Scale-up microbial fuel cell as waste-to-energy system for the Colorado Convention Center

Material Information

Scale-up microbial fuel cell as waste-to-energy system for the Colorado Convention Center
Zinner, Dania Joyce
Publication Date:
Physical Description:
xi, 61 leaves : illustrations ; 28 cm


Subjects / Keywords:
Microbial fuel cells -- Colorado -- Denver ( lcsh )
Waste products as fuel -- Colorado -- Denver ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Includes bibliographical references (leaves 55-61).
General Note:
Department of Civil Engineering
Statement of Responsibility:
by Dania Joyce Zinner.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
747102882 ( OCLC )
LD1193.L53 2011m Z56 ( lcc )

Full Text
Dania Joyce Zinner
B.S., University of Colorado Boulder, 2009
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Master of Science
Civil Engineering

This thesis for the Master of Science
degree by
Dania Zinner
has been approved
Zhiyong Jason Ren
Anu Ramaswami
Felix Flechas

Zinner, Dania Joyce. (M.S., Civil Engineering)
Scale-Up Microbial Fuel Cell as a Waste-to-Energy System for the Colorado
Convention Center.
Thesis directed by Assistant Professor Zhiyong Jason Ren.
Worldwide concerns of resource scarcity and climate change are driving the search
for carbon-neutral, renewable energy alternatives for fossil fuels. Organic wastes such
as food waste represent an abundant domestic resource for energy production.
Recognizing the potential embedded in organic waste, various energy conversion
technologies have been developed. Food waste from the Colorado Convention Center
in Denver, Colorado can be used as the substrate for a microbial fuel cell (MFC)
reactor, a newly developed technology that directly converts waste to energy.
Microbial fuel cells (MFCs) are bio-electrochemical reactors that use microbes as
biocatalysts and convert biodegradable resources into electricity. Previous lab-scale
experiments using Colorado Convention Center food waste showed a sustainable
power density of 155 mW/m with a waste reduction of approximately 70%. Based on
the encouraging preliminary data, a scaled-up version of a MFC reactor was designed
and constructed as a continuous flow-through system using an air cathode. The most
cost-effective materials were used to manufacture this reactor including activated
carbon cloth and stainless steel current collector to test the feasibility of scaling up a
MFC reactor. A new coating, polydimethylsiloxane (PDMS), was also used to

construct the scale-up MFC as a tubular reactor so that the cathode would have more
a holding tank for the scale-up MFC to assist with hydrolysis, which also helped to
compare a MFC with a closed biodegradation system like anaerobic digestion.
Several parameters including chemical oxygen demand (COD) loading and hydraulic
retention time (HRT) were optimized to achieve the best performance. The scale-up
MFC was discovered to also have a waste reduction of approximately 70%, with the
highest COD removal reaching about 90% and a power density around 19 mW/m or
1,865 mW/m A life cycle assessment (LCA) was also conducted to compare the
energy input of manufacturing the scale-up MFC with the fossil fuels displaced from
its electricity generation. This was also compared to food waste sent to a landfill and
composting. These results will be used to help determine the feasibility of an on-site
pilot scale MFC reactor for the Colorado Convention Center.
This abstract accurately represents the content of the candidates thesis. I recommend
its publication.
efficient oxygen reduction capabilities. In addition, a fermentation chamber served as

I dedicate this thesis to my boyfriend for all his help and support. I also dedicate this
thesis to my parents who always motivated me and encouraged me to work hard and
pursue my goals.

My thanks to my advisor, Dr. Zhiyong Jason Ren, for his contribution, guidance, and
support of my research. I also wish to thank all the members of my committee for
their valuable participation and insights. I would especially like to thank Clinton
Global Initiative University and the Wal-Mart Foundation for their grant to my
research and the support of EPA Grant No. EPA-HQ-OPPT-2010-02. Also, I am
grateful to the Colorado Convention Center for their support and donations of food
waste. In addition, I would like to thank the students in my research group who
helped with my research.

1. Introduction............................................................1
1.2 Purpose of Study....................................................... 2
1.3 Scope of Study........................................................4
1.3.1 Scaling Up Microbial Fuel Cells..................................4
1.3.2 Design...........................................................5 Stainless Steel vs. Titanium.................................5 PDMS vs. PTFE................................................5
1.3.3 Operation and Optimization.......................................6
1.3.4 Life Cycle Assessment (LCA)......................................6
2. Scale-Up MFC Construction and Operation..................................8
2.1 MFC Construction......................................................8
2.1.1 Cathode Structure................................................8
2.2 Food Waste...........................................................12
2.2.1 COD.............................................................13
2.2.2 Total Solids, Volatile Solids, Fixed Solids.....................14
2.3 MFC Operation........................................................15
2.3.1 Inoculum & Medium...............................................16
2.3.2 Fed-Batch Flow..................................................17

2.3.3 Continuous Flow
3. Preliminary Data.....................................................18
3.1 Titanium vs. Stainless Steel......................................18
3.2 PTFE vs. PDMS.....................................................19
3.3 Activated Carbon Cloth vs. Platinum-Coated Carbon Cloth...........20
3.4 Fermentation......................................................20
4. Results............................................................ 23
4.1 Baseline Measurements and Calculations............................23
4.1.1 Polarization and Power Density Curve.........................23
4.1.2 Calculations.................................................24
4.2 Simulated Food Waste..............................................26
4.3 COD Optimization..................................................26
4.4 HRT Optimization..................................................28
4.4.1 HRT Results..................................................29
4.4.2 Methane Production of Scale-Up MFC...........................31
5. Life Cycle Assessment for Scale-Up MFC...............................32
5.1 Boundaries of LCA.................................................32
5.1.1 Functional Unit..............................................33
5.1.2 System Boundaries of Scale-Up MFC and DADS Landfill..........33
5.2 Nominal Output/Input Energy Efficiencies..........................35
5.3 Life Cycle-Based Energy Efficiencies..............................37
5.3.1 Life Cycle-Based Energy Efficiency for DADS Landfill.........38
5.3.2 Economic Analysis of Scale-Up MFC............................40
5.3.3 Life Cycle-Based Energy Efficiency for the Scale-Up MFC......41
5.4 Energy/COD Removal of Scale-Up MFC vs. DADS Landfill..............44

6. Discussion........................................................45
6.1 Scaling Up a Microbial Fuel Cell...............................45
6.2 Life Cycle Energy Comparison of Scale-Up Microbial Fuel Cell with
DADS Landfill and Composting..................................... 47
A. Calculations for Experiment Results...............................50
B. LCA Methodology...................................................52

1.1 Diagram of a microbial fuel cell (Biomass, 2009)...........................3
2.1 Activated carbon cloth.....................................................9
2.2 Stainless steel mesh......................................................10
2.3 Diagram of scale-up MFC design............................................11
2.4 COD versus dilution factor................................................14
2.5 Picture of scale-up MFC set-up in lab.....................................16
3.1 Bottle reactor with brush anode...........................................18
3.2 Voltage profile for bottle reactors with PDMS and PTFE on
stainless steel mesh.....................................................19
4.1 Polarization and power density curve for scale-up MFC....................24
4.2 Maximum power density versus coulombic efficiency for different
influent COD concentrations..............................................28
4.3 Maximum power density vs. COD removal for different HRTs..................30
5.1 System boundaries of scale-up MFC and DADS landfill.......................34
5.2 Theoretical energy output for scale-up MFC and DADS landfill if
energy input ceased at one year..........................................37
6.1 Microbial fuel cell-powered lights on Xmas tree...........................46
A. 1 COD results for a continuous flow experiment with recycle using
scale-up MFC reactor with external resistance of 1,000 Q.................51

2.1 Electricity production and waste reduction from different food waste
sample (Kronoveter, 2010)...............................................12
2.2 Total solids, volatile solids, and fixed solids.........................14
3.1 Percentage of methane and carbon dioxide in holding tank and
scale-up MFC............................................................21
4.1 Results for different influent COD concentrations.......................27
4.2 Results for different hydraulic retention times.........................29
4.3 Methane production in scale-up MFC......................................31
5.1 Nominal output/input energy conversion efficiencies for the DADS
landfill and scale-up MFC...............................................36
5.2 Life cycle-based output/input energy conversion efficiency for
the DADS landfill.......................................................39
5.3 Economic analysis of materials used to manufacture scale-up MFC.........41
5.4 Primary energy input of materials and processes used to manufacture
and operate the scale-up MFC............................................42
5.5 Life cycle-based output/input energy conversion efficiency for
the scale-up MFC........................................................43
5.6 Comparison of primary energy output per kilogram COD removal............44
A. l SFW results for scale-up MFC with external resistance of 1000 Q.........50
B. l Economic Input-Output Life Cycle Assessment (EIO-LCA) example of
energy calculation for the specific sector of waste management services
(Carnegie Mellon, 2011)...........................................54

1. Introduction
Global concerns of environmental pollution, resource scarcity, and climate change are
driving the search for carbon-neutral, renewable energy alternatives for fossil fuels.
As cities try to reduce their fossil energy use and the associated carbon footprint,
bioenergy from waste materials is increasingly viewed as a win-win strategy, because
it offers significant economic and environmental benefits. Such strategies can
simultaneously accomplish waste treatment and generate energy locally, thus
reducing carbon emissions and benefiting the economy by providing jobs.
The U.S. produces 250 million tons of municipal solid waste (MSW) per year that has
an energy content of 0.85 terawatts per hour (TWh). More than half of this MSW is
biodegradable and can be used for bioenergy production. The U.S. EPA reports that
currently more than 500 landfills have biogas utilization projects to convert landfill
gas to electricity (US EPA, 2011).
Recognizing the potential embedded in organic waste, various energy conversion
technologies have been developed. These technologies range from anaerobic
digestion biogas energy recovery to combined heat/power systems, as well as newly
developed technologies, such as microbial fuel cells (MFCs). However, the energy
recovery efficiency and reduction of carbon emissions can vary from these
technologies depending on the nature of the bioresource, transportation, and pre-
processing that may be required.

