a membrane-less MFC ………………………………….…. 40

2.8 Power-current and polarization properties of an air-cathode MFC in the presence and absence of a PEM ………………………………………….……... 41

2.9 Effect of electrode spacing (2 and 4 cm) on power output of a membrane-less MFC ……………………………………………………….……. 42

2.10 Schematic of a membrane-electrode assembly MFC, an upflow MFC, and an upflow MFC with an interior cathode …………………………….…….. 44

2.11 Schematic of a sediment microbial fuel cell with a rotating cathode ….……….. 46

xvii 2.12 Effect of nitrate loading rate on the cathode potential and power density normalized by the net volume of cathode compartment of an MFC with an anaerobic denitrifying cathode compartment ……….………... 48

3.1 The experimental two-compartment MFC ………………………………………. 60

3.2 Electricity generation in two-compartment MFCs at 39±1°C ……………….….. 71

3.3 Polarization properties of an MFC with rumen microorganisms as biocatalysts and cellulose as the electron donor at 39±1°C ……………….……. 73

3.4 Electricity generation in MFCs at 39±1°C ………………………………….…… 75

3.5 DGGE profile of the community DNA extracted from anode-attached and bacteria suspended in the anode chamber of four MFCs ……………….….. 77

4.1 Schematic diagram of the microbial fuel cells used for the experiment …….…... 92

4.2 Effect of external resistance on polarization and power-current properties of MFCs with different external resistances ……………………….….. 99

4.3 Effect of external resistance on bacterial diversity in MFCs ………….…………. 101

4.4 Concentrations of SCFA produced in MFCs with different external resistances after 10 weeks of operation ……………………………………….….. 103

5.1 A schematic diagram of the microbial fuel cells used for the experiment ….…… 112

5.2 Current production from cellulose with rumen microorganisms in MFCs with two different external resistances over the course of operation ………….…. 116

5.3 Voltage-current and the power-current characteristics of MFCs ……………….... 117

5.4 Coulombic efficiency of MFCs during the operation period …………………….. 118

5.5 Concentrations of SCFA produced in MFCs with different external resistances over the course of operation ……………………………………….... 120

5.6 DGGE profiles of 16S rRNA genes amplified from planktonic and biofilm of bacterial and archaeal communities enriched in MFCs with two different external resistances over the time with cellulose ……………...... 123

xviii CHAPTER 1

INTRODUCTION

1.1 Bio-generation of electricity in microbial fuel cell

The world’s population is currently about 6.7 billion people, and at current rates

of increase is predicted to exceed 10 billion by the year 2050 (Bilgen et al. 2004;

Abulfotuh 2007; Martinot et al. 2007). This acute population growth leads to an inevitable and immediate consequence: a growing need for clean, safe, and sustainable

supplies of energy. Increasing concern over pollution, resource depletion, and climate

change implications of our continuing use of conventional fossil and nuclear fuels has prompted a growing interest in renewable energy sources. It is estimated that by 2030 about 15-20% of our energy needs will be met by renewable energy (Bilgen et al. 2004;

Abulfotuh 2007; Martinot et al. 2007). There is, therefore, a strict need for development

of new technologies that can make renewable resources accessible to supply this

increasing demand. In this regard, fuel cells, in which fuels are directly converted to

electrical energy by undergoing oxidation-reduction (redox) reactions at an anode and a

cathode, have been researched extensively (Srinivasan 2006).

1 A fuel cell is a device for the direct conversion of chemical energy to electrical

energy. It requires an anode, a cathode, a supporting electrolyte medium to connect the two electrodes, and an external circuit to utilize the energy (Larminie and Dicks 2003;

O'Hayre et al. 2005; Srinivasan 2006). Reactants must be supplied to both electrodes as a source for the electron transfer reactions; catalysts must be present to provide a rapid rate

of reaction at each electrode. Typically, a fuel cell works by the oxidation of a reduced

fuel (i.e., electron donor) at an anode with concomitant transfer of electrons via an

electrical circuit to a suitable electron acceptor (e.g., oxygen) at the cathode. In theory,

fuel cells are the most effective devices for the conversion of chemical energy to electrical energy because they avoid the limitations of the Carnot combustion cycle

(Larminie and Dicks 2003; O'Hayre et al. 2005; Srinivasan 2006). The simplest and most highly developed fuel cell currently is the hydrogen–oxygen fuel cell. In operation, the fuel (hydrogen gas) passes over the surface of the anode and is electrochemically oxidized to hydrogen ions, which enter the electrolyte and migrate towards the cathode.

Oxygen gas passes over the surface of the cathode and is reduced, combining with the hydrogen ions from the electrolyte to form water (Sopian and Daud 2006).

A second type of fuel cell is the bioelectrochemical fuel cell. In a

bioelectrochemical fuel cell, the electrode reactions are similar to those in the hydrogen–

oxygen fuel cell, except for the fact that the fuel is not hydrogen gas but a form of carrier-

bonded hydrogen, produced by physiological redox reactions. Bioelectrochemical fuel

cells are devices in which biological redox reactions provide power through intact

microorganisms or isolated enzymes (Bullen et al. 2006).

2 Despite the high-efficiency operation of fuel cells and the fact that they do not result in environmental pollution, their advantages are partially offset by the high cost of catalysts, highly corrosive electrolytes, and an elevated operation temperature (Srinivasan

2006). The discovery of electrode-reducing microorganisms with diverse catalytic abilities added to the high-efficiency design of fuel cells has resulted in the development of a third type of fuel cell system, known as microbial fuel cells (MFCs) (Rabaey and

Verstraete 2005; Lovley 2006; Du et al. 2007). MFCs provide alternative strategies for waste treatment and renewable energy production. MFCs are bioelectrochemical reactors in which microorganisms mediate the direct conversion of chemical energy stored in organic compounds from simple carbohydrates to waste organic matters into electrical energy. In MFCs, bacteria in the anodic compartment catalyze the oxidation of organic compounds and transfer the resulting electrons to the anode under anaerobic conditions

(Fig.1.1). These electrons then flow across a closed circuit, through a resistor or a device to be powered, to the cathode in the cathodic compartment. Protons released from the oxidation of organic compounds also migrate to the cathode, through a cation-exchange membrane separating anode and cathode, and combine with electrons and oxygen at the cathode surface to form water (Allen and Bennetto 1993).

Operational and functional advantages have distinguished MFCs from all the technologies currently used for generating energy from organic compounds. Such advantages are: 1) high conversion efficiency due to the direct conversion of chemical energy to electricity; 2) efficient operation at mild to low temperatures; 3) no off-gas treatment requirement, the off-gas of MFCs is enriched in carbon dioxide and normally has no useful energy content; 4) no external energy requirement, provided the cathode is

3 passively aerated; 5) low cost and ease of operation; and 6) MFCs have potential for

widespread application in locations lacking electrical infrastructures and also to expand the diversity of fuels we use to satisfy our energy requirements (Rabaey and Verstraete

2005).

Figure 1.1. A schematic of a microbial fuel cell containing a model bacterial cell

(BC). Electron transport involves oxidized and reduced electron carriers (ECred and ECox),

and mediators (Medox and Medred).

4 Progress made over the past two decades has considerably improved the power output and conversion efficiencies of MFCs, but substantial additional optimization is still required to enhance the performance of this technology. Such optimization requires fundamental understanding of the electrochemical and biological processes that influence the performance of these systems.

Electrochemical optimization can occur through enlarging the electrode surface, using different type of proton exchange materials or even omitting the proton exchange membrane, using different anode and cathode materials, adding redox mediators to either the electrode or the feed solution, and designing different configuration for the anode and cathode compartments ( Logan et al. 2006; Du et al. 2007).

Microorganisms that have been tested as biocatalysts for use in MFCs include pure cultures of obligately and facultatively anaerobic bacteria and consortia from sea floor sediments and municipal and industrial wastewater (Logan and Regan 2006a).

These microorganisms have shown the ability to oxidize organic compounds and transfer electrons directly to an electrode, a capability that has implications for harvesting energy from waste organic matter (Bond and Lovley 2003). As a result, MFCs have found application for wastewater treatment, powering marine electronic devices in remote locations, and used as biological sensors (Aelterman et al. 2006; Logan and Regan

2006b; Lovley 2006; Kumlanghan et al. 2007).

Organic compounds are typically used as electron donors for the microorganisms in MFCs. In previous studies the electron donors have been limited to readily metabolizable organic compounds such as simple carbohydrates, low molecular weight organic acids, starch and amino acids (Kim et al. 2000; Niessen et al. 2004; Rabaey et al.

5 2004; Liu et al. 2005b; Logan et al. 2005). Most of these compounds are of high-value

and have economically attractive alternative uses as chemical feed stocks in industry.

Search for electron donors has turned to low-value organic compounds such as municipal wastewater and marine sediment organic matter (Aelterman et al. 2006; Logan and Regan

2006b; Lovley 2006; Kumlanghan et al. 2007) . However, the function of MFCs on complex organic compounds has been impractical due to the limited catalytic activity of

the employed biocatalysts.

The most critical step in the MFC process is the transfer of electrons from the

bacteria to the electrode (Schröder 2007). Four concepts have been proposed to link

microbial catabolic activity to electrodes for in situ electricity generation. 1)

Fermentation products (e.g., hydrogen, methanol, or ethanol) have been used for in situ

electricity generation (Niessen et al. 2005). 2) Artificial electron transfer mediators serve

as electron shuttles that penetrate the bacterial cells, divert electrons from the respiration

chain and from internal metabolites, and transfer electrons to the fuel cell anode (Sund et

al. 2007). However, many of these mediators are toxic and also not cost-effective because

of the periodical replenishment requirement. 3) Direct electron transfer from the cell

surface to the electrode involves a variety of biomolecules (including cytochromes and

dehydrogenases); this method requires the physical contact of the bacteria with the electrode and has been studied in species such as Geobacter sulfurreducens and

Rhodoferax ferrireducens (Aelterman et al. 2006; Logan and Regan 2006b; Lovley 2006;

Kumlanghan et al. 2007). 4) Soluble redox mediators are excreted by bacteria to function

as electron shuttles between the bacteria and an electrode ( Hernandez and Newman 2001;

Gralnick and Newman 2007). These have been studied especially in Shewanella and

6 Geothrix species. Bacteria producing such mediators can be either attached to the

electrode or freely suspend in the anode compartment. Both of these mechanisms play an

important role in electricity generation in MFCs depending on the mode of operation

(Schröder 2007).

Bacterial metabolism in the anode compartment influences the performance of an

MFC, although exact details are yet to be completely elucidated. Two different metabolic

pathways can occur in the anode compartment: redox oxidative metabolism and

fermentation. Under the oxidative metabolism, bacteria gain energy by oxidizing the

electron donor and transferring the electrons to the anode or their characteristic electron

acceptors such as nitrate, sulfate, and/or Fe (III) oxide. Bacteria using the fermentation

metabolisms direct the flow of electrons toward the formation of reduced metabolites.

Neither the fermentation metabolism nor the bacterial electron transfer to their natural

electron acceptors are desirable in MFCs as these result in low power output and reduce

the substrate conversion efficiency (Rabaey and Verstraete 2005).

The selection of suitable microorganisms with enhanced capabilities for sustained

and efficient electricity production is also considered an important method for the

biological optimization of MFCs. Generation of electricity using relatively few numbers

of pure culture bacteria has been previously reported. However, for application purposes,

MFCs rely on the selection of natural microbial populations. Fuel cells have been used to enrich microbial consortia from a wide variety of environmental samples, including

anaerobic sludge, municipal and industrial wastewater, and seafloor sediments.

Enrichments have usually been made with passive techniques, in which an inoculum has

been added into the anode compartment of an MFC operated with replenishment of the

7 medium (Kim et al. 2004). In a few studies serial enrichments of the anode-colonizing bacteria into new fuel cells have been used and they have resulted in changes in the microbial community structures and increased power output and conversion efficiency

( Rabaey et al. 2004; Kim et al. 2005). The operation conditions of MFCs and the system engineering design also affect the result and effectiveness of the enrichment, a factor that has yet to be investigated.

1.2 Cellulosic biomass: A sustainable source of renewable energy

Cellulosic biomass that is produced as a by-product of agricultural and industrial activities has the largest potential contribution in the context of sustainable and secure energy production due to economic, scale of supply, and environmental considerations

(Bridgwater 2006; Chang 2007; Dunnett and Shah 2007; Himmel et al. 2007).

Conventionally, these waste products were dealt with in different ways including landfilling, composting, and incineration. These methods encounter technical and economic hurdles and result in environmental concerns such as green house gas emissions. As an alternative to the conventional methods, biotechnology provides means for successful conversion of these high energy compounds. Biotechnological methods are usually more efficient and environmentally friendly, use less energy and produce less secondary wastes as compared to conventional practices.

The major constituent of cellulosic biomass is cellulose, which is a linear polymer of glucose connected through β-1,4-linkages. Cellulose is usually arranged in microcrystalline structures, which can be very difficult to dissolve or hydrolyze under natural conditions. Bioconversion of cellulosic biomass has proved to be a practical and

8 promising processing method while producing value-added products and alternative fuels

(Knauf and Moniruzzaman 2004). Biotechnological processes of cellulosic biomass for

the production of hydrogen gas, methane, and ethanol have been well studied (Bridgwater

2006). The efficiency and advantages of these bioprocesses are partially offset by the

high cost of required cellulolytic enzymes, the generation of toxic compounds during

pretreatment of cellulose, and the lack of fermentative microorganisms that can

efficiently use all by-products. In addition, methane and ethanol still have to be

combusted, which reduces the energy efficiency, and separation, purification and storage

of hydrogen gas is costly. The development of cost-effective, energy-efficient and

environmentally-friendly methods of treating cellulosic biomass is, therefore, essential.

This will require fundamental scientific research and breakthroughs in technology.

The value of cellulosic biomass will increase if its degradation can be linked to

the generation of electricity through MFCs. In order to harvest the chemical energy in

cellulose, it is necessary to couple the hydrolysis of its compact crystalline complex with

reduction of the anode. This requires choosing anaerobic microorganisms or consortia that are capable of hydrolyzing cellulose and oxidizing the intermediate metabolites using the anode as an electron sink. The microbial populations in the rumen contain both strict and facultative anaerobes that can effectively hydrolyze cellulose (Krause et al. 2003) and conserve energy via anaerobic respiration or fermentation (Hobson and Stewart

1997). These considerations prompted the present study to focus on the development of an innovative microbial fuel cell for direct conversion of cellulose into electrical energy with rumen microorganisms as biocatalysts.

9 1.3 Research objectives

The objectives of this study were:

1) Test the possibility of generating electricity in an MFC with rumen

microorganisms as biocatalysts and cellulose as the electron donor.

2) Analyze the composition of the bacterial communities enriched in the cellulose-

fed MFCs.

3) Determine the effect of various external resistances on power output and

coulombic efficiency of cellulose-fed MFCs.

4) Evaluate bacterial diversity and cellulose metabolism under different circuit loads.

5) Assess the influence of methane formation on the performance of cellulose-fed

MFCs under long-term operation.

6) Characterize the diversity of methanogens in cellulose-fed MFCs.

1.4 Outline of the dissertation

The contents of the dissertation are structured as follows:

Chapter 1: An introduction to the subject and research aims of the dissertation.

Chapter 2: A review of cathodic limitations in MFCs

Chapter 3: Addresses objectives 1 and 2

Chapter 4: Presents experimental results and discussion on objectives 3 and 4

Chapter 5: Addresses objectives 5 and 6

Chapter 6: Concludes the dissertation, discusses the implications of this study,

and recommends future directions for research on microbial fuel cells.

10 1.5 Contribution of the dissertation

Chapters 2 and 3 of the dissertation have already been published in peer-reviewed journals as follows:

1) H. Rismani-Yazdi, A.D. Christy, B.A. Dehority, M. Morrison, Z. Yu and O.H. Tuovinen. 2007. Electricity generation from cellulose by rumen microorganisms in microbial fuel cells. Biotechnol. Bioeng. 97:1398-1407.

2) H. Rismani-Yazdi, S. Carver, A.D. Christy, and O.H. Tuovinen. 2008. Cathodic limitations and optimization in microbial fuel cells: An overview. J. Power Sources 180:683-694.

CHAPTER 2

CATHODIC LIMITATIONS IN MICROBIAL FUEL CELLS: AN OVERVIEW1

2.1. Introduction

With the increasing concern for alternative energy sources, waste management, global climate change, and non-edible feedstocks, the search for novel technological solutions continues. Fuel cells are one alternative energy technology being studied for full-scale implementation (Bullen et al. 2006). These can be classified into three subgroups: catalytic, enzymatic, and microbial. Since the turn of the century, the research on microbial fuel cells (MFCs) has experienced rapid increases. MFCs are unique in their ability to utilize microorganisms, rather than an enzyme or inorganic molecule, as catalysts for converting the chemical energy of feedstock directly into electricity.

MFCs often consist of two compartments, the anode and cathode, which are often separated by a proton exchange membrane (Fig. 1.1). The anode chamber contains microorganisms that oxidize the available substrate (i.e., the electron donor). The

1 This chapter was published as H. Rismani-Yazdi, S. Carver, A.D. Christy, and O.H. Tuovinen. 2008. Cathodic limitations and optimization in microbial fuel cells: An overview. J. Power Sources 180(2): 683-694. 12

anaerobic oxidation is coupled with liberation of electrons which are transported through the cellular respiratory chain ultimately to the anode. Substrates used in MFC research vary from sugars and organic acids such as glucose or acetate to complex polymers such as starch and cellulose. Domestic, industrial, and animal waste streams have been used as feedstock for generating electricity in MFCs. In the example blow, glucose is the electron donor:

+ - Anode: C6H12O6 + 6H2O → 6CO2 + 24H + 24e (2.1)

The anode acts as an artificial, external electron acceptor for the microorganisms. The

electrons travel through a resistor or a device to be powered, generating electricity until

reaching the cathode.

+ - Cathode: 6O2 + 24H + 24e → 12H2O (2.2)

While the electrons travel through the circuit, the corresponding protons migrate to the cathodic compartment through a proton-exchange membrane (PEM) to maintain charge neutrality. At the cathode an electron acceptor (e.g., oxygen) is reduced by the electrons via the circuit and the protons via the membrane. The electrochemical reactions in MFCs are comparable but the kinetics and coulombic efficiencies may vary depending on the physical, chemical and biological operating conditions.

The cathodic reduction can be classified into aerobic or anaerobic reactions depending on the source of the final electron acceptor available. In aerobic cathodes, oxygen is the terminal electron acceptor. The reduction of oxygen is the most dominant electrochemical reaction at the surface of cathode electrodes. Unlimited availability and high standard redox potential make oxygen an exceptional electron acceptor. Two

processes can occur during cathodic oxygen reduction. The desired reaction is the

production of water through a four-electron pathway:

+ - O2 + 4H + 4e → 2H2O (E´° = 0.816 V) (2.3)

The other pathway consists of a two-electron reaction with the production of hydrogen

peroxide:

+ - 2O2 + 4H + 4e → 2H2O2 (E´° = 0.295 V) (2.4)

Incomplete reduction of oxygen leads to low energy conversion efficiency and produces

reactive intermediates and free radical species which can be destructive. Permanganate

has also been used as an alternative electron acceptor to oxygen to support the cathodic

reduction reaction (You et al. 2006). The cathode compartment can also be maintained

under anaerobic conditions. In this case, microorganisms transfer the electrons from the

cathode to the final electron acceptor (e.g., nitrate) (Park et al. 2005; Clauwaert et al.

2007a).

An ideal MFC can produce current while sustaining a steady voltage as long as

the substrate is supplied. The theoretical ideal voltage, Ethermo (V), attainable from an

MFC can be thermodynamically predicted by the Nernst equation;

RT E = E 0 − ln(∏ ) (2.5) thermo nF where E0 is the standard cell potential (V), R is the ideal gas constant (8.314 J mol-1 K-1),

T is the temperature (K), n is the number of electrons transferred in the reaction

(dimensionless), F is the Faraday’s constant (96,485 C mol-1), and Π is the chemical

activity of products divided by those of reactants(dimensionless).

In practice, the actual voltage output of an MFC is less than the predicted thermodynamic ideal voltage due to irreversible losses (i.e., overpotentials). The three major irreversibilities that affect MFC performance are: activation losses, ohmic losses, and mass transport losses. These losses are defined as the voltage required to compensate for the current lost due to electrochemical reactions, charge transport, and mass transfer processes that take place in both the anode and cathode compartments (O'Hayre et al.

2005). The extent of these losses varies from one system to another. The real operational voltage output (Vop) of an MFC can be determined by subtracting the voltage losses associated with each compartment from the thermodynamically predicted voltage as follows:

Vop= Ethermo – [(ηact + ηohmic + ηconc)cathode + (ηact + ηohmic + ηconc)anode] (2.6) where, Ethermo is the thermodynamically predicted voltage, ηact is the activation loss due to reaction kinetics, ηohmic is the ohmic loss from ionic and electronic resistances, and ηconc is the concentration loss due to mass transport limitations. The above equation shows that cathode and anode overpotentials collectively limit the performance of MFCs and that the overall performance can be improved by optimizing both the anode and cathode.

The purpose of this paper is to review cathodic limitations in MFCs and recent studies that have addressed these problems and explored approaches for improvement.

These limitations are based on irreversible reactions and processes in the cathode compartment that can severely affect the performance of MFCs. Many reviews are available on MFC technology and operations (Table 2.1) but to this point, no in-depth overview of cathodic limitations exists in the literature.

Title Year Citation

Extracellular Electron Transfer 2001 (Hernandez and Newman 2001)

Production of Bioenergy and Biochemicals from Industrial 2004 (Angenent et al. 2004) and Agricultural Wastewater

Microbial Fuel Cells: Novel Biotechnology for Energy 2005 (Rabaey and Verstraete 2005) Generation

Applications of Bacterial Biocathodes in Microbial Fuel 2006 (He and Angenent 2006) Cells

Electricity-Producing Bacterial Communities in Microbial 2006 (Logan and Regan 2006a) Fuel Cells

Electrochemically Active Bacteria (EAB) and Mediator- 2006 (Chang et al. 2006) Less Microbial Fuel Cells

Microbial Fuel Cells: Novel Microbial Physiologies and 2006 (Lovley 2006) Engineering Approaches

Microbial Fuel Cells: Challenges and Applications 2006 (Logan and Regan 2006b)

Microbial Fuel Cells: Methodology and Technology 2006 (Logan et al. 2006)

Microbial Fuel Cells in Relation to Conventional Anaerobic 2006 (Pham et al. 2006) Digestion Technology

Microbial Fuel Cells for Wastewater Treatment 2006 (Aelterman et al. 2006)

Anodic Electron Transfer Mechanisms in Microbial Fuel 2007 (Schröder 2007) Cells and Their Energy Efficiency

A State of the Art Review on Microbial Fuel Cells: A 2007 (Du et al. 2007) Promising Technology for Wastewater Treatment and Bioenergy

Extracellular Respiration 2007 (Gralnick and Newman 2007)

Microbial Ecology Meets Electrochemistry: Electricity- 2007 (Rabaey et al. 2007) Driven and Driving Communities

Table 2.1. Summary list of recent (2001-2007) review papers on MFCs.

