using microscale flow cells to study the … · araz zarkhah, chris, timberg, karl siil, and saman...
TRANSCRIPT
USING MICROSCALE FLOW CELLS TO STUDY THE
ELECTROCHEMICAL REDUCTION OF CARBON DIOXIDE
BY
MICHAEL RICHARD THORSON
DISSERTATION
Submitted in partial fulfillment of the requirements
for the degree of Doctor of Philosophy in Chemical Engineering
in the Graduate College of the
University of Illinois at Urbana-Champaign, 2012
Urbana, Illinois
Doctoral Committee:
Professor Paul J. A. Kenis, Chair
Professor Charles F. Zukoski
Assistant Professor Charles M. Schroeder
Professor Andrew A. Gewirth
Professor Thomas B. Rauchfuss
ii
Abstract
An electrochemical process to reduce CO2 has potential to store electrical energy from
renewable sources such as wind and solar in chemical form via the production of chemical fuels
and feedstocks. When combined with a “green” renewable source, this technology provides a
means of (a) reducing dependence on foreign oil via enhancing the penetration of renewable
technologies into the transportation sector and (b) reducing CO2 emissions by moving away from
fossil fuels (Chapter 1). The goal of this research is to use a microfluidic platform to study the
influence of electrolytes, reactor conditions and catalysts to address the three primary
challenges for the electrochemical reduction of CO2 to CO: (1) low reaction energetic efficiency;
(2) low reaction rate; and (3) poor reaction selectivity.
This dissertation reports an investigation of roles electrolytes, operating conditions, and
catalysts can play in the reduction of CO2. The anion and cation size, along with pH, were found
to drastically influence both the current density and selectivity of the product distribution
(Chapter 2). Specifically, small anions and large cations were found to favor the production of
CO and hinder the evolution of H2.
The microfluidic reactor was used to look at the performance of a larger cation,
EMIM BF4, an ionic liquid. With the ionic liquid-based electrolyte, an early cathode onset
potential was observed at the expense of poor anode performance. The poor anode performance
was overcome via the addition of a secondary electrolyte stream. Using the modified reactor, in
the presence of an EMIM BF4 electrolyte solution, CO production was achieved at a reactor
potential of only 1.5 V, which constitutes a maximum cathode overpotential of 0.17 V
(Chapter 3). This low onset potential constitutes a drastic reduction in the onset potential for
CO production in an arrangement that achieves excellent selectivity for CO. While the energetic
iii
efficiency in the dual electrolyte setup was drastically improved, the achieved current density
was quite low because the EMIM+ cation poisoned the cathode. Because the poor anode
performance was from the BF4 anion, the membrane could be eliminated by substituting a BF4
anion in the ionic liquid electrolyte solution for either a OH or Cl anion. The substitution of the
BF4 anion for either a OH or Cl anion enabled drastic improvements in the partial current density
for CO when using either an EMIM OH or an EMIM Cl electrolyte solution (Chapter 4).
Further improvements were achieved via bolstering the conductivity with the addition of a
secondary salt, KOH.
Amine-based novel organometallic complexes have been developed for the
electrochemical reduction of CO2 to CO (Chapter 5). Specifically, several silver-based
organometallic catalysts, AgDAT, AgPc, and AgPz, have comparable or better current densities
with increased selectivity for CO as compared to Ag nanoparticle-based catalysts. Furthermore,
when comparing the performance relative to the silver loading, the amine-based organometallic
catalysts outperform the silver catalyst by more than 20 fold. Furthermore, when using copper-
based organometallic catalysts (i.e., CuDAT), the product selectivity changed. This resultant
change in selectivity and current density demonstrates that organometallic complexes have
potential to “tune” catalysts to produce a wide range of desired products.
The electrochemical CO2 conversion reactor used in the work described here (Chapters 2
through 5) is based on a microfluidic fuel cell developed earlier in the Kenis group. Chapter 6
reports on the influence of boundary layer depletion on performance in an alkaline, air-breathing
laminar flow fuel cell as a function of electrode length, electrode arrangement, and flow rates.
Higher current densities were achieved when using shorter electrodes. Furthermore, alternative
iv
electrode arrangements (aspect ratios and multiple electrodes in series) enable higher fuel
utilization and power output per catalyst area.
Looking forward, these studies will shed insight in regards to optimal electrolyte
composition and catalyst development with the aim of further improving the current density and
energetic efficiency of the reaction.
v
To Candi Thorson
vi
Acknowledgements
I am extremely grateful for the many people who have helped to make my dissertation
possible. First, I would like to thank my advisor, Dr. Paul J.A. Kenis for providing critical
feedback and much needed advice in my dissertation. I am very fortunate to have had an advisor
who was supportive of me finding an area of research that I was passionate about. I am
appreciative of my dissertation committee, Professors Thomas Rauchfus, Andrew Gewirth,
Charles Schroeder, and Charles Zukoski for the insight they provided and time they invested to
be part of my committee.
In addition to my mentors, I have been fortunate to have excellent co-workers within the
Kenis group. First, I would like to thank Dr. Fikile Brushett and Dr. Devin Whipple, who
provided mentorship. Specifically, I would like to thank Sachit Goyal, Matt Naughton,
Dr. Adam Hollinger, Molly Jhong, and Sichao Ma who I worked most closely with. Finally, I
am grateful for the other members of the Kenis group for their scientific insight, and comradery,
particularly, Dr. Benjamin Schudel, Sudipto Guha, Dr. Sarah Perry, Dr. Amit Desai, Josh Tice,
Ritika Mohan, Dr. Ashtamurthy Pawate, Vivek Kumar, Matthew Byrne, Dr. Matthew Cole, and
Dr. Pedro Lopez-Montesinos. In addition, thanks to the hard working undergraduate students,
Araz Zarkhah, Chris, Timberg, Karl Siil, and Saman Moniri, who assisted me in data acquisition
and whom I was given the privilege to mentor.
I am grateful for my collaborators, without whom much of this work would not have been
possible. Specifically, I am grateful for my collaborators at Dioxide Materials, Dr. Richard
Masel, Dr. Amin Salehi-Khojin, Dr. Wei Zhu, and Brian Rosen whom I worked alongside in the
testing of the ionic liquid, EMIM BF4. I am also very grateful for the assistance of Claire
vii
Tornow and Dr. Andrew Gewirth from the Chemistry Department at the University of Illinois for
catalyst synthesis, testing, and analysis.
I appreciate the generous funding for this work, which was made possible by the NSF,
DOE, I2CNER, and the Harry G. Drickamer fellowship program.
I especially thank my wife, Candi Thorson for her emotional support throughout my PhD.
I appreciate my parents for supporting me and instilling in me the perceptive that there is no
substitute for hard work.
viii
Table of Contents
Chapter 1: Introduction ................................................................................................................1
1.1 Grand Challenges .................................................................................................................1
1.2 Electrochemical Reduction of CO2 ....................................................................................10
1.3 Challenges for the Electrochemical Reduction of CO2......................................................12
1.4 Economic Analysis ............................................................................................................16
1.5 Remaining Challenges .......................................................................................................17
1.6 References ..........................................................................................................................19
Chapter 2: Effect of Ions on the Electrochemical Conversion of CO2 to CO ........................25
2.1 Introduction ........................................................................................................................25
2.2 Experimental ......................................................................................................................27
2.3 Results and Discussion ......................................................................................................30
2.4 Conclusions ........................................................................................................................38
2.5 References ..........................................................................................................................40
Chapter 3: Amine Based Electrolytes for the Electrochemical Reduction of CO2 ................44
3.1 Introduction ........................................................................................................................44
3.2 Experimental ......................................................................................................................46
3.3 Results and Discussion ......................................................................................................49
3.4 Conclusions ........................................................................................................................56
3.5 References ..........................................................................................................................57
Chapter 4: Increased Current Densities for Amine-Based EMIM+ Electrolytes ..................60
4.1 Introduction ........................................................................................................................60
4.2 Experimental ......................................................................................................................61
4.3 Results and Discussion ......................................................................................................61
4.4 Conclusions ........................................................................................................................65
4.5 References ..........................................................................................................................66
Chapter 5: Organometallic Catalysts for the Electrochemical Reduction of CO2 ................68
5.1 Introduction ........................................................................................................................68
5.2 Experimental ......................................................................................................................70
ix
5.3 Results and Discussion ......................................................................................................74
5.4 Conclusions ........................................................................................................................89
5.5 References ..........................................................................................................................90
Chapter 6: Design Rules for Electrode Arrangement in an Air-breathing Alkaline Direct
Methanol LFFC ............................................................................................................................93
6.1 Introduction ........................................................................................................................93
6.2 Experimental ......................................................................................................................95
6.3 Results and Discussion ......................................................................................................99
6.4 Conclusions ......................................................................................................................107
6.5 References ........................................................................................................................108
Chapter 7: Summary of Accomplishments and Future Directions .......................................112
7.1 Summary of Accomplishments ........................................................................................112
7.2 Future Directions .............................................................................................................115
7.3 References ........................................................................................................................117
Appendix A: Supplementary Section to Chapter 6 ................................................................119
1
Chapter 1
Introduction
1.1 Grand Challenges
Carbon dioxide is nature’s chemical of choice for energy storage. Most plants store solar
energy by converting CO2 from the air into glucose, which is temporarily stored in leaves, roots,
and stems [1]. Later, the glucose is turned into starch and plant matter [1]. Here we propose to
use CO2 as a storage medium for energy from renewable sources, such as the sun, to enable
“artificial photosynthesis”. Specifically, we propose that coupling renewable energy with a
process for the electrochemical conversion of CO2 can enable penetration of non-fossil fuel
sources into the transportation sector and thereby both decrease CO2 emissions and meet the
growing demand for liquid fuels within the transportation sector. The major technical challenge
to improving the efficiency of the process for the electrochemical conversion of CO2 is achieving
high current densities and high energetic efficiencies while being selective for the desired
product.
1.1.1 Increasing Demand for Oil
Figure 1.1 shows several projections predicting an increase in demand for petroleum for
the foreseeable future. Over the last 25 years, the demand for oil has steadily increased.
Furthermore, as seen in Figure 1.1, estimates for the growth rate in demand of petroleum vary
between 1.0 and 1.9 % per year, resulting in the predicted demand for petroleum to increase 35
to 76 % over the next 30 years [2]. Much of the historical and expected demand for petroleum
stems from a growing worldwide population and an increase in the percentage of the world
population living in petroleum reliant, developed communities [2]. The world population is
2
expected to grow from roughly 7 billion people to roughly 9 billion people between 2012 and
2030 [3]. However, most of the demand for petroleum comes from people living in developed
countries. The Organization for Economic Co-operation and Development (OECD), which
represents part of the developed world, uses half of the world’s energy to produce half of the
world’s GDP [4]. The high demand for energy of the OECD (50 % of worldwide demand) is
significant because it constitutes only a small fraction of the world’s population [4]. The fraction
of the world population living within developed countries is expected to increase from 50 to
80 % by 2030 [2]. Because both the worldwide population is growing, and a larger fraction of
the world population will rely on oil for their daily lives, the number of people worldwide
dependent on oil is expected to double, effectively doubling the demand for oil by 2030.
1.1.2 Limited Supply of Oil
The increasing demand for oil becomes complicated by limitations in supply. Current
Figure 1.1: Historical and Projected world energy demand for oil between 1980 and 2030. © National Petroleum
Council 2007, reprinted with permission from “Hard Truths: Facing the Hard Truths About Energy: A
Comprehensive View to 2030 of Global Oil and Natural Gas” [2].
3
petroleum capacity, in terms of oil in established wells, is expected to be mostly depleted by
2030, however, reserves and other advanced oil recovery means are expected to provide a
petroleum supply for the foreseeable future [5]. Nonetheless, because most of the oil reserves
which will be used over the next several decades have yet to be identified, all predictions of
when the world’s oil supplies will be depleted have significant uncertainty [5,6]. Furthermore,
as traditional wells become depleted, less efficient, unconventional means of oil recovery will
have to be used, which increases the cost due to the increased technology and equipment
necessary to recovery the oil. In some cases, unconventional means of oil recovery come with
increased risks (e.g., spills from deep water drilling), which must be considered in an energy
source analysis [7].
Even if petroleum is available for the foreseeable future, other factors such as geopolitical
influences, emissions of greenhouse gasses, and energy security issues can limit the extent of oil
Figure 1.2: Production rates of oil from the leading worldwide producers of petroleum. © National Petroleum
Council 2011 reprinted with permission from “Prudent Development: Realizing the Potential of North America's
Abundant Natural Gas and Oil” [4].
4
availability or drive up the price of petroleum. For example, political tensions or embargos can
disturb the free market and thereby, hinder the availability of the petroleum supply. This need
for free trade may explain some of the motivation for the United States to declare war against
Iraq. If this was indeed motivated by the availability of oil, funding this war provides quite
significant “hidden costs” to acquiring oil, and consequently, the cost of war should be
considered when considering the price of petroleum.
A major complication, which must be considered when securing future oil supplies, is the
prevention of potential energy security issues. As Figure 1.2 shows, the United States and
Europe rely heavily on a few countries within the Middle East, and to a lesser extent, Russia and
northwestern Africa, for the majority of their imported oil. Furthermore, this trend is expected to
become exaggerated through 2030 [2,8]. Being heavily reliant on a few countries for oil imports
Figure 1.3: Historical and Projected worldwide oil imports for 2005 and 2030. © National Petroleum Council 2007,
reprinted with permission from “Hard Truths: Facing the Hard Truths About Energy: A Comprehensive View to
2030 of Global Oil and Natural Gas” [2].
5
can rapidly affect Europe’s and United States’ economies if political situations change for the
worse between these countries. This potential energy security issue led the National Petroleum
Council to recommend that the United States diversify its imports, increase modernization, and
expand oil production [2]. However, as Figure 1.3 shows, the United States is already the
world’s 3rd
largest producer of oil and has realistic limitations regarding the vastness of its wells
[2]. The United States has large reserves of Natural Gas and Coal, which are much more
extensive than the United States’ oil reserves; however, at present, penetration of these sources
into the transportation sector is minimal.
1.1.3 Linking Supply and Demand
Figure 1.4 shows the current penetration of the most common sources of energy into
various sectors of energy demand [9]. Of note, the transportation sector relies almost entirely on
Figure 1.4: Primary U.S. energy consumption by source and sector for 2008. Reprinted (adapted) with permission
from: D. T. Whipple and P. J. A. Kenis, “Prospects of CO2 Utilization via Direct Heterogeneous Electrochemical
Reduction” Journal of Physical Chemistry Letters, 2010. 1 (24): p. 3451-3458.. Copyright (2010) American
Chemical Society [9].
6
petroleum as its energy source. The reliance of the transportation sector on petroleum
compounded with the aforementioned increasingly concentrated supply of petroleum exacerbates
potential national security issues because a shortage of oil immediately limits the extent of
transportation possible. Consequently, a need exists for technology that enables penetration of
other sources of energy into the transportation sector.
A simple means to enable penetration of other energy sources into the transportation
sector relies on improving automobile efficiency through the development of hybrid or all-
electric cars that use electricity from the electric grid. While electric powered cars are in early
stages of market penetration (e.g., Chevy Volt, Nissan Leaf) the majority of the transportation
infrastructure is already engineered to support automobiles that run on liquid fuels [10].
Consequently, liquid fuels have an inherent economic advantage with the current infrastructure.
Two methods that can enable penetration of natural gas or coal to the transportation
sector include natural gas to liquid (GTL) [11-15] and coal to liquid (CTL) [14-17] technology,
which yield syngas, a mixture of hydrogen and carbon monoxide, which can be further converted
to liquid fuels through an established technology, the Fischer-Tropsch process [18-23]. While
CTL and GTL are capable of addressing issues in the foreseeable future with limited fuel supply
and energy security, they are contributing towards global warming through the emission of CO2.
Carbon Capture and Sequestration (CCS) has been proposed as a means for long term CO2
storage in underground geologic formations away from the atmosphere [24,25]. However, issues
regarding long-term viability of this technology (e.g., leakage) could potentially negate the
positive effects of this technology [26].
7
1.1.4 Global Warming
A primary advantage of
alternative fuel sources, such as
wind and solar, is their low
emission levels of CO2. This is
particularly important because the
concentration of CO2 in the
atmosphere has been increasing
since the industrial revolution, as
shown in Figure 1.5 [27-29].
Increasing CO2 concentrations in the
atmosphere have been directly linked to
an increase in global temperature,
which has been theorized to lead to
catastrophic effects on the ecosystem
[30,31]. Furthermore, as shown in
Figure 1.6, the emission levels for CO2
are expected to increase roughly 40 %
over the next 20 years [2].
1.1.5 Alternative Energy Sources
The need to reduce emissions
has renewed interest in a means to
expand penetration of “low emission”
Figure 1.5: Historical CO2 concentrations in the atmosphere from
trapped CO2 in Ice (Ice Dome) [29] and direct observation of CO2 in the
atmosphere at Mauna Loa Observatory (Mauna Loa, Hawaii). Data
from Francey, 1999 and NOAA/ESRL 2012 [28].
270
290
310
330
350
370
390
1000 1250 1500 1750 2000
CO
2C
on
cen
trat
ion
(p
pm
)
Year (AD)
Mauna Loa
Ice Dome
Figure 1.6: Historical and predicted CO2 emissions between
1980 and 2030. © National Petroleum Council 2007, reprinted
with permission from “Hard Truths: Facing the Hard Truths
About Energy: A Comprehensive View to 2030 of Global Oil
and Natural Gas” [2].
8
renewable energy sources into the transportation sector. Several potential renewable energy
sources are making progress regarding improvements to the overall efficiency, including, biofuel
production form algae or biomass [32,33]. Additionally, wind, solar, and tidal power
technologies continue to become more efficient and less expensive per kW produced [34-36].
Furthermore, along with improved efficiencies, a favorable political environment for wind and
solar technology has resulted in large scale implementation of these technologies throughout the
United States [37].
While the efficiency, politics and economics of wind and solar have drastically improved
over the last century, the intermittent nature of the technology limits the extent of their
implementation. Specifically, the DOE recommended that the combined extent of wind and
solar implementation will max-out at 15 % of the United States’ electrical energy consumption
due to intermittent nature of these technologies [38]. Figure 1.7 shows examples from (a) a wind
turbine [39] and (b) a utility-scale photovoltaic system [40] to demonstrate typical fluctuations in
the power outputs from these sources.
While the electrical grid is the largest potential market for wind and solar, the electrical
grid does not respond well to intermittency. Electrical demand depends on many factors, such as
time of day, sun intensity and outside temperature. Consequently, traditional power plants must
scale their operation to match demand. This has several effects: First, power plants must be
designed to meet max-power demand, or else, entire cities will experience power “blackouts”
when demand for power is high. Second, the price of electricity is exceptionally high when
demand is at a maximum because power plants operate less efficiently to match high demand.
Third, power is often produced at a loss in off-peak times because power plants struggle to cut
back production and consequently receive a poor return from the market due an oversupply of
9
power. Figure 1.7 demonstrates how the aforementioned effects influenced Ontario’s power
prices in July of 2011 [41]. Specifically, Ontario’s power prices fluctuated between roughly
18 ¢/kWh and roughly -7 ¢/kWh [41].
The addition of renewable sources of electricity, such as wind and solar, can potentially
help in “peak-power” demand times. However, because renewable sources of electricity, such as
wind and solar, are intermittent by nature [39,40], traditional fossil-fuel or nuclear power plants
still need to be built to meet max power demand, unless a means of storage is developed. Also,
much of the power generated from renewable sources in off-peak periods, when demand is low,
receive a minimal, if not negative, financial return for the power provided. Because the grid
does not respond well to intermittent power and because the renewable power sources discussed
above (wind and solar) are intermittent by nature, in many cases these renewable energy sources
are only economically viable with external stimuli.
Figure 1.7: (a) Power output for a wind turbine over 10 days (860,000 seconds). Reprinted from “The Spectrum of
Power from Wind Turbines”, 169, Jay Apt, 369-374, Copyright (2007) [39], with permission from Elsevier.
(b) power output for a utility-scale photovoltaic system over 6 days. Reprinted from The character of power output
from utility-scale photovoltaic systems, 16, Aimee E. Curtright, Jay Apt, 241-247, Copyright (2007), with
permission from John Wiley and Sons [40].
(a) (b)
10
1.2 Electrochemical Reduction of CO2
1.2.1 Motivation for the Electrochemical Reduction of CO2
Here, we propose exploring the electrochemical reduction of CO2 as a means of
producing chemical fuels and feedstocks to store electrical energy from renewable sources such
as wind and solar in chemical form. When combined with a “green” renewable source, such as
wind or solar, this technology provides a means of (a) reducing our dependence on foreign oil
via enabling penetration of renewable technologies into the transportation sector and (b) reducing
CO2 emissions by moving away from fossil fuels.
An electrochemical process for CO2 reduction can be tied to renewable power sources in
several ways. First, an electrochemical reactor for chemical production can use energy produced
from a renewable power source in a continuous fashion. A second approach is to use a transient
process governed by the immediate economics in the following way: The energy from a
Figure 1.8: Historical electricity prices in Ontario for July 2011. Image courtesy of ieso “Power to Ontario. On
Demand” [40].
11
renewable source is fed to an electrochemical reduction reactor only in off-peak times. Storing
electrical energy from intermittent power sources in chemical form only when the demand (and
price) for electricity is low has potential to improve the overall process economics for renewable
energy sources and thereby enable wider penetration of renewable technology (well beyond
15 %). Lastly, an electrochemical reactor could be tied to a fuel cell to (a) store electrical energy
in chemical form in off-peak times and (b) convert the chemicals back to electricity when the
price of electricity is once again high.
1.2.2 Explanation of the Electrochemical Reduction of CO2
The electrochemical conversion of CO2 relies on an electrochemical reactor that balances
electrocatalytic reduction and oxidation reactions. Figure 1.9 depicts the overall reaction
process. Specifically, on the anode (e.g., Pt as the catalyst), water is oxidized to make oxygen
Figure 1.9: (Left) Schematic of an electrochemical reactor for CO2 reduction. (Right) Standard Reduction
Potentials for the common cathode and anode reactions.
Electrolyte
CO
2
H+
CO
, C
H4,
C2H
4,
H2
(Pro
du
cts
)
H+
H+
e-
O2
O2
O2
O2
O2
O2
O2
Anode (e.g., Pt) Reactions
H2O→2H++2e-+½O2
Cathode (e.g., Ag, Cu) Reactions
2H++2e-→H2
2H++2e-+CO2→H2O+CO
2H++2e-+CO2→HCOOH
8H++8e-+CO2→2H2O+CH4
12H++12e-+2CO2→4H2O+C2H4
Gas Diffusion Electrodes
Eo (vs. SHE)*
-0.41
-0.52
-0.61
-0.25
-0.34
Eo (vs. SHE)*
+0.82
*at pH 7
12
and protons. The evolved oxygen escapes the reactor via diffusion through micro-pores in the
anode gas diffusion electrode (GDE). The protons formed on the anode cross either a conductive
electrolyte stream or a conductive membrane to react on the cathode (e.g., Ag, Cu as the
catalyst). On the cathode, the protons can react to evolve H2, or, when reacting with CO2, to
form both water and either carbon monoxide, formic acid, methane, ethylene, or other C1, C2, or
C3 carbon products [42,43]. The carbon dioxide reaches the cathode via diffusion through the
cathode GDE to reach the cathode catalyst. Unlike a fuel cell, the reaction does not proceed
without an applied potential because the process is endothermic. Consequently, a potential must
be applied between the anode and cathode to overcome the activation energy barrier and generate
the potential necessary for the reaction to proceed. The theoretical potential necessary for the
reaction is the difference between the standard potentials for both the anode and cathode
reactions as shown in Figure 1.9 [44]. In reality, a higher potential must be applied, commonly
referred to as an overpotential, to overcome activation barriers to the reaction, often due to the
formation of high energy intermediates.