The aim of this study was to manufacture a scaled-up microbial fuel cell (MFC) using
Colorado Convention Center food waste as a substrate. This is a direct waste-to-
energy reactor with the ability to power a small strand of LED lights while treating
food waste. The goal was to create a sustainable food waste reduction system that will
offset its carbon footprint by generating electricity. A life cycle assessment (LCA)
was also prepared to calculate the amount of fossil fuels displaced from the
aforementioned electricity generation compared to the amount of greenhouse gases
released to manufacture the scale-up microbial fuel cell reactor. This was also
compared to food waste disposal at the Denver Arapahoe Disposal Site (DADS)
1.2 Purpose of Study
Organic solid wastes such as food waste represent an abundant domestic resource for
bioenergy production. New methods of energy production are being researched as a
result of the rising cost and limited amount of fossil fuel supplies. Microbial fuel cells
(MFCs) are a developing renewable energy system that can directly convert
biodegradable resources to electricity. This process entails the use of exoelectrogenic
bacteria as catalysts to degrade organic matter and transfer electrons to an electrode
(Logan, 2008). Microbial fuel cells can use food waste as the organic matter that fuels
the reactor. The organic matter stream flows through the anaerobic chamber that
contains an anode, which provides the surface area for the bacteria to consume the
substrate and deliver electrons to the anode. The cathode is positively charged and is
the equivalent of an oxygen sink at the end of an electron transport chain when using
an air cathode. The anode and cathode are connected with wires and a resistor to

complete the circuit (Logan, 2008). Figure 1 shows a pictorial diagram of the bio-
electrochemical process of a microbial fuel cell.
/R'lobwl Fuel Cell
Figure 1.1: Diagram of a microbial fuel cell (Biomass Magazine, 2009)
The purpose of this study was to develop a scale-up reactor that could handle a high
loading of food waste from the Colorado Convention Center. Different parameters
were tested to achieve the best and most stable performance of the reactor to
determine whether a pilot scale reactor would be feasible. The experiment also
explored the ability of a MFC reactor to power a strand of LED lights. The final
product would be a larger MFC reactor powered by food waste that could generate
enough electricity to be sustainable.

1.3 Scope of Study
The scope of this research involves many factors. First, a scaled-up design was
chosen to develop an on-site reactor for the Colorado Convention Center to handle
their food waste. The design of the cathode was also unique in that low-cost materials
were chosen to show that a cost-effective reactor could be constructed. Lastly, the
reactor was operated in both fed-batch and continuous-flow modes as well as
optimized by testing chemical oxygen demand (COD) loading and the different
hydraulic retention times (HRTs).
1.3.1 Scaling Up Microbial Fuel Cells
In practical applications, scaling up MFCs makes sense to achieve more power in an
attempt to make these reactors commercially available. To accomplish this, the
reactor needs to have a high performance and be relatively inexpensive to
Logan (2010) asserts that scale-up laboratory MFCS are those over 1 liter, where only
a few studies have been done because of the continuous pumping of large volumes of
medium. These studies have explored mediator- and membrane-less reactors, baffled
reactors convenient for stacking, and tubular MFCs (Jang et al., 2004; Li et al., 2008;
Scott et al., 2007). Due to the high surface area and exposure to oxygen, a tubular
MFC design was chosen for this scale-up MFC reactor to accommodate an air
cathode. The peak power performance for a scaled-up tubular MFC was shown to be
30 mW/m2 (Scott et al., 2007).

1.3.2 Design
An air cathode was chosen for this scale-up design because the reactor would only
need to have a diffusion layer that is oxygen permeable and exposed to the air.
Oxygen becomes the electron acceptor and is much more inexpensive than other
electron acceptors such as ferricyanide. A graphite fiber brush was chosen for the
anode and a combination of activated carbon cloth, stainless steel mesh, and
polydimethylsiloxane (PDMS) was chosen as the materials for the cathode. The
activated carbon cloth was used as the catalyst instead of platinum to minimize the
cost of manufacturing the reactor. Stainless Steel vs. Titanium
Stainless steel and titanium were compared as current collectors for the reactor. The
objective was to avoid precious metals such as platinum and titanium and to use
cheaper and widely available metals such as stainless steel to manufacture the scale-
Two different diffusion layers were tested to use as a cathode for the scale-up MFC.
PTFE has been commonly used with carbon cloth to act as a diffusion layer and
prevent water leakage. While this coating works well for bottle reactors and smaller
reactors, it could not be water sealed on a large scale because the weave of the carbon

mesh is too loose. Zhang et al. (2010) found this also to be true. Therefore,
polydimethylsiloxane (PDMS) was used for the cathode applied to stainless steel
mesh. This diffusion layer material proved to be water tight when applied in a
consistent fashion and is also oxygen permeable (Zhang et al., 2010).
1.3.3 Operation and Optimization
With this cost-effective design, the goal was to optimize performance to achieve the
highest power density. This was tested in both fed-batch mode and continuous-flow
mode. Different COD concentrations were tested to determine the optimal substrate
loading as well as different HRTs were evaluated to see what the best retention time
is when running this reactor continuously.
1.3.4 Life Cycle Assessment (LCA)
A life cycle assessment (LCA) was conducted to calculate the energy input and
greenhouse gas emissions released to manufacture and run the scale-up MFC. A LCA
is a methodology that examines the environmental impacts of a product, process, or
service from beginning to end or cradle to grave. These results were compared with
fossil fuels displaced by electricity generation from the scale-up MFC, which is a
carbon neutral energy source. This was also compared to other food waste disposal
alternatives such as landfills and composting.
Foley et al. (2011) compared MFCs with anaerobic treatment at wastewater treatment
plants with a LCA. They found that the positive environmental impact for both

technologies was the displacement of fossil-fuel based energy while the negative
environmental impacts for anaerobic treatment are electricity consumption and
transportation of biosolids and the negative impact for MFCs is the resource and
emissions-intensive materials needed to construct the MFC. Under the
IMPACT2002+LCA framework this study used, the positive benefits outweighed the
negative impacts, although the uncertainty for the MFC option was high and the
results were contingent on a high performance of the MFC. They state that achieving
the 1000 A-m target at 0.5 V net voltage will be a major challenge (Foley et al.,
2010). This is the only LCA study on MFCs that we know about thus far.