2.2. Cathodic limitations

2. 2.1. Activation losses

Current production in MFCs depends largely on the kinetics of the reduction that takes place at the cathode. The reaction kinetics is limited by an activation energy barrier which impedes the conversion of the oxidant into a reduced form (i.e., Eq 2.1). When current is drawn from a fuel cell, a portion of the cathode potential is then lost to overcome this activation barrier. The potential loss due to activation is called cathodic activation loss (ηact)cathode (i.e., activation overpotential) (O'Hayre et al. 2005).

Activation losses result in a characteristic, exponentially formed loss on the current-voltage curve at low current densities (Fig. 2.1). As more current is taken from the MFC, the activation loss increases and results in a lower cell potential (Larminie and

Dicks 2003). As with chemical and biological fuel cells, the cathodic activation losses dominate the performance of MFCs (Gil et al. 2003). The magnitude of cathodic activation overpotential depends on the reduction kinetics. Kinetic performance can be improved by decreasing the activation barrier and increasing the reaction interface area, temperature, or oxidant concentration.

Figure 2.1. A. Schematic potential losses for a cathodic reaction displaying activation, ohmic, mass transport and parasitic regions. B. A typical power-current curve.

2.2.2. Ohmic losses

MFC performance is also restricted by cathodic ohmic overpotentials, also known as internal resistances. This loss is the voltage that is required to drive the electron and proton transport processes. Since MFC conductors are not ideal, they have an intrinsic resistance to charge flow (Larminie and Dicks 2003). The ohmic overpotential (η ohmic),

therefore, represents the voltage which is lost in order to accomplish charge transport (i.e., electrons and protons). This loss generally follows the Ohm’s law;

η ohmic =i Rohmic (2.7) where i is the current (A) and Rohmic is the ohmic resistance (Ω) of the MFC. The cathodic ohmic resistance is a combination of both ionic, Rion, and electronic, Relec, resistances, and includes the resistance from the electrode, electrolytes and interconnections;

Rohmic = Rion + Relec (2.8)

Internal resistance is usually dominated by the electrolyte resistance since the ionic conductivity is orders of magnitude lower than the electrical conductivity of the electrode materials (O'Hayre et al. 2005). The ohmic resistance of the electrolyte, Rion, can be expressed by;

l Rion = (2.9) AK where l is the distance (cm) and A is the cross-sectional area (cm2) over which the ionic conduction occurs, and K is the specific conductivity(Ω cm)-1 of the electrolyte (Larminie and Dicks 2003).

The cathodic ohmic loss is more pronounced at medium current densities and, following the Ohm's law, the operating voltage decreases linearly as current increases (Fig. 2.1).

Reducing the cathodic ohmic losses is important for improving the performance of MFCs.

2.2.3. Mass transport losses

The process of supplying oxidants (i.e., O2) and removing products (i.e., H2O) at the cathode of an MFC is governed by mass transport. Insufficient mass transport causes

reactant depletion or product accumulation. Reactant depletion affects both the Nernstian cell voltage and the reaction rates, leading to a performance loss. This loss is the voltage required to drive mass transport processes at the cathode and is referred to as cathodic concentration loss or mass transport loss, (ηconc)cathode (O'Hayre et al. 2005). Mass transport losses occur at high current density, and the magnitude increases with increasing current density (Fig. 2.1).

Mass transport limitations due to oxidant transport in the cathode compartment are typically much more severe than transport limitations in the anode compartment.

Hence, when determining mass transport losses in fuel cell systems, only the limiting concentration for the oxidant is considered.

2. 3. Reducing cathodic activation losses

Oxygen reduction is the most common cathodic reaction in MFCs. The slow rate of oxygen reduction on the surface of graphite/carbon electrodes leads to a high reduction overpotential, which is among the most limiting factors in the performance of MFCs (Gil et al. 2003). Therefore, improved cathodic reaction rates impact the efficiency and power output of MFCs and represent a major challenge for research and development (Zhao et al.; 2006Morris et al. 2007). Different approaches have been explored in several studies to improve the performance of the cathode by lowering the cathodic overpotential for oxygen. These approaches include the use of mediators, electrode modification with catalysts, and optimizing operational conditions within the cathodic compartment.

2.3.1. Mediators

Mediators that undergo reversible redox reactions can reduce the cathodic overpotentials in MFCs. They act as electron shuttles and transfer electrons from the cathode electrode to the terminal electron acceptor. The rate of reduction of the mediator on the electrode surface is relatively faster than that of oxygen, thereby enhancing the kinetics of cathodic reactions. Despite the benefits, employing mediators is considered unsustainable due to the need for regular replacement and therefore not practical for long- term MFC applications (Logan and Regan 2006b). The most common soluble mediator used for the cathodic reaction in MFCs is ferricyanide (hexacyanoferrate) (Fig. 2.2A). It has faster reduction kinetics than that of oxygen on the cathode and a relatively large redox potential. In contrast to oxygen, its concentration in the solution is not also limited by the solubility.

3- - 4- [Fe(CN)6] + e → [Fe(CN)6] (E´° = 0.358 V) (2.10)

4- + 3- O2 + 4[Fe(CN)6] + 4H → [Fe(CN)6] + 2H2O (2.11)

However, due to the slow rate of re-oxidation of ferricyanide by oxygen, it functions as an electron acceptor rather than a mediator (Pham et al. 2004). Oh et al. (2004) described a 50-80% increase in maximum power using ferricyanide in the cathode compartment as compared to an oxygen-saturated aqueous cathode or a platinum coated air-cathode. The observed differences were attributed to high open circuit potential and a greater mass transfer efficiency of ferricyanide solution than that of dissolved oxygen. Similar results have also been reported by Oh and Logan (2006), Ringeisen et al. (2006), and Liu and Li

(2007). Although widely used in laboratory experiments, ferricyanide is not a suitable choice for sustainable electricity generation in MFCs. It is potentially toxic, requires

regular replenishing due to its low rate of regeneration by oxygen, and diffuses through the membrane over long-term operation which eventually reduces the overall performance of the MFCs (Logan and Regan 2006b).

Impregnating mediators into the cathode electrode materials has been demonstrated to eliminate the need for continuous addition or recycling of soluble mediators (Park and Zeikus 2003). Metal oxides incorporated into the cathode electrode as electron transfer mediators have been shown to improve power output of MFCs due to enhanced electron transfer kinetics (Park and Zeikus 2003). The redox couple Fe3+/Fe2+

has been used as an alternative mediator for cathodic oxygen reduction because of its fast redox reaction rate and relatively high standard potential (E0 = 0.77 V) (Terheijne et al. 2006). This reversible electron transfer reaction was shown to considerably decrease the cathodic overpotential. However, the performance of a ferric iron reducing cathode is limited by the low solubility of ferric iron at pH values higher than 2.5. MFCs commonly contain catholyte with near neutral pHs. Furthermore, the transport of cation species other than protons through the PEM also raises the pH of catholyte (Rozendal et al. 2006). Ter

Heijne et al. (2006) demonstrated that, by employing a bipolar membrane, the catholyte pH can be maintained sufficiently low to keep Fe3+ soluble.

The cathodic activation loss can also be minimized by using oxidants that have redox potentials higher than that of oxygen (Fig. 2.2A). Using permanganate as the cathodic electron acceptor (Eq. 2.12) under acidic conditions in a two-compartment MFC,

You et al. (2006) reported 4.5 and 11 times higher power density as compared to ferricyanide and oxygen, respectively, as the cathodic oxidant.

− + − 0 MnO4 + 4H + 3e → MnO2 + 2H2O (E = 1.70 V, pH 1.0) (2.12)

Figure 2.2. Different approaches to enhance cathodic kinetics in MFCs. A. The use of mediators (e.g., ferricyanide) and strong oxidants (permanganate); B. catalytic electrode;

C. bacteria (BC) catalyzing the oxidation of transition metals; and D. bacteria (BC) catalyzing the reduction of the final oxidant (i.e., electron acceptor) through hypothetical direct and indirect electron transport mechanisms.

The increase in power output was attributed to the higher open circuit potential of the permanganate cathode, and it was found to be pH dependent. The performance of permanganate was less than that of oxygen under alkaline conditions. The practical application of a permanganate catholyte is limited due to the need for replenishing the depleted permanganate solution and the acidic pH requirement for optimal operation

(You et al. 2006).

2.3.2. Catalysts

The use of catalysts on the cathode surface can lower the cathodic activation overpotential and increase the current output of MFCs. Such catalysts considerably decrease the activation energy barrier and improve the kinetics of oxygen reduction at the electrode surface (O'Hayre et al. 2005). Different chemical and biological catalysts have been tested in efforts to improve MFC performance.

2.3.2.1. Metal-based catalysts

Platinum (Pt) has been widely used as the most common precious metal catalyst in cathode materials of MFCs because it has a favorably low overpotential for oxygen reduction. Logan et al. (2005) reported one order of magnitude increases in the power output of two-compartment MFCs with a Pt-coated cathode as compared to a plain carbon electrode. Although Pt has excellent catalytic ability, its relatively high cost limits its application. Efforts have been made to address this problem by lowering the amount of Pt necessary to coat the cathode electrode. Cheng et al. (2006c) reduced the Pt loading

to as little as 0.1 mg cm-2 of the cathode surface area and reported no severe drop in MFC performance.

Gold (Au) has also been tested as a cathode catalyst in MFCs because of its low overpotential for oxygen reduction. Kargi and Eker (2007) reported electricity generation in a two-compartment MFC with Au-coated copper as a cathode electrode. The authors did not however elaborate on the performance of such an electrode in comparison to that of commonly used electrodes in MFCs.

Non-precious metals have also been studied as catalysts for improving the kinetics of oxygen reduction in the cathode. The main distinction of using non-noble metals is the cost consideration for future potential large-scale applications of MFCs

(Zhao et al. 2006). Morris et al. (2007) compared the catalytic performance of lead dioxide (PbO2) to Pt in two-compartment MFCs. Their results demonstrated up to four times improvement in the power output and 50% reduction in cost per unit of power with

PbO2-coated electrodes as compared to that obtained with Pt-cathodes. However, the possibility of lead leaching from the cathode is a potential hazard that limits the use of

PbO2 as a cathode catalyst in MFCs. Such a limitation can be overcome by improving coating techniques and binding materials that would enhance stability of the catalyst. It has been shown that MFC performance can be affected by the binding material used to apply the catalyst to the electrode. Cheng et al. (2006c) compared MFC performance using Nafion and polytetrafluoroethylene (PTFE) as a Pt binder and reported 12% and

14% higher cathode potential and maximum power density, respectively, with Nafion.

Cobalt- and iron-based materials have also been investigated as alternative catalysts for oxygen reduction in MFCs. Zhao et al. (2005; 2006) used pyrolyzed-Fe(II)

phthalocyanine (pyr-FePc) and cobalt-tetramethylphenylporphyrin (CoTMPP) as cathode catalysts and reported performances similar to that of Pt. These results were also confirmed in other studies in which CoTMPP was tested in air-cathode MFCs (Cheng et al. 2006c), and used for catalytic tubular membrane cathodes (Zuo et al. 2007).

Metal-based catalysts are generally susceptible to adverse environmental conditions that may occur in MFCs as a result of chemical reactions, biological activities, and changes in catholyte composition. For example, their catalytic activity is reduced in sulfide- or chloride-rich environments. Schmidt et al. (2001) used a thin-film rotating disk Pt-electrode in a liquid electrolyte to demonstrate that adsorbed Cl- ions act as site blocking species. This effect reduced the active area available for oxygen reduction and changed the reaction pathway toward the production of H2O2.

Figure 2.3. Effect of phosphate buffer concentration (pH 3.3) used as catholyte on the galvanodynamic polarization properties of pyr-FePc modified cathode for oxygen reduction. Modified from Zhao et al. (2006).

Metal-based catalysts are also sensitive to high cathodic pH values, a common phenomenon that occurs in MFCs due to crossover of cations through the membrane to the cathode compartment. Zhao et al. (2006) studied the influence of catholyte composition on the performance of iron- and cobalt-based cathode catalysts for oxygen reduction. They demonstrated that lowering the concentration of a phosphate buffer catholyte (pH 3.3) from 500 to 50 mM reduced the performance of a pyr-FePc modified electrode by 40% (Fig. 2.3).

Figure 2.4. Effect of catholyte pH (500 mM phosphate buffer) on the galvanodynamic polarization properties of CoTMPP modified cathode for oxygen reduction. Modified from Zhao et al. (2006).

An increase in the catholyte pH from 2.4 to 7 (500 mM phosphate) resulted in an 80% decrease in the rate of oxygen reduction by a CoTMPP-based electrode (Fig. 2.4).

Increasing the catalyst load could partially compensate for the unfavorable neutral pH and low buffering capacity and thus improve the cathodic performance (Zhao et al. 2006).

2.3.2.2. Biocatalysts

Microorganisms can also be used as catalysts and mediators in the cathode.

Microbial growth is inevitable in the cathodic compartment because it is not feasible to operate it as a sterile unit. Several studies have demonstrated the capability of some microorganisms to utilize the cathode as the sole source of electron donors. These biocatalysts retrieve electrons directly from the cathode (Fig. 2.2C), or from electron mediators impregnated into the cathode (Fig. 2.2D), using mechanisms that are not yet understood (He and Angenent 2006). The electrons are then transferred to a final electron acceptor such as oxygen or an alternative oxidant such as nitrate (Gregory et al. 2004;

Rhoads et al. 2005; Clauwaert et al. 2007a). Bergel et al. (2005) found increased performance with a stainless steel cathode colonized by marine bacteria as compared to a clean cathode. Clauwaert et al. (2007b) reported electricity generation in MFCs, in which the cathode was exposed to air and inoculated with a consortium of sludge and sediment microorganisms. Although the underlying mechanisms were not explained, these two studies suggested that the oxygen reduction on the cathode was directly catalyzed by the biofilm.

Bacteria have also been used to catalyze the re-oxidation of redox couples incorporated into the cathode electrode as electron mediators. Rhoads et al. (2005) used a

manganese-oxidizing bacterium, Leptothrix discophora, as the biocatalyst in the cathode.

The current output improved by two orders of magnitude in comparison to a plain graphite electrode. The cathodic reaction in this system involved the reduction of

2+ electrode-deposited MnO2 to manganese ion (Mn ) by electrons at the cathode surface.

2+ The concurrent re-oxidation of Mn to MnO2 was then mediated by the Mn-oxidizing

Leptothrix. Ter Heijne et al. (2007) used an iron oxidizing bacterium, Acidithiobacillus ferrooxidans, for continuous ferrous iron oxidation on the cathode in a bipolar-membrane

MFC. This biologically catalyzed Fe2+/Fe3+ cathodic reaction produced a power output

(1.2 W m-2) higher than that obtained with a similar cathodic reaction under abiotic conditions (0.86 W m-2) (Terheijne et al. 2006).

The research on using microorganisms as cathode catalysts is still in its infancy.

The performance of biocatalysts is constrained by high cathodic activation overpotentials

(Clauwaert et al. 2007a). Appropriate comparisons between the performance of biocatalysts and metal-based catalysts in MFCs have yet to be reported. The dynamics of environmental conditions in MFCs can be inhibitory to biocatalysts. The accumulation of metabolites and ions crossed over through the membrane can hinder the bacterial activity.

Metabolites can also compete against the cathode by acting as electron donors for bacteria, counteracting the biocatalyst effect and reducing the performance. Sustenance of bacteria in the cathode compartment requires a carbon source. It is not known whether bacteria acting as biocatalysts obtain electrons required for energy metabolism from the cathode or from the oxidation of the carbon sources. A potential benefit of using biocatalysts may be the reduction of pollutants such as nitrates or chloroorganics in the cathode compartment (He and Angenent 2006; Lovley 2006; Ter Heijne et al. 2007). In

addition the cost and properties of biocatalysts and their compatibility with operating conditions, as compared to metal-based catalysts, may be more favorable for some future

MFC applications.

2.3.3. Cathode surface area

The power output of MFCs is constrained by the surface area of the cathode electrode (Oh et al. 2004; Freguia et al. 2007a; Zuo et al. 2007). Increasing the cathode surface area provides more reaction sites available for oxidant reduction and improves the cathodic reaction rate. This provides, therefore, an approach for enhancing the power output of MFCs. Using two-compartment MFCs, Oh and Logan (2006) demonstrated that increasing the cathode surface area by eleven–fold (from 2 to 22.5 cm2), at a fixed surface area for the anode (22.5 cm2) and the PEM (30.6 cm2), improved the maximum power density by one order of magnitude (Fig. 2.5). Similar results have also been reported by others (Oh et al. 2004; Zuo et al. 2007; Fan et al. 2007; Kargi and Eker 2007).

The increase in the reactor volume required to accommodate electrodes with large surface areas remains a challenge, especially in large-scale applications (Logan and

Regan 2006b). Various electrode materials and design configurations have been investigated in attempts to increase the available reaction sites in the cathode compartment while maximizing the surface area per volume ratio, i.e., specific surface area. Since graphite plate and carbon paper electrodes have limited surface areas, woven graphite felt (Gil et al. 2003; Park and Zeikus 2003), woven graphite mat (Rabaey et al.

2005b), granular graphite (Freguia et al. 2007a), and reticulated vitreous carbon (He et al.

2005; 2006; Ringeisen et al. 2006) have been tested because they have a larger specific

surface area than a graphite plate of identical dimensions. Tubular cathodes containing high surface area to volume ratios have also been investigated in MFCs ( Rabaey et al.

2005b; He et al. 2005; 2006; Zuo et al. 2007). The scalable characteristic of tubular cathodes makes them a promising architecture for developing large-scale MFCs (Zuo et al. 2007).

Figure 2.5. Power output as a function of cathode surface area (2 to 22.5 cm2) in a two- compartment MFC with fixed surface areas for anode (22.5 cm2) and PEM (30.6 cm2).

Modified from Oh and Logan (2006).

2.3.4. Operational conditions

Increasing oxidant concentration at the cathode affects performance of MFCs through both the Nernst equation and the kinetics of reduction reaction (Srinivasan 2006).

The concentration of reactants and products at the reactions sites determines the ideal 31

thermodynamic voltage according to the Nernst equation (Eq. 2.5). The thermodynamic gain from increasing the oxidant concentration is, however, small due to the logarithmic nature of the Nernst equation. In contrast, increasing the oxidant concentration substantially improves the cathodic kinetics by increasing the reaction rates in a linear fashion (O'Hayre et al. 2005). For a general reduction; aO + e- → bR (2.13) where O and R are the oxidant and reduced species, and a and b are the corresponding stoichiometric coefficients, the reaction rate (r) can be calculated by; r = k [O]a (2.14) where [O] is the concentration of an oxidant and k is the rate constant.

The effect of reactant concentration also works in concert with concentration losses at high current densities and will be discussed further in the section on reducing mass transport limitations.

The operating temperature also controls the cathode performance by affecting the kinetics of oxidant reduction and mass and proton transfer (Amirinejad et al. 2006).

Using a single-chamber membrane-less MFC, Liu et al. (2005a) reported a 9% increase in power output when the operating temperature increased from 20 to 32ºC. This improvement was shown to be mainly a result of increases in the cathodic potential.

The mesophilic temperature range (20 to 40 °C), at which many MFCs are operated, is suboptimal for cathodic reduction and performance of metal-based catalysts.

Increasing the operating temperature of MFCs is, however, limited by the temperature tolerance of microorganisms employed as catalysts in the anode and cathode. Operation of MFCs at temperatures beyond this limit can adversely affect the MFC performance by

inactivating the microorganisms. Thermophilic bacteria have been shown to generate electricity at elevated temperatures up to 60ºC (Jong et al. 2006; Mathis et al. 2008). The upper temperature limit for extreme thermophiles is in excess of 100ºC. While thermophilic operation reduces the activation, mass transport, and ohmic overpotentials, it requires a considerable amount of energy input to maintain the elevated tempratures.

Thermophilic MFCs do not appear practical unless the resulting improvements compensate for the energy input requirement. There is no meaningful comparison of performance between mesophilic and thermophilic MFCs available in the literature at present.

2.4. Reducing cathodic ohmic losses

Cathodic ohmic losses can be minimized by increasing the conductivity of the electrolyte materials used as catholyte and proton exchange membrane. Improvement can also be introduced by reducing the path distance between the cathode and anode electrodes.

2.4.1. Catholyte

Several catholyte characteristics have been found to restrict MFC performance due to high ohmic resistance. These elements include low proton concentrations at neutral pH values (Jang et al. 2004) and low ionic conductivity of the employed catholyte in most MFCs . Optimization of catholyte composition and concentration can, therefore, enhance the performance of MFCs (Zhao et al. 2006).

Increasing the ionic strength of the catholyte improves the cathodic proton transfer rate and results in increased current output (Gil et al. 2003; Jang et al. 2004). Liu

and Logan (2004) reported 85% increase in the power output of a single-chamber MFC when the ionic strength of the electrolyte was increased from 100 to 400 mM with NaCl.

This effect was attributed to a decrease in the internal resistance. Increasing the conductivity of electrolyte from 10 to 60 mS cm-1 through the addition of 0.4 M KCl into a two-compartment MFC was shown to reduce the ohmic resistance by 42% (from 1087 to 625 Ω) (Oh and Logan 2006). Zhao et al. (2006) demonstrated that increasing the concentration of phosphate buffer (pH 7.0), used as the catholyte, from 50 to 500 mM in a two-compartment MFC reduced the ohmic resistance by three-fold, resulting in a 53% increase in the power output. Such improvements in the performance were due to a decrease in catholyte resistance to proton transfer. The cathodic ohmic resistance can also be minimized by active control of the bulk catholyte pH. Jang et al. (Jang et al. 2004) reported that acidification of the catholyte improved the current output of MFCs, suggesting that the H+ availability in the cathode compartment was limiting MFC performance. The authors did not, however, elaborate on the actual pH values.

Extremes of ionic strength and pH of the catholyte can adversely affect cathodic performance by resulting in the inactivation of metal-based catalysts and biocatalysts.

The inactivation effect is bound to vary from one type of catalyst to another. The tolerance of catalysts to ionic strength and pH has not been addressed in the MFC literature.

2.4.2. Membrane

The PEM functions as a solid electrolyte, permits the proton flux from the anode to the cathode, and is not conductive to electrons. The membrane has an inherent

resistance to the transport of protons which contributes to the ohmic losses. The magnitude of this resistance can be mitigated by optimizing the physical properties and the type of membrane (O'Hayre et al. 2005).

The thickness and surface area of the membrane affect its resistance to proton conduction. The effect of membrane thickness on the performance of MFCs has not so far been reported. Thinner membranes are expected to give a lower ohmic overpotential as a result of lower transfer resistance and faster flux. Thin membranes, however, tend to have a higher rate of substrate crossover (Liu et al. 2006).

Increasing the membrane surface area reduces the ionic resistance associated with the membrane. Oh and Logan (2006) showed that an increase in the membrane surface area from 3.5 to 30.6 cm2, in a two-compartment MFC with fixed anode and cathode surface areas (22.5 cm2), decreased the internal resistance (from 1,110 to 89 Ω) and resulted in power output improvement from 45 to 190 mW m-2. The surface area of a membrane should be compatible with the extent of proton flux available from the anodic reaction. At a fixed PEM surface area (3.5 cm2), an increase in the surface area of the anode (from 2 to 22.5 cm2) resulted in a negligible change in the power output, suggesting that the proton transfer to the cathode was a limiting factor (Fig. 2.6).