1.3 Challenges for the Electrochemical Reduction of CO2
As mentioned above, the electrochemical reduction of CO2 has potential to expand the
use of renewable energy sources and enable penetration of other sources into the transportation
sector. However, for large scale implementation of this technology, improvements are needed
regarding: (a) energetic efficiency, (b) reaction rate, and (c) product selectivity.
While all three of the aforementioned challenges will benefit from improved catalysts,
better catalysts are critical to improving the energetic efficiency of the reaction, as catalysts are
the only way to reduce the overpotentials for the reaction. Specifically for CO2 reduction, the
intermediate formed after the first electron transfer, CO2- requires an overpotential of -1.9 V vs.
13
SHE [45]. While transition metal catalysts reduce the overpotential for this reaction, as shown
schematically in Figure 1.10, further reduction in the overpotential can be achieved via either
improved catalysts of co-catalyst electrolytes. Specifically, Hori explored an extensive list of
transition catalysts and reported their product selectivity and onset potentials [42]. Furthermore,
Bocarsly et al. demonstrated that Pyridine can serve as a co-catalyst with Cu electrodes to reduce
the onset potential for methanol production [46-49].
Improving the reaction rate is another critical challenge for CO2 reduction as reaction rate
influences capital expenditures for both the reactor and the catalyst. Like the reaction
overpotential, the reaction rate is also drastically influenced by catalysis, however, other factors
such as electrolyte interactions, operation conditions, and reactor design influence the reaction
rate. Lister et al. demonstrated that both the partial pressure of CO2 and the reactor temperature
influence the partial reaction rates for both hydrogen and carbon monoxide formation [50].
Figure 1.10: Schematic of the overpotential for CO2 reduction. The red curve is a depiction of how a catalyst can
be used to lower the overpotential for the formation of a CO2 radical intermediate compared to the scenario without
a catalyst (red). Reprinted with permission from: D. T. Whipple and P. J. A. Kenis, “Prospects of CO2 Utilization
via Direct Heterogeneous Electrochemical Reduction” Journal of Physical Chemistry Letters, 2010. 1 (24): p. 3451-
3458.. Copyright (2010) American Chemical Society [9].
14
From this, an optimal temperature can be identified, beyond which and below which, the reaction
rate for CO formation decreases. Kaneco and others showed that electrolyte composition
drastically influences the partial reaction rates of multiple carbon products on mercury or copper
electrodes [51-53]. Whipple et al. showed that the electrolyte pH drastically influences the
partial reaction rate for hydrogen and formic acid formation on a tin electrode [54].
Furthermore, several reactor designs have been developed to improve the reaction rates [55-61].
The last requirement for large scale implementation of this technology is optimization of
the selectivity for a desired product. Many of the same factors influencing reaction rate
influence product selectivity. First, catalyst choice more so than virtually any other factor,
Figure 1.11: Cathode potentials, current densities and Faradiac efficiencies on various transition metal electrodes in
a 0.1 M KHCO3 electrolyte. Reprinted from “Electrocatalytic process of CO selectivity in electrochemical reduction
of CO2 at metal electrodes in aqueous media”, 39, Yoshio Hori, Hidetoshi Wakebe, Toshio Tsukamoto, Osamu
Koga, 1833-1839, Copyright (1994), with permission from Elsevier [62].
15
controls product selection. Figure 1.11 shows product distributions from the electrochemical
reduction of CO2 on a wide range of catalysts as tabulated by Hori et al. [62]. Four classes of
catalysts exist for the production of carbon products via the electrochemical reduction of CO2.
First, copper has been shown to be selective for the production of methanol [63,64], methane
[51,53,65-67], ethylene [51,67-70], and sometimes larger carbon products, such as propane,
propanol, etc. [43]. Ag, Au, Zn, Pd, and Ga are selective for the production of carbon monoxide
[42,55-57,71]. Pb, Hg, In, Sn, Cd, and Ti are selective for the production of formic acid
[9,42,72-79]. Lastly, Ni, Fe, Pt and Ti are selective for the production of H2.
While Hori tabulated the selectivity of a wide range of catalysts in a 0.1 M KHCO3
solution, the selectivity can be tuned by changing the electrolyte composition. Several
Figure 1.12: Current densities and Faradaic efficiencies of a copper electrode with a variety of electrolyte solutions.
Reprinted (adapted) with permission from: S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta,
“Electrochemical Reduction of High Pressure CO2 at a Cu Electrode in Cold Methanol.” Electrochimica Acta, 2006.
51 (23): p. 4880-4885. Kaneco 2006. Copyright (2006) American Chemical Society [53].
16
researchers have shown that various alkali cations influence the selectivity for methane and
ethylene on a copper electrode [53,80] as well as formic acid and hydrogen on a mercury
electrode [52]. Figure 1.12 shows the effect of a wide range of electrolyte solutions on product
selection with a copper electrode [51,53]. Additionally, pyridine in the electrolyte solution was
found to change the product distribution of a copper electrode from methane, ethylene, hydrogen
and carbon monoxide to methane [46,49,81].
1.4 Economic Analysis
The power required to produce a given mass of a product via electrochemical reduction
of CO2 can be calculated based on the applied reactor potential using the following equation:
(1.1)
where V is the potential applied to the reactor, F is Faraday’s constant, ci is the number of
electrons required for the reaction, and MW is the product molecular weight. At a typical reactor
potential of 2.5 V, the power required for the electrochemical reduction of CO2 to CO is
4,342 kWhr/ton. The price of CO on the market is roughly $1,000 per ton [82]. If ignoring
capital costs, transportation costs, and all other costs besides the direct electricity cost for the
reduction process, the financial return for producing a chemical per unit of electricity consumed
is the break-even price of electricity. The break-even price of electricity could be calculated
using the following equation:
(1.2)
Using this equation, the break-even cost of electricity is: 23 ¢/kWhr, which is greater than
the maximum price of electricity shown earlier in Ontario for an expensive summer month. As
long as electricity stays below this price, an electrochemical process for CO2 reduction could be
17
economically favorable with the other expenditures (e.g., capital, transportation, storage)
remaining low.
1.5 Remaining Challenges
Better catalysts and optimized reaction conditions have improved reaction kinetics and
thereby enabled current densities greater than 600 mA and energetic efficiencies greater than
70 % for the electrochemical reduction of CO2 to CO; however, an economical process requires
both high energetic efficiencies and current densities [9]. Figure 1.13 shows a comparison of
data in the literature for the electrochemical reduction of CO2 into formic acid, carbon monoxide,
and larger hydrocarbons, e.g., CH4, C2H4, with regards to energetic efficiency and current
Figure 1.13: Comparison of the energy efficiencies and partial current densities for CO2 reduction to formic acid,
syngas, and hydrocarbons (methane and ethylene) reported in literature with those of water electrolyzers.
Efficiencies of electrolyzers are total system efficiencies, while the CO2 conversion efficiencies only include
cathode losses and neglect anode and system losses. Reprinted with permission from: D. T. Whipple and P. J. A.
Kenis, “Prospects of CO2 Utilization via Direct Heterogeneous Electrochemical Reduction” Journal of Physical
Chemistry Letters, 2010. 1 (24): p. 3451-3458.. Copyright (2010) American Chemical Society [9].
18
density [9]. Figure 1.13 classifies the catalysts into three categories based on their products.
First, the reactions selective for larger hydrocarbons achieve high current densities at the expense
of energetic efficiency. While the direct conversion of CO2 to hydrocarbons is attractive due to
the immediate application of the product, the overall efficiency is far too low for
commercialization. Second, the reactions selective for formic acid achieve intermediate current
densities and intermediate energetic efficiencies. Third, the reactions selective for CO achieve
high energetic efficiencies at low current densities. Here, we chose to study the electrochemical
reaction for CO as it has the highest energetic efficiencies compared to other reactions in the
literature. An important consideration is how this technology compares to a more traditional
technology, water electrolysis. Figure 1.13 shows that the energetic efficiencies and current
densities for water electrolysis are much higher than those achieved for CO reduction. However,
using hydrogen as a fuel has inherent disadvantages due to challenges with both storage and
transportation of the compressed hydrogen gas. While transportation and storage issues exist
with water electrolysis, the efficiency and current density of a CO2 reduction process should be
competitive with water electrolyzers.
In summary, the electrochemical reduction of CO2 has tremendous promise as a means of
energy storage and enabling penetration of electrical energy into other energy sectors, such as
transportation. However, improvements are necessary primarily regarding the energetic efficiency
and current density of the reaction. The purpose of this dissertation is to improve the energetic
efficiency and current density for the electrochemical reduction of CO2 to CO using a microfluidic
reactor.
Chapter 2 details the influence of electrolytes on the electrochemical reduction of CO2 on a
Ag electrode. Chapter 3 details the development of an electrochemical platform to study the use of
an ionic liquid as a co-catalyst. Chapter 4 demonstrates a means to achieve high current densities
19
with ionic liquids. Chapter 5 explores the use of organometallic catalysts. Chapter 6 presents a study
of the role of local geometry on a methanol laminar flow fuel cell, a reactor configuration that can
also be applied to electrochemical CO2 conversion. Chapter 7 summarizes the key findings and
conclusions obtained from this work.
1.6 References
1. J.B. Reece, M.L. Cain, L.A. Urry, P.V. Minorsky and S.A. Wasserman, Campbell
Biology, Pearson Benjamin Cummings, 2010.
2. National Petroleum Council. Committee on Global Oil and Gas., Hard truths: Facing the
Hard Truths About Energy: A Comprehensive View to 2030 of Global Oil and Natural
Gas, National Petroleum Council, Washington, DC, 2007.
3. United Nations. Department of Economic Affairs: Population Division., World
Population to 2300: Economic & Social Affairs, United Nations, New York, NY, 2012.
4. Organization for Economic Co-orporation and Development (OECD): International
Energy Agency, World Energy Outlook, International Energy Agency, Paris, France,
2010.
5. R. W. Bentley, Global oil & gas depletion: an overview, Energy Policy, 2002, 30 (3), p.
189-205.
6. R. Guseo, A. Dalla Valle and M. Guidolin, World Oil Depletion Models: Price effects
compared with strategic or technological interventions, Technological Forecasting and
Social Change, 2007, 74 (4), p. 452-469.
7. J. S. Harvey, B. P. Lyons, T. S. Page, C. Stewart and J. M. Parry, An assessment of the
genotoxic impact of the Sea Empress oil spill by the measurement of DNA adduct levels
in selected invertebrate and vertebrate species, Mutation Research/Genetic Toxicology
and Environmental Mutagenesis, 1999, 441 (1), p. 103-114.
8. National Petroleum Council. Committee on Global Oil and Gas., Prudent Development:
Realizing the Potential of North America's Abundant Natural Gas and Oil, National
Petroleum Council, Washington, D.C., 2011.
9. D. T. Whipple and P. J. A. Kenis, Prospects of CO2 Utilization via Direct Heterogeneous
Electrochemical Reduction, Journal of Physical Chemistry Letters, 2010, 1 (24), p. 3451-
3458.
10. S. Beggs, S. Cardell and J. Hausman, Assessing the potential demand for electric cars,
Journal of Econometrics, 1981, 17 (1), p. 1-19.
11. R. Krishna, J. Ellenberger and S. T. Sie, Reactor development for conversion of natural
gas to liquid fuels: A scale-up strategy relying on hydrodynamic analogies, Chemical
Engineering Science, 1996, 51 (10), p. 2041-2050.
12. N. D. Parkyns, C. I. Warburton and J. D. Wilson, Natural gas conversion to liquid fuels
and chemicals: Where does it stand?, Catalysis Today, 1993, 18 (4), p. 385-442.
20
13. Buping Bao, Mahmoud M. El-Halwagi and Nimir O. Elbashir, Simulation, integration,
and economic analysis of gas-to-liquid processes, Fuel Processing Technology, 2010, 91
(7), p. 703-713.
14. Mikael Höök and Kjell Aleklett, A review on coal-to-liquid fuels and its coal
consumption, International Journal of Energy Research, 2010, 34 (10), p. 848-864.
15. Dragomir Bukur, Dr Mahmood Amani and Bashar Hassna, Reactors and Catalysts for
Gas-to-Liquids (GTL) and Coal-to-Liquids (CTL) Technologies, QNRS Repository,
2011, (2011), p. 3954.
16. Delanie Lamprecht, Reinier Nel and Dieter Leckel, Production of On-Specification Fuels
in Coal-to-Liquid (CTL) Fischer−Tropsch Plants Based on Fixed-Bed Dry Bottom Coal
Gasification, Energy & Fuels, 2009, 24 (3), p. 1479-1486.
17. X. Hao, G. Dong, Y. Yang, Y. Xu and Y. Li, Coal to Liquid (CTL): Commercialization
Prospects in China, Chemical Engineering & Technology, 2007, 30 (9), p. 1157-1165.
18. C. Oloman and H. Li, Electrochemical Processing of Carbon Dioxide, ChemSusChem,
2008, 1 (5), p. 385-391.
19. C. M. Sanchez-Sanchez, V. Montiel, D. A. Tryk, A. Aldaz and A. Fujishima,
Electrochemical Approaches to Alleviation of the Problem of Carbon Dioxide
Accumulation, Pure and Applied Chemistry, 2001, 73 (12), p. 1917-1927.
20. S. Choi, J. H. Drese and C. W. Jones, Adsorbent Materials for Carbon Dioxide Capture
from Large Anthropogenic Point Sources, ChemSusChem, 2009, 2 (9), p. 796-854.
21. X. Hao, M. E. Djatmiko, Y. Y. Xu, Y. N. Wang, J. Chang and Y. W. Li, Simulation
Analysis of a GTL Process Using ASPEN Plus, Chemical Engineering & Technology,
2008, 31 (2), p. 188-196.
22. M. Sudiro and A. Bertucco, Production of Synthetic Gasoline and Diesel Fuel by
Alternative Processes Using Natural Gas and Coal: Process Simulation and
Optimization, Energy, 2009, 34 (12), p. 2206-2214.
23. M. Sudiro, A. Bertucco, F. Ruggeri and A. Fontana, Improving Process Performances in
Coal Gasification for Power and Synfuel Production, Energy & Fuels, 2008, 22 (6), p.
3894-3901.
24. Soren Anderson and Richard Newell, Prospects for carbon capture and storage
technologies, 2003.
25. Heleen de Coninck, Jennie C. Stephens and Bert Metz, Global learning on carbon
capture and storage: A call for strong international cooperation on CCS demonstration,
Energy Policy, 2009, 37 (6), p. 2161-2165.
26. Michael L. Szulczewski, Christopher W. MacMinn, Howard J. Herzog and Ruben
Juanes, Lifetime of carbon capture and storage as a climate-change mitigation
technology, Proceedings of the National Academy of Sciences, 2012, 109 (14), p. 5185-
5189.
27. Charles D. Keeling, Robert B. Bacastow, Arnold E. Bainbridge, Carl A. Ekdahl, Peter R.
Guenther, Lee S. Waterman and John F. S. Chin, Atmospheric carbon dioxide variations
at Mauna Loa Observatory, Hawaii, Tellus, 1976, 28 (6), p. 538-551.
28. NOAA/ESRL (www.esrl.noaa.gov/gmd/ccgg/trends/) and Dr. Ralph Keeling Dr. Pieter
Tans, Scripps Institution of Oceanography (scrippsco2.ucsd.edu/).
21
29. R. J. Francey, C. E. Allison, D. M. Etheridge, C. M. Trudinger, I. G. Enting, M.
Leuenberger, R. L. Langenfelds, E. Michel and L. P. Steele, A 1000-year high precision
record of δ13C in atmospheric CO2, Tellus B, 1999, 51 (2), p. 170-193.
30. Terry L. Root, Jeff T. Price, Kimberly R. Hall, Stephen H. Schneider, Cynthia
Rosenzweig and J. Alan Pounds, Fingerprints of global warming on wild animals and
plants, Nature, 2003, 421 (6918), p. 57-60.
31. Malte Meinshausen, Nicolai Meinshausen, William Hare, Sarah C. B. Raper, Katja
Frieler, Reto Knutti, David J. Frame and Myles R. Allen, Greenhouse-gas emission
targets for limiting global warming to 2[thinsp][deg]C, Nature, 2009, 458 (7242), p.
1158-1162.
32. C. Knocke and J. Vogt, Biofuels - Challenges & Chances: How Biofuel Development Can
Benefit from Advanced Process Technology, Engineering in Life Sciences, 2009, 9 (2), p.
96-99.
33. P. J. L. Williams and L. M. L. Laurens, Microalgae as Biodiesel & Biomass Feedstocks:
Review & Analysis of the Biochemistry, Energetics & Economics, Energy &
Environmental Science, 2010, 3 (5), p. 554-590.
34. D. Abbott, Keeping the Energy Debate Clean: How Do We Supply the World's Energy
Needs?, Proceedings of the Ieee, 2010, 98 (1), p. 42-66.
35. O. Langhamer, K. Haikonen and J. Sundberg, Wave Power-Sustainable Energy or
Environmentally Costly? A Review with Special Emphasis on Linear Wave Energy
Converters, Renewable & Sustainable Energy Reviews, 2010, 14 (4), p. 1329-1335.
36. M. Lenzen, Current State of Development of Electricity-Generating Technologies: A
Literature Review, Energies, 2010, 3 (3), p. 462-591.
37. Holly Klick and Eric R. A. N. Smith, Public understanding of and support for wind
power in the United States, Renewable Energy, 2010, 35 (7), p. 1585-1591.
38. J.M. Duthie and H.W. Whittington, Securing Renewable Energy Supplies Through
Carbon Dioxide Storage in Methanol, Power Engineering Society Summer Meeting,
2002 IEEE, 2002, 1 p. 145-150.
39. Jay Apt, The spectrum of power from wind turbines, Journal of Power Sources, 2007, 169
(2), p. 369-374.
40. Aimee E. Curtright and Jay Apt, The character of power output from utility-scale
photovoltaic systems, Progress in Photovoltaics: Research and Applications, 2008, 16 (3),
p. 241-247.
41. Anonymous, Monthly Market Report: July 2011. ieso: Power to Ontario. On Demand.,
2011.
42. Y. Hori, H. Wakebe, T. Tsukamoto and O. Koga, Electrocatalytic Process of CO
Selectivity in Electrochemical Reduction of CO2 at Metal-Electrodes in Aqueous-Media,
Electrochimica Acta, 1994, 39 (11-12), p. 1833-1839.
43. Kendra P. Kuhl, Etosha R. Cave, David N. Abram and Thomas F. Jaramillo, New insights
into the electrochemical reduction of carbon dioxide on metallic copper surfaces, Energy
& Environmental Science, 2012, 5 (5), p. 7050-7059.
44. Y. Hori, Electrochemical CO2 Reduction on Metal Electrodes, Modern Aspects of
Electrochemistry, 2008, 42 p. 89-189.
22
45. Etsuko Fujita, Photochemical carbon dioxide reduction with metal complexes,
Coordination Chemistry Reviews, 1999, 185–186 (0), p. 373-384.
46. Emily Barton Cole, Prasad S. Lakkaraju, David M. Rampulla, Amanda J. Morris, Esta
Abelev and Andrew B. Bocarsly, Using a One-Electron Shuttle for the Multielectron
Reduction of CO2 to Methanol: Kinetic, Mechanistic, and Structural Insights, Journal of
the American Chemical Society, 2010, 132 (33), p. 11539-11551.
47. E. B. Cole, P. S. Lakkaraju, D. M. Rampulla, A. J. Morris, E. Abelev and A. B. Bocarsly,
Using a One-Electron Shuttle for the Multielectron Reduction of CO2 to Methanol:
Kinetic, Mechanistic, and Structural Insights, Journal of the American Chemical Society,
2010, 132 (33), p. 11539-11551.
48. A. J. Morris, R. T. McGibbon and A. B. Bocarsly, Electrocatalytic Carbon Dioxide
Activation: The Rate-Determining Step of Pyridinium-Catalyzed CO2 Reduction,
Chemsuschem, 2011, 4 (2), p. 191-196.
49. G. Seshadri, C. Lin and A. B. Bocarsly, A New Homogeneous Electrocatalyst for the
Reduction of Carbon-Dioxide to Methanol at Low Overpotential, Journal of
Electroanalytical Chemistry, 1994, 372 (1-2), p. 145-150.
50. Eric Dufek, Tedd Lister and Michael McIlwain, Bench-scale electrochemical system for
generation of CO and syn-gas, Journal of Applied Electrochemistry, 2011, 41 (6), p. 623-
631.
51. S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of
High Pressure CO2 at a Cu Electrode in Cold Methanol, Electrochimica Acta, 2006, 51
(23), p. 4880-4885.
52. Y. Hori and S. Suzuki, Electrolytic Reduction of Carbon-Dioxide at Mercury-Electrode
in Aqueous-Solution, Bulletin of the Chemical Society of Japan, 1982, 55 (3), p. 660-665.
53. S. Kaneco, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of CO2 to
Methane at the Cu Electrode in Methanol with Sodium Supporting Salts and its
Comparison with Other Alkaline Salts, Energy Fuels, 2006, 20 (1), p. 409-414.
54. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Microfluidic Reactor for the
Electrochemical Reduction of Carbon Dioxide: The Effect of pH, Electrochemical and
Solid State Letters, 2010, 13 (9), p. D109-D111.
55. T. Yamamoto, D. A. Tryk, A. Fujishima and H. Ohata, Production of Syngas Plus
Oxygen from CO2 in a Gas-Diffusion Electrode-Based Electrolytic Cell, Electrochimica
Acta, 2002, 47 (20), p. 3327-3334.
56. C. Delacourt, P. L. Ridgway, J. B. Kerr and J. Newman, Design of an Electrochemical
Cell Making Syngas (CO+H2) from CO2 and H2O Reduction at Room Temperature,
Journal of the Electrochemical Society, 2008, 155 (1), p. B42-B49.
57. F. Bidrawn, G. Kim, G. Corre, J. T. S. Irvine, J. M. Vohs and R. J. Gorte, Efficient
Reduction of CO2 in a Solid Oxide Electrolyzer, Electrochemical and Solid-State Letters,
2008, 11 (9), p. B167-B170.
58. S. D. Ebbesen and M. Mogensen, Electrolysis of Carbon Dioxide in Solid Oxide
Electrolysis Cells, Journal of Power Sources, 2009, 193 (1), p. 349-358.
23
59. S. H. Jensen, X. F. Sun, S. D. Ebbesen, R. Knibbe and M. Mogensen, Hydrogen and
Synthetic Fuel Production Using Pressurized Solid Oxide Electrolysis Cells, International
Journal of Hydrogen Energy, 2010, 35 (18), p. 9544-9549.
60. V. Kaplan, E. Wachtel, K. Gartsman, Y. Feldman and I. Lubomirsky, Conversion of CO2
to CO by Electrolysis of Molten Lithium Carbonate, Journal of the Electrochemical
Society, 2010, 157 (4), p. B552-B556.
61. Z. L. Zhan and L. Zhao, Electrochemical Reduction of CO2 in Solid Oxide Electrolysis
Cells, Journal of Power Sources, 2010, 195 (21), p. 7250-7254.
62. Yoshio Hori, Hidetoshi Wakebe, Toshio Tsukamoto and Osamu Koga, Electrocatalytic
process of CO selectivity in electrochemical reduction of CO2 at metal electrodes in
aqueous media, Electrochimica Acta, 1994, 39 (11–12), p. 1833-1839.
63. G. Arai, T. Harashina and I. Yasumori, Selective Electrocatalytic Reduction of Carbon-
Dioxide to Methanol on Ru-Modified Electrode, Chemistry Letters, 1989, (7), p. 1215-
1218.
64. D. P. Summers, S. Leach and K. W. Frese, The Electrochemical Reduction of Aqueous
Carbon-Dioxide to Methanol at Molybdenum Electrodes with Low Overpotentials,
Journal of Electroanalytical Chemistry, 1986, 205 (1-2), p. 219-232.
65. S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of
High Pressure Carbon Dioxide at a Cu Electrode in Cold Methanol with CsOH
Supporting Salt, Chemical Engineering Journal, 2007, 128 (1), p. 47-50.