2. Scale-Up MFC Construction and Operation
The following describes the scale-up MFC construction with specific details on
materials and the operation of a continuous-flow, tubular MFC with an air cathode.
This includes manufacture details and literature review comparisons of different
2.1 MFC Construction
The MFC consists of an anode, cathode, and various other materials to complete a
circuit and generate electricity. These materials are described in detail in the
following sections.
2.1.1 Cathode Structure
The materials chosen for the cathode were based on the factors of: most reliable
structure, highest potential power density, and cost-effectiveness. The key element to
a sustainable scale-up microbial fuel cell (MFC) reactor is the use of inexpensive
materials. Scaling up a MFC reactor means an increased size of reactor and thus,
more material. To make the construction of a large reactor feasible, the parts must be
relatively inexpensive and widely available. To date, many microbial fuel cell
reactors use platinum or titanium, which are precious metals and therefore, rare and

expensive. These help improve power density but may not be practical in the long run
due to the reasons above (Logan, 2010). Therefore, these metals were avoided as
catalysts and current collectors. PTFE was also considered when applied to carbon
cloth as a binder, but this material is also expensive (Logan, 2010), and was not
strong enough to water seal the carbon cloth, which has a loose weave (Zhang et al.,
The following three materials were selected to form the cathode structure:
Zorflex Activated Carbon Cloth (ACC) : This material was chosen because of its
high surface area and good adsorption capacity (Deng et al., 2009). Deng et al. (2009)
found that ACC performed better than carbon felt and Pt-coated carbon paper because
of its high surface area thus producing a low cathodic overpotential. The cost is also
relatively inexpensive at $40 per square meter (Chemviron Carbon, FM70). A piece
of activated carbon cloth is shown in Figure 2.1.
Figure 2.1: Activated Carbon Cloth

Stainless Steel Mesh: This material (Type 304, McMaster-Carr) was chosen because it
has been shown to an excellent current collector. Zuo et al. (2008) found that stainless
steel mesh applied to an Anion Exchange Membrane (AEM) increased the power by
28 percent. A roll of stainless steel mesh is shown in Figure 2.2.
Figure 2.2: Stainless Steel Mesh
Poly(dimethylsiloxane) (PDMS): This is a 10:1 mixture of base and curing agent
(Sylgard 184 Elastomer Kit, Dow Coming) used as a diffusion layer applied to the
air-side of the stainless steel mesh. PDMS has a flexible Si-0 structure with methyl
substituents. This means that PDMS is oxygen permeable, but also highly
hydrophobic therefore it is able to provide a watertight sealunlike the PTFE coating
(Zhang et al., 2010). Carbon black is also mixed with the base and curing agent for
extra conductivity resulting in the black color of the coating.
The materials above and a brush anode complete the design of the tubular MFC with
air cathode. The cathode structure is shown below in Figure 2.3.

Steel Mesh
Food Waste
Activated Carbon
Cloth (ACC)
Figure 2.3: Diagram of scale-up MFC design
The activated carbon cloth is on the inside with the PDMS-coated stainless steel mesh
wrapped around the cloth, which is all supported by a Plexigas cylinder with many
5/8 diameter holes. The food waste solution flows through the carbon fiber brush
anode. The above design is connected with tubing (3/8 ID, Tygon Tubing) to a

holding tank (2 liters) and attached to a datalogger with a resistor in between that
gives the voltage of the reactor.
2.2 Food Waste
To characterize the energy recovery potential from different residential wastes, a PhD
student (K. Kronoveter) at the University of Colorado Denver collected samples from
the Colorado Convention Center, Denver International Airport and Denver Botanical
Gardens, and tested the wastes for direct electricity production in microbial fuel cells
(see Table 2.1).
Table 2.1: Electricity production and waste reduction from different food waste
samples (Kronoveter, 2010)_____________________________________________^___
Sample Source SFW CCC DBG DIA Acetate
Power Density (mW/m2) 170 155 145 180 229
COD Removal (%) 74.1 73.1 69.9 74.0 73.2
SFW: Simulated Food Waste, CCC: Colorado Convention Center, DBG: Denver
Botanical Garden, DIA: Denver International Airport
According to the research done above and because of the closeness of proximity,
Colorado Convention Center food waste was chosen as the substrate for the scale-up
MFC. Food waste was collected from a banquet on December 16, 2009. The food
waste included ham, rolls, and salad with various sauces. The food waste was
characterized by measuring chemical oxygen demand (COD), total solids (TS),
volatile solids (VS), and fixed solids (FS). For a more thorough study on the
hydrolysis and characteristics of food waste as a substrate for a MFC, please refer to
K. Kronoveters work (Kronoveter, 2010).

Pant et al (2010) did a review of other studies that have used food wastewater as the
substrate for MFCs including brewery wastewater, chocolate wastewater, food
processing wastewater and meat processing wastewater (Feng et al., 2008; Wen et al.,
2009; Patil et al., 2009; Oh et al., 2005; Heilmann et al., 2006). The largest current
density at maximum power was the chocolate industry wastewater at 1.5 W/m (Pant
et al., 2010; Patil et al., 2009).
2.2.1 COD
The results for COD are below (Figure 2.4). Chemical oxygen demand is an indirect
measurement of the amount of organic matter in water. Food waste from the Colorado
Convention Center was diluted with deionized (DI) water to create the food waste
solution to be pumped into the MFC. Each point below is the average of tests done in

1000 l 1
Ml E 800 i
Q 600
O u 400 --
0 i 0
100 200 300 400 500 600 700
Dilution Factor
Figure 2.4: COD versus dilution factor
The COD for the food waste solution is in the optimal range when diluted by a factor
of 80 to 640. The COD curve decreases by a power of negative one with an increased
dilution factor because the food waste substrate is very complex. A dilution factor of
in the range of 80 to 160 was chosen for most of the experiments.
2.2.2 Total Solids, Volatile Solids, Fixed Solids
The results for TS, VS, FS are below in Table 2.2.
Table 2.2: Total solids, volatiles solids, and fixed solids
Substrate TS VS FS
Food Waste Solution (160x) 1.28 g/L 0.76 g/L 0.52 g/L

The above values are for a food waste solution diluted by a factor of 160; this is the
dilution factor used for food waste solution in the scale-up MFC. The above
parameters represent a rough approximation of the organic matter in the food waste
2.3 MFC Operation
The scale-up MFC operation begins with the food waste solution in a 2-liter holding
tank, which is then pumped through tubing into the 1.5-liter scale-up MFC (See
Figure 2.5). The food waste is blended and then diluted with deionized (DI) water,
because it is such a complex substrate and the COD would otherwise overwhelm the
reactor. Another student working with food waste as a substrate has discovered that
the optimal COD range for a microbial fuel cell reactor is between 500 and 1,500
mg/L (K. Kronoveter, personal communication, November 29, 2010). The following
describes the inoculum, buffer solution, and fed-batch vs. continuous flow operation.

Figure 2.5: Picture of scale-up MFC set-up in lab
2.3.1 Inoculum & Medium
The food waste solution was inoculated from anaerobic digester sludge from the
Littleton/Englewood Wastewater Treatment Plant in Colorado. The medium or food
waste solution mainly consists of 20% sludge or solution with adapted bacteria and
80% DI water. A phosphate buffer solution was added (4.58 g/L Na2HPC>4, 2.45 g/L
NaH2P04, 0.31 g/L NH4CI, 0.13 g/L KC1, 10 ml/L Vitamin solution, 5 ml/L Mineral
solution). These are nutrients for the bacteria or microbes and also help to regulate
pH. A typical pH for the final solution was about 6.68. For some experiments, 250
mg/L glucose was added to jumpstart the reactorthe substrate for these experiments
is henceforth referred to as simulated food waste (SFW).

2.3.2 Fed-Batch Flow
Batch flow is the complete replacement of media once the bacteria have consumed
the substrate. The food waste solution was replaced at the end of each fed-batch cycle
or a voltage of approximately 20 mV. About 20-30 percent of the previous solution is
retained so that the system has bacteria already adapted to food waste consumption.
2.3.3 Continuous Flow
Continuous flow is mainly characterized by the hydraulic retention time (HRT). The
theoretical HRT was calculated from the volume of the medium and the flow rate into
the reactor (Huang and Logan, 2008). For example, the HRT for a flow rate of 12
mL/min was calculated to be 2 hours for the scale-up MFC and 4.5 hours for the total
HRT (scale-up MFC and holding tank). The total HRT for scale-up MFC and holding
tank was used for conducting continuous flow experiments with recyclewe
recycled the effluent of the scale-up MFC back into the holding tank.
Di Lorenzo et al (2009) found that to achieve the highest energy recovery and COD
removal rate, the optimal HRT for their air cathode/disc anode stack MFC was 17
hours. In contrast, they calculated that an upflow anaerobic sludge blanket (UASB)
reactor could use HRTs as low as 4-8 hours. Other MFCs were operated at HRTs of
anywhere from 2.13 hours (Wen et al., 2010) to 24 hours (Zuo et al., 2007).

3. Preliminary Data
The data below informed the decision on which materials to use for the scale-up MFC
reactor. This data was gathered using a microbial fuel cell bottle reactor which as a
volume of 250 milliliters (See Figure 3.1). Bottle reactors are operated in fed-batch
f \
Figure 3.1: Bottle reactor with brush anode
3.1 Titanium vs. Stainless Steel
The first decision we made based on preliminary data was which metal to use for
current collection. The voltage results were very similarboth voltage profiles
showed a stable performance around 0.4 volts. The external resistance used for the
bottle reactors was 1,000 Ohms.