However, similar increases in the surface area of the anode when the membrane surface area was 30.6 cm2 improved the power output by four-fold (Fig. 2.6) (Oh and Logan

2006). Increasing the ratio of membrane surface area to the total MFC volume has also been shown to enhance the proton flux (He et al. 2005; Rabaey et al. 2005b).

Figure 2.6. Power output as a function of anode surface area (2 to 22.5 cm2) in a two- compartment MFC with different PEM surface areas. Modified from Oh and Logan

(2006).

Because of its relatively high conductivity to cations, Nafion, a sulfonated tetrafluorethylene synthetic polymer (pore size < 5 nm), has been extensively used as a

PEM in MFCs. The transport of protons in the membrane is accompanied by transport of

+ water through the formation of hydronium ion (H3O ). The long-term stability of Nafion in MFCs is unknown. The stability of Nafion can be compromised due to degradation by chemical and biological oxidative substances in the anode and cathode compartment.

Operating conditions such as ambient temperature, neutral pH, and the presence of

positively charged ions other than protons may repress the functionality of Nafion

(Biffinger et al. 2007b; Chae et al. 2008).

Efforts have been made to replace Nafion with other types of membranes that can function effectively under the operating conditions of MFCs (Min et al. 2005; Biffinger et al. 2007b; Kim et al. 2007). Grzebyk and Poźniak (2005) synthesized an interpolymer membrane from polyethylene/poly(styrene-co-divinylbenzene) and used it for electricity generation in MFCs. Comparison of the performance of the synthesized membrane versus conventional membranes was not reported. While the operating conditions are conducive to the use of alternative membranes, they often have a relatively high proton transport resistivity as compared to Nafion. In an MFC with tubular membrane-cathode assembly design, the use of a hydrophilic polysulfone ultrafiltration membrane (50 kDa molecular weight cutoff) accounted for up to 64% of the ohmic resistance (Zuo et al. 2007). Kim et al. (2007) compared the performance of cation exchange membrane (CEM), anion exchange membrane (AEM), and ultrafiltration membranes with that of Nafion in two- compartment MFCs. They showed that the internal resistance of MFCs with a CEM or an

AEM was relatively similar to that with Nafion. However, the power outputs of MFCs were 6% lower with CEM, and 19% higher with AEM compared to that achieved by

Nafion. The authors also reported more than one order of magnitude increase in the MFC internal resistance when an ultrafiltration membrane (0.5 kDa) was compared to Nafion.

The MFCs with more porous ultrafiltration membranes (1 and 3 kDa) had 10% more internal resistance and produced 10% less power output than MFCs with Nafion.

Increased membrane porosity, however, enhances the crossover of oxygen and the substrate.

Biffinger et al. (2007b) demonstrated that nanoporous nylon and polycarbonate membranes (0.2 µm pore size) could replace Nafion in miniature MFCs (1.2 ml anode and cathode compartment volume). The application of these membranes resulted in power output and stability similar to that of Nafion. These membranes are, however, non- specific and permit considerable electrolyte and substrate crossover due to larger pore size as compared to Nafion. While this may not be a problem for miniature MFCs it could be detrimental to MFC performance at larger scales. Cellulose nitrate membrane was also tested in miniature MFCs and was found to be more susceptible to physical degradation, resulting in a lower performance as compared to Nafion (Biffinger et al.

2007b).

MFC membranes allow diffusion of other cations in addition to protons from the anode to the cathode (Rozendal et al. 2006; Chae et al. 2008). It has been reported that the diffused cations inhibit the transfer of protons through the Nafion membrane by occupying the sulfonate groups of Nafion (Rozendal et al. 2006; Chae et al. 2008).

Cations crossover also results in the formation of a pH gradient across the membrane, the cathodic side being more alkaline (Rozendal et al. 2006; Chae et al. 2008). This condition has been shown to decrease the cathodic performance (Zhao et al. 2006). One reason for such an effect is reduced activity of the cathode catalysts at increasing pH values. Solutes from chemical reactions and biological activity as well as microbial adherence can foul the membrane and hinder the transfer of protons (Chae et al. 2008). Exclusion of PEM from MFC designs has been tested as a way to address these problems and reduce the ohmic resistance (Jang et al. 2004; Liu and Logan 2004; Cheng et al. 2006b). The design requires air-cathode MFCs, which have a single chamber and the cathode is exposed to

air on one side (Liu et al. 2004). This design can be operated with or without a PEM. If present, the PEM is attached to the cathode side facing the anolyte.

Membrane-less MFCs (Fig. 2.7A) have relatively high maximum power densities but at the expense of a somewhat reduced overall coulombic efficiency (Liu and Logan

2004; Cheng et al. 2006b; Fan et al. 2007). The removal of membrane increases the flux of protons, reducing the internal resistance and improving the power output of membrane-less MFCs (Cheng et al. 2006b). The power output of an air-cathode MFC was improved by 88% when the PEM was omitted (Fig. 2.8A) (Liu and Logan 2004). In this case, the higher power output was also shown to be due to increased cathode potential (Fig. 2.8B). Biofilm formation on the cathode in membrane-less MFCs has been reported (Fig. 2.7B) ( Liu and Logan 2004; Liu et al. 2005a). Biofilm may become a diffusion barrier to the H+ transfer to the cathode and lead to biofouling problems over a long-term operation.

2.4.3. Electrode-spacing

The ohmic resistance in MFCs scales with the distance between the anode and cathode (Eq. 9). Thus, if the space between the two electrodes is reduced, the protons have less distance to travel, and the ohmic resistance is lowered. Modification of electrode orientation has been, therefore, investigated as an effective approach to improve the performance of MFCs.

Figure 2.7. A. Schematic of a single-chamber membrane-less microbial fuel cell. The cathode is exposed to air on one side and to the anolyte containing the substrate on the other side. Biofilm is formed on the anode and the interior-side of the cathode. B. SEM image of bacterial biofilm on the interior-surface of the cathode electrode in a membrane- less MFC (Liu and Logan 2004). Reprinted with permission from ACS.

Figure 2.8. A. Power-current and B. polarization properties of an air-cathode MFC in the presence (p) and absence (a) of a PEM. Modified from Liu and Logan (2004).

Using a membrane-less MFC, Liu et al. (2005a) demonstrated that decreasing the spacing between the electrodes from 4 to 2 cm reduced the ohmic resistance and resulted in a 67% increase in the power output (Fig. 2.9). The effect of electrode spacing on performance of MFCs has also been verified in other studies (Jang et al. 2004; Fan et al.

2007; Ghangrekar and Shinde 2007; Kim et al. 2007). However, if the electrodes are spaced too close to each other in membrane-less MFCs, oxygen diffusion from the cathode to the anode increases. This can become inhibitory to anaerobic respiration and promote aerobic respiration, both of which reduce the coulombic efficiency.

Figure 2.9. Effect of electrode spacing (2 and 4 cm) on power output of a membrane-less

MFC. Modified from Liu et al. (2005a).

Cheng et al. (2006b) used continuous advective flow through the anode toward the cathode and reported reduced oxygen diffusion and increased power output. Fan et al.

(2007) separated the anode and cathode using J-cloth in a membrane-less MFC with 1.7 cm electrode spacing. They reported about a two-fold increase in coulombic efficiency

(from 35 to 71%) as compared to the MFC without a J-cloth, which was attributed to reduced oxygen diffusion to the anode.

Relatively high power outputs have also been achieved in MFCs with a membrane-electrode assembly (MEA) design in which the electrodes are placed against either side of the PEM (Fig. 2.10A) (Pham et al. 2005; Ringeisen et al. 2006; Biffinger et al. 2007a). The design minimizes the electrode spacing and reduces ohmic resistance.

Liang et al. (2007) compared the internal resistance of two air-cathode MFCs, one with an MEA design and the other one with a 4-cm electrode spacing. The authors showed that the cathodic internal resistance of the MFCs with MEA design (94 Ω) was 68% less than that of the MFCs with 4-cm distance between the anode and cathode (291 Ω), resulting in a more than three-fold increase in maximum power output (1180 vs. 354 mW m-2) achieved by MEA-MFCs. Using another MFC design with a tubular cathode, Zuo et al.

(2007) showed that the ohmic resistance was reduced by about 22% when the electrode spacing decreased from 3 to 5 cm, improving the cathode potential. He et al. (2005) reported that inefficient proton transfer over a relatively large distance between the anode and cathode hindered the power output of an upflow MFC (Fig. 2.10B) by more than three-fold from the theoretically predicted value.

Optimization of the upflow MFC configuration by introducing an interior cathode

(Fig. 2.10C) reduced the ohmic resistance by 80% and improved the power output by one order of magnitude (He et al. 2006).

Figure 2.10. Schematic of A. a membrane-electrode assembly MFC, B. an upflow MFC, and C. an upflow MFC with an interior cathode.

2.5. Reducing cathodic mass transport losses

Mass transport in the cathode compartment depends on convection and diffusion.

Mass transport in bulk catholyte is dominated by convection (i.e., macroscopic flow). In contrast, mass transport at the cathode surface is typically dominated by diffusive transport. Maintaining high bulk concentrations and an even distribution of oxidant (e.g.,

O2) across the cathode compartment can reduce mass transport losses. In addition, optimization of MFC operating conditions, electrode material, and cathode compartment geometry can minimize mass transport limitations and performance losses.

2.5.1. Oxidant concentration

Power output has been shown to be proportional to the concentration of dissolved

O2 in the catholyte of two-compartment MFCs (Gil et al. 2003). However, increasing the dissolved O2 concentration is limited by its solubility in water. In general, stirring and flushing the catholyte with air or pure O2 and recirculation of the catholyte have been tested in attempts to enhance the oxygen flux to the cathode. Using an upflow membrane- less MFC, Jang et al. (2004) demonstrated that a four-fold increase in the rate of cathode aeration doubled the current output. Jong et al. (2006) studied the effect of the retention time of air-saturated catholyte on the performance of a thermophilic MFC with continuous flow in the anode and the cathode compartments. They reported that the current output doubled when the retention time was reduced from 2.7 to 0.7 min. Further decreases in retention time, however, did not change the current output, suggesting that oxygen availability was no longer limiting the performance.

In a unique instance involving a bench-scale sediment MFC (Fig. 2.11) with exogenous supply of sucrose, He et al. (2007) employed a rotating cathode to enhance the oxygen flux from the air to the underlying water column. The rotating cathode was 50% immersed in water and 50% exposed to air. The power output improved during cathode rotation because of increased cathodic potential, suggesting O2 limitation of the cathodic reaction. However, the anodic potential also increased, perhaps because of O2 diffusion to the anode.

Figure 2.11. Schematic of a sediment microbial fuel cell with a rotating cathode.

Exogenous supply of oxygen to the cathode requires energy input and also increases the potential of oxygen diffusion to the anode. This may result in competition for electrons between the anode and dissolved O2. The electron scavenging effect of O2 in the anolyte reduces the current output and lowers the coulombic efficiency (Liu and

Logan 2004; Pham et al. 2004; Biffinger et al. 2007b; He et al. 2007).

The effect of increasing the cathodic oxidant concentration on MFC performance has also been studied using reactants other than oxygen. For example, using an upflow two-compartment MFC, Tartakovsky and Guiot (2006) compared cathode oxygenation by air and hydrogen peroxide. Oxygenation with hydrogen peroxide resulted in a three- fold increase in the power output as compared to aeration. The power output increased with loading rate of H2O2. Hydrogen peroxide is a strong oxidant that decomposes upon contact with organic compounds (including microbial biomass) and metals.

2H2O2 → O2 + 2H2O (15)

Clauwaret et al. (2007a) used an MFC in which a microbial consortium containing denitrifiers in the anaerobic cathode compartment reduced nitrate as the final electron acceptor. An increase in the loading rate of nitrate increased the cathodic potential and improved the power output (Fig. 2.12). You et al. (2006) studied the effect of permanganate concentration, used as the cathodic electron acceptor, on performance of

MFCs. They reported a three-fold increase in the current density when the permanganate concentration was increased from 0.02 to 0.2 g L-1.

In contrast to dissolved O2, the aqueous solubility of oxidants such as hydrogen peroxide and permanganate does not limit cathodic mass transport. The influence of oxidant concentration on MFC performance depends on the current density of fuel cells.

At high current output, the rate of oxidant depletion is higher than that at low current densities, aggravating the cathodic mass transport losses. Thus MFCs become more responsive to an increased supply of oxidant with an increase in current density.

Figure 2.12. Effect of nitrate loading rate on the cathode potential and power density normalized by the net volume of cathode compartment of an MFC with an anaerobic denitrifying cathode compartment. Modified from Clauwaert et al. (2007a).

2.5.2. Cathode electrode design

Research on optimizing electrode design and material selection in the MFC field and the resulting effect on mass transport processes has been rather limited and mainly focused on the anode. At the current status of technology, MFCs operate at relatively low current densities (<5 mA cm-2) making activation and ohmic overpotentials more pronounced than the mass transport losses. As those limitations are ameliorated, mass transport losses become a noticeable issue.

The design of the electrode to minimize cathodic mass transport losses in MFCs has not been addressed in the literature. Several design criteria should be considered including the thickness, porosity, composition, geometry, and high specific surface area of the electrode. Ideally these characteristics should improve hydrodynamic flow to facilitate the mass transport and prevent accumulation of water at the cathode.

2.5.3. Cathode compartment design

The cathodic mass transport overpotential is also a compartment configuration issue. In two-compartment MFCs with aqueous catholyte, the mass transport is limited by the lack of hydrodynamic flow and low oxidant solubility in the case of oxygen.

Moreover, biofilm formation on the cathode surface limits the oxidant transfer to the cathode. Such limitation increases as the thickness of the biofilm develops with time.

Chemicals used to provide catholyte buffering capacity and ionic conductivity may also have adverse effects on the cathode performance by reducing the active electrode surface area and limiting the activity of catalysts (Pham et al. 2005).

Air-cathode MFCs have been tested in efforts to overcome some of the problems associated with two-compartment MFCs (Park and Zeikus 2003; Liu and Logan 2004;

Liu et al. 2004; Pham et al. 2005). In air-cathode MFCs, the mass transport loss is minimized due to direct oxygen supply from ambient air to the electrode. This alleviates the energy requirement for providing hydrodynamic flow in the catholyte. The design also makes the MFC structure relatively simple and compact, and eliminates the catholyte and the inherent problems associated with it.

Air-cathodes have been reported to form salt accumulation through the crossover of cations and anions through the membrane, possibly reducing the activity of cathode catalysts (Pham et al. 2005). This problem has not yet been addressed in the literature.

Air-cathodes are also prone to flooding. The accumulation of water is due to oxygen reduction at the cathode and crossover of water from the anode compartment. Cathode flooding slows down oxygen replenishment via diffusion, leading to mass transfer losses.

This problem is particularly pronounced in membrane-less MFCs. The effect of cathode flooding on the MFC performance has not been investigated in detail. The accumulation of water at the cathode can be alleviated with forced air flow over the cathode, which has yet to be tested, and by employing physical barriers that decrease the crossover water flow. Coating the air-side of the cathode with polytetrafluoroethylene, a hydrophobic compound, has been shown to improve the MFC performance, presumably, by decreasing the water flooding of the cathode (Cheng et al. 2006a). However, the thickness of hydrophobic coatings could hinder oxygen diffusion to the reaction site and adversely affects the performance at high current densities. With the increasing current densities

being achieved by MFC technology, optimization of the coating material and thickness is required to balance adequate oxygen flux with minimized cathode flooding.

2.6. Other losses

Substrate crossover and unwanted side reactions in the cathode compartment, collectively called parasitic losses, affects the MFC performance negatively. Substrate crossover through the membrane from the anode to the cathode has been reported in

MFCs (Liu and Logan 2004; Zuo et al. 2007; Biffinger et al. 2007b; Kim et al. 2007;

Chae et al. 2008). The membrane should ideally not allow the transport of reactants between the anode and cathode compartments. However, substrate crossover occurs commonly because of molecular diffusion and electro-osmosis (Jiang and Chu 2004).

Reactant crossover is particularly severe in membrane-less MFCs because there is no physical barrier to separate the contents of the anode compartment from that of the cathode compartment.

Substrate crossover influences the cathodic performance by lowering its potential below the thermodynamically predicted value. It also affects the coulombic efficiency because the substrate is utilized and/or transported away from the anode. Further, the substrate and its oxidation products may result in structural changes on the cathode surface and poisoning of the cathode catalyst.

The effects of reactant crossover on cathode overpotentials could depend on factors such as the material and thickness of the membrane, concentration of the reactants, electrode material and spacing, and the current output. Increasing the current density, reducing the substrate concentration, and improvements in membrane materials, cathode

catalysts, and design of electrodes are possible approaches to decrease substrate crossover and minimize its associate effects. Substrate crossover has not been fully characterized in

MFCs and remains a target for further study.

2.7. Concluding remarks

The unit current output is low in present-day MFC design, and a successful experimental approach to improve this limitation has yet to be reported. In this review, we have addressed the performance limiting factors associated with MFC cathodes and the recent progress in researching how to overcome these limitations. As summarized here, cathodic overpotentials are the main contributor to overall performance losses in

MFCs. By elucidating the underlying mechanisms of the cathodic reaction, strategies for optimization can be better formulated. An effective strategy must address all of the cathodic overpotentials and their interactions: the reduction reaction (leading to activation losses), charge transfer (leading to ohmic losses), and mass transport processes (leading to mass transport losses).

Activation losses appear to be the dominant limitation in MFCs. Many approaches have been attempted to limit the loss through using more effective mediators and oxidants, catalysts and biocatalysts, and by optimizing the cathodic conditions. Materials with catalytic activity comparable to precious metal catalysts but with less sensitivity to operating conditions are needed. Catalysts must be made with material of high durability that can perform effectively under characteristic conditions of MFCs. The optimization of catalyst processing, electrode coating techniques, binder composition, and the catalyst/binder ratio remain a challenging target for further improvement.

Cathode ohmic limitations depend on the electrolyte conductivity and electrode spacing. An improved membrane should be developed with high ionic conductivity, selective permeability, and less-susceptibility to biological and chemical fouling. This membrane should have sufficient mechanical and chemical integrity to prevent pores or cracks from developing during long-term operation and under varying cathodic conditions. Other possible ways to decrease ohmic losses include optimizing catholyte composition to reach increased ionic conductivity.

Mass transport limitations become more prominent with continued MFC advancement and thus must also be minimized to achieve the best performance. Oxidants should consistently be added and products removed. More research is needed to investigate active modes of oxygen transport to the cathode. For a point of reference, this has been extensively researched in chemical fuel cells. The cost for such active aeration must be compensated by increases in MFC performance.

The cost of the materials for MFC construction has been brought up in numerous articles but an overall economic analysis is premature at this stage because the designs and choice of materials continue to rapidly improve. Expensive (e.g., Pt and Au) test materials, although not feasible in large-scale, may give useful insight into reaction mechanisms.

The suitability of less expensive materials has not been unequivocally examined. The research to date has been mostly empirical and only two papers (Marcus et al. 2007;

Picioreanu et al. 2007) have been published on systematic optimization and modeling of

MFCs. Modeling based on the published information poses a problem due to differences in design and operational conditions of the MFCs reported in the literature. Although numerous applications for MFCs have been proposed, none has been scaled up to a

demonstration scale. MFCs still face many challenges but with consistent advances in research and development, especially with respect to the cathode, performance can continue to improve.

CHAPTER 3

ELECTRICITY GENERATION FROM CELLULOSE BY RUMEN MICROORGANISMS IN MICROBIAL FUEL CELLS2

3.1. Introduction

Cellulosic biomass, including solid waste products of agricultural and industrial activities, is one of the most abundant renewable sources of energy on earth. Cellulose is a linear polymer of glucose connected through ß-1,4-linkages, and it is typically arranged in structures of varying crystallinity. In nature, cellulose is almost always associated with hemicellulose and/or lignin, the latter of which hinders its enzymatic hydrolysis (Himmel et al. 2007) by blocking access by cellulases. Chemical and biological approaches have been developed for sustainable energy production from cellulosic materials by converting them to ethanol, H2 and methane, but these approaches encounter technical and economical hurdles (Howard et al. 2003; Palmqvist and Hahn-Hagerdal 2000; Bridgwater

2006). An alternative strategy is direct conversion of cellulose to electrical energy in microbial fuel cells (MFCs). MFCs are bioelectrochemical reactors in which

2 This chapter was published as H. Rismani-Yazdi, A.D. Christy, B.A. Dehority, M. Morrison, Z. Yu and O.H. Tuovinen. 2007. Electricity generation from cellulose by rumen microorganisms in microbial fuel cells. Biotechnol. Bioeng. 97:1398-1407. 55

microorganisms mediate the direct conversion of chemical energy stored in organic matter or bulk biomass into electrical energy (Rabaey and Verstraete 2005). Organic compounds are typically used as electron donors, while the anode serves as an electron acceptor for the microorganisms in MFCs. In previous studies the electron donors have been limited to readily metabolizable organic compounds, including simple carbohydrates, low molecular weight organic acids, starch and amino acids (Niessen et al.

2004; Rabaey et al. 2004; Liu et al. 2005; Logan et al. 2005). Most of these compounds are of high-value and have economically attractive alternative uses as chemical feed stocks in industry. Low-value organic matter and waste materials can also be used in

MFCs. Municipal wastewater and marine sediment organic matter have been successfully tested as electron donors in MFCs (Bond et al. 2002; Min and Logan 2004). Niessen et al.

(2004; 2005; 2006) developed a fermentative bioprocess to produce H2 from cellulose and other polymeric carbohydrates by using anaerobic microbial consortia as well as pure cultures of Clostridium spp. in fuel cells. Biogenic H2 was then oxidized by an electrocatalytically active anode to electrons and protons, thus creating a current in a closed circuit. This system did not, however, represent a microbially-mediated electricity generation because the anodic reaction did not involve biological or biochemical catalysis.

Direct biocatalytically-mediated electricity generation from cellulose as the electron donor in MFCs has not previously been reported.

Microorganisms that have been tested as biocatalysts for use in MFCs include pure cultures of obligately and facultatively anaerobic bacteria (Kim et al. 2002; Bond and Lovley 2003; Chaudhuri and Lovley 2003; Bond and Lovley 2005) and mixed cultures from sea floor sediments (Bond et al. 2002; Logan et al. 2005) and municipal

and industrial wastewater (Kim et al. 2004; Min and Logan 2004; Rabaey et al. 2004).

For direct biological conversion of cellulose to electricity in an MFC, the ideal microorganism(s) must be able to hydrolyze cellulose anaerobically and be electrochemically active, utilizing an anode as an alternative electron acceptor while oxidizing metabolites of cellulose hydrolysis. They should not require exogenous redox mediators for electron transfer to the electrode, because some of these electron shuttles are toxic and have to be replenished periodically.

The rumen microbiota contains both strict and facultative anaerobes, which effectively hydrolyze cellulose (Krause et al. 2003), and conserve energy via anaerobic respiration or fermentation (Hobson and Stewart 1997). The objectives of this study were to (i) test the possibility of generating electricity in an MFC with rumen microorganisms as biocatalysts and cellulose as the electron donor, and (ii) analyze the composition of the bacterial communities enriched in the MFCs. This study demonstrates that rumen microbial consortia can be used as biocatalysts to generate electricity from cellulose in

MFCs. Phylogenetic diversity analysis of the enriched consortia in the MFCs showed the presence of hydrolytic and respiratory anaerobes that can couple cellulose hydrolysis to the reduction of the anode. To date, this is the first report of exploiting ruminal microbial communities for direct conversion of cellulose to electrical energy in an MFC.