66. S. Kaneco, R. Iwao, K. Iiba, K. Ohta and T. Mizuno, Electrochemical Conversion of
Carbon Dioxide to Formic Acid on Pb in KOH/Methanol Electrolyte at Ambient
Temperature and Pressure, Energy, 1998, 23 (12), p. 1107-1112.
67. R. L. Cook, R. C. Macduff and A. F. Sammells, High-Rate Gas-Phase CO2 Reduction to
Ethylene and Methane Using Gas-Diffusion Electrodes, Journal of the Electrochemical
Society, 1990, 137 (2), p. 607-608.
68. S. Kaneco, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of Carbon
Dioxide to Ethylene at a Copper Electrode in Methanol Using Potassium Hydroxide and
Rubidium Hydroxide Supporting Electrolytes, Electrochimica Acta, 2006, 51 (16), p.
3316-3321.
69. T. Mizuno, K. Ohta, M. Kawamoto and A. Saji, Electrochemical Reduction of CO2 on Cu
in 0.1 M KOH-methanol, Energy Sources, 1997, 19 (3), p. 249-257.
70. J. Yano and S. Yamasaki, Pulse-mode Electrochemical Reduction of Carbon Dioxide
Using Copper and Copper Oxide Electrodes for Selective Ethylene Formation, Journal of
Applied Electrochemistry, 2008, 38 (12), p. 1721-1726.
71. N. Furuya and S. Koide, Electroreduction of Carbon-Dioxide by Metal Phthalocyanines,
Electrochimica Acta, 1991, 36 (8), p. 1309-1313.
72. M. Azuma, K. Hashimoto, M. Hiramoto, M. Watanabe and T. Sakata, Electrochemical
Reduction of Carbon-Dioxide on Various Metal-Electrodes in Low-Temperature Aqueous
KHCO3 Media, Journal of the Electrochemical Society, 1990, 137 (6), p. 1772-1778.
73. N. Furuya, T. Yamazaki and M. Shibata, High Performance Ru-Pd Catalysts for CO2
Reduction at Gas-Diffusion Electrodes, Journal of Electroanalytical Chemistry, 1997, 431
(1), p. 39-41.
24
74. M. N. Mahmood, D. Masheder and C. J. Harty, Use of Gas-Diffusion Electrodes for
High-Rate Electrochemical Reduction of Carbon-Dioxide .1. Reduction at Lead, Indium-
Impregnated and Tin-Impregnated Electrodes, Journal of Applied Electrochemistry,
1987, 17 (6), p. 1159-1170.
75. Y. Akahori, N. Iwanaga, Y. Kato, O. Hamamoto and M. Ishii, New Electrochemical
Process for CO2 Reduction to Formic Acid from Combustion Flue Gases,
Electrochemistry, 2004, 72 (4), p. 266-270.
76. F. Koleli, T. Atilan, N. Palamut, A. M. Gizir, R. Aydin and C. H. Hamann,
Electrochemical Reduction of CO2 at Pb- and Sn-electrodes in a Fixed-bed Reactor in
Aqueous K2CO3 and KHCO3 Media, J. of Appl. Electrochem., 2003, 33 (5), p. 447-450.
77. H. Li and C. Oloman, Development of a Continuous Reactor for the Electro-reduction of
Carbon Dioxide to Formate - Part 2: Scale-up, Journal of Applied Electrochemistry,
2007, 37 (10), p. 1107-1117.
78. K. Subramanian, K. Asokan, D. Jeevarathinam and M. Chandrasekaran, Electrochemical
Membrane Reactor for the Reduction of Carbon Dioxide to Formate, Journal of Applied
Electrochemistry, 2007, 37 (2), p. 255-260.
79. B. Innocent, D. Liaigre, D. Pasquier, F. Ropital, J. M. Leger and K. B. Kokoh, Electro-
reduction of Carbon Dioxide to Formate on Lead Electrode in Aqueous Medium, Journal
of Applied Electrochemistry, 2009, 39 (2), p. 227-232.
80. A. Murata and Y. Hori, Product Selectivity Affected by Cationic Species in
Electrochemical Reduction of CO2 and CO at a Cu Electrode, Bulletin of the Chemical
Society of Japan, 1991, 64 (1), p. 123-127.
81. Amanda J. Morris, Robert T. McGibbon and Andrew B. Bocarsly, Electrocatalytic
Carbon Dioxide Activation: The Rate-Determining Step of Pyridinium-Catalyzed CO2
Reduction, ChemSusChem, 2011, 4 (2), p. 191-196.
82. Narasi Sridhar, Carbon Dioxide Utilization: Electrochemical Conversion of CO2 -
Opportunities and Challenges. Det Norske Veritas, 2011. 07 1-20.
25
Chapter 2
Effect of Ions on the Electrochemical Conversion of CO2 to CO1
2.1 Introduction
Carbon dioxide (CO2) concentrations in the atmosphere have been linked to climate
change [1]. A reduction in CO2 emissions and/ or an increase in CO2 capture may be needed to
stabilize CO2 concentrations in the atmosphere, and consequently, minimize the negative effects
of CO2 as a greenhouse gas [2-6]. Additionally, to reduce global dependence on fossil fuels,
approaches to ensure further penetration of alternative energy sources need to be developed.
Furthermore, uneven global distribution of oil reserves present potential energy security issues
for oil importing countries (e.g., US and Europe) [7]. Natural gas to liquid (GTL) and coal to
liquid (CTL) technologies are being developed to reduce dependence on imported petroleum [7].
While GTL and CTL can provide a means of decreasing dependence on imported oil, these
technologies do not address the non-renewable nature of fossil fuels. To reduce atmospheric
CO2 emissions, one technology being implemented is carbon capture and sequestration [5,8-10].
Additionally, carbon neutral energy sources such as wind and solar are becoming increasingly
important sources of energy. However, in the current infrastructure, the potential to replace
traditional non-renewable sources with the aforementioned renewable sources is limited due to
their intermittent nature. For example, in the US, the use of these energy sources will be limited
to a maximum of ~15 % of the US electricity supply without improved power output leveling
(smart grid) [11]. Conversion of CO2 to useable fuels provides a means of decreasing
dependence on foreign oil via the penetration of renewable energy sources into the transportation
1 Part of this work has been submitted for publication: Thorson, M. R, Siil K.I., Kenis, P. J. A., “Effect of Cations on
the Electrochemical Conversion CO2 to CO.” 2012
26
sector, while expanding the use of intermittent, renewable power sources (wind and solar). Since
transportation is responsible for about 33 % of CO2 emissions in the United States, introducing a
carbon neutral fuel would have a drastic impact on CO2 concentrations in the atmosphere [12].
Multiple methods (e.g., photochemical, thermochemical, and electrochemical) are being
explored for the conversion of CO2 into useful fuels. Specifically, electrochemical reduction of
CO2 combines electrical energy from a carbon neutral source with CO2 to yield many products,
including: methanol [13-15], formic acid [15-21], methane [15,22-25], ethylene [15,24-28], and
carbon monoxide (CO) [15,29-32]. A broad spectrum of metal catalysts has been tested for the
production of various fuels and feedstocks from CO2 reduction [15,16,21,25]. Carbon
monoxide, which is primarily produced using Ag and Au as the catalysts, is particularly
interesting as it can be combined with H2 to yield syngas, which can be further reacted in a
Fischer-Tropsch process to fuels such as methanol or “Fischer-Tropsch diesel” [33-35].
Consequently, several reactor designs have been explored for CO production [30-32,36-39]. As
we reported earlier, the primary hurdle to the further development of electrochemical CO2
reduction processes lies in the challenge to simultaneously maximize energetic efficiency and
throughput [7].
Previously, we reported a flow reactor for electrochemical CO2 conversion to formic acid
[18]. Using a similar reactor, we showed that amine based ionic liquid electrolytes can
drastically reduce the overpotential necessary for the electrochemical reduction of CO2 to CO,
albeit at the expense of throughput [40]. Pyridine has been shown to have similar effects
[41,42]. Additionally, several researchers have shown that various alkali cations influence the
selectivity for methane and ethylene on a copper electrode [23,43] as well as formic acid and
hydrogen on a mercury electrode [44]; however, little work has been published regarding the
27
influence of electrolyte choice with the most prominent catalysts for CO2 reduction, Ag or Au.
The study of electrolytes is further complicated by effects from the electrolyte solutions on
solution properties, such as conductivity, solubility and pH, which need to be de-convoluted.
To date, most efforts for the electrochemical reduction of CO2 to CO have focused on
researching different catalysts to optimize energetic efficiency and/or current density. Here we
investigate in detail the influence of alkali cations (i.e., Li+, Na
+, K
+, Rb
+, and Cs
+) with two
halogens (i.e., Cl- and I
-) and hydroxide (OH
-) on the electrochemical reduction of CO2 to CO
with regards to current density and selectivity.
2.2 Experimental
2.2.1 Electrochemical cell
A schematic of the electrochemical reactor used in this study is shown in Figure 2.1.
Stainless steel plates (6 x 3 cm) served as current collectors, which held the fuel cell together and
provided electrical contact between the gas diffusion electrode (GDE) and an external
potentiostat (PGSTAT-30, EcoChemie). The potentiostat was connected to the cell via banana
clips which plugged directly into precisely machined 3/16” holes in the current collectors. Two
1-mm thick poly(methyl methacrylate) (PMMA) windows with precisely machined 0.5-cm wide
by 2.0-cm long channels provided the electrolyte flow fields within the electrochemical reactor.
Figure 2.1: Schematic of an electrochemical reactor used to study influence of various electrolyte solutions on the
electrochemical reduction of CO2 to CO.
28
A 212 Nafion membrane (DuPont) separated the two PMMA windows. On the cathode side, the
catalyst was applied only to the first 1.5 cm and a polytetrafluoroethylene (PTFE) filter (10 µm,
Pall Life Sciences) was added to cover the last 0.5 cm. This facilitated the removal of bubbles
from oxygen evolution on the anode that entered the electrolyte stream and could otherwise
collect and block electrolyte contact with the catalyst. The cathode current collector had a
precisely machined 0.5-cm wide by 2.0-cm long by 0.25-cm deep window behind the GDEs to
allow for the flow of gases in and out of the reactor. The anode was left open to air for oxygen to
escape. Four insulated stainless steel bolts held the entire setup together.
2.2.2 Electrode preparation
The electrodes were prepared as previously reported [18] using Sigracet “35BC”-type
GDEs. In short, a suspension of catalyst and Nafion binder was prepared with a 50/50 mixture of
water and isopropyl alcohol, which was then sonicated and subsequently painted on a GDE
followed by hot-pressing. The cathodes consisted of 5 mg/cm2 Ag and 0.1 mg/cm
2 Nafion. The
anode was 2 mg/cm2 Pt black with 0.1 mg/cm
2 Nafion. All electrodes were hot-pressed at 130°C
and 2000 kPa for 15 min, similar to the procedure previously reported [18].
2.2.3 Electrolyte composition
All salts were purchased and used as provided. The following electrolyte solutions were
prepared in ultrapure water (18.2 MΩ∙cm) such that the final concentrations were 1.0 M: LiCl
(Product Number: L121; Purity: 98.5 %; Provider: Fischer Scientific), NaCl (S271; 99 %;
Fischer Scientific), KCl (P1597; 99 %; Sigma Aldrich), CsCl (289329; 99.9 %; Sigma Aldrich),
LiI (518018; 99.9 %; Sigma Aldrich), NaI (383112; 99.5 %; Sigma Aldrich), KI (P410; 99 %;
Fischer Scientific), CsI (202134; 99.9 %; Sigma Aldrich), NaOH (221465; 97 %; Sigma
Aldrich), KOH (44016; 90 %; Sigma Aldrich), RbOH (243698; 99 %; metal basis Sigma
29
Aldrich), and CsOH (C8518; 99.5 %; metal basis Sigma Aldrich). For the pH studies, a 1 M KCl
solution and the pH was adjusted with dilute solutions of KOH or HCl to achieve the desired
final pH levels.
2.2.4 Cell testing
An Autolab potentiostat (PGSTAT-30, EcoChemie) controlled the cell potential and
measured the resulting current as previously reported [18,45]. The individual electrode
potentials were measured using multimeters connected to each electrode and a Ag/AgCl
reference electrode (RE-5B, BASi) in the exit stream. The cell was allowed to reach steady state
for 200 s, after which, the gas flowed into a gas chromatograph (GC) (Trace GC, Thermo) for
composition analysis of H2 and CO in the affluent gas streams. The current was averaged for an
additional 150 s before stepping to the next potential. All experiments were run at ambient
conditions. The electrode potentials were not corrected for IR drop. A mass flow controller
(32907-80, Cole Parmer) was used to flow CO2 from a cylinder at 25 sccm Also, a syringe pump
(PHD 2000, Harvard Apparatus) supplied the electrolyte(s) at 0.5 mL/min. In the case of iodide
salt solutions (i.e., LiI, NaI, KI, CsI), the iodide electrolyte solutions were pumped between the
nafion membrane and the cathode at 0.5 mL/min and a 1.0 M solution of the corresponding
chloride salt solutions (i.e., LiCl, NaCl, KCl, CsCl), were pumped at 0.5 mL/min between the
nafion membrane and the anode. In the case of the chloride (i.e., LiCl, NaCl, KCl, CsCl) and
hydroxide (i.e., NaOH, KOH, RbOH, CsOH) salt solutions, the nafion membrane was removed
as well as one of the 1 mm flow channels and the electrolyte was flowed directly between the
cathode and the anode. The affluent gas stream flowed directly into a gas chromatograph
(Thermo Finnegan Trace GC) operating in the thermal conductivity detection (TCD) mode,
30
which uses a Carboxen 1000 column (Supelco) and a He carrier gas at a flow rate of 20 SCCM.
The column was held at 150°C and the TCD detector was held at 200°C.
2.3 Results and Discussion
2.3.1 Cation Size Effects
We first studied the effect of the cation size on the Faradaic yield of CO2 reduction and
on the partial current densities of the products, CO and H2. Specifically, we tested electrolyte
solutions with a variety of alkali cations (i.e., Li+, Na
+, K
+, Rb
+, Cs
+) for three anions (i.e., Cl
-, I
-,
Figure 2.2: Partial current densities for (a) H2, and (b) CO, as well as (c) Faradaic yield for CO in 1 M (1) chloride,
(2) hydroxide, and (3) iodide solutions, as a function of cathode potential. All conditions tested represent reactor
potentials between 2 and 3 V.
31
OH-) at a concentration of 1 M in an electrochemical reactor. Figures 2a and 2b show,
respectively, the partial current densities for H2 (iH2), the unwanted product, and for CO, the
desired product, as a function of cathode potential for applied cell voltages between 2.0 and
3.0 V for chloride, iodide, and hydroxide salts. First, we discuss the effects of different salts on
hydrogen evolution (Figure 2.2a). In the case of chloride salts (Figure 2.2a1), the smallest
cation (Li+) favored H2 evolution at large potentials. For example, in the presence of LiCl, iH2
was 11 mA/cm2 at a cathode potential of -1.85 V vs. Ag/AgCl. Whereas, with electrolyte
solutions with larger cations, (i.e., NaCl, KCl, CsCl), substantially less H2 was produced across
the range of potentials tested. For example, in the presence of NaCl, KCl, or CsCl, iH2 was
6 mA/cm2 at a cathode potential of -1.85 V vs. Ag/AgCl. In the case of iodide salts (Figure
2.2a2), the influence of the cation size was less pronounced; however, we observe the same trend
that the smallest cation (Li+) and the largest cation (Cs
+) have the highest and lowest iH2,
respectively. In the case of hydroxide salts (Figure 2.2a3), rather than testing the Li+ cation,
which was found to favor H2 evolution, we tested a larger cation (Rb+), which we expected to
favor CO production. Again, we observe the same trend that the smallest cation (Na+) and the
largest cation (Cs+) have the highest and lowest iH2, respectively. In summary, for the three sets
of electrolytes tested, we observed a strong trend that an increase in cation size led to a decrease
in iH2.
We also analyzed the effect of different electrolyte compositions on CO formation
(Figure 2.2b). In the case of chloride salts (Figure 2.2b1), a maximum iCO of 26 mA was
observed in the presence of the smallest cation (Li+); however, with larger cations, much more
CO is produced at the same potentials. In particular, with the largest cation (Cs+), an iCO of
60 mA was observed at a cathode potential of -1.85 V vs. a Ag/AgCl reference electrode. In the
32
case of iodide salts (Figure 2.2b2), the same trend was observed that the partial current density
of CO increased with an increase in cation size. In the case of hydroxide salts (Figure 2.2b3),
iCO increased as a function of increasing cation size for Na+, K
+, and Rb
+ whereas in the presence
of Cs+ iCO dropped relative to the iCO of Rb
+. The iCO of 175 mA/cm
2 at a cathode potential of -
2.0 V vs a Ag/AgCl reference electrode observed when using 1 M RbOH as the electrolyte is the
highest iCO reported in the literature at room temperature at this cell potential. In summary, we
observed a strong trend of larger cations increasing iCO without influencing the onset potential.
Figure 2.2c shows the Faradaic yield (selectivity) for CO as a function of cathode
potentials in the presence of chloride, iodide, and hydroxide salts. The high (or low) iCO and low
(or high) iH2 observed in the presence of large (or small) cations in the electrolyte solution,
correlate to high (or low) selectivity for conversion of CO2 to CO. In the presence of the
chloride, iodide, and hydroxide salts, the selectivity for CO was the highest for the largest cations
(e.g., Cs+, Rb
+).
Effects of cation size on product selectivity, similar to what is reported here, for the
electrochemical reduction of CO2 to methane, ethylene, methanol, and formic acid, has been
observed previously. For example, Hori et al. showed that large cations favor formic acid
production on a mercury electrode [44]. The role of ion selection for formic acid formation is
particularly interesting as this reaction has the same rate limiting step as the CO reaction, the
production of an unstable anion radical intermediate, “CO2∙ However, in this case, little .”־
explanation was given regarding the influence of cation size on selectivity of formic acid
formation over hydrogen evolution.
One would expect that the influence of cations on CO2 reduction is primarily controlled
by ion adsorption at the electrode surface. In general, smaller cations have larger hydration
33
powers, and consequently, have less adsorption on electrode surfaces. Ion adsorption can
stabilize some reactions while suppressing others on an electrode surface. This, in turn,
influences product selectivity, current density, and energetic efficiency.
In our work, the silver cathode used in the reaction is operated well below its point of
zero charge (pzc), and, as a result, cation adsorption is favored on the negatively charged
electrode surface [46]. Smaller cations have larger hydration powers (i.e.,
Li+>Na
+>K
+>Rb
+>Cs
+) [47]. Consequently, the smaller a cation is, the smaller is its propensity
to adsorb on an electrode surface.
Ions adsorbed on an electrode surface influence the reaction in two primary ways. First,
ion adsorption blocks surface sites and thereby decreases the current density [48]. Second, ions
on an electrode surface modify the potential of the outer Helmholtz layer (ϕ2) and the electrode
surface [48]. Because larger cations exhibit a higher propensity for adsorption on an electrode
surface, the electrode potential will be expected to become more positive for larger cations (i.e.,
Li+<Na
+<K
+<Rb
+<Cs
+) [48]. Maznichenko et al. validated this theory by showing that the
measured overpotentials associated with the H2 reduction reaction increased with increasing
cation size (i.e., Li+<Na
+<K
+<Rb
+<Cs
+) [49]. The resulting less negative ϕ2 causes a decrease in
[H+] at the electrode surface as described by the following equation: [49]
(2.1)
where F is Faraday’s constant, R is the gas constant, and T is the absolute temperature.
Since the rate limiting step in this reaction is: , increasing [H+] at the electrode
surface (for example in the presence of smaller and thus more hydrated cations) increases the
reaction rate and decreases the overpotentials for H2 production [44]. As a result, H2 production
34
increases with decreasing cation size (Li+>Na
+>K
+>Rb
+>Cs
+), as we also observed here
(Figure 2.2a).
The second way cations influence the surface reaction is by stabilizing anions on the
cathode surface [50]. For example, for the reaction of S2O8-2
to SO4-2
, large cations within the
double layer of the cathode stabilize the reactive anion and enable much higher current densities,
whereas in the discharge of H+ ions, cations within the double layer repel the reacting particle,
thereby hindering the reaction, resulting in much lower current densities [51]. Further
confirming the mechanism by which large cations stabilize anions on the surface, Frumkin et al.
observed an increase in the adsorption of anions on a cathode in the presence of large cations
(i.e., Cs+), thereby showing that ion-pairing is more pronounced with larger cations, and as a
result, drastically increases the interaction between cations and anions in the double layer [50].
While CO2 and CO are neutral products, the rate limiting step in the reaction is:
, which forms the negatively charged “CO2∙ on the cathode [7,52,53]. Cation adsorption ”־
on the cathode may stabilize “CO2∙ on the electrode surface through ion-pairing, thereby ”־
driving the reaction, improving the current density for a given overpotential. A similar trend was
observed by Ashworth et al. Similar to the situation here, they observed that cation adsorption
improved the reaction kinetics of converting a neutral species with an unstable anion
intermediate on the cathode [54]. Frumkin later proposed that the anion intermediate was
stabilized by the cations on the electrode surface, thereby enabling higher current densities from
electrolyte solutions with large cations [50].
Our systematic investigation of alkali cations on the electrochemical chemical reduction
of CO2 while using a Ag electrode indicates that the cation size plays a significant role in product
selectivity. Furthermore, the similar trends between this work and the aforementioned literature
35
survey, suggest that the influence of cation size on performance stems from cation adsorption on
a cathode surface. Larger cations have a higher propensity for electrode adsorption due to their
low hydration powers, whereas smaller cations are more likely to be hydrated in solution and
consequently, less likely to adsorb on an electrode surface. Cations absorbed on a cathode repel
H+ ions from the cathode. This explains the experimentally observed trend of decreased H2
evolution when larger cations are present in the electrolyte. Additionally, cation adsorption may
stabilize “CO2∙ on the cathode surface, similar to that reported by Ashworth et al., thereby ”־
driving the reaction of CO2 to CO. This trend was experimentally confirmed in Figure 2.2b.
2.3.2 Effect of Anion Size
In this study, we also studied the effect of different anions on CO2 reduction. Figure 2.3
shows the partial current density for CO and the individual polarization curves for the
electrochemical reduction of CO2 to CO in the presence of three anions (chloride, iodide and
hydroxide) paired with a potassium cation. These anions strongly influence the polarization of
both the anode and the cathode. On the anode, we observe a drastic improvement in the oxygen
evolution reaction in the presence of a hydroxide anion (Figure 2.3b). The trend of improved
anode kinetics in the presence of hydroxide solutions also was observed in the presence of all of
Figure 2.3: (a) Partial current densities of CO, and (b) Individual electrode polarization curves for KOH, KCl, and
KI for cell potentials between 2 and 3 V.
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
-150-100-500
Ele
ctro
de
Po
tent
ial
(vs.
Ag/
AgC
l)
Current Density (mA/cm2)
KOH
KCl
KI
-120
-100
-80
-60
-40
-20
0
-3.00-2.75-2.50-2.25-2.00
Cu
rre
nt
De
nsi
ty (
mA
/cm
2)
Cell Potential (V)
KOH
KCl
KI
(a) (b)
Anode
Cathode
36
the other cations tested. While many factors
are at play, the predominant effect from the
addition of a hydroxide anion probably stems
from an increase in the solution pH and,
consequently, improved oxygen evolution
kinetics. On the cathode, we observed a less
pronounced, yet consistent trend.
Specifically, we see in Figure 2.3b that small
anions seemed to reduce the cathode
overpotentials at a given current density. Furthermore, as seen in Figure 2.3a, the reduced
overpotentials result in higher partial current densities for CO across the range of conditions
tested for small anions (OH-<Cl
-<I
-). This observation is in agreement with simulations which
suggest that smaller anions improve the extent of cation adsorption on a negative electrode [55].
Further confirming this explanation, Frumkin, et al. showed that a decrease in cation size
improves anion absorption on the anode. In a similar fashion, our results suggest that decreasing
the anion size improves cation absorption on a cathode, and consequently reduces the
overpotentials (Figure 2.3b) and improves the current density for CO (Figure 2.3a) [50].