Stainless steel mesh was chosen as the current collector for the scale-up MFC because
it has the same performance as a titanium rod and is a less expensive metal. The cost
of type 304 stainless steel mesh is less than $50/m (Zhang et al., 2010) and can be
used as an integral part of the microbial fuel cell cathode by wrapping it around the
Plexiglas cylinder and coating it with polydimethylsiloxane (PDMS).
3.2 PTFE vs. PDMS
The bottle reactor results for PTFE on stainless steel mesh versus the PDMS on
stainless steel mesh bottle reactor results are below (Figure 3.2).
PDMS vs. PTFE Bottle Reactor
Figure 3.2: Voltage profile for bottle reactors with PDMS and PTFE on stainless steel mesh

The bottle reactor with PDMS on stainless steel mesh reached a maximum voltage of
0.35 V and kept a steady 0.2-0.25 V for 120 hours (5 days). This is similar to PTFE
on stainless steel mesh but not as high of voltage. However, PDMS was chosen for
the scale-up MFC reactor because of its ability to be a watertight coating over a large
surface area. In addition, PTFE is a more expensive binder at an average of $60/m2
(Logan, 2010) as opposed to PDMS, which has an estimated cost of $0.13/m (Zhang
et al., 2010).
3.3 Activated Carbon Cloth vs. Platinum-Coated Carbon Cloth
Another innovative material being used for oxygen reduction is activated carbon cloth
instead of a platinum catalyst. Activated carbon does not reduce oxygen as well as a
platinum coating, but its high surface area makes up for this by being a useful
material for MFCs where the current densities per electrode area are low (e.g. a
scaled-up MFC reactor). Also, platinum is a very expensive precious metal not
suitable to scale-up, while activated carbon cloth costs $40/m (P. Morgan,
Chemviron Carbon, personal communication, February 15, 2010). Compared to the
carbon cloth usually used for MFC applications with a cost of approximately
$l,000/m2 (Zhang et al., 2010), activated carbon cloth is clearly a much more cost-
effective choice for a scale-up MFC.
3.4 Fermentation
Fermentation is the process of anaerobic bacteria degrading a substrate to yield the
end products of methane (CH4) and carbon dioxide (CO2). Logan et al. (2006) states

that in a MFC, bacteria will use the substrate for fermentation and/or methanogenesis
if the bacteria are unable to use the electrode as an electron acceptor.
Other studies using a fermentation process in concert with a MFC include a carbon
monoxide fermentation chamber with MFC (Kim and Chang, 2009) and a process
combining dark fermentation, MFCs and a microbial electrolysis cell (MEC) for
hydrogen production (Wang et al., 2011).
The holding tank is technically a fermentation chamber since there are microbes in
this tank as well and the process of fermentation takes place in a closed chamber
yielding methane and carbon dioxide. To characterize the amount of methane and
carbon dioxide in the holding tank and headspace of the scale-up MFC, a sample of
gas was taken after the media had been changed and the MFC was running for a few
days. This sample was taken with a 100 \il syringe and analyzed by gas
chromatography (GC). The results can be seen in Table 3.1.
Table 3.1: Percentage of methane and carbon dioxide in holding tank and scale-up
CH4 C02 Other gases
Holding Tank -78% -22% N/A
Scale-Up MFC -71% -20% -9%
The holding tank headspace is comprised of approximately 78 percent methane and
22 percent carbon dioxide. The scale-up MFC is similar in that its headspace consists
of 71 percent methane and 20 percent carbon dioxide. Other gases were also detected
in the scale-up MFC headspacea hypothesis is that this could be hydrogen. COD
samples were also taken periodically from the holding tank. The COD gradually

declined in the holding tank from the process of fermentation. The fermentation
process would have 70 percent removal in about a week. The food waste solution
only would stay in the holding tank for a maximum of one day in this study, so the
fermentation process was assumed to have little to no effect in the holding tank.
Microbial fuel cells can remove COD at a much faster rate as can be seen in the next
chapter (Also see Appendix A).

4. Results
The following describes the preliminary results and calculations, COD optimization
results, and HRT results.
4.1 Baseline Measurements and Calculations
The polarization and power density curves are necessary to calculate the end results
and give us a depiction of the reactors performance at various external resistances.
The polarization curve compares voltage with current density while the power density
curve compares power density with current density.
4.1.1 Polarization and Power Density Curve
Below are the polarization and power density curves used to determine the optimal
external resistance for the scale-up MFC reactor in Figure 4.1. These were conducted
in batch mode.

Polarization/Power Density Curve
8 0.3
0 20 40 60 80
Current Density (mA/m2)
Power Density
Figure 4.1: Polarization and power density curve for scale-up MFC
As seen in Figure 4.1, at the highest power density, the optimal external resistance for
the maximum power density turned out to be 50 Ohms. This external resistance was
used in the following experiments.
4.1.2 Calculations
Several formulas are used to calculate the results for the scale-up MFC. These
parameters assist with evaluating the performance of a MFC. First, percent COD
removal is calculated from the influent and effluent COD (mg/L).

% COD Removal =
Influent COD Effluent COD
Influent COD
Then, the maximum power density (mW/m or mW/m ) is determined by the formula
below. The power can either be normalized by cathode surface area (SA) or volume
of wastewater (V). E stands for voltage (Volts) and R is resistance (Ohms).
Power Density =
Coulombic efficiency (CE) is the measurement of how well electrons are transferred
in an electrochemical reaction. The formula for fed-batch mode is below.
8 f / dt
F- &COD- V
CE is the integral of current or I (ampere) over time (s) divided by the Faradays
constant (F) times the change in COD (g/L) times the volume of wastewater (L).
Faradays constant is 96,485 coulombs, where 1 coulomb equals 1 ampere times 1

The coulombic efficiency above is the equation used for continuous flow mode.
V dot is the volumetric flow rate (L/s).
4.2 Simulated Food Waste
Experiments with simulated food waste (SFW) or food waste plus 250 mg/L glucose
were conducted to jumpstart the scale-up MFC and to help acclimate the microbes to
a MFC with food waste as the substrate. These experiments had a different
polarization and power density curve and used an external resistance of 1,000 Ohms.
Therefore, the results for these SFW experiments can be found in Appendix A (Table
A.l) since they are incongruous with the rest of the results in this section and
unnecessary to be included in the discussion of the scale-up MFC using food waste as
the substrate. The following experiments use pure food waste in the solution for the
4.3 COD Optimization
Based on the data collected for food waste characterization, food waste was diluted
with DI water to reach certain COD concentrations for the influent to the scale-up
MFC reactor. The different COD loading experiments were conducted in batch mode.

Table 4.1: Results for different influent COD concentrations
Influent COD (mg/L) Effluent COD (mg/L) COD Removal (%) Max Power Density (mW/rn2) Coulombic Efficiency (%)
763 166 78.2 10.31 7.30
1229 105 91.5 11.91 6.94
1420 216 84.8 15.28 6.48
*2085 263 87.4 19.44 2.11
*Hach spectrophotometer can only test COD up to 1500 mg/L so this is a rough estimate
provided by the analyzer
Table 4.1 shows the COD removal, max power density, and coulombic efficiency for
several different influent COD concentrations. The influent COD of 1229 mg/L had
the highest COD removal, but all of the COD removal percentages were around 80-
90%. The highest power density was at a COD of 2085 mg/L, but we cannot say for
sure that this is the correct value since the instrument only accurately reads to a COD
of 1500 mg/L. Therefore, we can assume that a COD above 1500 mg/L had the
highest power density, but its coulombic efficiency was the lowest. The correlation
between the max power density and coulombic efficiency for each influent COD
concentration is shown below in Figure 4.2.

CN 25
£ a 20
£ (/i c 15
Q u 0) 10
o CL X 5
£ 0
Max Power Density vs. CE
500 1000 1500 2000
-i 5.0
Max Power Density
Coulombic Efficiency
Influent COD (mg/L)
Figure 4.2: Maximum power density versus coulombic efficiency for different influent COD
The highest power density with a relatively high coulombic efficiency was the
influent COD concentration of 1420 mg/L, so we can conclude that the optimal
influent COD concentration is around 1400-1500 mg/L. The maximum power density
reached was 19 mW/m2.
4.4 HRT Optimization
The following results determine the optimal HRT for the scale-up MFC and the
methane gas generated from this ideal HRT was calculated and added to the total
power generation.

4.4.1 HRT Results
The scale-up MFC was operated in continuous flow mode to determine the ideal
hydraulic retention time (HRT). The HRT is the volume of the reactor (1.5 L) over
the flow rate so the faster the flow rate and water moves through the MFC reactor, the
lower the HRT. The results of these experiments at varying HRTs can be seen below
in Table 4.2. There was an average influent COD of 1224 mg/L.
Table 4.2: Results for different hydraulic retention times
HRT (hrs) Flow Rate (mL/min) % COD Removal Max Power Density (mW/m3) CE (%)
2 12 65.9 1600.7 0.7
3 8 69.4 1815.1 0.6
4 6 74.6 1865.0 0.6
8 3 69.9 1825.3 0.6
16 1.5 67.6 1222.3 0.5
The 4-hour HRT has the highest maximum power density (1865 mW/m ) closely
followed by the 8-hour and 3-hour retention times. The 4-hour HRT also had the
highest COD removal, although the percent COD removal for all of the HRTs were
around 70%. Also, the coulombic efficiency (CE) for all of the experiments was
about the same with a downward trend from the 2-hour HRT to the 16-hour HRT. A
24-hour HRT would have a flow rate of 1 mL/min so one could assume that it is
similar to fed-batch mode or the COD optimization results.