3.2. Materials and methods

3.2.1. Microorganisms

Standard anaerobic techniques were used throughout the study. Ruminal contents were collected via a cannula from the rumen of a fistulated Holstein cow maintained at the research farm of the Ohio State University in Columbus, Ohio. The cow was fed a ration once daily that contained alfalfa silage, corn grain, soybean meal, and a vitamin and mineral mixture. Rumen microorganisms for the inoculum (10% v/v) were obtained by squeezing the collected rumen contents through four layers of cheesecloth followed by immediate transfer in head space-free screw-cap bottles to the laboratory.

3.2.2. Medium

The medium (anolyte) used to grow the rumen bacteria in the anode compartment contained (per liter): 450 mg K2HPO4, 450 mg KH2PO4, 900 mg NaCl, 900 mg

(NH4)2SO4, 120 mg CaCl2·2H2O, and 90 mg MgSO4. Cysteine-HCl (500 mg/l) was added initially as an oxygen scavenger (Dehority 2003). The medium was supplemented with

40 % (v/v) autoclaved clarified rumen fluid (centrifuged for 5 min at 5,000 g at 4ºC) to simulate the rumen habitat and provide essential growth factors for the microorganisms.

Microcrystalline cellulose (7.5 g/l) (Sigmacell 20, Sigma Chemical Co., St. Louis, MO) was ball-milled for 24 h and used as a substrate and electron donor throughout the experiments unless otherwise stated. The medium was gassed with CO2 for 1 h to remove the oxygen. The medium was adjusted to pH 6.8 with NaOH, and Na2CO3 (4 g/l) was added prior to autoclaving in a sealed round bottom flask.

To determine the cellulolytic activity, microbial biomass suspended in the anode compartment and attached to the electrode was inoculated into the cellulose medium in stoppered culture tubes. The cultures were incubated anaerobically at 39±1ºC. Cellulose depletion was determined visually by comparing the inoculated tubes to an uninoculated control.

3.2.3. Microbial fuel cells

Replicate MFCs were constructed of two glass tubes (70 mm ID and 240 mm height) joined and clamped together across a tubular glass bridge (60 mm length, 30 mm

ID) (Fig. 3.1). An Ultrex proton-exchange membrane (CMI-7000, Membrane

International, Glen Rock, NJ) separated the anode and cathode compartments. The working volume in each chamber was 400 ml. Graphite plates (McMaster-Carr,

Cleveland, OH), 71×49×6 mm (~84 cm2) in dimensions, were pretreated as described by

Bond and Lovley (2003) and used as electrodes in both chambers. The two electrodes were connected via a wire in holes drilled directly into the graphite electrodes and the connections sealed using silver epoxy covered by nonconductive epoxy (McMaster-Carr,

Cleveland, OH).

Prior to inoculation, the anode and cathode chambers were autoclaved. The anode chamber was flushed vigorously with CO2 to remove the oxygen and then filled with 360 ml of the medium and 40 ml of the inoculum. Control experiments were performed in parallel but without inoculation or the addition of cellulose. The contents in the anode chambers were stirred during the start-up period using a magnetic stirrer until steady state current production was reached. Aerobic potassium ferricyanide solution (50 mM

K3Fe(CN)6) in phosphate buffer (100 mM K2HPO4, pH 7) was used as the catholyte to enhance oxygen reduction in the cathode compartment (Park and Zeikus 2003). All fuel cells were operated at 39±1ºC. During the MFC operation, the pH of each anode compartments was monitored using a pH electrode (Orion International, Germany) and adjusted to pH 6.8 with NaOH.

Figure 3.1. The experimental two-compartment MFC.

The electrical output of each fuel cell was monitored by measuring the potential difference (voltage) between the anode and the cathode across a known resistance (R)

(1000 ohms) using a data acquisition unit (DataQ instrument, Akron, OH) connected to a computer. The power density (W/m2) was calculated according to the equation, P =

I×V/A, where V is the voltage (V), I (I = V/R) the current (amps), and A the surface area of the electrode (m2). Polarization characteristics were determined by varying the resistance between the electrodes stepwise from 1 mega ohm to 9.9 ohms with a pause at each resistance to allow the voltage to reach a stable value. The internal resistance (Ri) of the fuel cells was estimated according to the equation, Ri = (Vo – Vr)/I, where Vo is the open circuit potential and Vr the potential across the external resistance.

3.2.4. DNA extraction and quantification

The anodes were removed from the MFCs and rinsed with sterile distilled water to remove debris and loosely attached bacteria. Attached biofilm was then scraped off each anode using a sterile microscope slide and washed into sterile Tris-EDTA buffer

([pH 8.0], 500 mM NaCl, 50 mM Tris-HCl, 50 mM EDTA). The suspended microorganisms in the anode compartment were also sampled from each MFC. The bacterial biomass of the anode-attached and the suspended bacteria were pelleted by centrifugation at 12,000 g for 15 min at 4°C. The pellets were stored at -80ºC to preserve the bacterial biomass until DNA extraction. Genomic DNA was extracted from the pellets using repeated bead beating in the presence of high concentrations of sodium dodecyl sulfate, salt, and EDTA, and with subsequent DNA purification by QIAamp columns (Qiagen, Valencia, CA), as described by Yu and Morrison (2004b). The DNA

recovered from each sample was quantified fluorospectrometrically using a Quant-iT broad range DNA assay kit Q33130 (Molecular Probes, Eugene, OR) and then diluted to

10 ng/μl with Tris-EDTA buffer.

3.2.5. DGGE

Differences in the anode attached and suspended microbial communities enriched in the MFCs were analyzed using denaturing gradient gel electrophoresis (DGGE). The substrates used in these experiments were (1) cellulose (7.5 g/l), (2) a mixture of soluble carbohydrates (glucose, xylose, cellobiose, and maltose; 1 g/l each), and (3) autoclaved clarified rumen fluid (40% v/v). Two inocula were examined; (1) fresh rumen microorganisms and (2) microorganisms enriched over two months in a cellulose-fed

MFC followed by subculturing twice in cellulose medium. All MFCs were inoculated once at the start of the experiments with either of the two inocula and operated for up to two months with intermittent replenishment of the substrates. The 16S rRNA genes were

PCR-amplified with universal primer sets GC EUB357f and EUB519r, and DGGE analysis of PCR products was performed as reported by Yu and Morrison (2004a). The initial rumen sample used to inoculate the MFCs was not available for DGGE and phylogenetic analysis.

3.2.6. Cloning, Sequencing and DNA Sequence Analysis

The nearly complete 16S rRNA genes from electrode-associated and suspended microorganisms, enriched for about two months in the anode compartment of the cellulose-fed MFC inoculated with fresh rumen microorganisms, were amplified with the

27f and 1525r universal primer sets (Lane 1991). The cloning and sequencing of the PCR precuts were performed as described previously, except for pre-confirmation of insert by

EcoRI digestion and omission of RFLP analysis (Larue et al. 2005). Base calling was manually assessed using BioEdit (Hall 1999). All the DNA sequences were then grouped into phylotypes by dereplication based on 98% identity using FastGroup (Yu et al. 2006).

The phylotypes were finally compared to sequences in GenBank using BLASTn. Only the sequence from a known bacterium was recorded to infer the diversity. These sequences were also classified using the Classifier program implemented in RDP II.

3.2.7. Nucleic acids accession numbers

The unique sequences determined in this study are listed under the GenBank accession numbers EF016572-EF016618 and EF016619–EF016642 for anode-attached and suspended microorganisms, respectively, and included in Tables 3.1 and 3.2.

Phylotype Prevalence Closest BLAST Identity RDP classifier Confidence Phylum2 (accession % match % assignment1 (%) #) (accession #)

AA-78 1.1 Clostridium-like 99 Firmicutes 61 (EF016608) species clone 16SX-1 AA-12 8.9 (U27711) 94 Clostridiaceae 72 (EF015681) AA-47 1.1 93 Clostridiales 75 Firmicutes (EF016596 AA-75 1.1 Unidentified 91 Clostridiaceae 94 (EF016606) Clostridiaceae pDH- A (U85415)

AA-24 1.1 Uncultured 91 Butyrivibrio 63 (EF016586) Clostridium sp. (AY330126)

AA-103 1.1 Clostridium sp. EBR- 94 Clostridiaceae 62 (EF016576 02E-0046 AA-65 1.1 (AB186360) 93 Clostridiales 87 (EF016601) AA-100 6.7 Clostridium-like 97 Clostridiaceae 70 (EF016574) species (AF433166)

AA-94 1.1 Clostridium 95 Acetivibrio 98 (EF016617) straminisolvens (AB125279)

Continued

Table 3.1. Bacterial composition as inferred from the cloned 16S rRNA gene fragments recovered from anode-associated bacteria (AB) in the cellulose-fed MFC. 1 In cases where the confidence level was not 100% at the genus level, a taxonomic rank was chosen with equal or greater than 60% confidence value. 2 The phylotypes were assigned to phyla based on RDP II taxonomy classifications.

Table 3.1. continued

AA-22 1.1 Clostridium 91 Acetivibrio 63 (EF016585) thermocellum DSM 1237 (L09173)

AA-79 1.1 Clostridium 90 Acetivibrio 92 (EF016609) papyrosolvens (X71852)

AA-33 1.1 Acetivibrio 91 Acetivibrio 98 (EF016589) cellulolyticus ATCC 33288 (L35516)

AA-27 1.1 Sedimentibacter sp. 93 Sedimentibacter 99 (EF016587) B4 (AY673993)

AA-11 5.6 Sedimentibacter 99 Sedimentibacter 73 (EF016580) hongkongensis KI AA-55 2.2 (AY571338) 98 Clostridiales 100 (EF016599) AA-67 1.1 97 Bacteria 100 (EF016602) AA-107 1.1 94 Soehngenia 73 (EF016578) AA-108 1.1 Sedimentibacter 99 Clostridiales 100 (EF016579) hongkongensis (AF433166)

AA-77 1.1 Desulfotomaculum 93 Clostridiales 83 (EF016607) sp. DSM 7475 AA-41 2.2 (Y11580) 89 Clostridiales 79 (EF016593) AA-92 1.1 88 Clostridiales 77 (EF16616) 1.1 Desulfitobacterium 89 Desulfitobacterium 100 hafniense Y51 (AP008230)

AA-98 1.1 Desulfosporosinus 91 Clostridiales 67 (EF016618) sp. STP12 (AJ582757)

AA-48 1.1 Ruminococcus sp. 89 Clostridiaceae 95 (EF016597) 16442 (AB075676)

Continued

Table 3.1. continued

AA-46 2.2 Ruminococcus 86 Clostridiaceae 73 (EF016595) flavefaciens strain JM1 (AY349157)

AA-50 1.1 Alicyclobacillus 82 Clostridia 65 (EF016598) acidoterrestris (AB042058)

AA-83 1.1 Rumen bacterium R- 85 Clostridia 75 (EF016612) 7 (AB239481)

AA-10 6.7 Geovibrio agilis 95 Geovibrio 100 (EF016573 (AJ299402) AA-34 2.2 88 Geovibrio 94 (EF016590) AA-104 16.7 Geovibrio 98 Geovibrio 100 (EF016577) ferrireductans

AA-91 1.1 (X95744) 93 Geovibrio 98 Deferribacteres (EF016615) AA-16 1.1 Clostridium 85 Geovibrio 66 (EF016583) straminisolvens (AB125279)

AA-84 1.1 Desulfovibrio 98 Desulfovibrio 71 (EF016613) desulfuricans F28-1 (DQ092636)

AA-1 2.2 Comamonas sp. 99 Comamonas 100 (EF016572) 23310 (AJ251577)

AA-18 1.1 Pseudoxanthomonas 99 Pseudoxanthomonas 78 (EF016584) mexicana (AF273082)

AA-90 1.1 Pseudomonas 91 Pseudoxanthomonas 87 (EF016614) boreopolis Proteobacteria (AB246809)

Continued

Table 3.1. continued

AA-7 1.1 Uncultured 91 Pseudomonadales 60 (EF016603) Pseudomonas sp. (AM259350)

Bacteroides Bacteroides

AA-80 1.1 Rhizobium sp. JEYF 98 Bacteria 91 (EF016611) (AB069650)

AA-38 3.3 Treponema bryantii 90 Treponema 98 (EF016591) (M57737) Spirochaetes

Phylotype Prevalence Closest BLAST Identity RDP classifier Confidence Phylum2 (accession #) % match % assignment1 % (accession #)

SB-36 1.1 Bacillales bacterium 83 Firmicutes 88 (EF016630) Gsoil 1105 (AB245375)

SB-78 1.1 Bacillus sp. TP50 91 Firmicutes 100 (EF016637) (AB066336)

SB-21 1.1 Butyrivibrio 83 Clostridiales 72 (EF016625) fibrisolvens OB156 (U41168)

SB-2 1.1 Clostridium-like 97 Clostridiaceae 71 (EF0166 27) species (U27711) SB-45 (EF016633) 1.1 93 Clostridiaceae 77 SB-34 (EF016629) 1.1 89 Clostridia 96 SB-14 1.1 Desulfosporosinus 90 Bacteria 100 (EF016623) orientis, DSMZ 7493

(AJ493052) Firmicutes

SB-84 1.1 Desulfotomaculum 82 Bacteria 100 (EF016639) putei SMCCW464 (AF053934)

SB-19 1.1 Bacterium L4M2 4- 99 Dethiosulfovibrio 70 (EF016624) 15 (AY862597)

SB-9 1.1 Moorella glycerini 89 Clostridiales 66 (EF016642) (U82327) SB-11 1.1 85 Sporanaerobacter 67 (EF016622) SB-91 1.1 Moorella 83 Sporanaerobacter 68 (EF016640) thermoacetica strain AMP (AY884087)

Continued

Table 3.2. Bacterial composition as inferred from the cloned 16S rRNA gene fragments recovered from suspended bacteria (SB) in the anode chamber in the cellulose-fed MFC. 1 In cases where the confidence level was not 100% at the genus level, a taxonomic rank was chosen with equal or greater than 60% confidence value. 2 The phylotypes were assigned to phyla based on RDP II taxonomy classifications.

Table 3.2. continued

SB-100 37.5 Comamonas sp. 99 Comamonas 100 (EF016619) 23310 (AJ251577) SB-10 1.1 98 Comamonas 100 (EF016621) SB-101 1.1 96 Comamonas 100 (EF016620) SB-28 2.3 95 Comamonas 100 (EF016626) SB-30 1.1 88 Comamonadaceae 69 (EF016628) SB-71 31 Comamonas sp. R- 96 Comamonas 99 Betaproteobacteria (EF016636) 25060 (AM084020) SB-82 1.1 91 Comamonas 90 (EF016638) SB-37 2.3 Pseudomonas 96 Stenotrophomonas 71 (EF016631) boreopolis (AB246809)

SB-68 1.1 Pseudoxanthomonas 98 Pseudoxanthomonas 82 (EF016635) mexicana AMX 26B (AF273082)

SB-93 2.3 Pseudomonas sp. 91 Burkholderiales 69 Gammaproteobacteria (EF016641) An18 (AJ551156)

SB-40 3.5 Unidentified rumen 99 Rikenellaceae 66 (EF016632) bacterium RF28 (AF001760)

SB-65 1.1 Unidentified rumen 95 Tannerella 82

(EF016634) bacterium RF36 Bacteroides (AF001767)

3.3. Results

3.3.1. Current production

The potential of rumen microorganisms as biocatalysts to produce current from cellulose was examined in replicate two-chamber MFCs. Electron flow commenced immediately upon inoculation of the cellulose-containing medium in the anodic chamber and initially reached 0.46 mA (Fig. 3.2.A). The current declined gradually as cellulose was depleted, but increased again when cellulose suspension was added to the anode chamber (at 125 h). This start-up phenomenon was observed in all the MFCs inoculated with rumen microorganisms.

The current production reached a steady state level of 0.47 ± 0.002 mA at 330 h, and this level was sustained with cellulose as the sole substrate in a fed-batch mode (Fig.

3.2.A). During the steady state period (330-430 h and 630-945 h), a stable power density of 26.7 ± 0.05 mW/m2 (475 ± 2 mV) was obtained. To test if electricity generation was dependent on cellulose as substrate, a control experiment was conducted in parallel in which cellulose was omitted from the medium. Although the current initiated immediately after the inoculation and decreased within the first 100 h in the same pattern as the cellulose-fed MFCs, it further dropped to 0.03 mA and never increased again (Fig.

3.2.B, [d]). This initial, unsustainable current was probably due to the reducing equivalents present in the rumen fluid (inoculum) added to the control MFC. These results indicate that cellulose was the essential substrate for the observed long-term electricity generation, and its hydrolytic metabolites supplied a sustainable source of electron donors for the anode-reducing bacteria.

Figure 3.2. Electricity generation in two-compartment MFCs at 39±1°C. Voltage was monitored across 1000 ohm resistance between two 84 cm2 graphite electrodes. A. With rumen microorganisms and cellulose. Manipulations: circuit opened (a), circuit closed (b), current fluctuation during polarization tests (c). B. Controls with rumen microorganisms and without cellulose (d), without rumen microorganisms and with cellulose (e), without rumen microorganisms and with cellulose but without cysteine in the medium (f).

Although metabolite analysis was not within the scope of this study, the control experiment along with the long-term current production from cellulose clearly indicated that the carry-over metabolites in the inoculum and the medium would be insufficient for sustaining electricity generation for a long time course.

There was negligible current production in the absence of the rumen microorganisms with or without using cysteine in the medium (Fig. 3.2.B, [e and f]).

Thus the chemical components of the medium were not electrochemically active in an abiotic system. Inoculation of the cellulose medium with the electrode-attached biofilm and the freely suspended microorganisms in separate culture tubes revealed that both populations possessed cellulolytic activity. Visual inspection clearly showed cellulose depletion after 48 h of incubation at 39±1ºC.

To assess the effect of long-term open circuit conditions on current generation by the enriched microorganisms in the MFCs, the circuit was kept open for about 5 days (Fig.

3.2.A). The current production resumed and returned to the original state (0.47 mA) upon re-connection of the circuit and continued to be steady except for two incidents of current fluctuation when the polarization properties of the MFCs were tested. This indicated that microbial activity was not compromised by system disturbances caused by open circuit conditions in which there was no current flow.

3.3.2. Polarization characteristics

Polarization properties of the MFCs are shown in Figure 3.3. The polarization curve reflecting voltage-current and power-current characteristics was consistent with the expected properties of a typical fuel cell (Srinivasan, 2006). Maximum power reached

0.46 mW at a current of 1.5 mA with 211 ohm resistance. At this resistance, the

2 2 maximum power and current density were 55 mW/m and 178 mA/m , respectively. The internal resistance (Ri) of the MFC was estimated to be 206 ohms, nearly equal to the external resistance of 211 ohms, at the maximum power. Such a relatively large internal resistance was likely due to the nature of the employed H-shape fuel cells. Modifications to the system configuration and operational conditions are required to reduce internal resistance and thus improve the power output.

Figure 3.3. Polarization properties of an MFC with rumen microorganisms as biocatalysts and cellulose as the electron donor at 39±1°C.

3.3.3. Substrate comparison

To test the ability of rumen microorganisms to generate electricity from simple and soluble substrates, a mixture of soluble carbohydrates, and autoclaved clarified rumen fluid (40% v/v) were used as electron donors in separate experiments. Current was produced from both sources (Fig. 3.4.A), and sustainable power densities as high as 26.2 mW/m2 and 19.5 mW/m2 were achieved by intermittently replenishing the carbohydrate mixture and clarified rumen fluid, respectively.

To reduce the duration of the start-up period, pre-enriched microorganisms were used as an inoculum (10% v/v) in a fresh MFC with cellulose as substrate. The current production started immediately after inoculation and increased to 0.49 mA at 30 h, followed by a drop to 0.40 mA at 45 h (Fig. 3.4.B). The current was sustainable at that level for about 11 days. With this pre-enriched inoculum, the start-up period was reduced to 45 h, as compared with an average of 300 h with fresh rumen fluid as the inoculum

(Fig. 3.2.A and 3.4.A). This indicates that using pre-enriched inoculum was a successful method to reduce the lag time between inoculation of MFCs and reaching a steady current production. The immediate flow of electrons and subsequent increase in current in this experiment was likely due to the presence of microbially produced soluble mediators in the pre-enriched inoculum and the presence of readily oxidizable metabolites in the medium.

Figure 3.4. Electricity generation in MFCs at 39±1°C. A. Rumen microorganisms with a mixture of soluble carbohydrates and autoclaved rumen fluid as substrates. B. Pre- enriched bacteria with cellulose as the substrate.

3.3.4. DGGE analysis of microbial community

Change in the diversity of microbial communities in response to the types of inocula and substrates used in MFCs operated for two months were analyzed with DGGE.

The DGGE profiles of anode-colonizing and suspended microorganisms sampled from the MFCs are summarized in Figure 3.5. The banding patterns of the anode-attached community were different from the suspended bacteria in all MFCs regardless of the type of substrate used as the electron donor, indicating different microbial compositions between the attached and the suspended microbial consortia. The DGGE profiles for the

MFCs operated with different substrates and inocula were also different in both the attached and the suspended consortia, demonstrating that microbial communities changed in response to the types of substrates as electron donors and the types of inocula used in each MFC.

3.3.5. Phylogenetic diversity revealed by cloning and sequencing

Two 16S rRNA gene clone libraries of the MFC fed with cellulose were constructed in this study. They were designated as AB for anode-attached bacteria and

BS for suspended bacteria in the anode chamber. Ninety-five clones were randomly selected for sequencing from each library, and 90 and 87 clones were successfully sequenced for AB and SB, respectively. There was a large difference in terms of phylogenetic diversity between the two libraries. The majority of the AB library sequences belonged to two phyla, the Firmicutes and the Deferribacteres, with 58.9% and 26.7% of the total clones sequenced, respectively (Table 2).

Figure 3.5. DGGE profile of the community DNA extracted from anode-attached and bacteria suspended in the anode chamber of four MFCs. Lane 1 = suspended bacteria, cellulose-MFC; 2 = anode-attached bacteria, cellulose-MFC; 3 = suspended bacteria, soluble carbohydrate mixture-MFC; 4 = anode-attached bacteria, soluble carbohydrate mixture-MFC; 5 = suspended bacteria, rumen fluid-MFC; 6 = anode-attached bacteria, rumen fluid-MFC; 7 = suspended bacteria, pre-enriched inoculum in cellulose-MFC; 8 = anode-attached bacteria, pre-enriched inoculum in cellulose-MFC.

The remaining 16S rRNA gene sequences belonged to a range of other phyla, including Proteobacteria (7.8%) (Alpha 1.1%, Beta 2.2%, Delta 1.1% and, Gamma

3.3%), Spirochaetes (5.6%), and Bacteroides (1.1%). Betaproteobacteria (75.9%) were the predominant cloned sequences in the SB library (Table 3). Representatives of

Firmicutes (12.6%), Gammaproteobacteria (5.7%), and Bacteroides (4.6%) were also found in the suspended community.