2.3.3 Effect of pH
Figure 2.4 summarizes the influence of the electrolyte solution pH on the partial current
density of CO. Small changes in the partial current density for CO are seen between pH 13 and
pH 2. However, at pH 1, the partial current density drastically decreases.
Because pH effects more than just the reduction of CO2 to CO on the cathode (e.g., H2
evolution on the cathode and O2 evolution on the anode), changing the pH can significantly
Figure 2.4: Effect of pH on the partial current density of
CO production.
-60
-50
-40
-30
-20
-10
0
-2-1.5-1-0.5
Cu
rre
nt D
en
sity
(m
A/c
m2 )
Cathode Potential (vs. Ag/AgCl)
pH 0
pH 1
pH 4
pH 10
pH 13
pH 14
0
20
40
60
80
100
-3.5-3-2.5-2
Fara
da
ic E
ffe
cie
ncy
(%
)
Cell Potential (V)
pH 0 pH 1 pH 2 pH 4 pH 10 pH 13 pH 14
0
20
40
60
80
100
-3.5-3-2.5-2
Fara
dai
c Ef
feci
ency
(%
)
Cell Potential (V)
pH 0 pH 1 pH 2 pH 4 pH 10 pH 13 pH 14
37
affect product selectivity. Figure 2.5 shows
the influence of the electrolyte solution pH
on the Faradaic efficiency of the reactor for
CO. Very high pH levels (i.e., pH 14)
improve the selectivity for CO production
even at low cell potentials. Also, low pH
levels decrease the selectivity for CO as
opposed to H2. Specifically, at a pH of 2, a
decrease in the Faradaic efficiency is
observed, at a pH of 1, the a further decrease in the Faradaic efficiency is observed, and, at a pH
of 0, the reactor is completely selective for H2.
The transition from high selectivity for CO in basic conditions to high selectivity for H2
in acidic conditions can be explained via individual electrode plots. Figure 2.6 shows the
cathode and anode polarization for a pH range of 0 to 14. In very basic conditions, an
improvement in anode performance is observed with little improvement on the cathode.
In acidic conditions, the cathode
overpotential is drastically reduced
because the cathode is flooded with
excess protons, which drive the
evolution of H2 as opposed to the
reduction of CO2 to CO. Furthermore,
neutral conditions experience “pH
sheltering” as the electrode surface pH
Figure 2.5: Effect of pH on the Faradaic efficiency for CO
production
0
20
40
60
80
100
-3.5-3-2.5-2
Fara
dai
c Ef
feci
en
cy (
%)
Cell Potential (V)
pH 0
pH 1
pH 2
pH 4
pH 10
pH 13
pH 14
0
20
40
60
80
100
-3.5-3-2.5-2
Fara
da
ic E
ffe
cie
ncy
(%
)
Cell Potential (V)
pH 0
pH 1
pH 2
pH 4
pH 10
pH 13
pH 14
0
20
40
60
80
100
-3.5-3-2.5-2
Fara
dai
c Ef
feci
en
cy (
%)
Cell Potential (V)
pH 0
pH 1
pH 2
pH 4
pH 10
Figure 2.6: Effect of pH on individual electrode polarization plots
for electrolytes with levels of pH ranging from 0 to 14.
-2.5-2
-1.5-1
-0.50
0.51
1.52
2.5
-150-125-100-75-50-250
Ele
ctro
de
Po
t. (
vs. A
g/A
gCl)
Current Density (mA/cm2)
pH 0
pH 1
pH 2
pH 4
pH 10
pH 13
pH 14
38
will be sheltered from the electrolyte solution pH because the anode and cathode reactions
produce and consume protons [17,56,57].
2.3.4 Comparison to Literature
Figure 2.7 summarizes the data obtained in this study. Specifically the measured current
densities plotted as a function of the Faradaic and energetic efficiencies for both the data
reported here and state-of-the-art data reported earlier [15,29-31,40,58-60]. We observe a drastic
improvement in the achieved resultant energetic (~50 %) and Faradaic (~100 %) efficiencies at
high current densities. In particular, with a 1 M RbOH electrolyte solution, current densities of
175 mA/cm2 and 65 mA/cm2 were achieved at cathode potentials of -2.0 V and -1.73 V,
respectively (Figure 2.2). This data corresponds to Faradaic efficiencies of 92 % and energetic
efficiencies of 43 and 41 %, respectively, as seen in Figure 2.4. This is a drastic improvement
compared to state-of-the-art electrochemical reduction of CO2, which tend to have drastically
reduced faradaic efficiencies at increased current densities.
Figure 2.7: Comparison of the (a) Faradaic and (b) energetic efficiencies and current densities for CO2 reduction to
carbon monoxide from literature and this work.[15,29-31,40,58-60].
2.4 Conclusions
Electrolyte composition influences many factors in an electrochemical reactor, including
conductivity, and pH. Here, cation size within electrolyte solutions was observed to play a
0
20
40
60
80
100
0 50 100 150 200
Fara
dai
c Ef
feci
ency
Current Density (mA/cm2)
This Data
Literature
0
20
40
60
80
100
0 50 100 150 200
Ener
geti
c Ef
feci
en
cy
Current Density (mA/cm2)
This Data
Literature
(a) (b)
0
10
20
30
40
50
60
70
80
90
100
0 200 400 600 800 1000 1200 1400 1600
Ene
rgy
Effi
cie
ncy
Current Density (mA/cm2)
This work
Data from Literature
Hydrocarbons0
20
40
60
80
100
0 50 100 150 200
Fara
dai
c Ef
feci
ency
Current Density (mA/cm2)
This Data
Literature
0
10
20
30
40
50
60
70
80
90
100
0 200 400 600 800 1000 1200 1400 1600
Ene
rgy
Effi
cie
ncy
Current Density (mA/cm2)
This work
Data from Literature
Hydrocarbons 0
20
40
60
80
100
0 50 100 150 200
Ener
geti
c Ef
feci
en
cy
Current Density (mA/cm2)
This Data
Literature
0
10
20
30
40
50
60
70
80
90
100
0 200 400 600 800 1000 1200 1400 1600
Ene
rgy
Effi
cie
ncy
Current Density (mA/cm2)
This work
Data from Literature
Hydrocarbons
0
10
20
30
40
50
60
70
80
90
100
0 200 400 600 800 1000 1200 1400 1600
Ene
rgy
Effi
cie
ncy
Current Density (mA/cm2)
This work
Data from Literature
Hydrocarbons
0%
20%
40%
60%
80%
100%
-3.25-3.00-2.75-2.50-2.25-2.00-1.75
Fara
dai
c Ef
fici
en
cy (%
)
Cell Potential
NaOH KOH RbOH CsOH
0%
20%
40%
60%
80%
100%
-3.25-3.00-2.75-2.50-2.25-2.00-1.75
Fara
dai
c Ef
fici
en
cy (%
)
Cell Potential
NaOH KOH RbOH CsOH
39
significant role in the electrochemical reduction of CO2 into CO. Specifically, small cations
(e.g., Li+, and Na
+) favor H2 evolution, and, consequently, have a low Faradaic yield for CO
production, whereas large cations (e.g., Rb+, Cs
+) favor CO production and suppress H2
evolution, resulting in a high Faradaic yield for CO, here up to 96 %. We observed a highest
partial current density of 175 mA/cm2 for CO at a cell potential of 3.0 V in the presence of
RbOH. Furthermore, we observed strong trends between anion choice and cell performance.
Anions influence both the anode and cathode polarization and consequently, the partial current
density for CO. Specifically, our data suggests that smaller anions reduce the overpotentials and
that the presence of OH- (high pH) further improves performance.
A major obstacle to the broader application of the electrochemical reduction of CO2 lies
in the challenge of simultaneously achieving high current densities and energetic efficiencies [7].
Here we demonstrate a major step towards improving both the energetic efficiencies and current
densities of the electrochemical reduction of CO2 to CO on a Ag cathode. In particular, we
observe a drastic improvement in the achieved resultant energetic (~50 %) and Faradaic (~90 %)
efficiencies at high current densities.
The work reported here suggest that further exploration of the effect of cations and anions
on the electrochemical conversion of CO2 to CO, or to other products, may lead to further
improvements in performance. In particular, further exploration of anion effects on the
electrochemical reduction of CO2 may be promising, as anions influence many parameters,
including product selectivity, cation adsorption (through ion-pairing), anode kinetics, and
conductivity.
40
2.5 References
1. National Petroleum Council. Committee on Global Oil and Gas., Hard truths: Facing the
Hard Truths About Energy: A Comprehensive View to 2030 of Global Oil and Natural
Gas, National Petroleum Council, Washington, DC, 2007.
2. G. Centi and S. Perathoner, Opportunities and Prospects in the Chemical Recycling of
Carbon Dioxide to Fuels, Catalysis Today, 2009, 148 (3-4), p. 191-205.
3. K. S. Lackner, Carbonate Chemistry for Sequestering Fossil Carbon, Annual Review of
Energy and the Environment, 2002, 27 p. 193-232.
4. M. Lenzen, Current State of Development of Electricity-Generating Technologies: A
Literature Review, Energies, 2010, 3 (3), p. 462-591.
5. S. I. Plasynski, J. T. Litynski, H. G. McIlvried and R. D. Srivastava, Progress and New
Developments in Carbon Capture and Storage, Critical Reviews in Plant Sciences, 2009,
28 (3), p. 123-138.
6. U.S.EPA, U.S. EPA, 2011, pp. EPA 430-R-411-005.
7. D. T. Whipple and P. J. A. Kenis, Prospects of CO2 Utilization via Direct Heterogeneous
Electrochemical Reduction, Journal of Physical Chemistry Letters, 2010, 1 (24), p. 3451-
3458.
8. S. Bachu, CO2 Storage in Geological Media: Role, Means, Status and Barriers to
Deployment, Progress in Energy and Combustion Science, 2008, 34 (2), p. 254-273.
9. K. van Alphen, P. M. Noothout, M. P. Hekkert and W. C. Turkenburg, Evaluating the
Development of Carbon Capture and Storage Technologies in the United States,
Renewable & Sustainable Energy Reviews, 2010, 14 (3), p. 971-986.
10. K. Hara and T. Sakata, Electrocatalytic Formation of CH4 from CO2 on a Pt Gas
Diffusion Electrode, Journal of the Electrochemical Society, 1997, 144 (2), p. 539-545.
11. J.M. Duthie and H.W. Whittington, Securing Renewable Energy Supplies Through
Carbon Dioxide Storage in Methanol, Power Engineering Society Summer Meeting,
2002 IEEE, 2002, 1 p. 145-150.
12. Paul McArdle, Emissions of Greenhouse Gases in the United States 2008, 2008.
13. G. Arai, T. Harashina and I. Yasumori, Selective Electrocatalytic Reduction of Carbon-
Dioxide to Methanol on Ru-Modified Electrode, Chemistry Letters, 1989, (7), p. 1215-
1218.
14. D. P. Summers, S. Leach and K. W. Frese, The Electrochemical Reduction of Aqueous
Carbon-Dioxide to Methanol at Molybdenum Electrodes with Low Overpotentials,
Journal of Electroanalytical Chemistry, 1986, 205 (1-2), p. 219-232.
15. Y. Hori, H. Wakebe, T. Tsukamoto and O. Koga, Electrocatalytic Process of CO
Selectivity in Electrochemical Reduction of CO2 at Metal-Electrodes in Aqueous-Media,
Electrochimica Acta, 1994, 39 (11-12), p. 1833-1839.
16. M. Azuma, K. Hashimoto, M. Hiramoto, M. Watanabe and T. Sakata, Electrochemical
Reduction of Carbon-Dioxide on Various Metal-Electrodes in Low-Temperature Aqueous
KHCO3 Media, Journal of the Electrochemical Society, 1990, 137 (6), p. 1772-1778.
17. M. N. Mahmood, D. Masheder and C. J. Harty, Use of Gas-Diffusion Electrodes for
High-Rate Electrochemical Reduction of Carbon-Dioxide .1. Reduction at Lead, Indium-
41
Impregnated and Tin-Impregnated Electrodes, Journal of Applied Electrochemistry,
1987, 17 (6), p. 1159-1170.
18. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Microfluidic Reactor for the
Electrochemical Reduction of Carbon Dioxide: The Effect of pH, Electrochemical and
Solid-State Letters, 2010, 13 (9), p. D109-D111.
19. Y. Akahori, N. Iwanaga, Y. Kato, O. Hamamoto and M. Ishii, New Electrochemical
Process for CO2 Reduction to Formic Acid from Combustion Flue Gases,
Electrochemistry, 2004, 72 (4), p. 266-270.
20. F. Koleli, T. Atilan, N. Palamut, A. M. Gizir, R. Aydin and C. H. Hamann,
Electrochemical Reduction of CO2 at Pb- and Sn-electrodes in a Fixed-bed Reactor in
Aqueous K2CO3 and KHCO3 Media, J. of Appl. Electrochem., 2003, 33 (5), p. 447-450.
21. N. Furuya, T. Yamazaki and M. Shibata, High Performance Ru-Pd Catalysts for CO2
Reduction at Gas-Diffusion Electrodes, Journal of Electroanalytical Chemistry, 1997, 431
(1), p. 39-41.
22. S. Kaneco, R. Iwao, K. Iiba, K. Ohta and T. Mizuno, Electrochemical Conversion of
Carbon Dioxide to Formic Acid on Pb in KOH/Methanol Electrolyte at Ambient
Temperature and Pressure, Energy, 1998, 23 (12), p. 1107-1112.
23. S. Kaneco, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of CO2 to
Methane at the Cu Electrode in Methanol with Sodium Supporting Salts and its
Comparison with Other Alkaline Salts, Energy Fuels, 2006, 20 (1), p. 409-414.
24. S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of
High Pressure CO2 at a Cu Electrode in Cold Methanol, Electrochimica Acta, 2006, 51
(23), p. 4880-4885.
25. R. L. Cook, R. C. Macduff and A. F. Sammells, High-Rate Gas-Phase CO2 Reduction to
Ethylene and Methane Using Gas-Diffusion Electrodes, Journal of the Electrochemical
Society, 1990, 137 (2), p. 607-608.
26. S. Kaneco, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of Carbon
Dioxide to Ethylene at a Copper Electrode in Methanol Using Potassium Hydroxide and
Rubidium Hydroxide Supporting Electrolytes, Electrochimica Acta, 2006, 51 (16), p.
3316-3321.
27. T. Mizuno, K. Ohta, M. Kawamoto and A. Saji, Electrochemical Reduction of CO2 on Cu
in 0.1 M KOH-methanol, Energy Sources, 1997, 19 (3), p. 249-257.
28. J. Yano and S. Yamasaki, Pulse-mode Electrochemical Reduction of Carbon Dioxide
Using Copper and Copper Oxide Electrodes for Selective Ethylene Formation, Journal of
Applied Electrochemistry, 2008, 38 (12), p. 1721-1726.
29. N. Furuya and S. Koide, Electroreduction of Carbon-Dioxide by Metal Phthalocyanines,
Electrochimica Acta, 1991, 36 (8), p. 1309-1313.
30. T. Yamamoto, D. A. Tryk, A. Fujishima and H. Ohata, Production of Syngas Plus
Oxygen from CO2 in a Gas-Diffusion Electrode-Based Electrolytic Cell, Electrochimica
Acta, 2002, 47 (20), p. 3327-3334.
31. C. Delacourt, P. L. Ridgway, J. B. Kerr and J. Newman, Design of an Electrochemical
Cell Making Syngas (CO+H2) from CO2 and H2O Reduction at Room Temperature,
Journal of the Electrochemical Society, 2008, 155 (1), p. B42-B49.
42
32. F. Bidrawn, G. Kim, G. Corre, J. T. S. Irvine, J. M. Vohs and R. J. Gorte, Efficient
Reduction of CO2 in a Solid Oxide Electrolyzer, Electrochemical and Solid-State Letters,
2008, 11 (9), p. B167-B170.
33. R. M. M. Abbaslou, J. S. S. Mohammadzadeh and A. K. Dalai, Review on Fischer-
Tropsch Synthesis in Supercritical Media, Fuel Processing Technology, 2009, 90 (7-8), p.
849-856.
34. M. E. Dry, High Quality Diesel via the Fischer-Tropsch Process - a Review, Journal of
Chemical Technology and Biotechnology, 2002, 77 (1), p. 43-50.
35. Mark E. Dry, Practical and theoretical aspects of the catalytic Fischer-Tropsch process,
Applied Catalysis A: General, 1996, 138 (2), p. 319-344.
36. S. D. Ebbesen and M. Mogensen, Electrolysis of Carbon Dioxide in Solid Oxide
Electrolysis Cells, Journal of Power Sources, 2009, 193 (1), p. 349-358.
37. S. H. Jensen, X. F. Sun, S. D. Ebbesen, R. Knibbe and M. Mogensen, Hydrogen and
Synthetic Fuel Production Using Pressurized Solid Oxide Electrolysis Cells, International
Journal of Hydrogen Energy, 2010, 35 (18), p. 9544-9549.
38. V. Kaplan, E. Wachtel, K. Gartsman, Y. Feldman and I. Lubomirsky, Conversion of CO2
to CO by Electrolysis of Molten Lithium Carbonate, Journal of the Electrochemical
Society, 2010, 157 (4), p. B552-B556.
39. Z. L. Zhan and L. Zhao, Electrochemical Reduction of CO2in Solid Oxide Electrolysis
Cells, Journal of Power Sources, 2010, 195 (21), p. 7250-7254.
40. Brian A Rosen, Amin Salehi-Khojin, Michael R Thorson, Wei Zhu, Devin T Whipple,
Paul J A Kenis and Richard I Masel, Selective Conversion of CO2 to CO at Low
Overpotentials, Science, 2011, p.
41. Emily Barton Cole, Prasad S. Lakkaraju, David M. Rampulla, Amanda J. Morris, Esta
Abelev and Andrew B. Bocarsly, Using a One-Electron Shuttle for the Multielectron
Reduction of CO2 to Methanol: Kinetic, Mechanistic, and Structural Insights, Journal of
the American Chemical Society, 2010, 132 (33), p. 11539-11551.
42. G. Seshadri, C. Lin and A. B. Bocarsly, A New Homogeneous Electrocatalyst for the
Reduction of Carbon-Dioxide to Methanol at Low Overpotential, Journal of
Electroanalytical Chemistry, 1994, 372 (1-2), p. 145-150.
43. A. Murata and Y. Hori, Product Selectivity Affected by Cationic Species in
Electrochemical Reduction of CO2 and CO at a Cu Electrode, Bulletin of the Chemical
Society of Japan, 1991, 64 (1), p. 123-127.
44. Y. Hori and S. Suzuki, Electrolytic Reduction of Carbon-Dioxide at Mercury-Electrode
in Aqueous-Solution, Bulletin of the Chemical Society of Japan, 1982, 55 (3), p. 660-665.
45. R. S. Jayashree, L. Gancs, E. R. Choban, A. Primak, D. Natarajan, L. J. Markoski and P.
J. A. Kenis, Air-breathing laminar flow-based microfluidic fuel cell, Journal of the
American Chemical Society, 2005, 127 (48), p. 16758-16759.
46. G. J. Clark, T. N. Andersen, R. S. Valentine and H. Eyring, A Comparison of the
Immersion and Open-Circuit Scrape Methods for Determining the Potential of Zero
Charge of Metal Electrodes, Journal of the Electrochemical Society, 1974, 121 (5), p.
618-622.
47. J.O.M. Bockris and A.K.N. Reddy, Modern electrochemistry, Plenum Press, 1998.
43
48. P. Delahay, Double layer and electrode kinetics, Interscience Publishers, 1965.
49. E. A. Maznichenko, B. B. Damaskin and Z. A. Iofa, Effect of alkali metal cations on the
overvoltage of hydrogen on mercury in acid solutions, Dokl. Akad. Nauk SSSR, 1961,
138 p. 1377.
50. A. N. Frumkin, Influence of cation adsorption on the kinetics of electrode processes,
Transactions of the Faraday Society, 1959, 55 p. 156-167.
51. Krjukova, Dokl. Akad. Nauk SSSR, 1949, 65 p. 517.
52. R. P. S. Chaplin and A. A. Wragg, Effects of Process Conditions and Electrode Material
on Reaction Pathways for Carbon Dioxide Electroreduction with Particular Reference to
Formate Formation, Journal of Applied Electrochemistry, 2003, 33 (12), p. 1107-1123.
53. Donald A. Tryk and Akira Fujishima, Global Warming Electrochemists Enlisted in War:
The Carbon Dioxide Reduction Battle, The Electrochemical Society Interface, 2001, p.
32-36.
54. Ashworth, Coll. Czech. Chem. Comm., 1948, 13 p. 229.
55. Paul S. Crozier, Richard L. Rowley and Douglas Henderson, Molecular-dynamics
simulations of ion size effects on the fluid structure of aqueous electrolyte systems
between charged model electrodes, Journal of Chemical Physics, 2001, 114 (17), p.
7513-7517.
56. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Microfluidic Reactor for the
Electrochemical Reduction of Carbon Dioxide: The Effect of pH, Electrochemical and
Solid State Letters, 2010, 13 (9), p. D109-D111.
57. H. Li and C. Oloman, Development of a Continuous Reactor for the Electro-reduction of
Carbon Dioxide to Formate - Part 2: Scale-up, Journal of Applied Electrochemistry,
2007, 37 (10), p. 1107-1117.
58. K. W. Frese and S. Leach, Electrochemical Reduction of Carbon-Dioxide to Methane,
Methanol, and CO on Ru Electrodes, Journal of the Electrochemical Society, 1985, 132
(1), p. 259-260.
59. T. Saeki, K. Hashimoto, A. Fujishima, N. Kimura and K. Omata, Electrochemical
Reduction of CO2 with High-Current Density in a CO2-Methanol Medium, Journal of
Physical Chemistry, 1995, 99 (20), p. 8440-8446.
60. Eric Dufek, Tedd Lister and Michael McIlwain, Bench-scale electrochemical system for
generation of CO and syn-gas, Journal of Applied Electrochemistry, 2011, 41 (6), p. 623-
631.
44
Chapter 3
Amine Based Electrolytes for the Electrochemical Reduction of CO21
3.1 Introduction
In Chapter 2, we showed that electrolyte solutions can drastically effect the
electrochemical reduction of CO2 both in terms of current density and faradaic efficiency.
Specifically, we identified that cation size, anion size, and pH play key roles in the product
distribution. Similarly, in the case of many other reactions, cation and anion selection has shown
to drastically effect the reaction kinetics and products [1-7]. We are interested in using
electrolyte solutions to further improve the selectivity, and furthermore, reduce the
overpotentials for the reaction.
1 Part of this work has been published: Rosen, B. A., Salehi-Khojin, A., Thorson, M. R., Zhu, W., Whipple, D. T.;
Kenis, P. J. A., Masel, R.I., “Ionic Liquid–Mediated Selective Conversion of CO2 to CO at Low Overpotentials.”
Science 2011, 334, 643-644. http://www.sciencemag.org/content/334/6056/643
Figure 3.1: Schematic of the proposed mechanism by which EMIM BF4 may stabilize the CO2-
intermediate on the electrode surface.
45
Bockris, etc. showed that the overpotential for CO2 conversion is quite large due to the
high potential to form a CO2 radical intermediate [8-10]. Specifically, the overpotential for the
formation of the “CO2 intermediate” is -1.9 V vs. RHE. If a complex can be formed between a
“co-catalyst” species (as proposed in Figure 3.1) and the intermediate on the catalyst surface, the
potential required to form the intermediate could be reduced (Figure 3.2), thereby reducing the
potential for the overall reaction.