The graph below (Figure 4.3) shows a comparison of percent COD removal and
maximum power density in mW/m3. One axis shows max power density and the other
shows the percentage of COD removal for the HRT.
Max Power Density vs. COD Removal
^ 1600
£, 1400
g 1200
S 1000
? 800
| 600
x 400
1 200
Max Power Density
% COD Removal
Figure 4.3: Maximum power density vs. COD removal for different HRTs
Since the 4-hour retention time has the highest power density and COD removal, it
seems that the ideal HRT would be around 4 hours. Or perhaps a range of 3 to 8 hours
could be used. It was noted that the longer the HRT, the more time the reactor should
have had for COD removal, but this was not the case. This may be the result of the
uncertainty of influent COD concentration to effluent COD concentration, which
varies throughout the continuous flow experiments. With a higher HRT than 4 hours,
the power density decreased along with COD degradation; this may be due to the
accumulation of recalcitrant substrate such as cellulosic material in the reactor that

inhibited continued microbial activity for electricity production. Since the coulombic
efficiencies are all about the same, this study found that the 4-hour HRT is the
optimal HRT, because a lower HRT is preferred so more food wastewater can be
treated in less time.
4.4.2 Methane Production of Scale-Up MFC
Since the optimal HRT is 4 hours, the amount of methane generated in the headspace
was measured. The volume of the headspace was multiplied by the change in
percentage of methane in the headspace per hour as seen below in Table 4.3.
Table 4.3: Methane production in scale-up MFC
Hour Percent Methane Volume of Gas Generated (scf)
1 0% 0
2 24% 0.0097
3 42% 0.0073
4 57% 0.0061
TOTAL 0.023
Using the EPA-recommended value for methane heat content of 1,012 Btu/scf
methane (US EPA, 2011) and assuming a 33% efficiency converting the energy of a
gas to electricity, the amount of electricity generated in 4 hours equals 0.0022 kWh.
This means that the amount of electricity from methane production the scale-up MFC
could produce in one day would be approximately 0.0132 kWh. These results are
used in the following LCA comparing a scale-up MFC to the DADS landfill.

5. Life Cycle Assessment for Scale-Up MFC
The life cycle assessment (LCA) for this project will consist of a life cycle energy
analysis that compares the energy input to manufacture the scale-up MFC and the
energy output of the MFC, or the amount of fossil fuels the MFC displacesthis is
measured by greenhouse gas emissions (GHGs). If we assume that any methane
produced will also be used as a source of energy, then a microbial fuel cell reactor is a
zero carbon source of energy. This will be compared to transporting the food waste to
the DADS landfill as an off-site food disposal option.
The DADS landfill is a regional MSW landfill owned by the City and County of
Denver and operated by the Waste Management of Colorado, Inc. DADS accepts
12,000 tons of waste per day or 3.7 million tons per year (City of Denver, 2011). The
DADS landfill along with the closed Lowry landfill produce methane gas to power
the 3.2 MW energy recovery plant. The DADS landfill has 150 gas wells and
generates 1,000 standard cubic feet per minute (scfm) of landfill gas (Colorado
SWANA, 2009).
5.1 Boundaries of LCA
The functional unit to compare the scale-up MFC with the DADS landfill and
boundaries for both systems are discussed below. Then the energy inputs and outputs
are compared to determine the efficiencies of both systems.

5.1.1 Functional Unit
In a LCA, the definition of the functional unit is how the outputs of the product
systems are defined. The functional unit for this LCA is the change of COD per
kilogram food waste. Percent COD removal is the measurement of how much organic
matter the system has degraded. The timeframe chosen is one year, although the
scale-up MFC removes COD in less time than the landfill. The amount of energy
produced (or GHG emissions displaced) will be the environmental parameter
5.1.2 System Boundaries of Scale-Up MFC and DADS Landfill
In this particular life cycle energy analysis, we did not include all small
parts/materials of the scale-up MFC (or landfill) that are not essential to the operation
(e.g. tubing connectors for the scale-up MFC). This includes secondary containment
and waste materials that were reused as a part of manufacturing the scale-up MFC,
that would have been disposed of anyway. Therefore, the life cycle assessment for the
scale-up MFC and DADS landfill can be considered a cradle-to-use assessment
neglecting environmental impacts of the end products. There are so many small parts
to a MFC that it would be hard to accurately portray the amount of energy consumed
by the transport of each individual piece to the place of construction. The exact
amount of energy needed to construct and operate a landfill is also hard to come by so
this process is also simplified to only include the main processes.

Figure 5.1 shows the system boundaries of the scale-up MFC and the DADS landfill.
For the scale-up MFC, the effluent is recycled to dilute the blended food waste
(shown in blue arrows) and is pumped back into the reactor. The energy inputs of the
scale-up MFC are mixing and pumping power while the outputs are biogas and
electricity (shown as black arrows).
Energy Flow Diagram of Scale-Up MFC System
Mixing Electncity Biogas Electricity
Energy Flow Diagram of DADS Landfill
Food Waste-
Energy for Biogas
1 J

Transportation of Landfill
Food Waste
0 0
Reapplied to Landfill
Figure 5.1: System boundaries of scale-up MFC and DADS landfill

The system boundaries of the DADS landfill include transportation of waste off-site
to the landfill plus the construction and O&M costs for the landfill. The effluent is
tested for heavy metals and other hazardous chemicals and is then reapplied to the
landfill for maintenance purposes (D. Nyiro, DADS Landfill, personal
communication, April 8, 2011). The energy output of the landfill is biogas production
which is converted to electricity and sold to the local community.
5.2 Nominal Output/lnput Energy Efficiencies
The nominal efficiencies of the DADS landfill and the scale-up MFC were calculated
as a baseline and a starting point to compare the two systems as can be seen in Table
5.1. The energy content of food waste is estimated as 5.2 million Btu/short ton (US
EIA, 2007). This is used to calculate the energy input. The energy output is calculated
by converting electricity to primary energy (dividing by 33% or the efficiency of a
power plant). The energy input and outputs are all standardized as primary energy, or
energy from primary sources such as coal, oil, and natural gas.


Table 5.1: Nominal output/input energy conversion efficiencies for the DADS landfill
and scale-up MFC____________________________
Primary Energy Input Assumptions Primary Energy Output Efficiency
DADS Landfill 2.7 x 1012 Btu per year Denver Arapahoe Disposal Site (DADS) receives 3.7 million short tons of total MSW per year, of which ~14% is easily biodegradable food waste (US EPA, 2011), with an energy content of 5.2 x 106 BTU/ton 2.6 x 1C)"1 Btu per year* 0.96%
Scale-Up MFC 6.3 x 104 Btu per year Assuming 11 kg per year of food waste with an energy content of 5.2 x 106 BTU/ton 4.98 x 104 Btu per year* 79%
*References: Colorado SWANA, 2009 and experimental data (see Chapter 4)
The efficiency is the energy output over the input. The energy output of the two
systems neglects the O&M energy requirements for both systems. Energy is defined
as primary energy (electricity divided by 33%) for both inputs and outputs.
The efficiencies above include a steady state assumption where we assume that the
input of food waste into the system directly affects the energy output of the system.
The following graph (Figure 5.2) shows the assumption of steady state for the two

Energy Output
0123456789 1011121314151617181920
Time (Months)
'Scale-Up MFC
DADS Landfill
Figure 5.2: Theoretical energy output for scale-up MFC and DADS landfill if energy input ceased
at one year
If the food waste or energy input ceased after 12 months, both systems would have a
gradual drop-off of energy output. Since the scale-up MFC is not as large as the
landfill and does not have as much waste input, the energy output would decrease
5.3 Life Cycle-Based Energy Efficiencies
A life cycle inventory aggregates all system-wide inputs and outputs from producing
a certain product or process. This was modified for this project to only include energy
to evaluate the primary energy output/input efficiencies. Life cycle-based energy

efficiencies were done for the scale-up MFC and the Denver Arapahoe Disposal Site
(DADS) landfill for comparison.
5.3.1 Life Cycle-Based Energy Efficiency for DADS Landfill
The timeframe for a life-cycle based energy efficiency for the DADS landfill was one
year. This was used to calculate the energy input released by calculating the economic
value of food waste disposal over this time period and then using the 2002 Purchaser
Price Model of the Economic Input-Output Life Cycle Assessment (EIO-LCA) tool
created by Carnegie Mellon University.
The life cycle inventory for the DADS landfill includes an estimated capital
construction cost of 42 million dollars over an operational lifetime of 25 years (US
EPA, 2000) and an estimate of $2 to $5 per ton O&M costs for the DADS landfill (D.
Nyiro, DADS Landfill, personal communication, April 8, 2011) as can be seen in
Table 5.2. The transportation was estimated by the EPA Waste Reduction (WARM)
Model with an input of 518,000 tons or the amount of food waste in DADS landfill
per year (US EPA, 2010).