The majority of the cloned 16S rRNA gene sequences of anode-attached and suspended consortia showed <97% homology with previously characterized and identified bacteria in the database. Of the 90 clones derived from anode-attached consortia, 30 (33.5%) showed more than 97% sequence similarity with known isolates,

47 (52%) clones had 90–97% similarity with known sequences, and for the remaining 13

(14.5%) clones, the similarity was less than 90%. Of the 87 clones sequenced from the library of suspended bacteria, 39 (44.8%) clones possessed more than 97% sequence identity with known bacteria, 40 (46%) clones showed 90-97% similarity and 8 (9.2%) clones fell under 90% similarity with existing GenBank sequences. The relatively low similarity of some of the clones to any sequences in the database suggests that some of the sequences from the MFC may represent novel microorganisms.

3.4. Discussion

This study demonstrates the generation of electricity in MFCs via the microbial degradation of cellulose coupled with anaerobic microbial oxidation of hydrolysis products with graphite plate as the terminal electron acceptor. As a reference point,

Niessen et al. (2004, 2005, 2006) reported on a fuel cell system that was tested with two

Clostridium spp. as well as undefined soil microbial populations that could produce H2 from cellulose and other polymeric carbohydrates. The biogenic H2 thus produced was then oxidized electrochemically by the catalytic action of a platinum- poly(tetrafluoroaniline)-coated anode. The majority of metabolites from cellulose biodegradation would not be coupled with electricity generation because the electrochemical anode catalyst was specific for H2. In the present study, selection of a biocatalyst for electron transfer at the anode concurrent with cellulose hydrolysis was the driving force to test rumen microorganisms in the MFCs. Ruminal microorganisms are well known for their extensive cellulolytic activity under anaerobic conditions, but their electrochemical activity has not been reported thus far. The results demonstrate unequivocally that microorganisms derived from cow rumen degrade cellulose under the experimental MFC conditions and reduce the anode as the electron acceptor.

Microorganisms from the inoculated MFCs could be maintained in a medium with cellulose as the sole carbon and energy source.

Power density as high as 55 mW/m2 was achieved using cellulose as the substrate and rumen microorganisms as the biocatalysts. This power output, although lower than that achieved in the previous MFC studies with modified electrodes, specially designed compartments, or well-adapted cultures (Park and Zeikus, 2003; Gil et al. 2003; Rabaey et al. 2003; Min and Logan, 2004; Pham et al. 2004), is comparable with the power generated from soluble and low-molecular weight organic compounds in two-chamber

MFCs under similar operating conditions (Bond and Lovley 2003; Lee et al. 2003; Phung et al. 2004; Logan et al. 2005). These results suggest that cellulose hydrolysis may not be the rate-limiting step for current generation in cellulose-fed MFCs. Rather, the

performance is likely restricted by the large internal resistance of the employed H-shape fuel cells imposed by the large distance between the anode and cathode, and the small surface area of the cation exchange membrane (Logan et al. 2006).

The MFCs demonstrated here exhibited constant power generation, without the need for exogenous electron transfer mediators, for at least 60 days with periodic cellulose replenishment. The system was insensitive to disturbances such as open circuit conditions. As expected, current was also produced when rumen organisms were supplied with soluble carbohydrates and rumen fluid.

The DGGE results showed that the outcome of the enrichment was dependent on the sources of the inocula and the types of substrates used in the MFCs. The data also showed that the attached and suspended microbial communities enriched in MFCs had different compositions. Rabaey et al. (2004) reported on the enrichment of electrochemically active suspended and attached bacteria in the anode compartment, consistent with the proposed anodic electron transfer mechanisms in MFCs. Suspended microorganisms produce soluble electron shuttles that transfer electrons from electron transport chain to the anode. Bacteria attached to the anode typically form a biofilm over time and transfer electrons using soluble shuttles or electron transport components associated with the membrane (Lovley 2006).

However, the extent of the flux of microorganisms governed by sorption and desorption in MFCs is unclear, and there may be a continuum of attachment and adherence to solid surfaces even for microorganisms unable to respire with the anode as their electron acceptor.

The phylogenetic diversity of the bacterial communities enriched under various conditions in MFCs inoculated with mixed consortia has been investigated in previous studies. Members of the family Geobacteraceae have been reported to couple the complete oxidation of acetate to reduction of the electrode in MFCs (Bond et al. 2002).

These organisms are members of the Deltaproteobacteria and have been identified as the predominant (> 70% of the 16S rRNA genes in clone libraries) electrode-colonizing microorganism in marine-sediment fuel cells (Bond et al. 2002). Gammaproteobacteria were the abundant cloned sequences (97%) in the anode-associated community of MFCs inoculated with anaerobic marine sediments and with cysteine as the sole substrate

(Logan et al. 2005). Shewanella spp. and Pseudomonas spp. of this subclass of

Proteobacteria are known to be able to use an electrode as their final electron acceptor in

MFCs (Kim et al. 2002; Rabaey et al. 2004). Phylogenetic analysis of electrode- associated microbes enriched with municipal wastewater in an MFC inoculated with activated sludge revealed a diverse population, comprised of mainly Alpha- (24%) and

Beta- (36%) Proteobacteria (Kim et al. 2004). In many of the previous studies, either the biofilm or a composite sample of suspended bacteria and the biofilm was analyzed.

In this study, to provide greater insight into the microbial ecology and diversity, bacterial 16S rRNA clone libraries were examined separately for the biofilm and suspended communities. The anode-associated clone library was dominated by

Firmicutes, which are also predominant in the rumen. The most abundant sequences in this phylum were related to the genera Clostridium and Sedimentibacter, with sequences similar to Desulfotomaculum and Ruminococcus comprising the next most frequently detected sequences. Clostridium and Ruminococcus comprise cellulolytic bacteria which

hydrolyze cellulose primarily via the complex cellulase system known as the cellulosome

(Krause et al. 2003). Previous studies have also reported the presence of Firmicutes in the anode-colonizing community of MFCs fed with artificial waste water containing glucose and glutamate (27%) (Chang et al. 2006), and in acetate-enriched MFCs inoculated with marine sediments (>20%) (Bond and Lovley 2003), and activated sludge (>6%) (Lee et al.

2003). The electrochemical activity of Gram-positive bacteria in biofuel cells has also been reported (Park et al. 2001). A strain of Clostridium butyricum was isolated from the anode biofilm of an MFC enriched with starch-processing wastewater. The isolate was electrochemically active and capable of transferring electrons to an electrode which was believed to be due to cell surface-localized cytochromes (Park et al. 2001). It is believed that some fermentative bacteria can direct the electron and carbon fluxes to the production of less reduced fermentation products by reducing electron acceptor compounds such as nitrate or Fe(III). This is corroborated by the observation that enriched consortia and isolates from the anode of an MFC produced more reduced metabolites when placed in an anaerobic vessel without an electrode (Rabaey et al. 2004).

Enrichment of fermentative bacteria on the anode surface of the cellulose-fed MFC reported here suggests that these bacteria could utilize the anode electrode as an electron sink in their metabolism, but it remains to be determined experimentally.

Close relatives of the genus Geovibrio, members of phylum Deferribacteres, were also found in the anode attached biofilm. Six clones (6.7%) were found >95% similar to

Geovibrio agilis, while another 15 (16.7%) clones were 98% similar to Geovibrio ferrireductans (Table 2), a Gram negative and strictly anaerobic bacterium known to couple either H2 or acetate oxidation to Fe(III) reduction (Caccavo et al. 1996).

G. ferrireductants can also use propionate, a key intermediate in anaerobic degradation of cellulose by rumen microorganisms. Colonization of the anode by metal-reducing bacteria Shewanella spp. and Geobacter spp. has been previously reported (Bond et al.

2002; Kim et al. 2002). However, members of the genus Geovibrio have not been identified as numerically important members of an anodic community prior to this study.

More studies are needed to assess the importance of Geovibrio spp. in cellulose-fed

MFCs.

The most predominant clones recovered from the suspended bacteria in the anode chamber appeared to be phylogenetically related (88-99% similarity) to the family

Comamonadaceae of the Betaproteobacteria. Most of the sequences within this group had high identity to cultured isolates and bacteria described as Comamonas spp., a group of Gram negative facultative anaerobes. Of particular interest, 33 clones (37.5%) were closely (99% identity) related to Comamonas sp. strain 23310 that can oxidize short chain fatty acids (SCFA) and amino acids with concomitant dissimilatory reduction of nitrate

(Etchebehere et al. 2001). The predominance of these bacteria in our MFCs suggests that these organisms could utilize the electrode, instead of nitrate, as an alternative electron acceptor while oxidizing the SCFAs, the main end product of cellulose degradation.

Previous research has shown that Rhodoferax ferrireducens, a member of the family

Comamonadaceae which is capable of using organic acids and glucose as substrates, could successfully transfer electrons to the anode in a glucose-based mediator-less MFC

(Chaudhuri et al. 2003). Sequences related to the Clostridium of phylum Firmicutes were also detected in the clone library of suspended bacteria. These organisms may be involved in cellulose hydrolysis and fermentation of the corresponding products,

consistent with their role in the rumen. Phylogenetic diversity analysis shows enrichment of consortia consisting of hydrolytic and respiratory anaerobes, in MFCs here. These data are consistent with the function of these microorganisms in coupling cellulose hydrolysis via anaerobic respiration to electron flow at the anode. In general, the microbial diversity found in this study was greater than that found in other MFC studies. This distinction is probably attributed to a high microbial diversity in the rumen microbiota (the original source of the inoculum), and the wide range of non-selective intermediates and metabolites derived from cellulose degradation.

3.5. Conclusions

This study demonstrates that rumen microorganisms are capable of hydrolyzing cellulose with concomitant transfer of electrons to an MFC electrode. This system has potential for generating electricity from a wide range of agricultural and industrial cellulosic wastes. The study also adds to the diversity of microorganisms that have been shown to produce electricity in MFCs and expands the range of suitable substrates to include cellulose, the most abundant plant biomass component readily available as a waste material in many parts of the world. Further study of the biological factors driving these systems is required and is currently under investigation.

CHAPTER 4

EXTERNAL RESISTANCE AFFECTS THE POWER OUTPUT AND BACTERIAL DIVERSITY OF CELLULOSE-FED MICROBIAL FUEL CELLS

4.1. Introduction

Microbial fuel cells (MFCs) technology provides an alternative route of energy production through the direct conversion of renewable sources of biomass to electricity.

The conversion is catalyzed by microorganisms capable of oxidizing the substrate and transferring the resulting electrons to an anode electrode (Logan and Regan 2006a;

Schröder 2007). Due to diverse catalytic abilities of microorganisms, a wide array of compounds ranging from low-molecular weight organic acids to complex carbohydrates can be used as substrates in MFCs (Rabaey et al. 2007).

The most abundant renewable resource with the potential to make a substantial alternative energy source is cellulose. Recently, attempts have been made to convert cellulose to electricity in MFCs by integrating the processes of saccharification (i.e., hydrolysis) and fermentation with electron transfer to the anode by microorganisms

( Rismani-Yazdi et al. 2007; Ishii et al. 2008b). This process requires microbial populations that can simultaneously carry out cellulose hydrolysis, fermentation, and

anaerobic respiration in the same bioreactor. To date, no single microbial species has been reported to possess all of these functions entirely. Previous studies have used mono- and co-cultures of bacterial isolates with catalytic capabilities specific to separate steps of the process (Niessen et al. 2005; Ren et al. 2007; Sund et al. 2007). Although this approach can be a tool for understanding underlying mechanisms, it is not practical for commercialization of MFC technology due to the technical and economic hurdles. In contrast, microbial communities in natural environments, such as the digestive track of ruminants, soil or wastewater, are more robust and efficient than pure cultures for catalysis of integrated processes. Microbial populations provide many collective metabolic activities that can function over a wide range of conditions and are inherently diverse metabolically and phylogenetically. MFCs operated with microbial consortia intrinsic to wastewater or anaerobic sludge produce substantially more electricity than those with defined bacteria cultures (Park and Zeikus 2003; Rabaey et al. 2004; Rabaey et al. 2005a; Cheng and Logan 2007). These communities perform a wide variety of enzymatic reactions including unwanted or inhibitory processes that might adversely affect electricity production and efficiency (Rabaey et al. 2007).

In order to improve system performance, microbial communities in MFCs can be managed to lead to an efficient and robust microbial assembly through manipulating engineering designs and operating conditions. Biological approaches can be used to control the diversity and metabolism of microbial communities in MFCs. Active enrichment methods have been shown to increase MFCs power output and coulombic efficiency (Rabaey et al. 2004; Kim et al. 2005; Rismani-Yazdi et al. 2007). Optimization of the operating conditions (e.g., temperature, pH, substrate loading rate) and the system

design configuration have also been effective in improving the performance (Logan et al.

2006; Du et al. 2007; Rismani-Yazdi et al. 2008a). In most cases, however, information is scarce on the effects of such manipulations on the diversity and function of microorganisms in MFCs. External resistance (circuit load) is one of the components of an MFC design which connects the anode and cathode electrodes. The external resistance is used to dissipate the electrical energy when MFCs are operated independent of an electrical device. The circuit load is used as an integrated part of an electrical grid that controls the characteristic outputs of fuel cells. The external resistance (R) controls the flow of electrons from the anode to the cathode, affecting voltage (V) and current (I) outputs of MFCs according to Ohm’s Law:

V = IR (4.1)

The power output (W) is consequently affected by the circuit load:

W = I2R (4.2)

As the rate of substrate consumption for electricity generation depends on the rate of electron flow in the circuit, the external resistance also influences the coulombic efficiency of the system. The circuit load also provides a tool to optimize the anode potential for enhanced growth and metabolic activity of microorganisms.

MFCs are usually operated using a fixed circuit load, the size of which varies considerably from one experiment to another (Table 4.1).

Only a few studies have tested the effect of varying the external load on the performance of MFCs (Gil et al. 2003; Jang et al. 2004; Clauwaert et al. 2007a). These studies are limited to the estimation of the coulombic efficiency, maximum power and

External Substrate Inoculum Catholyte Cathode Maximum Feeding MFC design Citation Resistance catalyst power mode Ω density (mW/m2) 100 Glucose Anaerobic Phosphate N/A 4310 Batch Two-chamber Rabaey et sludge buffer, al. 2004 Potassium ferricyanide 480 Cereal Sludge Nutrient Pt 81 Batch H-type, Two- Oh and wastewater mineral buffer chamber Logan 2005 493 Cysteine Marine Mineral salts Pt 39 Batch Two-chamber Logan et sediment medium, al. 2005 aerated 1000 Cellulose Rumen Phosphate N/A 55 Batch H-type, Two- Rismani- microorgan buffer, chamber Yazdi et isms Potassium al. 2007 ferricyanide

88 1000 Carboxymethyl Geobacter Phosphate Pt 143 Batch H-type, Two- Ren et al. cellulose sulfurredu- buffer, chamber 2007 Cellulose cens and Potassium 59.2 Clostridium ferricyanide e cellulolyt- icum 510 Cellulsoe Rice paddy Tris-HCl N/A 10 Batch H-type, Two- Ishii et al. field soil buffer, chamber 2008 Potassium ferricyanide

20 Cellulose Rumen Phosphate N/A 66 Batch Two-chamber This 249 microorgan buffer, 57.5 study 480 isms Potassium 53 1000 ferricyanide 47

Continued

Table 4.1. Maximum power out put in MFCs with different external resistances, designs and operational conditions. 88

Table 4.1. continued

1000 Cellulose Marine Buffer solution Pt 45 Batch Lab-scale (Rezaei et Chitin 20 sediment with minerals 80 Sediment fuel al. 2007) Chitin 80 and vitamins 87 cell 1000 Cellulose Marine Not specified Pt Not Batch One-chamber Mathis et sediment specified al. 2008 Acetate 207 465 Wastewater Wastewater N/A Pt 26 Continuous One-chamber Liu et al. 2004 1000 Wastewater Wastewater N/A Pt 494 Batch One-chamber, Liu and No PEM Logan, 2004 117 Wastewater Wastewater N/A Pt 371 Continuous One-chamber, Oh and No PEM Logan, 2005 1000 Swine Swine N/A N/A 261 Batch One-chamber, Min et al. Wastewater wastewater No PEM 2005 1000 Wastewater Wastewater N/A CoTMPP 369 Batch One-chamber, Cheng et No PEM al. 2006

89 1000 Wastewater Wastewater N/A Pt 480 Batch One-chamber, Cheng et No PEM al. 2006 25-1000 Acetate Domestic N/A Pt 1800 Continuous One-chamber, Fan et al., Wastewater No PEM 2007

current. However, the exposure time (4-50 hrs) to each resistance in these studies has not allowed for the system to reach stability before performance assessment.

Little is known about how MFCs perform at certain circuit loads over long-term operations. Moreover, the effect of various external resistances on the activity and composition of microorganisms has not yet been considered. An understanding of the relationship between circuit load and microbial diversity can be a useful tool to control microbial communities and enhance performance of MFCs. The objective of this study was to determine the effect of various external resistances on power output and coulombic efficiency of cellulose-fed MFCs. Furthermore, the bacterial diversity and metabolic intermediates from cellulose degradation were evaluated under different circuit loads over three months. To our knowledge, this study shows, for the first time, that a component of the system engineering design (i.e., external resistance) affects the bacterial diversity and therefore the performance of MFCs.

4.2. Materials and methods

4.2.1. Microorganisms and medium

Standard anaerobic techniques were employed throughout the experiments.

Microorganisms indigenous to a cow’s rumen were used as biocatalysts in MFCs.

Ruminal contents, collected as reported previously (Rismani-Yazdi et al. 2007), were diluted (50% v/v) using an anaerobic buffer solution modified from Dehority (Dehority

2003) (containing per liter of distilled water: 450 mg K2HPO4, 450 mg KH2PO4, 900 mg

NaCl, 730 mg (NH4)Cl, 120 mg CaCl2·2H2O, and 90 mg MgSO4, and 500 mg cysteine-

HCl), and blended for 3 min at high speed to detach bacteria from partially digested feed

fibers. The diluted suspension was then filtered through four layers of cheese cloth and used as the initial inoculum (10% v/v) in the MFCs. The bacteria in the anode compartment were grown in a medium (anolyte) made of the anaerobic buffer solution supplemented with 5 g/l yeast extract and 15g/l trypticase. Ball-milled microcrystalline cellulose (Sigmacell 20, Sigma, St. Louis, MO) was used as a substrate and the source of electron donor (Rismani-Yazdi et al. 2007).

4.2.2. MFCs construction and operating conditions

MFCs were constructed of two cubic-shaped polycarbonate compartments of equal dimensions (25 mm width, 60 mm height, and 62 mm length) (Fig. 4.1). The total volume of each compartment was 97 ml, of which 75 ml was the working volume. The two compartments were separated by an Ultrex proton-exchange membrane (CMI-7000,

Membrane International, Glen Rock, NJ). The membrane had an available surface area of

37 cm2 and was pretreated in the anaerobic buffer solution at 45ºC for 24 h prior to use.

Two equally-sized polished graphite plates, each with a surface area of 40 cm2, were used as the anode and cathode electrodes. The distance between the electrodes was 1.6 cm.

Electrode preparation and pretreatment were performed as described by Bond and Lovley

(Bond and Lovley 2003). The ohmic resistance of the electrodes and the connecting wires was 1.35 ± 0.5 ohms as measured with an ohm meter.

Four precision metal film resistors with 20, 249, 480, and 1000 ohms resistances and 0.1% resistance variation with temperature (Mouser Electronics, Mansfield, TX) were used in this study as the external resistances between anodes and cathodes. Eight

MFCs were randomly assigned so that each of the four resistors was tested in duplicate.

Specifications and experimental conditions were maintained the same for all MFCs; the only variable was the external resistance connecting the anodes and cathodes.

Anode compartment

Cathode compartment Proton exchange membrane

Figure 4.1. Schematic diagram of the microbial fuel cells used for the experiment.

The anode chamber of each pre-sterilized MFC was filled with the medium under a stream of oxygen-free CO2 and inoculated (10% v/v) with a suspension of rumen microorganisms. A potassium hexacyanoferrate solution (75 ml) (50 mM K3Fe(CN)6 in

100 mM K2HPO4, pH 7) (Park and Zeikus 2003) was used as the catholyte in the cathode

compartments. Fuel cells were incubated at 39±1ºC on a shaker with an agitation rate of

100 rpm.

The MFCs were initially operated under batch condition for 7 d with 5 g/l cellulose to allow build-up of biomass, and then operated under fed-batch condition for 9 consecutive weeks. During the fed-batch operation, fuel cells were fed 1 g cellulose/l every two days by removing 10 ml of the anolyte and replacing it with fresh medium.

The catholyte was also replenished simultaneously, and the anolyte pH was monitored and adjusted to pH 6.8 with NaOH and HCl.

After 10 weeks of enrichment and operation from the start of the experiment, the

MFCs were transferred into an anaerobic glove box and disassembled for microbial and metabolite sampling. The anolyte of each MFC was then used to inoculate new MFCs constructed as described above. The new fuel cells were inoculated with an equal amount of inoculum. The fresh anolyte medium was inoculated with 10% (v/v) of the cell suspension (optical density of 1.5 at 600 nm which corresponded to 3.3 ± 0.1 mg/ml cell protein concentration in used anolyte) of pre-enriched bacteria obtained from 10-week operated MFCs. The anodes and cathodes of the new fuel cells were connected with the same external resistance. These new MFCs were operated as fed-batch systems for 14 d under similar conditions as described above and used for comparative analysis of polarization and coulombic efficiency.

4.2.3. Electrical measurements

The electrical output of MFCs was monitored every 10 s by measuring the potential difference (voltage) between the anode and the cathode across the external

resistance (R) using an Agilent HP 34970 data acquisition unit (Agilent Technologies,

Santa Clara, CA) connected to a computer. Power density (P), expressed in terms of power per unit area of anode electrode (mW/m2), was calculated according to the equation (3);

P = I×V/A (4.3) where V is the voltage (V), I (I = V/R) the current (amps), and A the surface area of the anode electrode (m2). Polarization tests and determination of internal resistances were performed as reported previously (Rismani-Yazdi et al. 2007). The polarization test was conducted simultaneously with all eight MFCs in order to eliminate the effect of time as a variable on the performance of fuel cells. The absolute potential of the anode and cathode was also measured vs. an Ag/AgCl reference electrode (saturated KCl, 0.197 V vs. normal hydrogen electrode (NHE)) (CH Instruments, Austin, TX) with the anode or the cathode as the working electrode. All experiments were conducted in duplicate, and mean values are presented.

4.2.4. Coulombic efficiency

The coulombic efficiency (CE) (i.e., electron recovery from cellulose in the form of electricity) was calculated according to the equation (4):

CE = (Co/Ci) × 100% (4.4) where Co is the total coulombic output calculated by integrating the current over time, and

Ci is the theoretical amount of coulombs available from cellulose as calculated according to the equation (5):

Ci = F b v M (4.5)

where F is Faraday’s constant (96,485 C/mol-e-), b the number of moles of electrons produced per mol of equivalent glucose, v the anolyte volume (l), and M the concentration of glucose equivalent (mol/l). The equivalent glucose content of solids from cellulose fermentation was measured by using the following equation (6):

Geq = 180/162 × DW (4.6) where Geq is the glucose equivalent weight (g), and DW is the dry weight of the cellulose sample (g).