In this study, we explore the use of an ionic liquid, 1-ethyl-3-methylimidazolium
tetrafluoroborate, EMIM BF4, to stabilize the intermediate and thereby reduce the overpotential
necessary for the reaction. Ionic liquids are particularly interesting as they have the advantages
of having relatively high conductivities due to their ionic nature, high solubilities for CO2, and
“tune-ability” due to the ability to change the chemical structure or the cationic/ anionic species
[11,12]. The high solubility for CO2 inherent to many ionic liquids makes them potentially
Figure 3.2: . A schematic of how the free energy of the system changes during the reaction CO2 + 2H
+ +
2e– CO + H2O in water or acetonitrile (solid line) or EMIM-BF4 (dashed line). From Rosen, B. A.,
Salehi-Khojin, A., Thorson, M. R., Zhu, W., Whipple, D. T.; Kenis, P. J. A., Masel, R.I., “Ionic Liquid–
Mediated Selective Conversion of CO2 to CO at Low Overpotentials.” Science 2011, 334, 643-
644.Reprinted with permission from AAAS [2].
Intermediate
stabilized
46
excellent electrolytes because (a) they could be used for both CO2 capture and CO2 reduction and
(b) they may improve CO2 mass transport by increasing the local concentration of CO2 near the
cathode of an electrochemical [13-15].
This study plans to build on work done by our collaborators at Dioxide Materials, who
have identified that EMIM BF4 reduces the overpotential necessary for CO2 reduction to CO by
using a three electrode reactor. Here, in collaboration with Brian Rosen and Dr. Richard Masel
at Dioxide Materials, I build on their results of an early onset potential for CO2 reduction to CO
in an attempt to show a reduced overall reactor potential with this ionic liquid, EMIM BF4, in a
flowing electrochemical reactor. The electrochemical flow reactor discussed in-detail in
Chapter 2 was chosen for this study because (a) the small band gap between the cathode and
anode for the electrolyte stream makes for small ohmic losses, and (b) the flowing electrolyte
stream enables in-situ electrode characterization of both the anode and cathode electrode
polarization.
3.2 Experimental
3.2.1 Reactor Testing
The reactor operational procedure was identical to that used in Chapter 2. The following
electrolyte solutions were used in these studies: 18 mole% EMIM BF4 in 18.2 MΩ∙cm H2O,
1 M KCl in H2O, and 0.1 M H2SO4 in H2O. All of the experiments in this study used 5 mg/cm2
Ag on Sigracet 35BC cathodes. All of the anodes used in this study had a 2 mg/cm2 Pt loading
on Sigracet 35BC unless otherwise specified.
47
3.2.2. Reference Electrodes
Cathode potentials for all EMIM BF4 electrolyte solutions were measured relative to a
non-aqueous Ag/Ag+ reference electrode (BASi, MF-2062). The non-aqueous reference
electrode was prepared with an internal solution consisting of: 0.1 M tetrabutlyammonium
perchlorate (TBAP, Sigma Aldrich) and 0.01 M AgNO3 (BASi) in acetonitrile (Sigma Aldrich).
The Ag/Ag+ reference electrode was calibrated relative to the potential of the
ferrocene/ferrocenium couple in a 2.25 mM ferrocene (Sigma Aldrich), 18 mole% EMIM BF4
aqueous solution, and a 2.25 mM ferrocene in acetonitrile solution. Figure 3.3 shows anodic and
cathodic waves for ferrocene using a Ag/Ag+ reference electrode in acetonitrile and
18 mole% EMIM BF4 solutions. The potential difference between the ferrocene/ferrocenium
couple relative and the Ag/Ag+ reference electrode was +100 mV in the acetonitrile solution and
-23 mV in the EMIM BF4 solution. The shift in potential is believed to be due to differences in
the junction between the reference electrode and the acetontile or EMIM BF4 solutions. Based
on this data, we conclude that the Ag/Ag+ acetonitrile reference electrode will read a potential
Figure 3.3: Anodic and cathodic waves for Ferrocene using a Ag/Ag+ reference electrode in acetonitrile
and 18% EMIM BF4 solutions.
-100
-50
0
50
100
150
200
-0.6 -0.4 -0.2 0 0.2 0.4 0.6
Cu
rre
nt
(uA
)
Potential (vs Ag/Ag+)
Acetonitrile
EMIM BF4
48
which is 123 mV higher in the acetonitrile solution as compared to the EMIM BF4 solution.
Pavlishchuk and Addison report that the Ag/Ag+ reference electrode used in this study is 537 mV
off of SHE [16]. Consequently, the Ag/Ag+ reference electrode is 631 mV (537+123) below
SHE.
For all individual electrode potential measurements relative to an aqueous solution, a
Ag/AgCl reference electrode (BASi, RE-5B) was used in the aqueous stream.
3.2.3 Cyclic Voltammetry
Cyclic voltammetry was carried out in-situ within the electrochemical reactor relative to a
Ag/AgCl or a Ag/Ag+ reference electrode for aqueous or non-aqueous electrolyte solutions,
respectively. A reference electrode was placed upstream in the electrolyte solution and the
cathode potential was controlled relative to reference electrode at a scan rate of 20 mV/s. For all
of the tests, either N2 (an inert gas) or CO2, a reactant gas was flowed over the cathode chamber
at a flow rate of 5 SCCM. The electrolytes were pumped at 0.5 mL/min using a syringe pump
(Harvard Apparatus, PHD 2200).
3.2.4 Electrochemical impedance spectroscopy
Electrochemical impedance spectroscopy (EIS) was used to determine the internal reactor
resistance of the electrochemical reactor used in these studies with the various combinations of
electrolytes and membranes. The impedance spectra were obtained using a Frequency Response
Analyzer (FRA) module controlled by the potentiostat (Autolab PGSTAT-30, EcoChemie). EIS
was done at a potential of 3V to ensure that the reaction was selective for the desired product,
carbon monoxide.
49
3.3 Results and Discussion
3.3.1 Cyclical Voltammetry
Figure 3.4 shows in-situ cyclic
voltammograms (CV) using the single
channel electrochemical reactor described in
Chapter 2 [17] with a Ag cathode with either
a 1 M KCl aqueous electrolyte or a
18 mole% EMIM BF4 in H2O electrolyte
flowed at 0.5 mL/min and with either N2 or
CO2 flowed behind the cathode at 5 SCCM. Because N2 is an inert gas at the potentials applied,
when flowing N2 over the cathode, the only potential product is H2 from water electrolysis [18].
The results obtained when using an ionic liquid electrolyte, EMIM BF4, were compared to results
obtained when using a more traditional electrolyte stream, 1 M aqueous KCl [17,19-21]. In
Figure 3.4, as the cathode potential was scanned in the more negative direction, an increase in
the cathode current density was observed for all four conditions tested, corresponding to an
increase in reduction current. When N2 is flowed over the cathode, for both the EMIM BF4 and
KCl solutions, the observed current is assumed to go towards the evolution of H2 as no CO2 is
available for reduction, and N2 is inert at the potentials applied here [18]. An increase in current
with the substitution of CO2 for N2 suggests that the addition of CO2 results in the reduction of
CO2 to carbon fuels (i.e., CO, CH4, CH3OH, CO2H2, or C2H4). Specifically, the substitution of
CO2 for N2 with both the aqueous and ionic liquid solutions resulted in a considerable increase in
current density, thereby suggesting that the main products in the presence of CO2 are from the
reduction of CO2. Furthermore, the onset potential (as defined as the potential at which a current
Figure 3.4: Cyclic voltammograms at a scan rate of
20 mV/s over a Ag cathode with EMIM BF4 or
aqueous catholytes and a N2 or CO2 gas stream
flowing over the cathode.
-50
-45
-40
-35
-30
-25
-20
-15
-10
-5
0
-1.75 -1.5 -1.25 -1 -0.75 -0.5
Cu
rre
nt d
en
sity
(m
A/c
m2 )
Potential (vs. Ag/AgCl)
1M KCl with N2
1M KCl with CO2
18% EMIM BF4 with N2
18% EMIM BF4 with CO2
50
density of 5 mA/cm2) with the CO2
gas stream is decreased in the case of
the ionic liquid electrolyte solution (-
1.08 V) as compared to the aqueous
solution (-1.246 V). However, the
EMIM BF4 solution exhibited
hysteresis on the return scan, thereby
suggesting that the system may be
mass transfer limited. The decreasing
slope beyond -1.35 for the EMIM BF4
solution also suggests that the system may become mass transport limited at higher current
densities, which is expected due to the poor viscosity of the EMIM BF4 solution. In the presence
of the EMIM BF4 solution, the rate of hydrogen evolution compared to that in the aqueous
solution also increases, but to a lesser extent than the CO2 reduction. Because different reference
electrodes were used for the ionic liquid solution and the aqueous solution, the early onset
potential needed to be confirmed via the total applied potential in a flow cell.
3.3.2 Individal Electrode Potentials
Using the single electrolyte channel reactor with a Ag cathode and a Pt anode and a CO2
gas stream (5 SCCM), the individual electrode performance of the reactor with a 18 mole%
EMIM BF4 electrolyte solution and a 1M KCl aqueous solutions, both flowed at 0.5 ml/min, are
compared in Figure 3.5. Relative to the aqueous stream, the ionic liquid solution exhibits poor
anode performance, which is hypothesized to be due to the ionic liquid, EMIM BF4, poisoning
the anode. Regarding the cathode performance, the overpotentials at current densities appear to
Figure 3.5: Individual electrode potentials of 18% EMIM
BF4 and 1M KCl for a Ag cathode and a Pt Anode in a single
electrolyte channel reactor
51
be lower when using the EMIM BF4 electrolyte as compared to when using the aqueous KCl
electrolyte, however, a greater overpotential is required at higher current densities. The
increased cathode overpotentials confirms that the reactor may be mass transport limitations. In
further support of this theory, the diffusivity of CO2 in EMIM BF4 is ~10 times worse than the
diffusivity of CO2 in an aqueous liquid [2]. To independently address and study the influence of
the electrolyte on the cathode and anode, we developed a dual electrolyte reactor.
3.3.3 Dual Electrolyte Reactor
The dual electrolyte reactor, as shown in Figure 3.6 and originally discussed in Chapter 2,
was originally developed to enable the study of the influence of EMIM BF4 on the anode and
cathode performance. This design is similar to the single electrolyte channel electrochemical
reactor used previously [17,19]. The reactor has a separate anode and cathode GDEs backed by
aluminum current collectors. The anode is exposed to the atmosphere to enable the O2 to escape
the reactor. A gas channel above the cathode enables both the feed of CO2 and the escape of CO
through the vapor phase. Two electrolytes, an anolyte and a catholyte, flow over the anode and
cathode, respectively. The electrolyte streams are separated via a membrane. The dual
Figure 3.6: Schematic diagram of the dual electrolyte microfluidic reactor for CO2 conversion.
52
electrolyte streams enable separate
control of conditions at each electrode
and can thereby prevent the ionic liquid
from poisoning the anode via physical
separation. Reference electrodes are
placed in each of the outlet streams
(Ag/Ag+ in the ionic liquid and
Ag/AgCl in the aqueous stream). The
reference electrodes enable
measurement of the potentials of the
individual electrodes as well as the
junction potential between the two electrolytes (the difference between the reference electrodes
at open circuit).
3.3.4 Membranes
Membrane selection is critical for maintaining ionic contact between the two electrolyte
streams (thereby reducing reactor resistance), preventing membrane poisoning, limiting the
junction potential and preventing mixing of the electrolyte streams.
Here, several potential membranes were explored, a cation exchange membrane, an anion
exchange membrane, and a porous separator. The cation exchange membrane, Nafion (Dupont),
has the advantage of drastically reducing fuel crossover while allowing for the transfer of H+
cations in a low resistance environment, and is consequently the most commonly used fuel cell
membrane. Anionic membranes (i.e., Acta Membrane) allow the transfer of an anion, while
limiting cation crossover. Lastly, Hollinger et al. identified that filter paper (Track-Etch
Figure 3.7: Individual electrode potentials of EMIM BF4
and KCl for in a single electrode reactor and a combination
of EMIM BF4 and KCl in a dual electrolyte reactor with
Nafion or tracketch membranes.
53
membranes) can be used in-lieu of a
more traditional membrane to reduce
crossover of unwanted species through
the reduction of the effective cross-
sectional area [22]. Figure 3.7 shows
overlays the individual cathode (as
measured via Ag/Ag+ reference
electrode in catholyte) and anode (as
measured via Ag/AgCl reference
electrode in anolyte) polarization
curves for the aforementioned three
membranes (Nafion, Tracketck, and
Acta membrane) in the dual electrolyte reactor (18 % EMIM BF4 at the Ag cathode and 1M
H2SO4 at the Pt Anode) as compared to a single channel reactor with a 18 mole% EMIM BF4
electrolyte and a 1 M KCl aqueous electrolyte. The anode polarization curves show that (1) a
Nafion membrane prevents an EMIM BF4 solution from poisoning the anode, (2) a Track-Etched
membrane reduced, but didn’t eliminate the poisoning on the anode, and (3) the anion exchange
membrane (Acta), reduces the poisoning but has less of an effect at high current densities, where
the anion crossover rate is high. The effectiveness of the cation exchange membrane (which
prevents anion diffusion across the membrane) and the ineffectiveness of the anion exchange
membrane (which allows anion diffusion across the membrane) at high current densities, where
the anion crossover rate is high, suggests that the anion, BF4-, is responsible for the anode
poisoning.
Figure 3.8: Nyquist plots from EIS of 3 different reactor
electrolyte arrangements. Aq. is a 1 M KCl aqueous
solution, IL is an 18 mole% EMIM BF4 solution, and
IL_N_Aq is an 18 mole% EMIM BF4 catholyte and a 1 M
KCl anolyte separated by a Nafion membrane
0.00
2.00
4.00
6.00
8.00
10.00
0 5 10 15-l
m Z
(Ω
-cm
2)
Re Z (Ω-cm2)
Aq.
IL
IL_N_Aq.
54
While a Nafion membrane exhibits superior behavior in terms of preventing anode
poisoning, issues with membrane poisoning from ionic liquids have previously been reported
[23]. To study the EMIM BF4 / Nafion interaction, impedance was carried out with and without
the presence of the membrane. Nyquist plots of the EIS data from the reactor with various
electrolytes and membranes are shown in Figure 3.8. The distance between the vertical axis and
the high frequency points (the points closest to the origin) represents the ohmic resistance (RΩ)
or reactor resistance (Rreactor) of the system and is composed of the electrolyte resistance and any
resistance in the components of the reactor.
The membrane resistance can be estimated from the Nyquist plots by measuring the
resistance with the membrane and subtracting the resistance without the membrane. From
Figure 3.8, we identify the resistance of the reactor with and without the membrane.
Specifically, with a single channel reactor using an 18 mole% EMIM BF4 electrolyte or a 1 M
KCl aqueous electrolyte, the reactor resistance is 3.41 or 2.65 Ω cm2, respectably. Based on
these resistances, the combined resistance of an arrangement with a 212 Nafion membrane
separating an 18 mole% EMIM BF4 electrolyte and a 1 M KCl electrolyte (with no poisoning or
no junction potential) could be approximated by the mathematical addition of these resistances,
6.06 Ω cm2. However, the Nyquist plots shows that for the aforementioned
electrolyte/membrane arrangement, the total reactor resistance is 12.73 Ω cm2. In conclusion,
the ionic liquid, EMIM BF4, appears to be swelling or blocking sites in the Nafion membrane
and thereby increasing the reactor resistance.
Additionally, we measured the reactor resistance in the single channel reactor with only
18 % EMIM BF4 and with only 1 M KCl. The reactor resistances in these two cases were
expected to be quite similar as both solutions have very similar conductivities. However, with
55
the EMIM BF4 electrolyte solution in the reactor, an additional resistance of 0.85 Ω cm2 was
recorded, which suggests that the EMIM BF4 electrolyte solution is interfering with one of the
electrodes, and thereby increasing the reactor resistance. We hypothesize that the additional
reactor resistance originates from EMIM BF4 swelling the Nafion binder used to hold the catalyst
metal (Ag or Pt) to the GDE (Sigracet 35BC) because a similar phenomena was observed by
Schmidt, et al. Specifically, he observed that ionic liquids can swell Nafion [23]. With the
EMIM electrolyte, the reactor is resistance limited.
3.3.5 Early Onset Potenial in Operational Reactor
Figure 3.9 shows the faradaic efficiency for CO and H2, total energetic efficiency, and
reaction turnover number between reactor potentials of 1.5 and 2.5 V reported by my
collaborator, Brian Rosen, at Dioxide Materials when using a EMIM BF4 – Nafion – H2SO4 dual
electrolyte setup [2]. The onset of CO production is achieved at a total reactor potential of only
Figure 3.9: Performance for EMIM BF4 – Nafion – H2SO4 setup a dual electrolyte setup in terms of
faradaic efficiency for CO, and H2, total energetic efficiency, and reaction turnover number. From
Rosen, B. A., Salehi-Khojin, A., Thorson, M. R., Zhu, W., Whipple, D. T.; Kenis, P. J. A., Masel, R.I.,
“Ionic Liquid–Mediated Selective Conversion of CO2 to CO at Low Overpotentials.” Science 2011, 334,
643-644.Reprinted with permission from AAAS [2].
56
1.5 V in the presence of an
18 % EMIM BF4 catholyte
This low cell potential,
1.5 V, constitutes a drastic
reduction in the onset
potential for CO production
(0.17 V overpotential),
which theoretically occurs at
1.33 V [6,7,19].
Furthermore, the results
show that excellent selectivity for CO can be achieved between reactor potentials of 1.5 and
2.5 V.
3.3.6 Low Current Densities
Figure 3.10 shows the current densities achieved with a single electrolyte channel reactor
with a 1 M KCl aqueous electrolyte solution compared to the current density achieved with a
18 %EMIM BF4 –Nafion – 1M H2SO4 electrolyte setup in the dual electrolyte reactor. While
EMIM BF4 drastically reduces the onset potential, the current densities are too low in the present
setup for immediate application. Specifically, drastic reductions in the cell resistance (or
membrane poisoning) are necessary to capitalize on the low onset potentials observed in the
presence of EMIM BF4.
3.4 Conclusions
EMIM BF4 solutions show tremendous potential for reducing the required overpotentials
for reaction. Specifically, the cyclical voltammetry confirms that earlier onset potential are
Figure 3.10:Current density output for an electrochemical cell with an
18% EMIM BF4 solution as compared to an electrochemical cell with a
1 M KCl electrolyte solution.
-120
-100
-80
-60
-40
-20
0
-3.25-3.00-2.75-2.50-2.25-2.00-1.75
Cu
rre
nt D
en
sity
(mA
/cm
2)
Cell Potential
EMIM BF4
KCl
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
-4.0-3.5-3.0-2.5-2.0-1.5
Cu
rre
nt D
en
sity
(mA
/cm
2)
Cell Potential (V)
57
possible in the flowing reactors. However, both the CV and single electrode plots suggest that
mass transport may hinder the reaction rate on the electrode surface.
The individual electrode polarization plots clearly show that EMIM BF4 poisons the Pt
anode. However, the development of the dual electrolyte reactor, with a Nafion cation-exchange
membrane, enables a drastic reduction in the extent of cathode poisoning observed. Because the
poisoning was only eliminated via the use of a cation exchange membrane eliminates, we believe
the BF4 group is responsible for the anode poisoning. While a Nafion membrane helps to
overcome anode poisoning, the Nafion membrane appears to be poisoned by an EMIM BF4
solution as is evident by an increased reactor resistance with the addition of the membrane.
In spite of the high reactor resistances, the dual electrolyte reactor enabled the
identification of the onset potential for CO. The onset of CO production is achieved at a total
reactor potential of only 1.5 V in presence of 18 % EMIM BF4, which constitutes a total reactor
overpotential of 0.17 V. The low onset potential observed here constitutes a drastic reduction in
the onset potential for CO production in an arrangement that achieves excellent selectivity for
CO. The reactor developed in this study served the purpose of analyzing the overall reactor
onset potential, however, practical applications require much higher current densities, which
result in large overpotentials in the present setup.
3.5 References
1. S. Kaneco, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of CO2 to
Methane at the Cu Electrode in Methanol with Sodium Supporting Salts and its
Comparison with Other Alkaline Salts, Energy Fuels, 2006, 20 (1), p. 409-414.
2. Brian A. Rosen, Amin Salehi-Khojin, Michael R. Thorson, Wei Zhu, Devin T. Whipple,
Paul J. A. Kenis and Richard I. Masel, Ionic Liquid–Mediated Selective Conversion of
CO2 to CO at Low Overpotentials, Science, 2011, 334 (6056), p. 643-644.
3. E. B. Cole, P. S. Lakkaraju, D. M. Rampulla, A. J. Morris, E. Abelev and A. B. Bocarsly,
Using a One-Electron Shuttle for the Multielectron Reduction of CO2 to Methanol:
Kinetic, Mechanistic, and Structural Insights, Journal of the American Chemical Society,
2010, 132 (33), p. 11539-11551.
58
4. A. J. Morris, R. T. McGibbon and A. B. Bocarsly, Electrocatalytic Carbon Dioxide
Activation: The Rate-Determining Step of Pyridinium-Catalyzed CO2 Reduction,
Chemsuschem, 2011, 4 (2), p. 191-196.
5. G. Seshadri, C. Lin and A. B. Bocarsly, A New Homogeneous Electrocatalyst for the
Reduction of Carbon-Dioxide to Methanol at Low Overpotential, Journal of
Electroanalytical Chemistry, 1994, 372 (1-2), p. 145-150.
6. Y. Hori, Electrochemical CO2Reduction on Metal Electrodes, Modern Aspects of
Electrochemistry, 2008, 42 p. 89-189.
7. Yoshio Hori, Hidetoshi Wakebe, Toshio Tsukamoto and Osamu Koga, Electrocatalytic
process of CO selectivity in electrochemical reduction of CO2at metal electrodes in
aqueous media, Electrochimica Acta, 1994, 39 (11–12), p. 1833-1839.
8. J. O. Bockris and J. C. Wass, The Photocatalytic Reduction of Carbon-Dioxide, Journal
of the Electrochemical Society, 1989, 136 (9), p. 2521-2528.
9. J.O.M. Bockris and A.K.N. Reddy, Modern electrochemistry, Plenum Press, 1998.
10. K. Chandrasekaran and L. O'M Bockris, In-situ spectroscopic investigation of adsorbed
intermediate radicals in electrochemical reactions: CO2− on platinum, Surface Science,
1987, 185 (3), p. 495-514.
11. J. E. Bara, T. K. Carlisle, C. J. Gabriel, D. Camper, A. Finotello, D. L. Gin and R. D.
Noble, Guide to CO2Separations in Imidazolium-Based Room-Temperature Ionic
Liquids, Industrial & Engineering Chemistry Research, 2009, 48 (6), p. 2739-2751.
12. S. J. Zhang, Y. H. Chen, F. W. Li, X. M. Lu, W. B. Dai and R. Mori, Fixation and
Conversion of CO2Using Ionic Liquids, Catalysis Today, 2006, 115 (1-4), p. 61-69.
13. A. Yokozeki, M. B. Shiflett, C. P. Junk, L. M. Grieco and T. Foo, Physical and Chemical
Absorptions of Carbon Dioxide in Room-Temperature Ionic Liquids, Journal of Physical
Chemistry B, 2008, 112 (51), p. 16654-16663.
14. L. E. Barrosse-Antle and R. G. Compton, Reduction of Carbon Dioxide in 1-butyl-3-
methylimidazolium Acetate, Chemical Communications, 2009, (25), p. 3744-3746.
15. S. S. Moganty and R. E. Baltus, Diffusivity of Carbon Dioxide in Room-Temperature
Ionic Liquids, Industrial & Engineering Chemistry Research, 2010, 49 (19), p. 9370-
9376.
16. Vitaly V. Pavlishchuk and Anthony W. Addison, Conversion constants for redox
potentials measured versus different reference electrodes in acetonitrile solutions at
25°C, Inorganica Chimica Acta, 2000, 298 (1), p. 97-102.
17. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Microfluidic Reactor for the
Electrochemical Reduction of Carbon Dioxide: The Effect of pH, Electrochemical and
Solid State Letters, 2010, 13 (9), p. D109-D111.