Table 5.2: Life cycle-based output/input energy conversion efficiency for the DADS
Primary Energy Input Assumptions Primary Energy Output (Btu per year) Efficiency
Life Cycle Activities Energy Input (Btu per year)
Food Energy Content 2.7 x 1012 See Table 5.1 2.6 x 1010 0.96%
Transportation 2.0 x 1010 EPA WARM Model (20-mile distance)
O&M (EIO-LCA Sector #562000: Waste management and remediation services) 7.6 x 109 Average of $3.50 per ton (D. Nyiro, personal communication)
Capital Construction (EIO-LCA Sector #230103: Other nonresidential structures) 1.4 x 109 Annualized value ($1.3 million in 2011$) with discount rate of 10% for an operational lifetime of 25 years x 14% food waste
TOTAL 2.7 x 1012
The economic value of each of these sectors was converted to 2002$, multiplied by
14% (EPA-estimated food waste percentage in landfills) and used in EIO-LCA to
calculate the total primary energy input of the landfill. The energy output is from
methane recovery from the DADS landfill (Colorado SWANA, 2009) also multiplied
by 14%. The electrical power output from the landfill was divided by 33% efficiency
to calculate the primary energy output (Btu per year) above.

Different sectors of EIO-LCA were used to calculate the energy input from a landfill
(Carnegie Mellon, 2011). This included construction and O&M for the landfill as well
as the energy to transport the food waste to an off-site facility. The energy efficiency
of output versus input is the same as the nominal output/input energy conversion
efficiency (See Table 5.1).
5.3.2 Economic Analysis of Scale-Up MFC
Table 5.3 shows the cost in 2011 dollars of the materials used to construct the scale-
up MFC.
Table 5.3: Economic analysis of materials used to manufacture scale-up MFC
Material Size/Amount Estimated Cost ($2011) Manufacturer
Plexiglas (Thermoplastic) ~500 grams $104 Colorado Plastics
Activated Carbon Cloth 1 square foot $4 Chemviron Carbon
Stainless Steel Mesh 1 square foot $78
PDMS (Adhesive, similar to glue) 250 mL $50 Dow Coming, Inc.
Carbon Black 2 grams $1 Fuel Cell Store
Graphite Fiber Brush 2" x 20" brush $85 Gordon Brush
Tubing 4 feet $10 Tygon Tubing
Acetone (Plus other chemicals) 1 Liter $22 Fisher Scientific Inc.
Food Waste 2,360 cm3 $0 Colorado Convention Center
Pump Medium Flow $207 Control Company
TOTAL $561

Approximately $560 was spent to manufacture a scale-up MFC with holding tank for
a continuous flow system with a capacity of about two liters. A pilot scale MFC
would be an on-site system.
This cost compared to costs associated with landfill disposal and composting will be
discussed in Chapter 6.
5.3.3 Life Cycle-Based Energy Efficiency for the Scale-Up MFC
The Economic Input-Output Life Cycle Assessment (EIO-LCA) was used for all
materials. This model considers an entire economy, or all activities of all industry
sectors, to estimate the material or energy resources consumed by that industry sector.
In this way, a specific industrial sector was chosen to estimate the energy and
greenhouse gas emissions of a certain material (Carnegie Mellon, 2011). To use EIO-
LCA, all 2011 prices were converted to 2002$ by dividing by a factor of 1.212 (Sahr,
2010). The dollar amounts in Table 5.3 were converted to 2002 dollars to use in the
EIO-LCA purchaser price model. An example of the EIO-LCA cradle-to-gate
calculation is shown in Appendix B (Table B.l). The total energy input by
manufacturing and operating the scale-up MFC is shown in Table 5.4 assuming an
operational lifetime of one year.

Table 5.4: Primary energy input of materials and processes used to manufacture and operate the scale-up MFC
Construction/Materials Size/Amount Purchaser Price (2002$)/Time Primary Energy (Btu per year)
Plexiglas (Thermoplastic)1 -500 grams $85.81 3.2 x 106
Activated Carbon Cloth2 1 ft2 (~20 g) $3.67 5.4 x 104
Stainless Steel Mesh3 1 ft2(-200 g) $64.36 1.9 x 106
PDMS (Adhesive, binder)4 250 mL $41.25 6.7 x 10s
Carbon Black5 2 grams $0.83 3.8 x 101
Graphite Fiber Brush6 2" x 20" brush $70.13 2.8 x 106
Tubing7 8 4 feet $8.25 9.5 x 104
Acetone (Plus other chemicals) 1 Liter $18.15 7.1 x 105
TOTAL 9.4 x 106
O&M Blending Food Waste9 10 300 W 20 seconds 1.7 x 101
O&M Pumping10 4.8 W 1 year 4.3 x 105
1. EIO-LCA Sector #325211 (Plastics material and resin manufacturing)
2. EIO-LCA Sector #313210 (Broadwoven fabric mills)
3. EIO-LCA Sector #331110 (Iron and steel mills)
4. EIO-LCA Sector #325520 (Adhesive manufacturing)
5. EIO-LCA Sector #325182 (Carbon black manufacturing)
6. EIO-LCA Sector #331110 (Iron and steel mills) Carbon Fiber/Titanium Steel Assume carbon fiber manufacturing is similar
7. EIO-LCA Sector #326220 (Rubber and plastics hose and belting manufacturing)
8. EIO-LCA Sector #325190 (Other basic organic chemical manufacturing) Includes all chemicals added to food waste solution
9. Average power of a blender (ABS Alaskan, 2011)
10. Peristaltic pump (12 V, 0.4 A)

The above table gives the total energy input in Btu per year to manufacture and
operate the scale-up MFC.
Table 5.5 compares the energy produced from the scale-up MFC with the values
generated from Table 5.4. The measurements for the highest amount of electricity
output were calculated in Chapter 4 and divided by 33% efficiency to determine
primary energy output.
Table 5.5: Life cycle-based output/input energy conversion efficiency for the scale-up
Primary Energy Input Assumptions Primary Energy Output (Btu per year) Efficiency
Life Cycle Activities Energy Input (Btu per year)
Food Energy 6.3 x 104 See Table 5.1 4.98 x 104 0.50%
Transportation 0 On-site facility
O&M Mixing 1.7x10' Blending food waste (See Table 5.3)
O&M Pumping 4.3 x 105 See Table 5.3
Construction- Materials (Various Sectors of EIO- LCA) 9.4 x 106 See Table 5.3, Assuming an operational lifetime of one year
TOTAL 9.9 x 106
Since the power (3.2 mW) of the scale-up MFC is so small, the embodied energy
calculated by the comprehensive LCA is a lot larger. The longer the reactor runs
though, the more electricity (and methane gas) will be generated so if there were a
longer timeframe, the scale-up MFC would produce more energy.

5.4 Energy/COD Removal of Scale-Up MFC vs. DADS Landfill
Since the embodied energy or energy input of both of these systems vastly
overwhelms the energy output or energy produced, we will just compare the energy
production of each system against COD removal. Table 5.6 shows the primary energy
(in kilojoules) over kilograms COD removed by the scale-up MFC versus the primary
energy generated by food waste from the DADS landfill over kg COD removed from
this waste.
Table 5.6: Comparison of primary energy output per kilogram COD removal
Scale-Up MFC DADS Landfill
(kJ energy produced/ kg COD removed) (kJ energy produced/ kg COD removed)
6,860 84.3
The DADS landfill calculation used the amount of primary energy produced in one
year divided by the amount of COD treated in one year at a 70% removal rate. The
scale-up MFC calculations used the amount of primary energy produced in one year
over the amount of COD treated in one year, which was significantly higher than the
DADS landfill. Therefore, the amount of GHG emissions displaced per kilogram
COD removed would also be much higher for the scale-up MFC.