4.2.5. Analysis of cellulose degradation

Anolyte was transferred into serum vials chilled on ice, amended by an equal volume of neutral detergent solution (containing per liter of distilled water: 30 g SDS,

14.6 g EDTA, 4 g NaOH, 3.6 g Na2B4O7, 4.56 g Na2HPO4, and 10 ml 2-ethoxyethanol)

(Goering and Van Soest 1970), and stored at -20°C prior to analysis. Residual cellulose was determined by the neutral detergent method as described by Weimer and co-workers

(Weimer et al. 1990). Briefly, vials were thawed, sealed with flanged rubber stoppers, crimp sealed, and autoclaved for 45 min at 121°C. The contents of the hot vials were then immediately vacuum-filtered through pre-weighed P5 filter papers (I.D., 47-mm and nominal pore size, 5 μm) (Fisher Scientific). The filters were dried at 105°C for 3 days and weighed after equilibration at 22±2 ºC. The dry weight of residual cellulose in each

MFC was calculated after deducting the weight of inocula (from neutral detergent-treated samples obtained from similarly inoculated and operated MFCs lacking substrate) and accounting for the predetermined moisture content of the fresh cellulose (1.5% by weight). The weight loss of cellulose due to MFC metabolism was determined by

subtracting the residual dry weight of cellulose from the dry weight of cellulose fed into the fuel cells.

4.2.6. Analysis of bacterial diversity

Planktonic and anode-attached bacteria enriched in MFCs were harvested after 10 weeks from the beginning of the experiments. Biomass sampling and extraction, purification, and quantification of DNA were carried out as reported previously

(Rismani-Yazdi et al. 2007). Analysis of bacterial community structure was performed using denaturing gradient gel electrophoresis (DGGE). PCR amplification of the V3- region of 16S rRNA with primer sets Eub357f (5'-CCTACGGGAGGCAGCAG-3') with a GC clamp and Eub519r (5'-GWATTACCGCGGCKGCTG-3'), and DGGE analysis of

PCR products were conducted as described by Yu and Morrison (2004a). The DGGE gels were analyzed using the band-searching algorithm of BioNumerics software

(BioSystematica, Tavistock, Devon, U.K.). Bands with intensity < 0.02 were excluded from the analysis. In order to determine the differences between DGGE fingerprints, dendrogram trees were constructed using the unweighted pair-group method with arithmetic mean (UPGMA) in combination with Jaccard’s similarity coefficient. The

Jaccard’s coefficient is based on the presence of bands in pairs of different fingerprints; absence of a band in both fingerprints is not accounted for in the analysis.

4.2.7. Analysis of fermentative metabolites

Analysis of short chain fatty acids (SCFA) was performed using a Shimadzu LC-

10AT liquid chromatograph (Shimadzu, Columbia, MD) equipped with an automatic

sampler/injector, an ultraviolet detector, a cation-exchange HPLC column, 30 cm × 7.8 mm I.D. (Supelcogel C-610H, Supelco), and a guard column, 5 cm × 4.6 mm I.D.

(Supelguard H, Supelco) (Peu et al. 2004). Samples of thoroughly mixed anolyte obtained from each MFC were centrifuged at 13,000 × g for 10 min at 4°C, and the supernatants were stored at -80ºC. Prior to analysis, the supernatants were thawed on ice, diluted using HPLC grade water (Honeywell Burdick & Jackson, Morristown, NJ), and filtered through a 0.22 µm syringe-tip filter (Fisher Scientific) into 2-ml glass vials with septum lids. A volume of 100 μl of each sample was injected into the column at ambient

temperature. The separation of SCFA was obtained with a mobile phase of 0.1% H3PO4 at a flow rate of 0.5 ml/min with detection at 210 nm. Two samples were analyzed in duplicate for each MFC. Acetic, propionic, butyric, isobutyric, valeric, isovaleric, lactic, pyruvic and fumaric acids, used for preparing standard curves, were obtained from

Sigma-Aldrich (St. Louis, MO).

4.2.8. Analysis of bacteria cell protein

The collected anolyte from each MFC was centrifuged at 16,000 × g for 20 min at

4ºC. Cell pellets were resuspended in lysis buffer ([pH 8], 50 mM Tris-HCl, 50 mM

EDTA, 2% Chaps (w/v), 1% protease inhibitor cocktail (Sigma, St. Louis, MO) (v/v),

0.06% RNase (w/v)), and disrupted by passing two times through a French Pressure Cell press (SLM-Aminco; Spectronic Instruments, Rochester, NY) at 68.9 MPa (10,000 psi).

The protein concentration of the lysates was determined by UV spectroscopy at 562 nm

(Molecular Devices, Sunnyvale, CA) using bicinchoninic acid (BCA) protein assay

(Pierce, Rockford, IL) according to manufacturer instructions.

4.3. Results

4.3.1. MFC performance as a function of external resistance

The four external resistances tested in this study were 20, 249, 480, and 1000 ohms. Each MFC was loaded with a fixed resistor that was not changed during the study except after 12 weeks of operation when the resistance between the anode and cathode was varied to perform the polarization test. Figure 4.2 shows the comparison of polarization (i.e., voltage-current) and power-current curves for MFCs with different external resistances. A maximum power density of 66 mW/m2 was achieved by MFCs with 20 ohms circuit load, while lower power densities, 57.5, 53 and 47 mW/m2, were obtained with 249, 480 and 1000 ohms external resistances, respectively.

The absolute anode and cathode potentials of the MFCs measured vs. an Ag/AgCl reference electrode are presented in Table 4.2. The results showed relatively similar cathode potentials at different external resistances. This was expected since the cathodic reaction in each MFC was based on the ferricyanide reduction. However, the anode potential varied under different circuit loads employed. MFCs with lower external resistances resulted in higher anode potentials. Open circuit potentials also varied in

MFCs operated with different circuit loads. MFCs with 20 ohms external resistance showed the highest open circuit potential of 631 mV, whereas MFCs with higher circuit load resulted in lower open circuit potentials (Table 4.2).

Analysis of coulombic efficiency revealed differences in electron recovery in the form of electricity from the cellulose substrate under various circuit resistances (Table

4.2). Higher coulombic efficiencies were achieved in MFCs with lower external resistance. A maximum coulombic efficiency of 19 % was obtained in MFCs with 20

ohms external resistance. MFCs with 249 and 480 ohms circuit loads had equal coulombic efficiency of 14%, while in MFCs with 1000 ohms external resistance only 12

% of the electrons available from cellulose were recovered as electricity.

700 A 600 20 Ohms 249 Ohms 500 480 Ohms 1000 Ohms 400

300

Potential (mV) Potential 200

100

0 100 200 300 400 500 600 700

70 B 60

) 50 -2

Power density (mW m 10

0 100 200 300 400 500 600 700 Current density (mA m-2)

Figure 4.2. Effect of external resistance on A: polarization (i.e., voltage-current) and, B: power-current properties of MFCs with different external resistances.

External Potential Open Coulombic Total resistance (mV vs. Ag/AgCl) circuit efficiency concentration (Ω) potential (%) of SCFA (mV) (mg/l) Anode Cathode Whole cell

20 261±6 268±17 14±3 631±2 19±1.6 4422.7±10

249 149±24 252±36 107±8 622±5 14±0.5 5653.9±163

480 134±0.6 260 128 608±1.5 14±2.0 6157.4±20

1000 21.5±12 250±6 236±38 579±0.3 12±0.7 7759.7±493

Table 4.2. Effect of external resistance on performance characteristics of MFCs a Data were obtained after 12 weeks of operation under a fixed load from the start of experiment. b Data are presented as mean ± SD, (n = 3).

4.3.2. Bacterial diversity as a function of external resistance

The effect of different external resistances on the structure of bacterial communities enriched in MFCs was investigated using the DGGE analysis of 16S rRNA genes following 10 weeks of continuous electricity generation under a fixed circuit load.

Figure 4.3.A presents the DGGE profiles of both anode-attached and planktonic bacteria for the four resistances tested. UPGMA dendograms with Jaccard’s coefficient revealed differences in the DGGE profiles between anode-attached vs. planktonic samples and the various external resistances (Fig. 4.3.B). A greater similarity was observed among the 100

DGGE fingerprints of anode-attached communities (65-75%) compared to those from planktonic population (48-78%). Within the anode-attached populations, samples from

MFCs with lower circuit loads (20 and 249 Ω) had more similar (75%) bacterial community than those with higher external resistances (65% similarity between 480 and

1000 Ω).. In the planktonic communities, MFCs with higher external resistances resulted in more similar DGGE banding pattern (78%) than those from lower circuit loads (48%).

A B

Figure 4.3. Effect of external resistance on bacterial diversity in MFCs. A: DGGE profiles of 16S rRNA genes amplified from planktonic and anode-attached bacterial communities enriched in MFCs with different external resistances for 10 weeks with cellulose. B: UPGMA dendograms constructed using Jaccard’s similarity coefficient generated from the DGGE profiles.

101

4.3.3. Effect of different external resistance on microbial metabolism

Cellulose degradation and production of short chain fatty acids (SCFA) were analyzed after 10 weeks of operation. Cellulose degradation was complete (< 0.1% residual) and there were no discernible differences among the MFCs. Figure 4.4 shows the profile of SCFA produced in MFCs under different external resistances. The main metabolite at all resistances was acetic acid. Anaerobic degradation of cellulose was also accompanied by production of propionic, butyric, isobutyric, valeric, isovaleric, and lactic acids. Production of fumaric and pyruvic acid were below the detection limit (10 mg/l). The SCFA profiles differed between the four external loads tested here (Fig. 4.4).

The total SCFA concentration of the anolyte was higher in MFCs with larger circuit loads

(Table 4.2). The 1000 ohm-MFCs had the highest total SCFA concentration whereas the concentration in MFCs with 20 ohms resistance was the lowest. The same trend was also observed for the concentration of individual fatty acids, except for lactic acid which remained relatively constant across all four resistances.

102

103

Figure 4.4. Concentrations of SCFA produced in MFCs with different external resistances after 10 weeks of operation.

103

4.4. Discussion

This study showed that the external resistance affects the performance, bacterial diversity, and metabolites produced in cellulose-fed MFCs. The external resistance affects the characteristic outputs of the MFCs by controlling the flow of electrons from the anode to the cathode. In previous studies, MFCs have been operated at case-specific fixed circuit loads that have widely varied in size (Table 4.1). Consequently the performance has varied considerably over the range of resistances used. A comparison of the performance of MFCs in previous studies with respect to the external resistances is not possible due to multiple differences in operational conditions and system engineering design. The effect of varying the external resistance on the performance of MFCs has been previously tested in limited studies (Gil et al. 2003; Jang et al. 2004; Clauwaert et al.

2007a). However, the exposure time to each resistance in these studies has not been sufficiently long (4-50 h) to allow for the microbial community to respond to changes in the circuit load. Thus the outcomes have been due to electrochemical effects rather than changes in the activity and diversity of microorganisms. In this study, the effect of external resistance on performance was tested with a systematic approach of operating parallel MFCs independently at constant circuit loads for three months. This experimental approached was designed to eliminate all other variables except for the circuit load.

The results showed that power output was a function of circuit load as MFCs with lower external resistance produced more power. The performance of an MFC depends on the diversity and activity of its microorganisms. Any change in fuel cell design and operational conditions that affect the bacterial dynamics and metabolism results in a

104

change in the performance of MFC. Understanding these effects would provide a tool for controlling the microbial communities in MFCs to achieve enhanced performances.

Cluster analysis with Jaccard’s coefficient showed here that external resistance affects the outcome of long-term bacterial enrichment in MFCs. The effect was induced mainly by the impedance of the circuits to electron flow and differences arose in the anode potential under different circuit load. Specifically, high anode potential was associated with low circuit load. The anode potential depends on the external resistance and electron transfer activity from the bacteria to anode. It is possible that operating the MFCs at low circuit loads selected for bacterial community with relatively higher anodic-electron transfer activity. This interpretation is supported by previous studies based on poised potential experiments, suggesting that a more positive anode potential induces greater colonization of anode-reducing microorganisms on the electrode ( Rabaey et al. 2004;

Finkelstein et al. 2006), and with the reports from Aelterman et al. (2008) indicating that anode potential influences the respiratory activity and growth of bacteria in the anode compartment.

Circuit load also affected the coulombic efficiency. MFCs with lower external resistance resulted in higher coulombic efficiency which was consistent with other performance characteristics. Coulombic efficiency here was calculated on the basis of equivalent glucose content of cellulose consumption and did not take into account the amount of metabolites such as SCFA produced from cellulose degradation. Coulombic efficiencies were relatively low at all circuit loads due to accumulation of SCFA, loss of electrons to secondary reactions in the anode compartment such as methanogenesis

(unpublished data) , aerobic oxidation due to O2 diffusion from the cathode to the anode

105

compartment, and bacterial growth (Freguia et al. 2007b). Substrate crossover through the membrane could also contribute to low coulombic efficiencies achieved here (Rismani-

Yazdi et al. 2008a).

Analysis of SCFA also showed differences among the four circuit loads tested.

The cellulose loading rates in all MFCs were the same but the concentrations of SCFA were higher at larger circuit loads. This can be explained by 1) increased SCFA production under higher external resistance possibly due to a shift in metabolism from respiration to fermentation, and 2) increased consumption of metabolites in MFCs with lower circuit loads perhaps because of higher rates of respiration and anodic-electron transfer activity. The concentration of the substrate in the MFCs should be compatible with the circuit load such that nearly all of the metabolites are consumed. Accumulation of SCFA here suggests metabolite oxidation using the anode as an electron sink was a rate limiting step in the conversion of cellulose to electricity in MFCs.

This study demonstrates that the external resistance affects the bacterial diversity and their metabolism in MFCs which consequently influences the power output and the coulombic efficiency of the system. The external resistance provides a useful tool to control the anode potential and select for robust microbial communities to achieve enhanced MFC performance. Understanding the effect of circuit load on growth, metabolic activity and structure of microorganisms remains a target for future research.

The results also show that analysis of microbial communities and their changes in MFCs provides valuable insights into electrochemical performance characteristics. To our knowledge, this is the first report indicating the interaction between microorganisms and a component of the system engineering design (i.e., external resistance). More research is

106

required to identify and study other operational and design factors that trigger assembly of the desired microbial community with the desired functions in MFCs for various applications.

107

CHAPTER 5

EVALUATION OF METHANE FORMATION IN MICROBIAL FUEL CELLS GENERATING ELECTRICITY FROM CELLULOSE

5.1. Introduction

Microbial fuel cells (MFCs) present an attractive alternative to conventional electricity generation. Efforts have been made in recent years to improve the power output of MFCs by optimizing design factors and electron transfer and associated electrochemical and redox reactions in the anode and cathode compartments (Rismani-

Yazdi et al. 2008a). Relatively less effort has been devoted to the improvement of substrate oxidation and direction of electron disposal away from undesirable side reactions (He et al. 2005; Kim et al. 2005). Electron bypasses are particularly an issue in

MFCs that are operated with mixed microbial populations, inevitable with complex substrates and pilot scale systems. As the technology advances and performance of MFCs improves these reactions and design strategies for their proper control need to be addressed.

Methanogenesis acts as a sink for electrons and competes against anodic electron transfer, thereby decreasing the coulombic efficiency of MFCs (Freguia et al. 2007b). 108

It can indirectly affect the metabolic pathways and growth of some bacteria in

MFCs through interspecies hydrogen transfer (Rychlik and May 2000). Methanogens are generally able to utilize H2 + CO2, C1 compounds such as fomate and methanol, and acetate for growth and methane production. Many of the substrates for methanogenic archaea are central metabolites that are formed in the anaerobic degradation and fermentation of organic matter such as cellulose in MFCs. Methanogenesis from the above-mentioned substrates proceed according to the following reactions (Blaut et al.

1992):

Reaction ΔG°´ (J/mol CH4)

CO2 + 4 H2 → CH4 + 2 H2O –130

4 HCOOH → 3 CO2 + CH4 + 2 H2O –119

4 CH3OH → 3 CH4 + 1 CO2 + 2 H2O –106

– + CH3COO + H → CH4 + CO2 –36

Methane production has been previously reported in MFCs with fermentable substrates when inoculated with samples of waterlogged soils, anaerobic sludge or sewage treatment plants (Rabaey et al. 2004; Kim et al. 2005; He et al. 2005; Ishii et al.

2008a; Ishii et al. 2008b). Methanogenesis is also a common process in fermentations occurring in the rumen ecosystem, a potential drawback when using rumen fluid as inoculum for cellulose-fed MFCs (Rismani-Yazdi et al. 2007). However, little is known about the effects of methanogenesis on the performance of MFCs ( He et al. 2005; Kim et al. 2005; Freguia et al. 2007b; Ishii et al. 2008a). It was hypothesized in this study that 109

methanogens are present in cellulose-fed MFCs because they were originally inoculated with rumen fluid. The purpose of this study was to characterize the diversity of methanogens and assess their influence on the performance of MFCs upon long-term operation.

5.2. Materials and methods

5.2.1. Microbial fuel cell construction and operation

Replicate MFCs comprised of two compartments, anode and cathode of equal dimensions and volume separated with a proton exchange membrane, were constructed as reported previously (Rismani-Yazdi et al. 2008b) (Fig. 5.1). Detailed specifications of the

MFCs are presented in Table 5.1. The anode compartments were filled with a nutrient medium under anaerobic conditions and inoculated with a bovine ruminal microbial consortium prepared as previously described (Rismani-Yazdi et al. 2008b). The cathode compartments contained an aerobic solution of potassium hexacyanoferrate (Park and

Zeikus 2003), which was replenished every two days. Colloidal cellulose was used as the sole substrate in the anode compartments. The MFCs received 5 g/l cellulose during the first week of the experiment, and then were fed 1 g/l cellulose every other day (day 8 through 90) by replacing 10 ml of the anolyte with fresh medium. The MFCs were operated at 39±1ºC on a shaker with an agitation rate of 100 rpm. The pH of the anolyte was monitored and adjusted to pH 6.8 with NaOH and HCl.

110

Component or condition Description Manufacturer Compartments Transparent polycarbonate McMaster-Carr, plastic Cleveland, OH Dimensions: 25 mm width, 60 mm height, and 62 mm length Total volume: 97 ml Working volume: 75 ml Electrodes Polished graphite plates (40 McMaster-Carr, cm2) Cleveland, OH Proton exchange Ultrex (37 cm2) CMI-7000, Membrane membrane International, Glen Rock, NJ External resistance 20 and 100 Ω Mouser Electronics, Mansfield, TX Anode catalysts Rumen microorganisms (10% - v/v) Cathode mediator Potassium hexacyanoferrate (50 Fisher Scientific

mM K3Fe(CN)6 in 100 mM

K2HPO4, pH 7) Substrate Cellulose Sigmacell 20, Sigma, Feeding schedule: 1-7 d (5 g/l), St. Louis, MO 8-91 d (1 g/l every other day) Incubator temperature 39±1ºC Shaker agitation rate 100 rpm SMS Optical, Hauppage, NY

Table 5.1. Microbial fuel cell specifications and experimental conditions.

111

Cathode compartment Anode compartment

Figure 5.1. A schematic diagram of the microbial fuel cells used for the experiment.

(http://digitalunion.osu.edu/r2/summer07/nskrinak/assembly.html)

5.2.2. Electrical measurements

The circuit between the anode and the cathode of each MFC was connected with a

20 Ω or 100 Ω external resistance. The potential difference (voltage) was measured every

10 s using a computer controlled Agilent HP 34970 data acquisition unit (Agilent

112

Technologies, Santa Clara, CA). The current (I, amps) was calculated according to the

Ohm’s law (I = V/R), where V is the voltage (V) and R (Ω) is the known external resistance. The volumetric power density (W/m3) was computed as the product of the voltage, current and the reciprocal of the working volume of the anode compartment.

Polarization tests were performed by varying the external resistance between the electrodes stepwise from 1 MΩ to 5Ω using a resistance decade box (Extech Instruments,

Waltham, MA) with a pause at each resistance to allow voltage equilibrium. The absolute potential of the anode and cathode was measured vs. an Ag/AgCl reference electrode

(saturated KCl, 0.197 V vs. normal hydrogen electrode (NHE)) (CH Instruments, Austin,

TX) with the anode or the cathode as the working electrode. Coulombic efficiency (CE) was calculated as the ratio of the coulombic output computed by integrating the electrical current over time to the theoretical amount of coulombs available from glucose equivalents through cellulose hydrolysis (Rismani-Yazdi et al. 2008b).

5.2.3. Analytical techniques

The headspace gas of the anode compartment of MFCs was collected in Tedlar gas sampling bags (Zefon International, Ocala, FL) attached to the anode chamber using a three-way valve. The methane concentrations of the headspace gas were analyzed using a Shimadzu GC-2014 gas chromatograph (GC) equipped with a stainless steel column (2 m × 1.8 in × 2.0 mm I.D.) packed with 80-100 mesh Porapak N (Varian, Palo Alto, CA) and a thermal conductivity detector, with nitrogen as carrier gas. Injection, detection and column temperatures were 110, 110 and 180°C, respectively. Short chain fatty acids of the anolyte from MFCs were separated and quantified by HPLC as previously described

113

(Rismani-Yazdi et al. 2008b). Analysis of cellulose degradation was performed as reported by Rismani-Yazdi et al. (2008b).

5.2.4. DNA extraction and amplification

The biomass of planktonic and anode-colonizing microorganisms was harvested by sampling the anolyte and scraping the biofilm off the anode, respectively, as described previously (Rismani-Yazdi et al. 2007). Extraction of genomic DNA was performed using a repeated bead beating method followed by DNA purification with QIAamp columns (Qiagen, Valencia, CA) as described by Yu and Morrison (2004b). The DNA quantity from each sample was determined fluorospectrometrically using a Quant-iT broad range DNA assay kit Q33130 (Molecular Probes, Eugene, OR). The V3-region of the 16S rRNA genes was PCR-amplified using universal bacterial primer sets (Eub357f

(5'-CCTACGGGAGGCAGCAG-3') with a GC-clamp and Eub519r (5'-

GWATTACCGCGGCKGCTG-3'), and archaeal specific primers (Arc-344f with a GC- clamp and Eub-519r) (Lane 1991). DGGE analysis of PCR products was performed as reported by Yu and Morrison (2004a).

5.2.5. Recovery and sequencing of DGGE bands

The dominant and unique bands from the archaeal DGGE gels were cut out with a sterile scalpel blade, and placed in a 1.5-ml Eppendorf tube containing 20 µl 50 mM Tris-

HCl, pH 8, and stored at -20°C. DNA was eluted from the gel fragments by freezing (at -

80°C for 5 min) and thawing (at 65°C for 5 min) in three cycles followed by centrifugation at 16,000 g. A 1-µL aliquot of the resulting solution was re-amplified by

114

PCR with Arc-344f /Eub-519r primers. Products were then purified using a PCR purification kit (Qiagen, Valencia, CA), and sequenced directly with the Arc-344f primer with no GC clamp on the automated 3730 DNA analyzer (Applied Biosystems,

Framingham, MA) with BigDye terminator cycle sequencing chemistry. Sequencing of

PCR products was performed twice for each band position to check if there was hidden diversity at each band position.

5.2.6. Band sequencing

The archaeal 16S rRNA gene sequences were analyzed by means of the BLASTn program (http://www.ncbi.nlm.nih.gov/BLAST/) (Altschul et al. 1990) and close relatives of the sequences were retrieved from the GenBank database. Phylogenetic associations were assigned using the Classifier program of the Ribosomal Database

Project II (http://rdp.cme.msu.edu/classifier) (Wang et al. 2007). The sequences of 12 bands excised from the DGGE gels have been deposited in the GenBank database under accession numbers EU553823 to EU553835 as listed in Table 5.2.