18. H. De Jesus-Cardona, C. del Moral and C. R. Cabrera, Voltammetric Study of
CO2Reduction at Cu electrodes Under Different KHCO3 Concentrations, Temperatures
and CO2 Pressures, Journal of Electroanalytical Chemistry, 2001, 513 (1), p. 45-51.
19. D. T. Whipple and P. J. A. Kenis, Prospects of CO2 Utilization via Direct Heterogeneous
Electrochemical Reduction, Journal of Physical Chemistry Letters, 2010, 1 (24), p. 3451-
3458.
59
20. S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of
High Pressure CO2at a Cu Electrode in Cold Methanol, Electrochimica Acta, 2006, 51
(23), p. 4880-4885.
21. Y. Hori, A. Murata and R. Takahashi, Formation of Hydrocarbons in the
Electrochemical Reduction of Carbon-Dioxide at a Copper Electrode in Aqueous-
Solution, Journal of the Chemical Society, Faraday Transactions 1, 1989, 85 p. 2309-
2326.
22. A. S. Hollinger, R. J. Maloney, R. S. Jayashree, D. Natarajan, L. J. Markoski and P. J. A.
Kenis, Nanoporous separator and low fuel concentration to minimize crossover in direct
methanol laminar flow fuel cells, Journal of Power Sources, 2010, 195 (11), p. 3523-
3528.
23. C. Schmidt, T. Glück and G. Schmidt-Naake, Modification of Nafion Membranes by
Impregnation with Ionic Liquids, Chemical Engineering & Technology, 2008, 31 (1), p.
13-22.
60
Chapter 4
Increased Current Densities for Amine-Based EMIM+ Electrolytes
4.1 Introduction
As discussed in Chapter 3, one of the biggest challenges for the electrochemical reduction
of CO2 to CO lies in achieving high energetic efficiencies [1-4] at reasonable current densities
[5,6]. Towards achieving this goal, several electrolyte solutions have shown much promise [4,7-
13]. Specifically, pyridine has been shown to change product selectivity [7-9], other electrolyte
solutions have been shown to improve partial current densities for the desired products
[10,14,15] and EMIM BF4 has been shown to reduce the onset potential for CO production [4].
While EMIM BF4 drastically reduced the onset potential for CO production, the current
densities achieved with EMIM BF4 are quite low in comparison to more traditional electrolyte
solutions due to the large reactor resistance with EMIM BF4 in the dual electrolyte reactor [4].
As discussed in Chapter 3, the primary reason for the low current densities lies in the high
reactor resistances stemming from membrane poisoning.
This study explores substituting smaller anions (i.e., OH- and Cl
-) in place of BF4
- on an
EMIM+ cation. We expect this substitution to (a) address anode poisoning without the use of a
dual electrolyte reactor and (2) improve both reaction selectivity (for CO) and current density
based on correlations identified in Chapter 2 between decreased anion size and improved
performance. Furthermore, we study the effect of bolstering the electrolyte conductivity with
additional salts to improve reaction rates, as well as the effect of decreasing the pH via the
addition of hydroxide to improve anode kinetics [16].
61
4.2 Experimental
4.2.1 Reactor Testing
This study used the same equipment and procedures discussed in Chapter 2. For all of
the experiments carried out in this study, a 1 mg/cm2 Ag cathode loading and a 6.67 mg/cm
2 Pt
black anode loading was used. All gas flow rates were 5 SCCM and all liquid flow rates were
0.5 mL/min. All product analysis was carried out with the Thermo GC discussed in Chapter 2.
4.2.2 Ionic Liquid Solution Preparation
EMIM BF4 (Product Number:711721; Provider: Sigma Aldrich), EMIM Cl (30764;
Sigma Aldrich), KCl (P1597; Sigma Aldrich) and KOH (44016; Sigma Aldrich) were obtained
and used as provided unless otherwise specified. EMIM OH was produced via ion exchange.
Specifically, 26 g of EMIM BF4 was mixed with 8 g of KOH and 20 mL of H2O. KBF4 formed
a gel immediately upon mixing. The solution was mixed for 30 minutes. The mother liquor was
separated from the gel via vacuum filtration (2x) using 10 µm pore-size filter paper.
4.3 Results and Discussion
4.3.1 Propensity for Precipitation
Table 4.1 shows the results from the addition of 1 mL of various ionic liquid solutions
with various 1 mL salt solutions. Both the BMIM+ BF4
- and EMIM
+ BF4
- ionic liquid solutions
precipitated with most additive solutions. The precipitant solutions were analyzed with
elemental analysis for both the KCl / EMIM BF4 and NaCl / EMIM BF4 solutions and the
precipitant was found to contain only trace amounts of C, H or N, thereby confirming that the
precipitates were salts of the BF4 anion. While many BMIM+ PF4
- solutions did not precipitate
with many solutions, the majority of these solutions phase separated. However, the EMIM+ Cl
-
62
ionic liquid solution was stable with the addition of most salts. This suggests that the EMIM+ Cl
-
ionic liquid solution’s conductivity could be bolstered with the addition of other electrolytes.
Table 4.1: Solid form data from the addition of salt “additive”
solution to various ionic liquid solutions
Additives 18 mole% Ionic Liquid Solutions
EMIM Cl EMIM BF4 BMIM BF4 BMIM PF4
KOH -a Gel Crystal -
NaOH - Crystal Crystal -
Choline
Ace. - - - -
Choline Cl - - - -
Choline OH - - - Gel
NaCl - Crystal Crystal -
CsCl - Gel Gel -
CaCl2 Crystal Gel Gel -
MgSO4 - - - -
NaHCO3 - Crystal Crystal -
KHCO3 - Crystal Crystal -
K2CO3 - Crystal Crystal Gel
NH4HCO3 - - - -
CsOH - Gel Gel - a “-“ indicates that no precipitate was observed upon the
addition of an additive to an ionic liquid solution
4.3.2 Individual Electrode Characterization of EMIM+ Cl
- Solution
A dual electrolyte reactor developed
for the EMIM BF4 electrolyte solution
overcame anode poisoning from the addition
of the ionic liquid solution as discussed in
Chapter 3, however, at the expense of low
current densities due to high membrane
resistances. Because the source of the anode
poisoning (from the BF4- or the EMIM
+ ion)
was not previously confirmed, we analyzed
Figure 4.1: Individual electrode polarization plots for a
single channel reactor with either an EMIM Cl solution or
an aqueous KCl solution and for a dual electrolyte reactor
with an EMIM BF4 solution.
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
63
the individual electrode polarization of an EMIM Cl solution in a single electrolyte reactor.
Figure 4.1 shows the cathode and anode electrode polarization in the single channel flow reactor
with a 4.7 mole% EMIM Cl electrolyte solution compared to that of a 1 M aqueous KCl
electrolyte solution in the single channel flow reactor and an 18 mole% EMIM BF4 electrolyte
solution in the dual channel flow reactor. With the EMIM Cl solution, no anode poisoning was
observed, confirming that the BF4 anion is responsible for the anode poisoning. Consequently, a
second electrolyte stream, the anolyte, is not necessary when using an EMIM Cl electrolyte
solution. Furthermore, the current densities achieved with the EMIM Cl solution are comparable
to that of an aqueous solution.
4.3.3 EMIM+ Cl
- Performance in the Single Channel Electrolyte Reactor
Figure 4.2a shows the partial current densities for CO and H2 when using a 4.7 mole%
EMIM Cl electrolyte pumped through the reactor at 0.5 mL/min with a Ag cathode and a Pt
anode. High selectivity for CO is maintained with the substitution of the BF4 anion with the Cl
anion on the EMIM complex. Figure 4.2b shows a comparison of the current densities in a
single channel flow reactor with an aqueous 1 M KCl electrolyte, an 18 mole% EMIM BF4
Figure 4.2: (a) Partial current densities for CO and H2 for a 4.7 mole% EMIM Cl solution flowed at 0.5 mL/min with
a Ag cathode and a Pt anode. (b) comparison of comparison of the current densities in a single channel flow
reactor with an aqueous KCl electrolyte, an EMIM BF4 electrolyte and an EMIM Cl electrolyte.
0
20
40
60
80
-3.25-2.75-2.25-1.75-1.25
Cu
rre
nt D
en
sity
(mA
/cm
2 )
Cell Potential (V)
CO
H2
-120
-100
-80
-60
-40
-20
0
-3.3-2.8-2.3-1.8-1.3
Cu
rre
nt D
en
sity
(mA
/cm
2 )
Cell Potential
EMIM BF4
EMIM Cl
KCl
(a) (b)
0
20
40
60
80
-3.25-2.75-2.25-1.75-1.25
Cu
rre
nt D
en
sity
(mA
/cm
2)
Cell Potential (V)
CO
H2
-120
-100
-80
-60
-40
-20
0
-3.3-2.8-2.3-1.8-1.3
Cu
rre
nt D
en
sity
(mA
/cm
2)
Cell Potential
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
64
electrolyte and a 4.7 mole% EMIM Cl electrolyte. When using an EMIM Cl solution, the partial
current density for CO drastically increases such that its performance is comparable to that with
an aqueous electrolyte stream with the BF4 for Cl substitution.
Figure 4.3a shows the conductivity of the EMIM Cl solution with various concentrations
of EMIM Cl. The conductivity can only be improved roughly 7% by changing the concentration
of EMIM Cl because the conductivity of the earlier trial (at a concentration of 4.7 mole%) was
close to the maximum conductivity achievable with an EMIM Cl / aqueous mixture.
4.3.4 EMIM+ Cl
- Conductivity Studies
While EMIM Cl performs well compared to EMIM BF4, the performance was further
improved via bolstering the conductivity with a secondary salt, KOH, as was shown to be
possible in section 4.3.1. Figure 4.4 shows the performance of the reactor with (a) 4.7 mole%
EMIM OH and (b) 4.7 mole% EMIM Cl with a secondary ion, KOH at 1 M. When using the
EMIM OH electrolyte, the current density was expected to be greater than when using the
EMIM Cl electrolyte because of the effect of pH shown in Chapter 2. However, while the
current density with the EMIM OH electrolyte was better than with the EMIM BF4 electrolyte,
with the EMIM OH electrolyte, the current density was worse than with the EMIM Cl
Figure 4.3: (a) Conductivity for various concentration EMIM Cl solutions and (b) individual polarization plots for
EMIM Cl electrolyte solutions with concentrations between 4.7 and 10.9 mole%.
90.0
92.0
94.0
96.0
98.0
4 5 6 7 8 9 10 11 12
Co
nd
uct
ivit
y (m
S/cm
)
mole% EMIM Cl
-2.00
-1.50
-1.00
-0.50
0.00
0.50
1.00
1.50
-40-35-30-25-20-15-10-50
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
10.9% EMIM Cl
8.9% EMIM Cl
7.5% EMIM Cl
5.8% EMIM Cl
4.7% EMIM Cl
(a) (b)
90.0
92.0
94.0
96.0
98.0
4 5 6 7 8 9 10 11 12
Co
nd
uct
ivit
y (m
S/cm
)
mole% EMIM Cl
-2.00
-1.50
-1.00
-0.50
0.00
0.50
1.00
1.50
-40-35-30-25-20-15-10-50
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
10.9% EMIM Cl
8.9% EMIM Cl
7.5% EMIM Cl
5.8% EMIM Cl
4.7% EMIM Cl
90.0
92.0
94.0
96.0
98.0
4 5 6 7 8 9 10 11 12
Co
nd
uct
ivit
y (m
S/cm
)
mole% EMIM Cl
-2.00
-1.50
-1.00
-0.50
0.00
0.50
1.00
1.50
-40-35-30-25-20-15-10-50
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
10.9% EMIM Cl
8.9% EMIM Cl
7.5% EMIM Cl
5.8% EMIM Cl
4.7% EMIM Cl
65
electrolyte. This is believed to be due to the lower conductivity (70 mS/cm vs. 90 mS/cm) with
the hydroxide salt as opposed the chloride salt. The electrolyte consisting of a mixture of
EMIM Cl and KOH salts outperformed all other electrolyte solutions shown here due to the
higher conductivity (206 mS/cm) of the electrolyte and the improved anode performance at basic
conditions.
4.4 Conclusions
Drastic improvements have been achieved in the partial current density for CO with
EMIM+ based electrolytes. Specifically, substituting the BF4 anion for either a OH anion or a Cl
anion enabled partial current densities for CO greater than 60 mA/cm2 at a reactor potential of
3 V. Furthermore, the substitution of BF4 for OH or Cl eliminated anode poisoning from the BF4
anion, thus enabling the use of a single channel electrolyte stream. The elimination of the
membrane and the second electrolyte stream reduced the reactor resistance (from EMIM+
poisoning the Nafion as discussed in Chapter 3), which led to the higher current densities.
The conductivity of an EMIM Cl / aqueous mixture is limited to <100 mS/cm, however,
the conductivity of the electrolyte steam can be improved with the addition of a secondary salt
(i.e., KOH). Specifically, the addition of KOH to an EMIM Cl solution, drastically increased the
Figure 4.4: Partial current density for (a) CO and (b) H2 with various ionic liquid electrolyte streams.
0
20
40
60
80
100
120
-3.25-2.75-2.25-1.75-1.25
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2 )
Cell Potential (V)
EMIM BF4EMIM ClEMIM Cl & KOHEMIM OH
0
5
10
15
20
-3.25-2.75-2.25-1.75-1.25
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cell Potential (V)
EMIM BF4EMIM ClEMIM Cl & KOHEMIM OH
(a) (b)
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
-150-100-500
Ele
ctro
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
EMIM BF4
EMIM Cl
KCl
66
electrolyte conductivity (206 mS/cm), and consequently, enabled partial current densities for CO
greater than 100 mA/cm2 at a reactor potential of 3V.
4.5 References
1. Y. Hori, Electrochemical CO2 Reduction on Metal Electrodes, Modern Aspects of
Electrochemistry, 2008, 42 p. 89-189.
2. A. T. Bell, Basic Energy Needs: Catalysis for Energy. Department of Energy, 2008.
PNNL17214
3. M. R. Dubois and D. L. Dubois, Development of Molecular Electrocatalysts for CO2
Reduction and H(2) Production/Oxidation, Accounts of Chemical Research, 2009, 42
(12), p. 1974-1982.
4. Brian A. Rosen, Amin Salehi-Khojin, Michael R. Thorson, Wei Zhu, Devin T. Whipple,
Paul J. A. Kenis and Richard I. Masel, Ionic Liquid–Mediated Selective Conversion of
CO2 to CO at Low Overpotentials, Science, 2011, 334 (6056), p. 643-644.
5. D. T. Whipple and P. J. A. Kenis, Prospects of CO2 Utilization via Direct Heterogeneous
Electrochemical Reduction, Journal of Physical Chemistry Letters, 2010, 1 (24), p. 3451-
3458.
6. Narasi Sridhar, Carbon Dioxide Utilization: Electrochemical Conversion of CO2 -
Opportunities and Challenges. Det Norske Veritas, 2011. 07 1-20.
7. E. B. Cole, P. S. Lakkaraju, D. M. Rampulla, A. J. Morris, E. Abelev and A. B. Bocarsly,
Using a One-Electron Shuttle for the Multielectron Reduction of CO2 to Methanol:
Kinetic, Mechanistic, and Structural Insights, Journal of the American Chemical Society,
2010, 132 (33), p. 11539-11551.
8. A. J. Morris, R. T. McGibbon and A. B. Bocarsly, Electrocatalytic Carbon Dioxide
Activation: The Rate-Determining Step of Pyridinium-Catalyzed CO2 Reduction,
Chemsuschem, 2011, 4 (2), p. 191-196.
9. G. Seshadri, C. Lin and A. B. Bocarsly, A New Homogeneous Electrocatalyst for the
Reduction of Carbon-Dioxide to Methanol at Low Overpotential, Journal of
Electroanalytical Chemistry, 1994, 372 (1-2), p. 145-150.
10. S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of
High Pressure CO2 at a Cu Electrode in Cold Methanol, Electrochimica Acta, 2006, 51
(23), p. 4880-4885.
11. S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of
High Pressure Carbon Dioxide at a Cu Electrode in Cold Methanol with CsOH
Supporting Salt, Chemical Engineering Journal, 2007, 128 (1), p. 47-50.
12. S. Kaneco, R. Iwao, K. Iiba, K. Ohta and T. Mizuno, Electrochemical Conversion of
Carbon Dioxide to Formic Acid on Pb in KOH/Methanol Electrolyte at Ambient
Temperature and Pressure, Energy, 1998, 23 (12), p. 1107-1112.
13. S. Kaneco, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of CO2 to
Methane at the Cu Electrode in Methanol with Sodium Supporting Salts and its
Comparison with Other Alkaline Salts, Energy Fuels, 2006, 20 (1), p. 409-414.
67
14. A. Murata and Y. Hori, Product Selectivity Affected by Cationic Species in
Electrochemical Reduction of CO2 and CO at a Cu Electrode, Bulletin of the Chemical
Society of Japan, 1991, 64 (1), p. 123-127.
15. Y. Hori and S. Suzuki, Electrolytic Reduction of Carbon-Dioxide at Mercury-Electrode
in Aqueous-Solution, Bulletin of the Chemical Society of Japan, 1982, 55 (3), p. 660-665.
16. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Microfluidic Reactor for the
Electrochemical Reduction of Carbon Dioxide: The Effect of pH, Electrochemical and
Solid State Letters, 2010, 13 (9), p. D109-D111.
68
Chapter 5
Organometallic Catalysts for the Electrochemical Reduction of CO2
5.1 Introduction
Storage of electrical energy in chemical bonds (i.e., chemical fuels) provides an excellent
storage medium for intermittent renewable energy sources, thereby enabling “artificial
photosynthesis”. However, artificial photosynthesis, using the electrochemical reduction of CO2,
is obstructed by the lack of catalysts for CO2 reduction with (a) sufficiently low overpotentials to
be energetically efficient [1-4] and (b) high enough current densities to be profitable [2,5].
According to a recent DOE report, "electron conversion efficiencies of greater than 50 percent
can be obtained, but at the expense of very high overpotentials (~1.5 V)", consequently more
efficient and higher throughput catalysts need development [1,6-8].
The catalytic activity of all commonly used catalysts’ comprised of a single metal have
previously been studied [3,9-11]. From this, four distinct classes of catalysts have been
identified: catalysts selective for carbon monoxide (e.g., Ag, Au, Zn), catalysts selective for
formic acid (e.g., Sn, Cd, Ti), catalysts selective for hydrogen (e.g., Pt, Ni, Fe) and copper, which
produces a wide range of hydrocarbons [3,9].
A primary challenge lies in the formation of the “CO2- intermediate” as discussed in
Chapter 1 [14-16]. Using one of the catalysts selective for CO, Ag, Rosen et al. (see Chapter 2)
demonstrated the use of an amine in the electrolyte solution, which can serve the role of co-
catalyst and reduce the onset potential for CO2 production [1]. Specifically, EMIM BF4 was
shown to reduce the onset potential for CO2 reduction such that the overpotential is < 0.17 V;
however, at the expense of low current densities [1]. Similarly, another amine, Pyridine, has
69
been shown to make the electrochemical reduction of
CO2 to CO on a copper electrode more selective for
methanol at a low overpotential [17-19]. Clearly,
amines can play a significant role in assisting the
electrochemical reduction of CO2 in terms of reducing
the overpotentials and increasing the product
selectivity.
Here, we are interested in developing an
organometallic complex to incorporate amines on the
electrode surface and thereby, develop a catalyst with low overpotentials, high selectivity, and
high current densities. Thorum et al. previously developed a copper based amine, copper
complex with 3,5-diamino-1,2,4-triazole
(Figure 5.1), for oxygen reduction
[12,13]. The copper complex was found
to have similar performance to that of
Pt/C (Figure 5.2), which constitutes a
huge breakthrough in the development
of organometallic complexes.
Here, we explored the use of this
copper complex, CuDAT, and
modifications of this complex for CO2
reduction. Specifically, we explored
several Ag and Cu based amine
Figure 5.1: A novel carbon-supported copper
complex of 3,5-diamino-1,2,4-triazole
(CuDAT) developed by the Gewirth lab.
Reprinted with permission from: Matthew S
Thorum, Jessica Yadav and Andrew A Gewirth,
“Oxygen Reduction Activity of a Copper
Complex of 3,5-Diamino-1,2,4-triazole
Supported on Carbon Black.” Angewandte
Chemie International Edition, 2009. 48 (1): p.
165-167. Copyright (2008) John Wiley and
Sons. [12]
Figure 5.2: Individual electrode polarization curves for a
microfluidic H2/O2 fuel cell with a CuDAT catalyst compared t
omore traditional catalysts. Reprinted with permission from:
Fikile R. Brushett, Matthew S. Thorum, Nicholas S. Lioutas,
Matthew S. Naughton, Claire Tornow, Huei-Ru “Molly” Jhong,
Andrew A. Gewirth and Paul J. A. Kenis, “A Carbon-Supported
Copper Complex of 3,5-Diamino-1,2,4-triazole as a Cathode
Catalyst for Alkaline Fuel Cell Applications.” Journal of the
American Chemical Society, 2010. 132 (35): p. 12185-12187.
Copyright (2010) American Chemical Society.[13]
0
0.2
0.4
0.6
0.8
1
0 100 200 300 400
Current Density (mA/cm2)
Po
ten
tial
(V)
vs R
HE
Pt/C cathode
Cu-tri/C cathode
Ag/C cathode
Pt/C anode
1 M KOH
70
complexes synthesized in the Gewirth lab. Ag was chosen because, as a bare metal, Ag has
excellent selectivity for CO at high current densities and because of the Kenis lab’s extensive
experience with Ag [20-24]. Cu was chosen based on the experience Dr. Gewirth’s lab has with
Cu-based organometallic catalyst development. Furthermore, Cu is the only bare-metal which
has been demonstrated to be capable of producing large hydrocarbons from CO2; however, the
potentials necessary for direct conversion of CO2 to hydrocarbons are energetically inefficient
due to their high overpotentials [21,25-28].
At present, one challenge with using copper catalysts is improving the reaction
selectivity. Copper tends to produces a wide range of products and is unselective for any one
product due to its complex reaction pathway [29-32]. Specifically, methane requires an 8-
electron pathway with many potential alternative reaction pathways, which result in the
formation of different chemicals. For example, ethylene, methanol, carbon monoxide, and
formic acid deviate from the methane reaction pathway after the 4th
, 4th
or 5th
, 2nd
, and 1st
electron transfer, respectively, en route to their own 12, 6, 2 and 2 electron-transfer reaction
pathways. Here we explore the use of Cu-based organometallic complexes to influence the
reaction and change the product selectivity.
5.2 Experimental
5.2.1 Catalyst Synthesis
All catalysts were synthesized by Claire Tornow in the Gewirth Lab. Below are the
synthesis procedures for the catalysts used in this study.
Silver phthalocyanine (75 % dye content, Sigma-Aldrich) (AgPc) was used as received.
Silver phthalocyanine (61.7 mg) was readily dissolved in H2O. Vulcan carbon (162.7 mg) was
71
added to the solution and the mixture was sonicated for 30 minutes and allowed to sit overnight.
The sample was centrifuged once and placed in oven at 120ºC for 24 hours.
Silver pyrazole (AgPz) was synthesized according to Masciocchi, et al. in the presence of
Vulcan [33]. AgNO3 (35.5 mg) was dissolved in H2O. To this, ~0.5 mL NH3 was added. Vulcan
(160.8 mg) was added and the mixture was sonicated for 10 minutes. H20 solvated pyrazole was
added dropwise and the resulting mixture was sonicated for 30 minutes. This was allowed to sit
overnight. The mixture was centrifuged and placed in a vacuum oven at 90ºC for 4 hours.
Silver complex of tris(2-pyridylmethyl)amine (AgTPA) was synthesized using a 1:1
molar ratio of TPA to Ag. The TPA ligand (44.4 mg) was dissolved in EtOH. Ag2SO4 (30.4 mg)
was dissolved in H2O. Vulcan (156.3 mg) was added to the solution of Ag2SO4 and the solution
was sonicated for 10 minutes. TPA solution was added dropwise to the carbon mixture. The
resulting mixture was sonicated for an additional 30 minutes and allowed to sit overnight. The
mixture was centrifuged once and placed in a vacuum oven at 90ºC for 4 hours.