6. Discussion
The following provides a discussion on the results of the scale-up MFC experiments
and the comparison between a MFC system with food waste and disposal of food
waste in the DADS landfill.
6.1 Scaling Up a Microbial Fuel Cell
Figure 6.1 shows a Christmas tree with LED lights powered by microbial fuel cells.
Dr. Jae Do Park, Electrical Engineer, experimented with several MFCs and was able
to store the electricity generated from these MFCs to be able to light up a small Xmas

Figure 6.1: Microbial fuel cell-powered lights on Xmas tree
With the scale-up MFC, we found that microbial fuel cells can scale-up to treat more
COD or have a higher percentage of COD removed, but the energy production cannot
be simply scaled up by manufacturing a larger microbial fuel cell. The voltage
generated was about the same as a smaller, lab-scale MFC. Perhaps, more electricity
can be generated by using a stackable microbial fuel cell or using several lab-scale
Wang and Han (2008) found that stacking MFCs in parallel produced twice as much
power as stacking them in series. Also, Cusick et al. (2011) constructed a pilot scale
microbial electrolysis cell (MEC) fed with winery wastewater. This MEC reactor
contained 144 electrode pairs in 24 modules. A MEC is a device similar to a MFC,
but developed to convert wastewater into storable energy such as hydrogen or
methane (Cusick et al., 2011).

6.2 Life Cycle Energy Comparison of Scale-Up Microbial Fuel Cell
with DADS Landfill and Composting
The Colorado Convention Center spends approximately $15,500 per year to have
their food waste hauled to A1 Organics (L. Smith, personal communication, March 2,
2011). In 2010, the Colorado Convention Center diverted 178.6 tons of food waste to
be composted. The food waste was then combined with other waste to make
Ecogro, a Class I compost (A. Graff, A1 Organics, personal communication,
February 22, 2011). In this way, the food waste is recycled, reused, and then has an
economic value to be sold (per cubic yard) for landscaping. Lundie and Peters (2005)
have explored the issue of food waste management options by comparing LCAs of a
household sink processor, home composting, landfilling the food waste, and
centralized composting. They found that centralized composting and landfilling are
not the best options because of the energy-intense waste collection activities they
require (transport to facility). Lundie and Peters (2005) concluded that home
composting was the best option, but of course, this is not an option for the Colorado
Convention Center because they most probably do not have the space for their own
composting facility. This is a good segue into our research on a scale-up MFC that
could be used as an on-site waste-to-energy system for the Colorado Convention
The above discussion on composting and having an on-site food waste disposal
system leads to our comparison of the scale-up MFC and DADS landfill LCAs. These
life cycle energy analyses compare the energy conversion efficiencies for both

systems. The embodied energy consumed to operate a landfill such as the DADS
landfill was much higher as shown by EIO-LCA as opposed to the energy produced
from methane recovery resulting in an energy conversion efficiency of about 1%. On
the other hand, the scale-up MFC also had an embodied energy consumption much
higher than the energy output with a life-cycle based energy conversion efficiency of
about 0.5%. However, it was found that it only takes the scale-up MFC one day (24
hours) to remove 70% of the COD while it takes a landfill a lot longer to reach the
same percent COD removal.
The amount of energy in kilojoules per kilogram COD removed was a magnitude of
50 times higher for the scale-up MFC (6,860 kJ/kg COD) compared to the DADS
landfill (84.3 kJ/kg COD). This is probably because the potential for energy
conversion of a MFC is higher and it has a much lower COD loading than a landfill.
In summation, the energy input for a life cycle energy analysis conducted by EIO-
LCA (which may overestimate the values) for both the DADS landfill and the scale-
up MFC outweighed the energy output. However, the MFC cannot handle the amount
of food waste generated by the Colorado Convention Center, while a landfill and
composting center can handle approximately 180 tons per year. One would need to
scale-up the MFC a lot more to handle that amount of food waste. Nevertheless,
because the energy consumption and thus, the GHG emissions released, are so high
for landfilling and centralized composting (mostly due to transportation and O&M), it
would make sense that an on-site system would be the best option in terms of a LCA.
Therefore, the next steps for this research would be to design an on-site pilot system
for the Colorado Convention Center that would accurately portray the benefits of

having an on-site system in a LCA and would use a substance that is considered a
waste to generate some carbon-neutral energy for the Colorado Convention Center.
In addition, some lessons learned that would be helpful for the design of a pilot scale
reactor is that several MFCs in parallel work better than a large MFC, an air cathode
MFC is the most cost-effective, and a gravity-fed reactor would reduce the energy
needed to pump the food waste solution into the reactor. Also, it is useful to have a
stock of reactor effluent with microbes already acclimated to food waste on hand to
add to the food waste solution to improve performance or to jumpstart the reactor if
there are operation and maintenance issues. This knowledge would have helped with
the experiments for the scale-up MFC.
In conclusion, the Colorado Convention Center spends $15,500 annually on a system
that hauls their food waste to another location, which as we can see uses a lot of fuel
and releases carbon emissions into the air. Instead, they could use the funds above to
manufacture about 28 scale-up MFC reactors ($560 each), which could use their food
waste to generate electricity on-site. We recommend that the next steps for this
research would be to design a pilot scale MFC to help the Colorado Convention
Center reduce their waste and add to their use of carbon-neutral energy.

APPENDIX A. Calculations for Experiment Results
Below in Table A.l are the simulated food waste (SFW) results for the first scale-up
MFC manufactured that uses an external resistance of 1,000 Ohms.
Table A.l: SFW results for scale-up MFC with external resistance of 1000 Q
Experiment HRT (if applicable) Maximum Power Density (mW/m2) Coulombic Efficiency <%) COD Removal (%>
SFW Batch 2.54 1.6 78.5
SFW Continuous Flow 4.5 hrs 2.69 1.5 72.8
The amount of COD removal in a continuous flow experiment with recycle is shown
below in Figure A.l. These results were also from the scale-up MFC with an external
resistance of 1,000 Ohms. This graph shows that 70% removal for the scale-up MFC
is at 24 hours. This experiment shows that the scale-up MFC can be operated with a
recycle element that would lessen the water footprint of the reactor.

Continuous Flow Experiment with
0 10 20 30 40 50
Time (hr)
Figure A.1: COD results fora continuous flow experiment with recycle using scale-up MFC
reactor with external resistance of 1,000 Q
The COD removal, maximum power density, and CE were calculated using the
equations in Section 4.1.2. The assumptions made were that the volume in the scale-
up MFC was 1.5 liters and the volume of the holding tank was 2 liters so with a flow
rate of 12 mL/min, the total HRT would be 4.5 hours. If the experiment was
conducted without recycle, only the scale-up MFC volume was used, so since the
HRT equals the volume divided by the flow rate, then the HRT using a flow rate of
12 mL/min would be 2 hours.

APPENDIX B. LCA Methodology
Below find the calculations for the life cycle assessment of the scale-up MFC and the
comparison to the DADS Landfill.
LCA Methodology for DADS Landfill
1. Energy input: EIO-LCA. Construction, O&M, and transportation cost for
landfill; used EIO-LCA to compute the energy input (one can also use to
calculate GHG emissions released).
2. Energy output: Total power production x (tons food waste/total tons of
waste). The 18,000 MWh per year produced by the methane recovery plant
from the DADS landfill (Colorado SWANA, 2009) was multiplied by the
ratio of tons of food waste (EPA-estimated national average of 14%) per year
over total tons of municipal solid waste per year in the DADS landfill (City of
Denver, 2011).
3. Energy output per functional unit comparison: kJ/kg COD for one year.
The amount of energy from food waste for one year over 70% COD removal
times tons of food waste in one year.

LCA Methodology for Scale-Up MFC
1. Energy input: EIO-LCA. Material costs were input into EIO-LCA and
O&M (mixing and pumping) was calculated using average wattage multiplied
by the time of energy input.
2. Energy output: Electricity plus biogas (x 33% efficiency to convert to
electricity). P = E2/R = (0.4 V)2/(50 Q) = 3.2 mW. Then converted to kWh
(and GJ), after that one can use an emission factor of 1.75 lb CChe/kWh for
Denver (Ramaswami et al., 2008) to get the GHG emissions displaced.
Methane energy calculated in Chapter 4. The total electricity production
calculated in Chapter 4 was then divided by 33% to get primary energy.
3. Energy output per functional unit comparison: kJ/kg COD for one year.
The amount of energy from food waste produced in 24 hours x (365 days/24
hours) over 70% removal of food waste in 24 hours x (365 days/24 hours).
Table B.l shows a calculation of energy from EIO-LCA. As you can see, EIO-LCA is
a rough estimate of all the embodied energy that goes into a particular product or
service in an industry sector. EIO-LCA can also calculate total GHG emissions
released from the economic activity.