5.3. Results and discussion

5.3.1. Methane production and MFC performance

Methane production was evaluated in four MFCs inoculated with rumen microorganisms as biocatalysts to produce electricity from cellulose. Two of the MFCs were each operated with a 20-Ω external resistance (R20Ω) and the other two had 100-Ω external resistances (R100Ω). Methane concentration in the headspace of the anode compartment of MFCs was analyzed during the first week (1-7 d) and the last two weeks

115

(77-90 d) of the experiment. During the first week of operation 0.31 ± 0.004 (± SD) and

0.44 ± 0.004 mmol of methane was produced in R20Ω and R100Ω MFCs, respectively. The concentration of methane in the headspace decreased to below the detection limit (< 0.5

×10-3 mmol) during the last two weeks of the operation. This finding was consistent with the report from Rabaey et al. (Rabaey et al. 2004) that methanogenesis in a two- compartment glucose-fed MFC inoculated with anaerobic sludge was suppressed to

(0.4%±0.9% in the headspace) after 71-d of operation. The decrease in methane production here was accompanied with an increase in the performance of the MFCs. Fig.

5.2 shows the current production in MFCs over the course of operation (90 days).

4.5 4 3.5 R 20Ω R 100Ω 3 2.5 2

Current (mA) 1.5 1 0.5

0 0 20406080100 Time (day)

Figure 5.2. Current production from cellulose with rumen microorganisms in MFCs with two different external resistances over the course of operation.

116

Both R20Ω and R100Ω MFCs had a relatively similar current output during the first five weeks of the operation. Subsequently the R20Ω MFCs produced more current than the

R100Ω MFCs. Current production reached 0.1 mA after one week and then slightly increased to 0.2 mA on day 34 from the start of the experiment. Thereafter the current output of MFCs was enhanced sharply and reached 3.3 and 2.2 mA on day 90 for the

R20Ω and R100Ω MFCs, respectively. The polarization curves reflecting voltage-current and the power-current characteristics of MFCs are presented in Fig. 5.3.

700 4 V-I (R 20Ω) 3.5 600 V-I (R 100Ω) )

P-I (R 20Ω) 3 -3 500 P-I (R 100Ω) 2.5 400 2 300 1.5 Voltage (mV) Voltage 200 1 Power densityPower m (W

100 0.5

0 0 0 100 200 300 400 500 600 700 Current density (mA m-2)

Figure 5.3. Voltage-current (V-I) and the power-current (P-I) characteristics of MFCs.

117

The power output of MFCs with a 20-Ω external resistance was higher as compared to the MFCs with a 100-Ω circuit load. A maximum volumetric power density of 3.5 W m-3 was achieved in the R20Ω MFCs, which was over three times greater than that obtained

-3 with R100Ω MFCs (1.03 W m ). Figure 5.4 shows the change in the coulombic efficiency of MFCs during the operation period. The coulombic efficiency was relatively higher in

R20Ω MFCs than R100Ω MFCs, and ranged from 13.0 ± 1.0 (± SD) and 11.5 ± 1.0 % on day

15 to 28.6 ± 3.2 and 25.3 ± 2.1%, respectively, after 90 days from the start of the experiment.

30 R 20Ω 25 R 100Ω

Coulombic efficiency (%) Coulombic efficiency 5

0 0 20406080100 Time (day)

Figure 5.4. Coulombic efficiency of MFCs during the operation period.

118

The differences in the performance of MFCs here can be attributed to the differences in the external resistance of systems because all other operating conditions were the same for the MFCs. This is in agreement with our previous findings that the highest power output was associated with the lowest circuit load (range 20-Ω) in cellulose-fed MFCs (Rismani-Yazdi et al. 2008b). The improved current output and coulombic efficiency of MFCs over the course of operation may be attributed to an increase in the abundance and activity of bacteria shuttling electrons to the anode and to the suppression of methanogenesis. The suppression is consistent with studies by Kim et al. (2005) and He et al. (2005). These authors showed that the inhibition of methanogenesis in MFCs with a selective metabolic inhibitor (2-bromoethanesulfonate) improved the power output and coulombic efficiency. The contribution of methanogenesis to reduced coulombic efficiency in MFCs has also been demonstrated by

Freguia et al. (2007b).

5.3.2. Effect of methanogens on metabolites

Samples of the anolyte from MFCs were analyzed to track the production and concentration of metabolites from cellulose fermentation over the time. Fig. 5.5 shows changes in the concentration of short chain fatty acids (SCFA) after 6, 16, 49, 73, and 90 days from the start of the experiment. The concentration of SCFA was generally lower in

R20Ω MFCs than R100Ω MFCs. This is due to the effect of external resistance on microbial dynamics and consumption of metabolites in MFCs as previously discussed (Rismani-

Yazdi et al. 2008b). Acetic acid was the major metabolite of cellulose metabolism in

MFCs. After acetate, isovaleric, propionic, butyric, valeric, isobutyric, and lactic acids

119

had the next highest concentrations, respectively. The concentrations of all SCFA, except acetic and lactic acids, decreased during the first 73 d of MFCs operation, and then leveled off. The acetic acid concentration increased constantly over the first seven weeks of the experiment, followed by a decrease on day 73 and then remained constant until the end of the experiment.

7000.00 )

-1 6000.00 Acetic R20 5000.00 Acetic R100 4000.00 Propionic R20 3000.00 Propionic R100 2000.00 Isobutyric R20 Isobutyric R100 1000.00 Consentration (mg l (mg Consentration 0.00 0 20406080100 Time (day)

2500.00 Isovaleric R20 ) -1 2000.00 Isovaleric R100 Butyric R20 1500.00 Butyric R100

1000.00 Valeric R20 Valeric R100 500.00 Lactic R20 Consentration (mg l (mg Consentration Lactic R100 0.00 0 20406080100 Time (day)

Figure 5.5. Concentrations of SCFA produced in MFCs with different external resistances over the course of operation.

120

Lactic acid was the only fatty acid that increased in its concentration slightly over time (Fig. 5.5). Pyruvic and fumaric acids were not detected. Changes in SCFA with time may be due to i) increases in the rate of metabolite consumption coupled with electron transfer to the anode as evident from increased performance of MFCs, ii) alteration in the diversity of microbial communities (Fig. 5.6), and/or iii) the suppression of methanogenesis over time.

Methanogenesis can affect the profile of SCFA because it influences the growth and metabolism of some species of hydrogen- producing cellulolytic bacteria (Rychlik and May 2000). Hydrogen is the major intermediate in cellulose degradation. In the rumen, the origin of inoculum used in this study, hydrogen is a central metabolite in cellulose degradation. Increased partial pressure of hydrogen reduces the activity and changes fermentation pathways of some cellulose-hydrolyzing microorganisms because it inhibits the re-oxidation of NADH produced during glycolysis. Utilization of hydrogen by the methanogens reduces the hydrogen tension. Thus the interspecies hydrogen transfer between cellulolytic bacteria and methanogens may be beneficial for enhanced hydrolysis of cellulose in MFCs. Rychlik and May (2000) reported that in the presence of methanogens the metabolism of cellulose-hydrolyzing microorganism was oriented toward a greater production of acetate instead of more reduced metabolites. This explanation may account for the decrease of acetate concentration in the anolyte as methanogenesis was suppressed in this study.

121

5.3.3. Bacterial and archaeal diversity

To assess links between performance of MFCs, methane production, and the diversity of bacteria and archaea, DNA samples from MFCs were characterized by

DGGE. Samples of planktonic population were obtained from the anolyte of each MFC after 16, 49, 73 and 90 days from the start of the experiment. The anode biofilm was sampled from each MFC after 73, and 90 days of operation. The DGGE profiles of the archaeal and bacterial community during the operation course of MFCs are shown in Fig.

5.6. The DGGE banding pattern changed greatly over time, indicating an alteration in the diversity of bacterial and archaeal communities. The DGGE profiles of the inoculum were clearly different from those of the samples obtained from the MFCs during the course of operation. In keeping with results from previous studies (Rismani-Yazdi et al.

2007, 2008b), there were differences between the biofilm and planktonic populations for both the archaeal and bacteria communities.

The diversity of archaeal population in MFCs decreased over time as indicated by the reduction in the number of bands in the DGGE profile (Fig. 5.6B). The phylogenetic identification of sequences from intense bands excised from the archaeal DGGE gels are presented in Table 5.2. Out of the 12 sequences analyzed, one sequence showed similarity (96%) with a previously characterized methanogen, seven sequences (1, 3-5, 8,

9 and 12) were close relatives (>96% identity) of uncultured methanogens, and four sequences (6, 7, 10 and 11) fell under 96% similarity with existing GenBank sequences.

122

A Ladder Inoculum

R20Ω Planktonic day 16

R100Ω Planktonic day 16

R20Ω Planktonic day 49

R100Ω Planktonic day 49 R Planktonic day 73 20Ω R Biofilm day 73 20Ω R Planktonic day 73 100Ω R Biofilm day 73 100Ω R20Ω Planktonic day 90

R20Ω Biofilm day 90

R100Ω Planktonic day 90

R100Ω Biofilm day 90 Ladder

Ladder B 1 Inoculum R Planktonic day 16 20Ω 3 2 R100Ω Planktonic day 16 6 5 4 R20Ω Planktonic day 49 8 7 R100Ω Planktonic day 49

R20Ω Planktonic day 73

R20Ω Biofilm day 73

R100Ω Planktonic day 73

R100Ω Biofilm day 73

10 9 R20Ω Planktonic day 90 R Biofilm day 90 20Ω R Planktonic day 90 100Ω 12 11 R Biofilm day 90 100Ω Ladder

Figure 5.6. DGGE profiles of 16S rRNA genes amplified from planktonic and biofilm. (A) bacterial and (B) archaeal communities enriched in MFCs with two different external resistances over the time with cellulose. Bands cut and sequenced from archaeal DGGE profile are indicated with arrows and labeled 1 to 12.

123

Band (#) Closest BLAST match Identity RDP classifier Confidence Phylum2 (accession #) (accession #) % assignment1 %

1 (EU553823) Methanobrevibacter sp. 98 Methanobrevibacter 79 (AY196673) 2 (EU553824) Methanobrevibacter 96 Methanobrevibacter 96 smithii (U55235.1) 3 (EU553825) Uncultured methanogenic 96 Picrophilus 67 archaeon (EU413649.1)

4 (EU553826) Uncultured methanogenic 98 Euryarchaeota 64 archaeon (EU413660.1)

5 (EU553827) Uncultured methanogenic 96 Methanobrevibacter 96 archaeon (EU413665.1) 6 (EU553828) Uncultured methanogenic 91 Thermofilum 60

archaeon (EU284795.1) Archaea 7 (EU553829) Uncultured archaeon 94 Methanobrevibacter 84 (AJ576161.1) 8 (EU553830) Uncultured methanogenic 98 Euryarchaeota 68 archaeon (EU413597.1)

9 (EU553831) Uncultured 98 Methanoculleus 100 Methanomicrobiaceae archaeon (AB364338.1) 10 Uncultured methanogenic 94 Archaea 61 (EU553832) archaeon (EU413660.1) 11 Uncultured archaeon 94 Thermoplasmatales 62 (EU553834) (UAR493123) 12 Uncultured methanogenic 97 Methanobrevibacter 96 (EU553835) archaeon (EU413665.1)

Table 5.2. Phylogenetic identification of the bands excised and sequenced from DGGE gels of 16S rRNA genes PCR-amplified using archaea-specific primers. 1 In cases where the confidence level was not 100% at the genus level, a taxonomic rank was chosen with equal or greater than 60% confidence value. 2 The phylotypes were assigned to phyla based on RDP II taxonomy classifications.

124

The relatively low homology of sequences in this study with those in the database may indicate that some of the sequences from the MFCs may represent novel methane- producing archaea. These archaea have yet to be recovered by cultivation methods for further characterization.

During the first seven weeks of operation, DGGE bands belonging to the order of

Methanobacteriales were detected as the major types (Table 5.2). After 13 weeks of operation, the major DGGE bands detected were shown to originate from members of the

Thermoplasmatales, non-methanogenic archaea. A unique band (9) was observed in the

DGGE profile of planktonic archaeal communities in R20Ω MFCs. This band did not, however, occur in other profiles. Sequence analysis identified this band as a member of the order Methanomicrobiales. These results suggest that there was a major shift in the archaeal population during the operation of MFCs.

DGGE bands corresponding to methanogenic archaea were identified in both planktonic and biofilm communities of the R20Ω and R100Ω MFCs after 13 weeks of operation. This presence of methanogens coincided with the lack of methane detection in the MFCs. The lack of methane may be explained as follows: 1) Methane production was below the detection limit, 2) methanogens were present but not active possibly due to unfavorable redox conditions in MFCs, and/or 3) anaerobic methane oxidation by methane-oxidizing archaea. Recent studies have shown that relatives of methane- producing archaea have the capacity to reverse methanogenesis and thereby consume methane to produce cellular carbon and energy (Valentine 2007). This metabolic pathway requires an electron acceptor (e.g., sulfate). Colonization of the anode by methane- producing archaea, which was also observed by He et al. (2005) using fluorescence in

125

situ hybridization, suggests that these microorganisms may use the electrode as an alternative electron sink in their metabolism. However, this speculation remains to be verified experimentally. The presence of methanogens on the electrode surface is unfavorable for electricity generation in MFCs, because archaeal biofilm would block the access to the electrode by anode-reducing bacteria.

Methanogens detected in MFCs here were hydrogenotrophic and thus restricted to

H2 + CO2 and formate (Ferry 1993). The interspecies hydrogen transfer between this group of methanogens and cellulolytic bacteria may be beneficial for enhanced hydrolysis of cellulose in MFCs. This is one reason that electricity generation from cellulose using pure cultures as reported by Ren et al. (2007) is not sustainable unless the employed cellulolytic bacteria do not have hydrogen sensitive hydrogenases and/or the electricity producing microorganism is capable of utilizing the produced hydrogen. The acetotrophic methanogens can, however, utilize acetate as the major electron donor for methane production, which results in reduced coulombic efficiency in MFCs. Clearly, further research is required to elucidate the impact of different groups of methanogens in

MFCs.

Metabolic inhibitors have been used in two separate studies to inhibit the process of methanogenesis during the start-up period in MFCs (He et al. 2005; Kim et al. 2005).

The use of inhibitors is, however, not sustainable practice due to the cost and possible inhibitory influences of the compounds on growth and metabolism of the anode-reducing microorganisms. Results here show that the operation of MFCs over a prolonged period can actively suppress methanogenesis, and thus improve the performance of the system.

Further research of microbial communities, interspecies interactions and processes

126

involved in electricity generation is essential to effectively design and control cellulose- fed MFCs for enhanced performance.

127

CHAPTER 6

CONCLUSIONS

6.1. Implications of this research

Microbial fuel cell (MFC) technology provides an alternative route of energy production through the direct conversion of renewable sources of biomass to electricity.

In this research it was discovered, to my knowledge for the first time, that microorganisms indigenous to bovine rumen could couple cellulose hydrolysis and fermentation to anaerobic respiration utilizing an anode as the final electron acceptor.

This finding was an important breakthrough in converting cellulose directly into electrical energy in MFCs. This represents sustainable and environmentally-friendly energy. The results have expanded the range of suitable substrates to include cellulose, the most abundant biomass resource in many parts of the world with the potential to provide a substantial alternative energy source. The conversion of cellulose to electricity was shown to be a multiple-step process, accomplished with a consortium consisting of cellulolytic, fermentative and anode-reducing microorganisms. The performance of

MFCs depends on the diversity and activity of microorganisms. Any change to fuel cell

128

design and operating conditions that affect the bacterial dynamics and metabolism results in a change in the performance of MFCs. DGGE and phylogenetic diversity analysis showed that the outcome of the bacterial enrichment in MFCs depended on the source of the inoculum, the type of substrate, and the external resistance of the circuit used in the

MFCs. The rate of electricity production in cellulose-fed MFCs is related to the rate of consecutive-step reactions: hydrolysis, fermentation, oxidation, anodic electron transport

(electricigenesis), and electron flow through the circuit. Identifying possible constraints to this process could help in designing strategies to improve the performance of MFCs.

It was shown that the circuit load influences power output and rate of electron recovery from cellulose in MFCs. The impact of circuit resistance on microbial populations in MFCs can consequently affect the bacterial metabolism in the system.

This was shown here by the changes in the profile and concentrations of SCFA metabolites through the cellulose-to-electricity process. The loading resistance thus can be used as a tool to control microbial communities in MFCs to achieve efficient conversion of the substrate into electricity and to boost the power output. The results also showed that analysis of microbial communities and their changes in MFCs can provide valuable insights into electrochemical performance characteristics.

Methanogenesis was shown to affect the performance of MFCs during the early stages of the operation by reducing current output and coulombic efficiency. It affected growth and metabolism of hydrogen-producing cellulolytic bacteria, and changed cellulose metabolism pathways. Archaea occupied the anode surface, possibly competing for physical space in the biofilm. However, the operational conditions of MFCs suppress the methanogenic activity over the time. The interspecies hydrogen transfer between

129

cellulolytic bacteria and methanogens might be beneficial for enhanced hydrolysis of cellulose in MFCs. As the technology advances and performance of MFCs is improved, research is required to understand these complex interplays.

6.2. Suggestions for future research

MFCs are valued as a promising technology for simultaneous production of renewable electricity and treatment of organic wastes. The scientific and technical knowledge and experience obtained from this research expand the applications of this technology and the diversity of the fuels that could be utilized. However, application of cellulose-fed MFCs requires further improvements in the power output and efficiency of the systems. Such improvements necessitate identification, characterization and optimization of biological, electrochemical and engineering factors that affect the performance of MFCs.

Efforts have been made and must continue on evaluation and optimization of structural designs, cathode catalysts and configuration, and characterization of existing membranes while also searching for alternatives. The goal in these efforts must be to overcome activation, ohmic, and mass transport overpotentials in both the anode and the cathode compartments.

Advancement of cellulose-fed MFCs also requires fundamental understanding of the biological factors that influence the performance of these systems. Research must address the selection and characterization of robust microorganisms to enhance electricity generation from cellulose. To select microbial consortia that function at the maximum rate of converting cellulose to electricity with high power output, enrichment strategies

130

need to be improved that consider the effect of system engineering design and operational factors. More work is also needed to obtain detailed fundamental information on the physiology and ecology of the microorganisms enriched in MFCs. The most critical step in the MFC process is the transfer of electrons from the bacteria to the electrode. Thus the mechanisms associated with anodic electron transfer and the metabolic pathways governing the conversion of cellulose to electricity in MFCs need to be investigated.

Understanding the prevailing side reactions in the anode compartment (e.g., methanogenesis and aerobic oxidation) that bypass anodic electron transfer and designing strategies for their proper control and inhibition are required. These strategies should lead to improvement in substrate oxidation and coulombic efficiency of MFCs.

Research should also focus on model microorganisms and the isolation, identification, and characterization of key microorganisms responsible for electricity generation from cellulose. Development of a single biocatalyst, using recombinant DNA technology and novel isolation methods, that can accomplish both cellulose metabolism and anode-reduction remains a target for future research. The information obtained from single organisms can then be applied to complex microbial communities.

The current level of knowledge is scarce on interspecies relationship among microorganisms and their interactions with various design and operational factors in

MFCs. There are yet many questions unanswered at the interface of biology and electrochemistry in MFCs; 1) How do microbial communities evolve to respire with the anode as an electron acceptor and maintain that function?, 2) What are the underlying principles of microbial enrichment in MFCs?, 3) What community structure and functional group are optimal for electricity production?, 4) What operational conditions

131

promote formation of the desired community with its desired function?, 5) Are the conditions and outcomes predictable, reproducible, and controllable?, and, 6) What specific methods and procedures are required to control the conditions?

The cellulose-fed MFC is an emerging technology with environmental and sustainability advantages that is yet far from its potential capacity. As with any promising technology, such as solar cells when they emerged 30 years ago, it is most likely that breakthrough advances in science and technology will make the goal of economically viable MFC-electricity generation a reachable goal.

132

REFERENCES

Abulfotuh F. 2007. Energy efficiency and renewable technologies: the way to sustainable energy future. Desalination 209(1-3):275-282.

Aelterman P, Rabaey K, Clauwaert P, Verstraete W. 2006. Microbial fuel cells for wastewater treatment. Water Science and Technology 54(8):9-15.

Aelterman P, Freguia S, Keller J, Verstraete W, Rabaey K. 2008. The anode potential regulates bacterial activity in microbial fuel cells. Applied Microbiology and Biotechnology 78(3):409-418

Allen RM, Bennetto HP. 1993. Microbial fuel-cells - electricity production from carbohydrates. Applied Biochemistry and Biotechnology 39:27-40.

Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic local alignment search tool. Journal of Molecular Biology 215(3):403-410.

Amirinejad M, Rowshanzamir S, Eikanic MH. 2006. Effects of operating parameters on performance of a proton exchange membrane fuel cell. Journal of Power Sources 161(2):872-875.

Angenent LT, Karim K, Al-Dahhan MH, Domiguez-Espinosa R. 2004. Production of bioenergy and biochemicals from industrial and agricultural wastewater. Trends in Biotechnology 22(9):477-485.

Bergel A, Feron D, Mollica A. 2005. Catalysis of oxygen reduction in PEM fuel cell by seawater biofilm. Electrochemistry Communications 7(9):900-904.

Biffinger JC, Pietron J, Ray R, Little B, Ringeisen BR. 2007a. A biofilm enhanced miniature microbial fuel cell using Shewanella oneidensis DSP10 and oxygen reduction cathodes. Biosensors & Bioelectronics 22(8):1672-1679

133

Biffinger JC, Ray R, Little B, Ringeisen BR. 2007b. Diversifying biological fuel cell designs by use of nanoporous filters. Environmental Science & Technology 41(4):1444-1449.

Bilgen S, Kaygusuz K, Sari A. 2004. Renewable energy for a clean and sustainable future. Energy Sources 26(12):1119-1129.

Blaut M, Muller V, Gottschalk G. 1992. Energetics of methanogenesis studied in vesicular systems. Journal of Bioenergetics and Biomembranes 24(6):529-546.

Bond DR, Holmes DE, Tender LM, Lovley DR. 2002. Electrode-reducing microorganisms that harvest energy from marine sediments. Science 295(5554):483-485.

Bond DR, Lovley DR. 2003. Electricity production by Geobacter sulfurreducens attached to electrodes. Applied and Environmental Microbiology 69(3):1548- 1555.

Bond DR, Lovley DR. 2005. Evidence for involvement of an electron shuttle in electricity generation by Geothrix fermentans. Applied and Environmental Microbiology 71(4):2186-2189.

Bridgwater T. 2006. Biomass for energy. Journal of the Science of Food and Agriculture 86(12):1755-1768.

Bullen RA, Arnot TC, Lakeman JB, Walsh FC. 2006. Biofuel cells and their development. Biosensors & Bioelectronics 21(11):2015-2045.

Caccavo F, Coates JD, RosselloMora RA, Ludwig W, Schleifer KH, Lovley DR, McInerney MJ. 1996. Geovibrio ferrireducens, a phylogenetically distinct dissimilatory Fe(III)-reducing bacterium. Archives of Microbiology 165(6):370- 376.