The silver complex of 3,5-diamino-1,2,4-triazole (AgDAT) was synthesized using a 2:1
molar ratio of DAT to Ag. The DAT ligand (502.4 mg) was dissolved in H2O and Ag2SO4
(393.3 mg) was also dissolved in H2O. Vulcan (2.0325 mg) was added to the solution of Ag2SO4
and sonicated for 10 minutes. Solution of DAT was added dropwise to the carbon suspension
and the resulting mixture was sonicated for 30 minutes and allowed to sit overnight. The mixture
was centrifuged once and placed in a vacuum oven at 90ºC for 4 hours.
The following procedures are for the copper organometallic catalysts synthesized for this
study: The ligands 3-amino-4-methyl-1,2,4-triazole (Sigma-Aldrich), 4-amino-3,5-dipyridyl-
1,2,4-triazole (Sigma-Aldrich), 3,4,5-triamino-1,2,4-triazole (BIONET Research Intermediates,
Key Organics), and 4-amino-3,5-dimethyl-1,2,4-triazole (Alfa Aesar) were used as received. Cu
72
complexes were synthesized by addition of the respective triazole ligand to an aqueous solution
containing Cu(SO4) · 5 H2O (Sigma-Aldrich, 99.999 %) in a 2:1 molar ratio. The resulting
precipitates were collected via vacuum filtration. The appropriate amounts of the Cu complexes
(to generate loadings of a 30 % complex on carbon) were placed in an aqueous solution
containing Vulcan XC-72 (Cabot Corporation) to create a suspension which was further
sonicated for 1 hour. The material was collected by first centrifuging, then drying the sample
under vacuum at 90ºC to produce the carbon-supported Cu catalysts.
5.2.2 Composition
The organometallic silver catalysts used in this study were analyzed via elemental
analysis by Claire Tornow for their silver content and the AgDAT, AgPc, AgTPA, and AgPz
catalysts were found to have silver mass% loadings of 8.62, 6.25, 4.32, and 2.51 %, respectively.
Furthermore, the copper catalysts, CT01-301, and CT01-289 have metal mass% loadings of 1.21,
and 1.26 %.
5.2.3 UV-Vis Analysis
The analysis of the rate of formic acid production was carried out based on the
concentrations of formic acid in the effluent electrolyte stream. The formic acid concentration
was measured using a modified version of a procedure developed by Sleat and Mah [34].
0.4 mL of a fresh reactant solution (0.5 g citric acid, 10 g acetamide, and 100 mL isopropyl
alcohol) was mixed with 0.2 mL of the effluent sample containing formic acid, 0.02 mL of 30 %
sodium acetate, and 1.4 mL acetic anhydride. Following mixing, these solutions were allowed to
react for 24 hrs before being placed in cuvettes (BrandTech Scientific, Brand-UV Cuvettes 7591)
for UV-Vis (Perkin Elmer, Lambda 650) analysis at a wavelength of 515 nm. For each
experiment, a new calibration curve was prepared using known standard concentrations (1, 5, 10,
73
20, 40, and 60 mM formic acid in 1 M KOH). A line was fit to the intensity of the UV peak at
515 nm relative to the formic acid concentration using a method of least squares of the error to
make a linear calibration curve. The UV peak intensity of the samples prepared from the
effluent stream was compared to the linear calibration to determine the formic acid concentration
in the effluence electrolyte sample. Figure 5.3 shows vials filled with a mixture of the formic
acid stream and the reactant stream for the analysis of the formic acid concentration.
5.2.4 Three Electrode Cell Operation
The three electrode cell experiments were carried out using a CH Instruments
bipotentiostat. The electrochemical cell consisted of a Pt gauze counter electrode and a “no-
leak” Ag/AgCl reference electrode (Cypress), separated from the working electrode by means of
a Luggin capillary. The catalysts for the three electrode cell experiments were prepared as
follows: catalyst inks containing the carbon supported Ag complex (1.0 mg mL-1) and Nafion
Figure 5.3: (top) Formic acid control solutions mixed with reactant solutions for UV-Vis calibration and (bottom)
effluent electrolyte mixed with reactant solution to analyze in UV-Vis for formic acid concentrations.
1 mM 5 mM 10 mM 20 mM 40 mM 60 mM
2.0 V 2.25 V 2.5 V 2.75 V 3.0 V
74
solution (4 μL mL-1, 5 wt%, Aldrich) were prepared in a 5% butanol-water mixture and
sonicated prior to electrode preparation. A 20 μL drop of the catalyst ink was deposited on a
rotating ring-disk electrode (Pine Instruments) comprised of a polished (0.05 micron alumina)
glassy carbon disk electrode (0.196 cm2) with a Pt ring.
5.2.5 Reactor Operation
The reactor operational procedure was identical to that used in Chapter 2. Experimental
data, in which current densities were above 200 mA/cm2, was not reported here because, at these
high current densities, CO, and H2 are produced at sufficiently high rates to quickly form bubbles
in the electrolyte solution, partially blocking the electrode and resulting in inconsistent reactor
performance. In all of the studies outlined here, unless otherwise stated, a 1 M KOH electrolyte
solution was flowed through the electrolyte chamber at 0.5 mL/min. All of the experiments in
this study used a 1 mg/cm2 cathode loading on Sigracet 35BC. All of the anodes used in this
study had a 2 mg/cm2 Pt loading on Sigracet 35BC. For the analysis of the copper and
organometallic copper-complexes, a single GC sample (rather than the normal three) was used
for the analysis at each potential applied because the elution times for ethylene requires the GC
to run for 12 minutes for each sample.
5.3 Results and Discussion
5.3.1 Cyclical Voltammetry in a flow reactor
One of the primary advantages of the electrochemical reactor used in this study is the
ability to carry out cyclical voltammetry in-situ for the investigation of catalyst performance.
Figure 5.4 shows cyclical voltammetry scans taken within the flow reactor of the cathode relative
to a Ag/AgCl reference electrode between -0.4 and -1.4 V vs. Ag/AgCl for a silver electrode
(Ag), a carbon supported silver triazole electrode (AgDAT), a carbon supported silver pyrazole
75
electrode (AgPz), a carbon supported silver tripyridylamine electrode (AgTPA), and a carbon
supported silver phthalocyanine electrode (AgPc). A 1 M KOH electrolyte was flowed through
the electrolyte chamber at 0.5 mL/min and either CO2 or N2 was flowed behind the cathode in
Figure 5.4: Cyclic voltammetry reduction activity measurements for (a) AgPc, (b) AgDAT, (c) AgPz, (d) Ag, and
(e) AgTPA with a CO2 feed compared to their respective activity under Ar. Data was recorded in a flow reactor with
a 1 M KOH electrolyte solution flowed at 0.5 mL/min scanned at 25 mV/s
0
5
10
15
20
25
30
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt D
en
sity
(mA
/cm
2)
Potential (V vs Ag/AgCl)
N2
CO2
0
10
20
30
40
50
60
70
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt D
en
sity
(mA
/cm
2 )
Potential (V vs Ag/AgCl)
N2
CO2
0
5
10
15
20
25
30
35
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt D
en
sity
(mA
/cm
2 )
Potential (V vs Ag/AgCl)
CO2
N2
0
5
10
15
20
25
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt D
en
sity
(mA
/cm
2)
Potential (V vs Ag/AgCl)
CO2
N2
-4
-2
0
2
4
6
8
10
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt D
en
sity
(mA
/cm
2)
Potential (V vs Ag/AgCl)
CO2
N2
(a) (b)
(c) (d)
(e)
0
5
10
15
20
25
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt D
en
sity
(mA
/cm
2)
Potential (V vs Ag/AgCl)
CO2
N2
0
5
10
15
20
25
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt D
en
sity
(mA
/cm
2)
Potential (V vs Ag/AgCl)
CO2
N2
76
the gas chamber at a flow rate of 5
SCCM. These graphs provide
information regarding (a) the onset
potentials for various products and (b) the
product selectivity. An increase in current
density resulting from a change in
reactant gas (N2 to CO2) suggests that the
catalyst is active for CO2 reduction and
will produce a carbon product (e.g., CO,
CH4, C2H4, CHOH, or HCO2H). The
more drastic the change in current density, the more selective the catalyst is.
As seen in Figure 5.4, when CO2 is substituted for Ar as the reactant gas, with Ag,
AgDAT, AgPz, or AgPc as the cathode catalysts (a) an earlier onset potential and (b) an increase
in the current density are observed. The improved performance in the presence of CO2 suggests
that Ag, AgDAT, AgPz, and AgPc are active for CO2 reduction and aid the production of a
carbon product (e.g., CO, CH4, C2H4, CHOH, or HCO2H). However, with AgTPA as the
cathode catalysts, when CO2 is substituted for Ar as the reactant gas, a small decrease change in
the current density is observed. This change in catalyst activity suggests CO2 reduction may be
active on this catalyst. However, the decrease in current density also suggests that the kinetics
for CO2 reduction for the AgTPA catalyst may be quite low. Compared to the other catalysts,
AgTPA appears to suppress CO2 reduction kinetics.
The current densities achieved with three of the electrodes with the Ag-amine complex
catalyst (DAT, Pz, Pc) were comparable to that of the Ag catalyst when compared relative to the
Figure 5.5: Cyclic voltammetry measurements comparing the
onset potentials of CO production for Ag particles, AgDAT/C ,
AgPz/C, and AgPc/C. Data was recorded in 1 M KOH, under
CO2, at 25 mV/s. Figure courtesy of Claire Tornow.
0.00
10.00
20.00
30.00
40.00
50.00
60.00
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt d
en
sity
(mA
/cm
2 )
Potential (V vs Ag/AgCl)
TPA
Ag
Pc
DAT
Pz
0.00
10.00
20.00
30.00
40.00
50.00
60.00
-1.5 -1.3 -1.1 -0.9 -0.7 -0.5
Cu
rre
nt d
en
sity
(mA
/cm
2)
Potential (V vs Ag/AgCl)
TPA
Ag
Pc
DAT
Pz
77
total mass loading of catalyst. The high performance with these organometallic catalysts is very
promising as Ag is the best catalyst in the literature for CO production.[3]
Figure 5.5 compares the cyclical voltammetry scans for the catalysts determined from the
initial CV scans to be selective for CO (AgTPA, Ag, AgPC, AgDAT, and AgPz). Earlier onsets
are seen with AgPz, AgPc, and AgDAT in the presence of CO as compared to Ag
(Pz<DAT<Pc<Ag). The decrease in current density with AgTPA further suggests that while the
complex may be active for CO2 reduction, the kinetics for CO2 reduction are reduced.
5.3.2 Cyclical Voltammetry in a 3-electrode cell
Figure 5.6 shows cyclical voltammetry taken by Tornow from Dr. Gewirth’s lab in a 3-
electrode cell with a 1 M KOH electrolyte and a scan rate of 25 mV/s. Six catalysts were tested
including: silver (Ag), carbon supported silver (Ag/C), carbon supported silver triazole
(AgDAT), carbon supported silver pyrazile (AgPz), carbon supported silver tripiradine
(AgTPA), and carbon supported silver pthyocyline (AgPc) with either CO2 or Ar bubbled
through the electrolyte prior to the cyclical voltammetry. Similar to the CVs in-situ for the
reactor, these results provide information regarding the onset potentials and the relative current
densities for various catalysts. Furthermore, information can be obtained regarding the activity
for CO2 reduction based any changes in performance from when CO2 (rather than Ar) is bubbled
through the electrolyte.
78
Figure 5.6: Cyclic voltammetry measurements of reduction activity of Ag catalysts with a CO2 feed compared to
their respective activity under Ar. Data was recorded in 1 M KOH at 25 mV/s. Figure courtesy of Claire Tornow.
(a) (b)
(c) (d)
(e) (f)
79
The Ag, AgDAT, AgPz, and AgPc
catalysts show a decrease in the onset
potential and an increase in the current
density in presence of CO2 as opposed to
Ar, confirming what was observed earlier in
the flow reactor that these catalysts are
active for CO2 reduction and will produce a
carbon product (e.g., CO, CH4, C2H4,
CHOH, or HCO2H). However, the Ag/C
and AgTPA catalysts show no change in
both the onset potential and the current density in presence of CO2 as opposed to Ar, suggesting
that these catalysts may not have activity for CO2 reduction. Of note, in the flow reactor (section
5.3.1), the addition of CO2 decreased the observed current density for the AgTPA catalyst,
thereby suggesting CO2 reduction activity, but also poor kinetics. However, here, with the three
electrode cell, we do not observe any change in activity from bubbling CO2 into the liquid
electrolyte of the three electrode cell.
The current densities achieved with the Ag-amine complex (DAT, Pz, Pc) catalysts are
comparable (at the same total mass loading) to that observed with the Ag catalyst, which is the
best catalyst in the literature for CO production. While higher current densities are achieved
with Ag/C as opposed to Ag, they have poor selectivity for CO2 reduction.
Similarly to Figure 5.5, we compare the cyclical voltammetry scans for the catalysts
determined from the initial CV scans to be selective for CO (Ag, AgDAT, AgPz, and AgPC) in
Figure 5.7. Table 5.1 compares the onset potentials as determined by the potential at which a
Figure 5.7: Cyclic voltammetry measurements comparing the
onset potentials of CO production for Ag particles, AgDAT/C ,
AgPz/C, and AgPc/C. Data was recorded in 1 M KOH, under
CO2, at 25 mV/s. Figure courtesy of Claire Tornow.
80
specific current density is achieved. Earlier onsets are seen with the Pz, Pc, and DAT in the
presence of CO as compared to Ag (Pz<DAT<Pc<Ag).
Table 5.1. Comparison of onset potentials for catalysts under CO2 and Ar, reported as the potentials at
which the current density reaches -0.05, -0.1, and -0.2 mA/cm2.
Catalyst
Onset Potentials under CO2 (V) Onset Potentials under Ar (V)
-0.05 mA/cm2 -0.1 mA/cm
2 -0.2 mA/cm
2 -0.05 mA/cm
2 -0.1 mA/cm
2 -0.2 mA/cm
2
Ag -1.20 -1.22 -1.25 -1.38 NA NA
Ag/C -0.86 -0.92 -1.00 -0.93 -1.00 -1.09
AgDAT/C -1.03 -1.14 -1.21 -1.08 -1.20 -1.29
AgTPA/C -1.09 -1.16 -1.24 -1.14 -1.20 -1.26
AgPz/C -1.01 -1.13 -1.21 -1.06 -1.19 -1.30
AgPc/C -1.05 -1.15 -1.24 -1.18 -1.26 -1.34
The onset for H2 is as follows: Ag/C<Pz<DAT<TPA<Pc<Ag. Ag/C has a very early
onset potential in the presence of Ar, thereby suggesting a reduction in the onset potential for H2
production. Because no improvement is observed for the Ag/C catalyst from the bubbling of
CO2 in the electrolyte solution carbon products (e.g., CO, CH4, C2H4, CHOH, or HCO2H) are not
expected in the flow reactor tests.
5.3.3 Operation in a Flow Reactor
Figure 5.8 shows the partial current densities for CO and H2, the cathode overpotentials
and the partial current normalized to the silver loading for the AgDAT, AgPz, AgPc, and AgTPA
organometallic catalysts compared to both Ag and Ag/C in the flow reactor with a 1 M KOH
electrolyte solution (0.5 mL/min) and a CO2 gas feed (5 SCCM). Regarding the partial current
density for CO relative to the absolute loading of catalyst, we observe, a higher current density
for AgDAT and AgPz as compared to Ag. Additionally, we observe a slightly lower current
density for AgPc as compared to Ag and we observe CO suppression with AgTPA. In terms of
81
the partial current density for CO normalized to the mass of the Ag loading, we observe that the
current density per Ag loading is improved by more than a factor of 10 with the organometallic
catalysts as compared to Ag. This is very promising as the silver content of these organometallic
complexes are quite low in comparison to the pure silver catalyst.
Figure 5.8b shows that the Ag/C catalyst has a large partial current density for H2, which
is the undesired reaction because other means are more efficient means of making H2,. However,
with the amine-based catalysts, low H2 partial current densities are observed. The partial current
densities for H2 increase in the following order: Ag<DAT<Pc<TPA<Pz<Ag/C.
In the cathode electrode polarization plot, Figure 5.8d, the overpotentials for the catalysts
Figure 5.8: Partial current densities for (a) CO and (b) H2 relative to electrode surface area as well as (c) partial
current density for CO relative to the Ag loading and (d) cathode polarization curves for electrodes with 1 mg.cm2
Ag and Ag-ligand catalyst loadings in a flow reactor with a 1 M KOH electrolyte flowed at 0.5 mL/min and CO2
flowed over the cathode at 5 SCCM..
0
20
40
60
80
100
120
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2)
Cathode vs Ag/AgCl (V)
AgDAT/CAgPz/CAgPc/CAgTPA/CAgAg/C
0
500
1,000
1,500
2,000
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/mg
Ag)
Cathode vs Ag/AgCl (V)
AgDAT/CAgPz/CAgPc/CAgTPA/CAgAg/C
0
5
10
15
20
25
30
35
40
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2)
Cathode vs Ag/AgCl (V)
-2.5
-2.3
-2.1
-1.9
-1.7
-1.5
-1.3
-1.1
0 25 50 75 100 125
Cat
ho
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
0
20
40
60
80
100
120
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2)
Cathode vs Ag/AgCl (V)
AgDAT/CAgPz/CAgPc/CAgTPA/CAgAg/C
0
500
1,000
1,500
2,000
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/mg
Ag)
Cathode vs Ag/AgCl (V)
AgDAT/CAgPz/CAgPc/CAgTPA/CAgAg/C
0
5
10
15
20
25
30
35
40
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2)
Cathode vs Ag/AgCl (V)
-2.5
-2.3
-2.1
-1.9
-1.7
-1.5
-1.3
-1.1
0 25 50 75 100 125
Cat
ho
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
(a)
(d)(c)
(b)
82
increase in the following order: DAT<Pz<Ag<Pc<Ag/C<TPA. The large overpotential for
Figure 5.9: (a), (c), and (e) SEM images of the electrode surface before application in the flow reactor. (b) and (d)
SEM images of the electrode surface after application in the flow reactor with a 1 M KOH electrolyte flowed over
the electrode at 0.5 ml/min and a potential of 3 V applied across the reactor. The following magnifications were
used (a) 50,000×, (b) 100,000×, (c) 10,000×, (e) 10,000×, and (e) 10,000×.
(a) (b)
(c) (d)
(e)
83
AgTPA explains why both the hydrogen evolution and the CO2 reduction reactions are slow: the
overall kinetics for AgTPA are quite slow relative to the other catalysts. The overall kinetics are
best for AgDAT and AgPz as is evidenced by their low cathode overpotentials.
5.3.4 Imaging of the Catalyst Surface
XPS and SEM analysis enables us to better understand the charge and surface structure of
silver-based organometallic catalysts. Figure 5.9 shows SEM images of a Sigracet 35BC
electrode with AgDAT/C painted on the electrode at a loading of 1 mg/cm2 both before operation
in a flow cell and after operation in a flow cell. The SEM images shown in Figure 5.9 provide
several very interesting findings. First, as seen in Figure 5.9a and b, the particle size of the
catalyst on the surface is roughly 50 microns, which is quite similar to what is observed for the
Ag nanoparticles used earlier in this study. The similar size of the nanoparticles suggests that the
difference between the Ag and AgDAT samples is not a result of a change in particle size.
Second, at a magnification of 10,000×, 50,000× or 100,000×, the catalyst particles on the surface
do not exhibit significant agglomeration after a potential is applied (comparing Figure 5.9 insets
(a) and (c) to (b) and (d)), which suggests the silver particles are not free to agglomerate. Third,
before using the AgDAT electrode in the reactor, crystals can be observed on the electrode
surface. This can be seen in Figure 5.9c and e. However, after flowing a 1 M KOH electrolyte
over the cathode and applying 3 V across the reactor, the crystals go away. SEM elemental
analysis showed that the composition of the crystals was high in Ag, C, H and N, thereby
suggesting that some of the DAT crystallized on the surface and was washed away as DAT has
high solubility by itself in water. However, from Figure 5.9c, it can be seen that the majority of
the surface is covered by nanoparticles, rather than crystals. The crystals are believed to be rich
in excess triazole components which wash away when the reactor is used.
84
XPS provides insight regarding the charge of the silver on the surface. By looking at the
electrode surface we can see that the silver goes from a Ag+ to Ag after being used in the flow
cell. This suggests that silver on the surface is being reduced, presumably after a potential of 3 V
is applied across the reactor.
5.3.5 Confirmation of Amine Activity
The XPS data showing the transition of the charge on the silver from Ag+ to Ag suggests
that the silver-amine complex for AgDAAT does not stay intact. To confirm the activity of the
Amine, two experiments were performed.
First, 3,5-diamino-1,2,4-triazole (DAT) was added to a Ag/C catalyst solution, which was
shown to not be active for CO2 reduction in Sections 5.3.1, 5.3.2, and 5.3.3, and sonicated before
being hand-painted on a Sigracet 35BC electrode, and tested for CO2 reduction activity in a flow
cell. Specifically, DAT was added to the Ag/C catalysts solution such that the DAT to Ag/C
ratio was 30 to 100 on a per mass basis. In Figure 5.10, the performance of an electrode with
Ag/C catalyst was compared to the performance of an electrode with Ag/C catalyst combined
Figure 5.10: (a) Faradaic efficiency, and (b) partial current density for CO production with a Ag/C electrode as
compared to a Ag/C electrode with the addition of DAT to the catalyst mixture before hand-painting the catalyst on
the Sigercet 35BC electrode.
0
5
10
15
20
25
30
-3.25-3.00-2.75-2.50-2.25-2.00-1.75
Par
tial
Cu
rre
nt
De
nsi
ty fo
r C
O (
mA
)
Cathode (vs Ag/AgCl)
Ag/C+DAT
Ag/C
0
10
20
30
40
50
60
70
80
90
100
-3.25-3.00-2.75-2.50-2.25-2.00-1.75
Fara
dai
c Ef
fici
en
cy (%
)
Cell Potential (V)
Ag/C + DAT
Ag/C
(a) (b)
85
with DAT. The addition of DAT to the catalyst mixture prior to painting drastically improved
the selectivity of the catalyst for CO as opposed to H2; specifically, the partial current density for
CO increased by more than 20 fold. This suggests that the amine may act as a co-catalyst on the
electrode surface, thereby driving CO production.
Secondly, the activity of the Amine was explored without any silver. Five milligrams of
3,5-diamino-1,2,4-triazole was mixed with 6.5 µl of Nafion in a 1:1 H2O to 2-propanol solution,
sonicated and hand-painted on a Sigracet 35BC electrode. When the electrolyte flowed over the
reactor, some of the DAT was presumed to have come off into solution because the effluent
electrolyte had a yellow tint. After flowing the electrolyte through the reactor for several
minutes, the yellow tint went away and the reactor was tested for CO2 reduction activity.
Figure 5.11 compares the activity of a bare Sigracet 35BC electrode in the flow reactor as
compared to the activity of a Sigracet 35BC electrode with DAT in the flow reactor. The
addition of DAT enabled a small, yet significant, amount of CO production, thus showing that
the DAT is active for CO2 reduction, however, with very poor overall current densities. The
observation that DAT is active for CO2 reduction confirms the theory that the amine is involved
Figure 5.11: The partial current densities for CO and H2 production in the flow cell with (a) DAT painted on a
Sigercet 35BC electrode, and (b) a bare Sigercet 35BC electrode.
0
10
20
30
40
50
60
-2.0-1.8-1.6-1.4-1.2
Part
ial C
urr
en
t D
en
sit
y (
mA
/cm
2)
Cathode Potential (vs Ag/AgCl)
CO
H2
0
10
20
30
40
50
60
-2.0-1.8-1.6-1.4-1.2
Pari
tal C
urr
en
t D
en
sit
y (
mA
/cm
2)
Cathode Potential (vs AgAgCl)
CO
H2
(a) (b)
86
the CO2 reduction reaction. The low
current densities with the Amine suggest
that it may be acting as a co-catalyst
because the amine alone is unlikely to
achieve high current densities.