Table B.l: Economic Input-Output Life Cycle Assessment (EIO-LCA) example of energy calculation for the
specific sector of waste management services (Carnegie Mellon, 2011)
Sector #562000: Waste management and remediation services
Economic Activity: $1.5 Million Dollars
Displaying: Energy
Number of Sectors: Top 10
Sector Total Enerav 11 _ i Nat Coal TJ 535 Petro 1 11 Bio/Waste NonFossEle 11 U
Total for all sectors 7.97 1.48 1.98 3.49 0.256 0.760
562000 Waste management and remediation services 2.14 0.011 0.387 1.54 0 0.200
221100 Power generation and supply 1.68 1.23 0.358 0.059 0 0.039
S00700 General state and local government services 0.833 0.008 0.250 0.575 0 0
324110 Petroleum refineries 0.387 0.000 0.103 0.251 0.019 0.014
331110 Iron and steel mills 0.227 0.135 0.062 0.002 0.001 0.027
481000 Air transportation 0.223 0 0 0.223 0 0.000
484000 Truck transportation 0.204 0 0 0.202 0 0.002
211000 Oil and gas extraction 0.201 0 0.164 0.017 0 0.020
325190 Other basic organic chemical manufacturing 0.161 0.020 0.061 0.022 0.048 0.009
492000 Couriers and messengers 0.150 0 0 0.148 0 0.002

ABS Alaskan. (2011). Power consumption table. Retrieved on February 15, 2011
from http: //
Biomass Magazine. (2009). Microbial fuel cells. Retrieved on December 1, 2009
Carnegie Mellon University Green Design Institute. (2011). Economic Input-Output
Life Cycle Assessment (EIO-LCA) US 2002 (428) model. Retrieved on
February 15, 2011 from
City of Denver. (2011). Denver Arapahoe Disposal Site. Retrieved on April 5, 2011
Colorado Solid Waste Association of North America (SWANA). (2009).
DADS/Lowry gas to energy plant. Retrieved on April 5, 2011 from

Cusick, R.D., Bryan, B., Parker, D.S., Merrill, M.D., Mehanna, M., Kiely, P.D., Liu,
G., Logan, B.E. (2011). Performance of a pilot-scale continuous flow
microbial electrolysis cell fed winery wastewater. Appl. Microbiol.
Biotechnol., DOI 10.1007/s00253-011 -3130-9
Deng, Q., Li, X., Zuo, J., Ling, A., Logan, B.E. (2009). Power generation using an
activated carbon fiber felt cathode in an upflow microbial fuel cell. J. Power
Sources, DOI 10.1016/j.jpowsour.2009.08.092
Di Lorenzo, M., Scott, K., Curtis, T.P., Katuri, K.P., Head, I.M. (2009). Continuous
feed microbial fuel cell using an air cathode and a disc anode stack for
wastewater treatment. Energy Fuels, 25, 5707-5716.
Feng, Y., Wang, X., Logan, B.E., Lee, H. (2008). Brewery wastewater treatment
using air-cathode microbial fuel cells. Appl Microbiol Biotechnol, 78, 873-
Finnveden, G., Johansson, J., Lind, P., Moberg A. (2005). Life cycle assessment of
energy from solid wastepart 1: general methodology and results. Journal of
Cleaner Production, 13, 213229.
Foley, J.M., Rozendal, R.A., Hertle, C.K., Lant, P.A., Rabaey, K. (2010). Life cycle
assessment of high-rate anaerobic treatment, microbial fuel cells, and
microbial electrolysis cells. Environ. Sci. Technol, 44, 3629-3637.

Huang, L., Logan, B.E. (2008). Electricity production from xylose in fed-batch and
continuous-flow cells. Appl Microbiol Biotechnol, 80, 655-664.
Heilmann, J., Logan, B.E. (2006). Production of electricity from proteins using a
microbial fuel cell. Water Environ. Res., 78, 531537.
Jang, J.K., Pham, T.H., Chang, I.S., Kang, K.H., Moon, H., Cho, K.S., Kim, B.H.
(2004). Construction and operation of a novel mediator- and membrane-less
microbial fuel cell. Process Biochem., 39 (8), 1007-1012.
Kim, D., Chang, I.S. (2009). Electricity generation from synthesis gas by microbial
processes: CO fermentation and microbial fuel cell technology. Bioresource
Technology, 100, 4527-4530.
Kronoveter, K. (2010). Untitled. Unpublished work, University of Colorado Denver.
Li, Z., Yao, L., Kong, L., Liu, H. (2008). Electricity generation using a baffled
microbial fuel cell convenient for stacking. Bioresour. Technol., 99, 1650-
Logan, B.E. (2008). Microbial fuel cells. New Jersey: John Wiley & Sons, Inc.
Logan, B.E. (2010). Scaling up microbial fuel cells and other bioelectrochemical
systems. Appl. Microbiol. Biotechnol., DOI 10.1007/s00253-009-2378-9

Logan, B.E., Hamelers, B., Rozendal, R., Schroeder, U., Keller, J., Freguia, S.,
Aelterman, P., Verstraete, W., Rabaey, K. (2006). Microbial fuel cells:
Methodology and technology. Environ. Sci. Tech., 40 (17) 5181-5192.
Lundie, S., Peters, G.M. (2005). Life cycle assessment of food waste management
options. Journal of Cleaner Production, 13, 275-286.
National Renewable Energy Laboratory (NREL). (2011). U.S. Life Cycle Inventory
Database. Retrieved on February 18, 2011 from
Oh, S., Logan, B.E. (2005). Hydrogen and electricity production from a food
processing wastewater using fermentation and microbial fuel cell
technologies. Water Res. 39, 4673^1682.
Pant, D., Van Bogaert, G., Diels, L., Vanbroekhoven, K. (2009). A review of the
substrates used in microbial fuel cells (MFCs) for sustainable energy
production. Bioresour. Technol., doi:10.1016/j.biortech.2009.10.017.
Patil, S.A., Surakasi, V.P., Koul, S., Ijmulwar, S., Vivek, A., Shouche, Y.S.,
Kapadnis, B.P., (2009). Electricity generation using chocolate industry
wastewater and its treatment in activated sludge based microbial fuel cell and
analysis of developed microbial community in the anode chamber. Biores.
Technol., 100, 5132-5139.

Ramaswami, A., Hillman, T., Janson, B., Reiner, M., & Thomas, G. (2008). A
demand-centered, hybrid life-cycle methodology for city-scale greenhouse gas
inventories. Environmental Science Technology, 42 (17), 6455-6461.
Retrieved on February 16, 2011 from
http://pubs.acs.0rg/d0i/abs/l 0.1021 /es702992q
Sahr, Robert C. (2010). Consumer Price Index (CPI) conversion factors 1774 to
estimated 2020 to convert to dollars of 2002. Oregon State University.
Retrieved on February 15, 2011 from
Scott, K., Murano, C., Rimbu, G. (2007). A tubular microbial fuel cell. J. Appl.
Electrochem., 37, 1063-1068.
Sponza, D.L., Agdag, O.N. (2003). Impact of leachate recirculation and recirculation
volume on stabilization of municipal solid wastes in simulated anaerobic
bioreactors. Process Biochemistry, 39, 2157-2165.
US Energy Information Administration (ElA). (2007). Calculations to Obtain
Average Million Btu Per Ton for Municipal Solid Waste (MSW). Retrieved
on April 24, 2011 from

US Environmental Protection Agency (EPA). (2000). Superfimd Fact Sheet.
Retrieved on April 15, 2011 from
US Environmental Protection Agency (EPA). (2011). Landfill Methane Outreach
Program: Basic Information. Retrieved on April 15, 2011 from
US Environmental Protection Agency (EPA). (2010). Waste Reduction Model
(WARM). Retrieved on May 3, 2011 from
Wang, A., Sun, D., Cao, G., Wang, H., Ren, N., Wu, W., Logan, B.E. (2011).
Integrated hydrogen production process from cellulose by combining dark
fermentation, microbial fuel cells, and a microbial electrolysis cell.
Bioresource Technology, 102, 4137-4143.
Wang, B., Han, J. (2008). A single chamber stackable microbial fuel cell with air
cathode. Biotechnology Letters, 31, 387-393.
Wen, Q., Wu, Y., Cao, D., Zhao, L., Sun, Q. (2009). Electricity generation and
modeling of microbial fuel cell from continuous beer brewery wastewater.
Biores. TechnoL, 76/0,4171-4175.

You, S., Zhao, Q., Zhang, J., Liu, H., Jiang, J., Zhao, J. (2007). Increased sustainable
electricity generation in up-flow air-cathode microbial fuel cells. Biosensors
and Bioelectronics, (23), 1157-1160.
Zhang, F., Saito, T., Cheng, S., Hickner, M.A., Logan, B.E. (2010). Microbial Fuel
Cell Cathodes With Poly(dimethylsiloxane) Diffusion Layers Constructed
around Stainless Steel mesh Current Collectors. Environ. Sci. Technol. 44 (4)
Zuo, Y., Cheng, S., Call, D., Logan, B.E. (2007). Tubular membrane cathodes for
scalable power generation in microbial fuel cells. Environ. Sci. Technol., 41,
Zuo, Y., Cheng, S., Logan, B.E. (2008). Ion exchange membrane cathodes for
scalable microbial fuel cells. Environ. Sci. Technol., 42 (18), 6967-6972.