Chae KJ, Choi M, Ajayi FF, Park W, Chang IS, Kim IS. 2008. Mass transport through a proton exchange membrane (Nafion) in microbial fuel cells. Energy & Fuels 22(1):169-176.

134

Chang IS, Moon H, Bretschger O, Jang JK, Park HI, Nealson KH, Kim BH. 2006. Electrochemically active bacteria (EAB) and mediator-less microbial fuel cells. Journal of Microbiology and Biotechnology 16(2):163-177.

Chang MCY. 2007. Harnessing energy from plant biomass. Current Opinion in Chemical Biology 11(6):677-684.

Chaudhuri SK, Lovley DR. 2003. Electricity generation by direct oxidation of glucose in mediatorless microbial fuel cells. Nature Biotechnology 21(10):1229-1232.

Cheng S, Liu H, Logan BE. 2006a. Increased performance of single-chamber microbial fuel cells using an improved cathode structure. Electrochemistry Communications 8(3):489-494.

Cheng S, Liu H, Logan BE. 2006b. Increased power generation in a continuous flow MFC with advective flow through the porous anode and reduced electrode spacing. Environmental Science & Technology 40(7):2426-2432.

Cheng S, Liu H, Logan BE. 2006c. Power densities using different cathode catalysts (Pt and CoTMPP) and polymer binders (Nafion and PTFE) in single chamber microbial fuel cells. Environmental Science & Technology 40(1):364-369.

Cheng S, Logan BE. 2007. Sustainable and efficient biohydrogen production via electrohydrogenesis. Proceedings of the National Academy of Sciences of the United States of America 104(47):18871-18873.

Clauwaert P, Rabaey K, Aelterman P, De Schamphelaire L, Ham TH, Boeckx P, Boon N, Verstraete W. 2007a. Biological denitrification in microbial fuel cells. Environmental Science & Technology 41(9):3354-3360.

Clauwaert P, Van der Ha D, Boon N, Verbeken K, Verhaege M, Rabaey K, Verstraete W. 2007b. Open air biocathode enables effective electricity generation with microbial fuel cells. Environmental Science & Technology 41(21):7564-7569.

Dehority BA. 2003. Rumen microbiology. Thrumpton, Nottingham: Nottingham University Press. 372 p.

135

Du ZW, Li HR, Gu TY. 2007. A state of the art review on microbial fuel cells: A promising technology for wastewater treatment and bioenergy. Biotechnology Advances 25(5):464-482.

Dunnett AJ, Shah N. 2007. Prospects for bioenergy. Journal of Biobased Materials and Bioenergy 1(1):1-18.

Etchebehere C, Errazquin MI, Dabert P, Moletta R, Muxi L. 2001. Comamonas nitrativorans sp nov., a novel denitrifier isolated from a denitrifying reactor treating landfill leachate. International Journal of Systematic and Evolutionary Microbiology 51:977-983.

Fan YZ, Hu HQ, Liu H. 2007. Enhanced Coulombic efficiency and power density of air- cathode microbial fuel cells with an improved cell configuration. Journal of Power Sources 171(2):348-354.

Ferry JG. 1993. Methanogenesis: ecology, physiology, biochemistry and genetics. New York, N.Y.: Chapman & Hall. 536 p.

Finkelstein DA, Tender LM, Zeikus JG. 2006. Effect of electrode potential on electrode- reducing microbiota. Environmental Science & Technology 40(22):6990-6995.

Freguia S, Rabaey K, Yuan Z, Keller J. 2007a. Non-catalyzed cathodic oxygen reduction at graphite granules in microbial fuel cells. Electrochimica Acta 53(2):598-603.

Freguia S, Rabaey K, Yuan ZG, Keller J. 2007b. Electron and carbon balances in microbial fuel cells reveal temporary bacterial storage behavior during electricity generation. Environmental Science & Technology 41(8):2915-2921.

Ghangrekar MM, Shinde VB. 2007. Performance of membrane-less microbial fuel cell treating wastewater and effect of electrode distance and area on electricity production. Bioresource Technology 98(15):2879-2885.

Gil GC, Chang IS, Kim BH, Kim M, Jang JK, Park HS, Kim HJ. 2003. Operational parameters affecting the performance of a mediator-less microbial fuel cell. Biosensors & Bioelectronics 18(4):327-334.

136

Goering HK, Van Soest PJ. 1970. Forage fiber analysis (apparatus, reagents, procedures and some applications). Agriculture Handbook No. 379. Washington, USA: Agriculture Research Service, United States Department of Agriculture.

Gralnick JA, Newman DK. 2007. Extracellular respiration. Molecular Microbiology 65(1):1-11.

Gregory KB, Bond DR, Lovley DR. 2004. Graphite electrodes as electron donors for anaerobic respiration. Environmental Microbiology 6(6):596-604.

Grzebyk M, Pozniak G. 2005. Microbial fuel cells (MFCs) with interpolymer cation exchange membranes. Separation and Purification Technology 41(3):321-328.

Hall TA. 1999. BioEdit: A user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl Acids Symp Ser 41:95–98.

He Z, Angenent LT. 2006. Application of bacterial biocathodes in microbial fuel cells. Electroanalysis 18(19-20):2009-2015.

He Z, Minteer SD, Angenent LT. 2005. Electricity generation from artificial wastewater using an upflow microbial fuel cell. Environmental Science & Technology 39(14):5262-5267.

He Z, Shao HB, Angenent LT. 2007. Increased power production from a sediment microbial fuel cell with a rotating cathode. Biosensors & Bioelectronics 22(12):3252-3255.

He Z, Wagner N, Minteer SD, Angenent LT. 2006. An upflow microbial fuel cell with an interior cathode: Assessment of the internal resistance by impedance Spectroscopy. Environmental Science & Technology 40(17):5212-5217.

Hernandez ME, Newman DK. 2001. Extracellular electron transfer. Cellular and Molecular Life Sciences 58(11):1562-1571.

Himmel ME, Ding SY, Johnson DK, Adney WS, Nimlos MR, Brady JW, Foust TD. 2007. Biomass recalcitrance: Engineering plants and enzymes for biofuels production. Science 315(5813):804-807.

137

Hobson PN, Stewart CS. 1997. The rumen microbial ecosystem. New York: Blackie Academic & Professional.

Howard RL, Abotsi E, Jansen van Rensburg EL, Howard S. 2003. Lignocellulose biotechnology: Issues of bioconversion and enzyme production. Afr J Biotechnol 2:602-619.

Ishii S, Hotta Y, Watanabe K. 2008a. Methanogenesis versus electrogenesis: Morphological and phylogenetic comparisons of microbial communities. Bioscience Biotechnology and Biochemistry 72(2):286-294.

Ishii S, Shimoyama T, Hotta Y, Watanabe K. 2008b. Characterization of a filamentous biofilm community established in a cellulose-fed microbial fuel cell. BMC Microbiol 8(1):6.

Jang JK, Pham TH, Chang IS, Kang KH, Moon H, Cho KS, Kim BH. 2004. Construction and operation of a novel mediator- and membrane-less microbial fuel cell. Process Biochemistry 39(8):1007-1012.

Jiang RZ, Chu DR. 2004. Comparative studies of methanol crossover and cell performance for a DMFC. Journal of the Electrochemical Society 151(1):A69- A76.

Jong BC, Kim BH, Chang IS, Liew PWY, Choo YF, Kang GS. 2006. Enrichment, performance, and microbial diversity of a thermophilic mediatorless microbial fuel cell. Environmental Science & Technology 40(20):6449-6454.

Kargi F, Eker S. 2007. Electricity generation with simultaneous wastewater treatment by a microbial fuel cell (MFC) with Cu and Cu-Au electrodes. Journal of Chemical Technology and Biotechnology 82(7):658-662.

Kim BH, Park HS, Kim HJ, Kim GT, Chang IS, Lee J, Phung NT. 2004. Enrichment of microbial community generating electricity using a fuel-cell-type electrochemical cell. Applied Microbiology and Biotechnology 63(6):672-681.

Kim HJ, Park HS, Hyun MS, Chang IS, Kim M, Kim BH. 2002. A mediator-less microbial fuel cell using a metal reducing bacterium, Shewanella putrefaciense. Enzyme and Microbial Technology 30(2):145-152.

138

Kim JR, Cheng S, Oh SE, Logan BE. 2007. Power generation using different cation, anion, and ultrafiltration membranes in microbial fuel cells. Environmental Science & Technology 41(3):1004-1009.

Kim JR, Min B, Logan BE. 2005. Evaluation of procedures to acclimate a microbial fuel cell for electricity production. Applied Microbiology and Biotechnology 68(1):23-30.

Kim N, Choi Y, Jung S, Kim S. 2000. Effect of initial carbon sources on the performance of microbial fuel cells containing Proteus vulgaris. Biotechnology and Bioengineering 70(1):109-114.

Knauf M, Moniruzzaman M. 2004. Lignocellulosic biomass processing: A perspective. International Sugar Journal 106(1263):147-150.

Krause DO, Denman SE, Mackie RI, Morrison M, Rae AL, Attwood GT, McSweeney CS. 2003. Opportunities to improve fiber degradation in the rumen: microbiology, ecology, and genomics. Fems Microbiology Reviews 27(5):663-693.

Kumlanghan A, Liu J, Thavarungkul P, Kanatharana P, Mattiasson B. 2007. Microbial fuel cell-based biosensor for fast analysis of biodegradable organic matter. Biosensors & Bioelectronics 22(12):2939-2944.

Lane DJ. 1991. 16S/23S rRNA sequencing. In: Stackebrandt E, Goodfellow MD, editors. Nucleic acid techniques in bacterial systematics. New York: John Wiley and Sons.

Larminie J, Dicks A. 2003. Fuel cell systems explained. New York: John Wiley & Sons. 406 p. p.

Larue R, Yu ZT, Parisi VA, Egan AR, Morrison M. 2005. Novel microbial diversity adherent to plant biomass in the herbivore gastrointestinal tract, as revealed by ribosomal intergenic spacer analysis and rrs gene sequencing. Environmental Microbiology 7(4):530-543.

Lee JY, Phung NT, Chang IS, Kim BH, Sung HC. 2003. Use of acetate for enrichment of electrochemically active microorganisms and their 16S rDNA analyses. Fems Microbiology Letters 223(2):185-191.

139

Liang P, Huang X, Fan MZ, Cao XX, Wang C. 2007. Composition and distribution of internal resistance in three types of microbial fuel cells. Applied Microbiology and Biotechnology 77(3):551-558.

Liu H, Cheng SA, Logan BE. 2005a. Power generation in fed-batch microbial fuel cells as a function of ionic strength, temperature, and reactor configuration. Environmental Science & Technology 39(14):5488-5493.

Liu H, Cheng SA, Logan BE. 2005b. Production of electricity from acetate or butyrate using a single-chamber microbial fuel cell. Environmental Science & Technology 39(2):658-662.

Liu H, Logan BE. 2004. Electricity generation using an air-cathode single chamber microbial fuel cell in the presence and absence of a proton exchange membrane. Environmental Science & Technology 38(14):4040-4046.

Liu H, Ramnarayanan R, Logan BE. 2004. Production of electricity during wastewater treatment using a single chamber microbial fuel cell. Environmental Science & Technology 38(7):2281-2285.

Liu JG, Zhao TS, Liang ZX, Chen R. 2006. Effect of membrane thickness on the performance and efficiency of passive direct methanol fuel cells. Journal of Power Sources 153(1):61-67.

Liu ZD, Li HR. 2007. Effects of bio- and abio-factors on electricity production in a mediatorless microbial fuel cell. Biochemical Engineering Journal 36(3):209-214.

Logan BE, Hamelers B, Rozendal R, Schrorder U, Keller J, Freguia S, Aelterman P, Verstraete W, Rabaey K. 2006. Microbial fuel cells: Methodology and technology. Environmental Science & Technology 40(17):5181-5192.

Logan BE, Murano C, Scott K, Gray ND, Head IM. 2005. Electricity generation from cysteine in a microbial fuel cell. Water Research 39(5):942-952.

Logan BE, Regan JM. 2006a. Electricity-producing bacterial communities in microbial fuel cells. Trends in Microbiology 14(12):512-518.

140

Logan BE, Regan JM. 2006b. Microbial Fuel Cells: Challenges and Applications. Environmental Science & Technology 40(17):5172-5180.

Lovley DR. 2006. Microbial fuel cells: novel microbial physiologies and engineering approaches. Current Opinion in Biotechnology 17(3):327-332.

Marcus AK, Torres CI, Rittmann BE. 2007. Conduction-based modeling of the biofilm anode of a microbial fuel cell. Biotechnology and Bioengineering 98(6):1171- 1182.

Martinot E, Dienst C, Weiliang L, Qimin C. 2007. Renewable energy futures: Targets, scenarios, and pathways. Annual Review of Environment and Resources 32:205- 239.

Mathis B, Marshall C, Milliken C, Makkar R, Creager S, May H. 2008. Electricity generation by thermophilic microorganisms from marine sediment. Appl Microbiol Biotechnol 78(1):147-55.

Min B, Logan BE. 2004. Continuous electricity generation from domestic wastewater and organic substrates in a flat plate microbial fuel cell. Environmental Science & Technology 38(21):5809-5814.

Min BK, Cheng SA, Logan BE. 2005. Electricity generation using membrane and salt bridge microbial fuel cells. Water Research 39(9):1675-1686.

Morris JM, Jin S, Wang JQ, Zhu CZ, Urynowicz MA. 2007. Lead dioxide as an alternative catalyst to platinum in microbial fuel cells. Electrochemistry Communications 9(7):1730-1734.

Niessen J, Harnisch F, Rosenbaum M, Schröder U, Scholz F. 2006. Heat treated soil as convenient and versatile source of bacterial communities for microbial electricity generation. Electrochemistry Communications 8(5):869-873.

Niessen J, Schröder U, Harnisch F, Scholz F. 2005. Gaining electricity from in situ oxidation of hydrogen produced by fermentative cellulose degradation. Letters in Applied Microbiology 41(3):286-290.

141

Niessen J, Schröder U, Scholz F. 2004. Exploiting complex carbohydrates for microbial electricity generation - a bacterial fuel cell operating on starch. Electrochemistry Communications 6(9):955-958.

O'Hayre R, Cha S-W, Colella W, Prinz FB. 2005. Fuel cell fundamentals. New York: John Wiley & Sons. 409 p. p.

Oh S, Min B, Logan BE. 2004. Cathode performance as a factor in electricity generation in microbial fuel cells. Environmental Science & Technology 38(18):4900-4904.

Oh SE, Logan BE. 2006. Proton exchange membrane and electrode surface areas as factors that affect power generation in microbial fuel cells. Applied Microbiology and Biotechnology 70(2):162-169.

Palmqvist E, Hahn-Hagerdal B. 2000. Fermentation of lignocellulosic hydrolysates. I: inhibition and detoxification. Bioresource Technology 74(1):17-24.

Park DH, Zeikus JG. 2003. Improved fuel cell and electrode designs for producing electricity from microbial degradation. Biotechnology and Bioengineering 81(3):348-355.

Park HI, Kim DK, Choi YJ, Pak D. 2005. Nitrate reduction using an electrode as direct electron donor in a biofilm-electrode reactor. Process Biochemistry 40(10):3383- 3388.

Park HS, Kim BH, Kim HS, Kim HJ, Kim GT, Kim M, Chang IS, Park YK, Chang HI. 2001. A novel electrochemically active and Fe(III)-reducing bacterium phylogenetically related to Clostridium butyricum isolated from a microbial fuel cell. Anaerobe 7(6):297-306.

Peu P, Beline F, Martinez J. 2004. Volatile fatty acids analysis from pig slurry using high-performance liquid chromatography. International Journal of Environmental Analytical Chemistry 84(13):1017-1022.

Pham TH, Jang JK, Chang IS, Kim BH. 2004. Improvement of cathode reaction of a mediatorless microbial fuel cell. Journal of Microbiology and Biotechnology 14(2):324-329.

142

Pham TH, Jang JK, Moon HS, Chang IS, Kim BH. 2005. Improved performance of microbial fuel cell using membrane-electrode assembly. Journal of Microbiology and Biotechnology 15(2):438-441.

Pham TH, Rabaey K, Aelterman P, Clauwaert P, De Schamphelaire L, Boon N, Verstraete W. 2006. Microbial fuel cells in relation to conventional anaerobic digestion technology. Engineering in Life Sciences 6(3):285-292.

Phung NT, Lee J, Kang KH, Chang IS, Gadd GM, Kim BH. 2004. Analysis of microbial diversity in oligotrophic microbial fuel cells using 16S rDNA sequences. Fems Microbiology Letters 233(1):77-82.

Picioreanu C, Head IM, Katuri KP, van Loosdrecht MCM, Scott K. 2007. A computational model for biofilm-based microbial fuel cells. Water Research 41(13):2921-2940.

Rabaey K, Boon N, Hofte M, Verstraete W. 2005a. Microbial phenazine production enhances electron transfer in biofuel cells. Environmental Science & Technology 39(9):3401-3408.

Rabaey K, Boon N, Siciliano SD, Verhaege M, Verstraete W. 2004. Biofuel cells select for microbial consortia that self-mediate electron transfer. Applied and Environmental Microbiology 70(9):5373-5382.

Rabaey K, Clauwaert P, Aelterman P, Verstraete W. 2005b. Tubular microbial fuel cells for efficient electricity generation. Environmental Science & Technology 39(20):8077-8082.

Rabaey K, Rodriguez J, Blackall LL, Keller J, Gross P, Batstone D, Verstraete W, Nealson KH. 2007. Microbial ecology meets electrochemistry: electricity-driven and driving communities. ISME Journal 1(1):9-18.

Rabaey K, Verstraete W. 2005. Microbial fuel cells: novel biotechnology for energy generation. Trends in Biotechnology 23(6):291-298.

Ren ZY, Ward TE, Regan JM. 2007. Electricity production from cellulose in a microbial fuel cell using a defined binary culture. Environmental Science & Technology 41(13):4781-4786.

143

Rezaei F, Richard TL, Brennan RA, Logan BE. 2007. Substrate-enhanced microbial fuel cells for improved remote power generation from sediment-based systems. Environmental Science & Technology 41(11):4053-4058.

Rhoads A, Beyenal H, Lewandowski Z. 2005. Microbial fuel cell using anaerobic respiration as an anodic reaction and biomineralized manganese as a cathodic reactant. Environmental Science & Technology 39(12):4666-4671.

Ringeisen BR, Henderson E, Wu PK, Pietron J, Ray R, Little B, Biffinger JC, Jones- Meehan JM. 2006. High power density from a miniature microbial fuel cell using Shewanella oneidensis DSP10. Environmental Science & Technology 40(8):2629-2634.

Rismani-Yazdi H, Carver SM, Christy AD, Tuovinen OH. 2008a. Cathodic Limitations in Microbial Fuel Cells: An Overview. Journal of Power Sources 180(2):683-694.

Rismani-Yazdi H, Christy AD, Dehority BA, Morrison M, Yu Z, Tuovinen OH. 2007. Electricity generation from cellulose by rumen microorganisms in microbial fuel cells. Biotechnology and Bioengineering 97(6):1398-1407.

Rismani-Yazdi H, Christy AD, Yu Z, Carver SM, Dehority BA, Tuovinen OH. 2008b. External resistance affects the power output and bacterial diversity of cellulose- fed microbial fuel cells. Environmental Science & Technology In review process.

Rozendal RA, Hamelers HVM, Buisman CJN. 2006. Effects of membrane cation transport on pH and microbial fuel cell performance. Environmental Science & Technology 40(17):5206-5211.

Rychlik JL, May T. 2000. The effect of a methanogen, Methanobrevibacter smithii, on the growth rate, organic acid production, and specific ATP activity of three predominant ruminal cellulolytic bacteria. Current Microbiology 40(3):176-180.

Schmidt TJ, Paulus UA, Gasteiger HA, Behm RJ. 2001. The oxygen reduction reaction on a Pt/carbon fuel cell catalyst in the presence of chloride anions. Journal of Electroanalytical Chemistry 508(1-2):41-47.

Schröder U. 2007. Anodic electron transfer mechanisms in microbial fuel cells and their energy efficiency. Physical Chemistry Chemical Physics 9(21):2619-2629.

144

Sopian K, Daud WRW. 2006. Challenges and future developments in proton exchange membrane fuel cells. Renewable Energy 31(5):719-727.

Srinivasan S. 2006. Fuel cells: From fundamentals to applications. New York: Springer. 691 p.

Sund CJ, McMasters S, Crittenden SR, Harrell LE, Sumner JJ. 2007. Effect of electron mediators on current generation and fermentation in a microbial fuel cell. Applied Microbiology and Biotechnology 76(3):561-568.

Tartakovsky B, Guiot SR. 2006. A comparison of air and hydrogen peroxide oxygenated microbial fuel cell reactors. Biotechnology Progress 22(1):241-246.

Ter Heijne A, Hamelers HVM, Buisman CJN. 2007. Microbial fuel cell operation with continuous biological ferrous iron oxidation of the catholyte. Environmental Science & Technology 41(11):4130-4134.

Terheijne A, Hamelers HVM, De Wilde V, Rozendal RA, Buisman CJN. 2006. A bipolar membrane combined with ferric iron reduction as an efficient cathode system in microbial fuel cells. Environmental Science & Technology 40(17):5200-5205.

Valentine DL. 2007. Adaptations to energy stress dictate the ecology and evolution of the Archaea. Nature Reviews Microbiology 5(4):316-323.

Wang Q, Garrity G, Tiedje J, Cole J. 2007. Naive bayesian classifier for rapid assignment of rRNA sequences into the new bacterial taxonomy. Applied and Environmental Microbiology 73(16):5261-7.

Weimer PJ, Lopezguisa JM, French AD. 1990. Effect of cellulose fine-structure on kinetics of its digestion by mixed ruminal microorganisms invitro. Applied and Environmental Microbiology 56(8):2421-2429.

You SJ, Zhao QL, Zhang JN, Jiang JQ, Zhao SQ. 2006. A microbial fuel cell using permanganate as the cathodic electron acceptor. Journal of Power Sources 162(2):1409-1415.

145

Yu Z, Yu M, Morrison M. 2006. Improved serial analysis of V1 ribosomal sequence tags (SARST-V1) to support examination of bacterial diversity and community composition. Environmental Microbiology 8:603-611.

Yu ZT, Morrison M. 2004a. Comparisons of different hypervariable regions of rrs genes for use in fingerprinting of microbial communities by PCR-denaturing gradient gel electrophoresis. Applied and Environmental Microbiology 70(8):4800-4806.

Yu ZT, Morrison M. 2004b. Improved extraction of PCR-quality community DNA from digesta and fecal samples. Biotechniques 36(5):808-812.

Zhao F, Harnisch F, Schröder U, Scholz F, Bogdanoff P, Herrmann I. 2005. Application of pyrolysed iron(II) phthalocyanine and CoTMPP based oxygen reduction catalysts as cathode materials in microbial fuel cells. Electrochemistry Communications 7(12):1405-1410.

Zhao F, Harnisch F, Schrorder U, Scholz F, Bogdanoff P, Herrmann I. 2006. Challenges and constraints of using oxygen cathodes in microbial fuel cells. Environmental Science & Technology 40(17):5193-5199.

Zuo Y, Cheng S, Call D, Logan BE. 2007. Tubular membrane cathodes for scalable power generation in microbial fuel cells. Environmental Science & Technology 41(9):3347-3353.

146