5.3.6 pH Perturbation
In Figure 5.12, the influence of
pH is investigated on individual
electrode polarization and partial current
densities for the electrochemical
reduction of CO2 to CO on a AgDAT
cathode catalyst and a Pt anode catalyst
with a 0.5 mL/min 1 M KOH electrolyte
flow rate. Similar to what was seen in
Chapter 2, basic conditions are the most
favorable for CO production.
As shown in Figure 5.12a, the
CO production is unaffected by basic
conditions. However, with acidic
conditions, the partial current density for
CO decreases. Specifically, starting at
pH 4, the partial current density for CO
begins to decrease, and the trend Figure 5.12: Partial current densities for (a) CO, and (b) H2 as well
as individual electrode plots for electrolyte pH solutions between 0
and 14 flowed at 0.5 mL/min in a flow reactor.
0
20
40
60
80
100
120
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2)
Cathode (vs Ag/AgCl)
pH 14
pH 10
pH 4
pH 2
pH 0
0
1
2
3
4
5
6
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode vs Ag/AgCl (V)
0
20
40
60
80
100
120
140
160
180
-2.0-1.5-1.0-0.50.0
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2)
Cathode (vs Ag/AgCl)
0
1
2
3
4
5
6
-2.0-1.8-1.6-1.4-1.2-1.0
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode (vs Ag/AgCl)
-2.0
-1.5
-1.0
-0.5
0.0
0.5
1.0
1.5
2.0
2.5
0 25 50 75 100 125
Cat
ho
de
Po
ten
tial
(vs
Ag/
AgC
l)
Current Density (mA/cm2)
(a)
(b)
(c)
87
continues as the electrolyte solution
becomes more acidic (i.e., pH 2 and
pH 0).
In Figure 5.12b, the pH can be
seen to drastically effects H2 evolution
on the cathode, and consequently,
change the selectivity for CO to H2. In
acidic conditions, H2 evolution
drastically increases because the cathode
is flooded with excess protons, which
drive the evolution of H2 as opposed to the reduction of CO2 to CO. This is confirmed by the
decrease in cathode overpotential under acidic conditions as seen in Figure 5.12c. Furthemore,
neutral conditions experience “pH sheltering” as the electrode surface pH will be sheltered from
the electrolyte solution pH because the anode and cathode reactions produce and consume
protons [35-37].
The anode is also influenced by pH as shown in Figure 5.12c. Specifically, the anode
overpotential is decreased in basic conditions. Consequently, basic conditions are the most
beneficial for CO2 reduction because smallest overall cell potential is required.
5.3.7 Copper Organometallic Complexes
Figure 5.13 shows the cathode polarization curves for four copper-based organometallic
catalysts compared to a copper catalyst, all on electrodes with a 1 mg/cm2 catalyst loading in an
electrochemical flow reactor with a 1 M KOH electrolyte flowed at 0.5 mL/min and a CO2
Figure 5.13: Cathode polarization curves for a copper and four
copper-based organometallic catalysts on electrodes with a
1 mg/cm2 catalyst loading in an electrochemical flow reactor
with a 1 M KOH electrolyte flowed at 0.5 mL/min and a CO2
reactant stream flowed behind the cathode at 5 SCCM.
-2.5
-2.3
-2.1
-1.9
-1.7
-1.5
-1.3
-1.1
-0.9
0 25 50 75 100 125
Cat
ho
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
301 21
40 289
Cu
-2.5
-2.3
-2.1
-1.9
-1.7
-1.5
-1.3
-1.1
-0.9
0 25 50 75 100 125
Cat
ho
de
Po
ten
tial
(V v
s A
g/A
gCl)
Current Density (mA/cm2)
301 21
40 289
Cu
88
reactant stream flowed behind the cathode at 5 SCCM. The overpotentials for the
organometallic complexes are much higher than that for the lone copper catalyst.
Figure 5.14 shows the product distribution for the same experiment as Figure 5.13. The
partial current densities for H2, formic acid, CO, and CH4 are much lower with the
organometallic complexes as compared to the copper electrode. Furthermore, the onset
potentials for many of the carbon products are delayed. Specifically, the onset potential for
methane is much later with the organometallic complexes as opposed to the copper electrode.
This same trend is seen with carbon monoxide and formic acid, but to a lesser extent. Because of
the later onset potential for the carbon products, the faradaic efficiency for H2, generally the
Figure 5.14: Product selectivity for (a) H2, (b) formic acid, (c) CO, and (d) CH4 for a copper and four copper-based
organometallic catalysts on electrodes with a 1 mg/cm2 catalyst loading in an electrochemical flow reactor.
0
10
20
30
40
50
60
70
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
10
20
30
40
50
60
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
5
10
15
20
25
30
35
40
45
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty F
A (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
1
2
3
4
5
6
7
8
9
10
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
H4
(mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
(a)
(d)(c)
(b)
0
10
20
30
40
50
60
70
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
10
20
30
40
50
60
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
5
10
15
20
25
30
35
40
45
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty F
A (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
1
2
3
4
5
6
7
8
9
10
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
H4
(mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
(a)
(d)(c)
(b)
0
10
20
30
40
50
60
70
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
10
20
30
40
50
60
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
5
10
15
20
25
30
35
40
45
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty F
A (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
1
2
3
4
5
6
7
8
9
10
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
H4
(mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
(a)
(d)(c)
(b)
0
10
20
30
40
50
60
70
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
10
20
30
40
50
60
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
5
10
15
20
25
30
35
40
45
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty F
A (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
1
2
3
4
5
6
7
8
9
10
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
H4
(mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
(a)
(d)(c)
(b)
0
10
20
30
40
50
60
70
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
10
20
30
40
50
60
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
5
10
15
20
25
30
35
40
45
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty F
A (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
1
2
3
4
5
6
7
8
9
10
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
H4
(mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
(a)
(d)(c)
(b)
0
10
20
30
40
50
60
70
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
10
20
30
40
50
60
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
5
10
15
20
25
30
35
40
45
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty F
A (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
1
2
3
4
5
6
7
8
9
10
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
H4
(mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
(a)
(d)(c)
(b)
0
10
20
30
40
50
60
70
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty H
2(m
A/c
m2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
10
20
30
40
50
60
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
O (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
5
10
15
20
25
30
35
40
45
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty F
A (
mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
0
1
2
3
4
5
6
7
8
9
10
-2.3-2.1-1.9-1.7-1.5-1.3-1.1
Par
tial
Cu
rre
nt
De
nsi
ty C
H4
(mA
/cm
2 )
Cathode vs Ag/AgCl (V)
301
21
40
289
Cu
(a)
(d)(c)
(b)
89
undesired product, is higher with the organometallic complexes. While the organometallic
complexes negatively affect the reaction, specifically by exhibiting a lower current density, a
later onset for carbon products, and a higher selectivity for H2, our observation indicates that
organometallic catalysts can potentially be “tuned” with respect to product selectivity.
5.4 Conclusions
Amines have previously been shown to influence both product selectivity and kinetics for
the electrochemical reduction of CO2 to CO via incorporating amines in the electrolyte solution.
Here, we report that development of amine based novel organometallic complexes for the
electrochemical reduction of CO2 to CO, which reduce the onset potential by ~0.1V and increase
the current densities as compared to a more traditional catalyst, Ag. Specifically, AgDAT,
AgPc, and AgPz have comparable or better current densities as compared to Ag with increased
selectivity for CO. Furthermore, when comparing the performance relative to the silver loading,
the amine based organometallic catalysts outperform the silver catalyst by roughly 20 fold.
A major obstacle to the broader application of the electrochemical reduction of CO2 lies
in the challenge of simultaneously achieving high current densities and energetic efficiencies.
Here we demonstrate a major step towards improving both the energetic efficiencies and current
densities of the electrochemical reduction of CO2 to CO on a Ag cathode. In particular, we
observe a drastic improvement in the current densities achieved on a per-metal loading.
Furthermore, we demonstrate the potential to influence the selectivity of other catalysts.
Specifically, CuDAT and other amines influence the current density and product selectivity for a
copper catalyst, which makes a broad range of products. This work demonstrates that
organometallic complexes can be used to “tune” catalyst selection for the desired products.
90
The work reported here suggest that further exploration of the effect of amine-based
catalysis on the electrochemical conversion of CO2 to CO, or to other products, may lead to
further improvements in performance.
5.5 References
1. Brian A. Rosen, Amin Salehi-Khojin, Michael R. Thorson, Wei Zhu, Devin T. Whipple,
Paul J. A. Kenis and Richard I. Masel, Ionic Liquid–Mediated Selective Conversion of
CO2 to CO at Low Overpotentials, Science, 2011, 334 (6056), p. 643-644.
2. D. T. Whipple and P. J. A. Kenis, Prospects of CO2 Utilization via Direct Heterogeneous
Electrochemical Reduction, Journal of Physical Chemistry Letters, 2010, 1 (24), p. 3451-
3458.
3. Y. Hori, Electrochemical CO2 Reduction on Metal Electrodes, Modern Aspects of
Electrochemistry, 2008, 42 p. 89-189.
4. M. R. Dubois and D. L. Dubois, Development of Molecular Electrocatalysts for CO2
Reduction and H2 Production/Oxidation, Accounts of Chemical Research, 2009, 42 (12),
p. 1974-1982.
5. Narasi Sridhar, Carbon Dioxide Utilization: Electrochemical Conversion of CO2 -
Opportunities and Challenges. Det Norske Veritas, 2011. 07 1-20.
6. A. T. Bell, Basic Energy Needs: Catalysis for Energy. Department of Energy, 2008.
PNNL17214
7. G. Centi and S. Perathoner, Opportunities and Prospects in the Chemical Recycling of
Carbon Dioxide to Fuels, Catalysis Today, 2009, 148 (3-4), p. 191-205.
8. C. S. Song, Global Challenges and Strategies for Control, Conversion and Utilization of
CO2 for Sustainable Development Involving Energy, Catalysis, Adsorption and Chemical
Processing, Catalysis Today, 2006, 115 (1-4), p. 2-32.
9. Yoshio Hori, Hidetoshi Wakebe, Toshio Tsukamoto and Osamu Koga, Electrocatalytic
process of CO selectivity in electrochemical reduction of CO2 at metal electrodes in
aqueous media, Electrochimica Acta, 1994, 39 (11–12), p. 1833-1839.
10. Masashi Azuma, Kazuhito Hashimoto, Masahiro Hiramoto, Masahiro Watanabe and
Tadayoshi Sakata, Electrochemical Reduction of Carbon Dioxide on Various Metal
Electrodes in Low-Temperature Aqueous KHCO[sub 3] Media, Journal of the
Electrochemical Society, 1990, 137 (6), p. 1772-1778.
11. Nagakazu Furuya, Tsuyoshi Yamazaki and Masami Shibata, High performance RuPd
catalysts for CO2 reduction at gas-diffusion electrodes, Journal of Electroanalytical
Chemistry, 1997, 431 (1), p. 39-41.
12. Matthew S Thorum, Jessica Yadav and Andrew A Gewirth, Oxygen Reduction Activity of
a Copper Complex of 3,5-Diamino-1,2,4-triazole Supported on Carbon Black,
Angewandte Chemie International Edition, 2009, 48 (1), p. 165-167.
13. Fikile R. Brushett, Matthew S. Thorum, Nicholas S. Lioutas, Matthew S. Naughton,
Claire Tornow, Huei-Ru “Molly” Jhong, Andrew A. Gewirth and Paul J. A. Kenis, A
Carbon-Supported Copper Complex of 3,5-Diamino-1,2,4-triazole as a Cathode Catalyst
91
for Alkaline Fuel Cell Applications, Journal of the American Chemical Society, 2010,
132 (35), p. 12185-12187.
14. J. O. Bockris and J. C. Wass, The Photocatalytic Reduction of Carbon-Dioxide, Journal
of the Electrochemical Society, 1989, 136 (9), p. 2521-2528.
15. J.O.M. Bockris and A.K.N. Reddy, Modern electrochemistry, Plenum Press, 1998.
16. K. Chandrasekaran and L. O'M Bockris, In-situ spectroscopic investigation of adsorbed
intermediate radicals in electrochemical reactions: CO2− on platinum, Surface Science,
1987, 185 (3), p. 495-514.
17. E. B. Cole, P. S. Lakkaraju, D. M. Rampulla, A. J. Morris, E. Abelev and A. B. Bocarsly,
Using a One-Electron Shuttle for the Multielectron Reduction of CO2 to Methanol:
Kinetic, Mechanistic, and Structural Insights, Journal of the American Chemical Society,
2010, 132 (33), p. 11539-11551.
18. A. J. Morris, R. T. McGibbon and A. B. Bocarsly, Electrocatalytic Carbon Dioxide
Activation: The Rate-Determining Step of Pyridinium-Catalyzed CO2 Reduction,
Chemsuschem, 2011, 4 (2), p. 191-196.
19. G. Seshadri, C. Lin and A. B. Bocarsly, A New Homogeneous Electrocatalyst for the
Reduction of Carbon-Dioxide to Methanol at Low Overpotential, Journal of
Electroanalytical Chemistry, 1994, 372 (1-2), p. 145-150.
20. C. Delacourt, P. L. Ridgway, J. B. Kerr and J. Newman, Design of an Electrochemical
Cell Making Syngas (CO+H2) from CO2 and H2O Reduction at Room Temperature,
Journal of the Electrochemical Society, 2008, 155 (1), p. B42-B49.
21. S. Kaneco, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of CO2 to
Methane at the Cu Electrode in Methanol with Sodium Supporting Salts and its
Comparison with Other Alkaline Salts, Energy Fuels, 2006, 20 (1), p. 409-414.
22. R. L. Cook, R. C. Macduff and A. F. Sammells, High-Rate Gas-Phase CO2 Reduction to
Ethylene and Methane Using Gas-Diffusion Electrodes, Journal of the Electrochemical
Society, 1990, 137 (2), p. 607-608.
23. N. Furuya and S. Koide, Electroreduction of Carbon-Dioxide by Metal Phthalocyanines,
Electrochimica Acta, 1991, 36 (8), p. 1309-1313.
24. T. Yamamoto, D. A. Tryk, A. Fujishima and H. Ohata, Production of Syngas Plus
Oxygen from CO2 in a Gas-Diffusion Electrode-Based Electrolytic Cell, Electrochimica
Acta, 2002, 47 (20), p. 3327-3334.
25. S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of
High Pressure CO2 at a Cu Electrode in Cold Methanol, Electrochimica Acta, 2006, 51
(23), p. 4880-4885.
26. S. Kaneco, K. Iiba, H. Katsumata, T. Suzuki and K. Ohta, Electrochemical Reduction of
High Pressure Carbon Dioxide at a Cu Electrode in Cold Methanol with CsOH
Supporting Salt, Chemical Engineering Journal, 2007, 128 (1), p. 47-50.
27. S. Kaneco, R. Iwao, K. Iiba, K. Ohta and T. Mizuno, Electrochemical Conversion of
Carbon Dioxide to Formic Acid on Pb in KOH/Methanol Electrolyte at Ambient
Temperature and Pressure, Energy, 1998, 23 (12), p. 1107-1112.
92
28. Kendra P. Kuhl, Etosha R. Cave, David N. Abram and Thomas F. Jaramillo, New insights
into the electrochemical reduction of carbon dioxide on metallic copper surfaces, Energy
& Environmental Science, 2012, p. in press.
29. William J. Durand, Andrew A. Peterson, Felix Studt, Frank Abild-Pedersen and Jens K.
Nørskov, Structure effects on the energetics of the electrochemical reduction of CO2 by
copper surfaces, Surface Science, 2011, 605 (15–16), p. 1354-1359.
30. Andrew A. Peterson, Frank Abild-Pedersen, Felix Studt, Jan Rossmeisl and Jens K.
Norskov, How copper catalyzes the electroreduction of carbon dioxide into hydrocarbon
fuels, Energy & Environmental Science, 2010, 3 (9), p. 1311-1315.
31. Andrew A. Peterson and Jens K. Nørskov, Activity Descriptors for CO2 Electroreduction
to Methane on Transition-Metal Catalysts, The Journal of Physical Chemistry Letters,
2012, 3 (2), p. 251-258.
32. W. Tang, A. A. Peterson, A. S. Varela, Z. P. Jovanov, L. Bech, W. J. Durand, S. Dahl, J.
K. Norskov and I. Chorkendorff, The importance of surface morphology in controlling
the selectivity of polycrystalline copper for CO2 electroreduction, Physical chemistry
chemical physics : PCCP, 2012, 14 (1), p. 76-81.
33. Norberto Masciocchi, Massimo Moret, Paolo Cairati, Angelo Sironi, G. Attilio Ardizzoia
and Girolamo La Monica, The Multiphase Nature of the Cu(pz) and Ag(pz) (Hpz =
Pyrazole) Systems: Selective Syntheses and Ab-Initio X-ray Powder Diffraction
Structural Characterization of Copper(I) and Silver(I) Pyrazolates, Journal of the
American Chemical Society, 1994, 116 (17), p. 7668-7676.
34. R. Sleat and R. A. Mah, Quantitative Method for Colorimetric Determination of Formate
in Fermentation Media, Applied and Environmental Microbiology, 1984, 47 (4), p. 884-
885.
35. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Microfluidic Reactor for the
Electrochemical Reduction of Carbon Dioxide: The Effect of pH, Electrochemical and
Solid State Letters, 2010, 13 (9), p. D109-D111.
36. H. Li and C. Oloman, Development of a Continuous Reactor for the Electro-reduction of
Carbon Dioxide to Formate - Part 2: Scale-up, Journal of Applied Electrochemistry,
2007, 37 (10), p. 1107-1117.
37. M. N. Mahmood, D. Masheder and C. J. Harty, Use of Gas-Diffusion Electrodes for
High-Rate Electrochemical Reduction of Carbon-Dioxide .1. Reduction at Lead, Indium-
Impregnated and Tin-Impregnated Electrodes, Journal of Applied Electrochemistry,
1987, 17 (6), p. 1159-1170.
93
Chapter 6
Design Rules for Electrode Arrangement in an
Air-breathing Alkaline Direct Methanol LFFC1
6.1 Introduction
An ever increasing demand for high power density, portable electronic devices (e.g.,
smartphones, laptops) with a fast recharge has motivated the development of many micro-scale
fuel cells [1-5]. Small-scale fuel cells demonstrate superior energy densities compared to
rechargeable batteries; offering smaller, and lighter alternatives for applications requiring
portable power sources [6]. Specifically, developments in both novel fuel cell catalysts and
electrode assemblies, and advances in fabrication have enabled the miniaturization of fuel cells
capable of integration in small portable applications [2-4,7,8]. Furthermore, liquid organic feed
sources, such as methanol, have a distinct advantage compared to gaseous hydrogen feed sources
as the fuel can be stored safely at low pressure, in an energy dense form. In comparison to
traditional batteries, the liquid nature of methanol also enables nearly instantaneous recharging
via refilling or changing-out of a fuel cartridge. Additionally, the use of a membraneless laminar
flow fuel cell (LFFC) enables the use of alkaline media, avoiding the membrane-associated
issues typically encountered in alkaline fuel cells. Furthermore, in alkaline media, the reaction
kinetics are better at both the anode [9] and the cathode [10], and high performance can be
achieved with abundant and cheap catalysts such as Ag [11-15].
LFFCs have several further attractive characteristics. The flowing liquid electrolyte
stream mitigates water management issues and enables fuel flexibility [16]. Alternative fuel feed
1 Part of this work has been published: Thorson, M. R, Brushett, F. B., Kenis, P. J. A., “Design Rules for Electrode
Arrangement in an Air-breathing Alkaline Direct Methanol LFFC.” Journal of Power Sources 2012
94
mechanisms [17,18], as well as modifications to electrode location and method of catalyst
deposition [19-23] have led to improved current densities. Alternative cell designs [24-26] (e.g.,
F, Y, and T channel geometries) and further modifications to LFFC configurations [27] (e.g.,
herringbone mixers, multiple inlets) have improved both fuel conversion and current density,
while helping to mitigate fuel crossover. Recently, nanoporous separators have been employed
to minimize the interfacial area between the fuel and electrolyte streams, thereby drastically
reducing fuel crossover [28].
A major limitation to obtaining higher current densities in present LFFC configurations is
the depletion of fuel along the electrode surface. Several computational studies have reported on
this boundary layer depletion issue. Specifically, using a CFD model of the formic acid reaction
within a membraneless LFFC, Bazylak et al. showed that a ~15 µm thick boundary layer formed
within 1 mm, which drastically increased the mass transport limitations [29]. Yoon, et al.
modeled a similar scenario to examine the electrode performance for formic acid oxidation with
a membraneless LFFC [27] in which he showed that depletion quickly becomes a major
limitation within the fast forming boundary layer and that the current density will decrease by a
factor of 4 within the first centimeter of the electrode length. Additionally, Khabbazi et al.
modeled the influence of electrode length (8 to 30 mm) on electrode performance and showed
that the power density could be increased by ~80 % by reducing the electrode length [30]. These
aforementioned models all assume well developed flow profiles with smooth electrodes. Actual
electrode surfaces will have inherent roughness, which will influence the diffusion of fuel to
reactive sites on the electrode. Moreover, the local geometries and the orientation of the inlets
for the electrolyte and fuel streams will affect the formation of the boundary layer on the fuel
stream and thereby, change the influence of electrode length on performance.
95
Here, we experimentally studied the influence of electrode length on performance in an
air-breathing alkaline direct methanol LFFC with the aim of formulating design rules for
multichannel LFFCs with optimum specific energy. The effects of aspect ratio (length-to-width)
on electrode performance, and thus overall fuel cell performance, were investigated.
Furthermore, similar experiments were used to develop an empirical model, which predicts the
optimal electrode configurations for a multichannel LFFC.
6.2 Experimental
6.2.1 Fuel Cell Assembly
Fig. 6.1 shows the LFFC we used in this study, a configuration that is similar to those we
have reported on previously [28,31]. Instead of graphite, we used two stainless steel backing
plates (6 x 3 cm2) as current collectors. We chose to use stainless steel plates backing plates to
improve the structural strength (prevent cracking of the graphite), reduce contact resistance
between the wires from the potentiostat and the backing plate, and improve the pressure
distribution of the bolts. In agreement with prior work [32], we did not observe corrosion of the
steel backing plates of our setup. Wires with banana clips plugged directly into precision-
machined 3/16” holes in the anode and cathode current collectors connected the cell with an in-
house built load box. The cathode current collector had a precision-machined 3-mm wide by 40-
mm long window that enabled diffusion of oxygen from air to the cathode for an air-breathing
configuration. Two 1-mm thick poly(methyl methacrylate) (PMMA) sheets with precision
machined 3-mm wide by 50-mm long windows served as the the channels for the electrolyte
flow field within the fuel cell. We used Teflon sheets with precision cut windows (3-mm wide
and 2.25, 4.5, 9, 18, or 36-mm long), placed between the electrode and the flow channel, to vary
the area of the cathode and anode exposed to air and fuel, respectively. A polycarbonate track-
96
etch membrane (6-µm thick, 0.05-µm pore size, PCT0059030 Sterlitech Corporation) was placed
between the two PMMA flow channels to reduce the liquid-liquid interfacial area thereby
reducing the effects of fuel crossover [28]. The cell was held together by four insulated bolts
(McMaster Carr).
6.2.2 Electrode Preparation
The anode was prepared via painting a 1:1 (V/V) H2O:2-propanol solution containing
dispersed catalyst onto the hydrophobized side of a sheet of Toray carbon paper (TGP-H-120,
Fuel Cell Earth) such that the final catalyst loading was 2 mg cm-2
of unsupported Pt/Ru alloy
catalyst (50:50 wt %, Alfa Aesar) and 0.1 mg cm-2
of Nafion binder (LIQUION LG-1105, Ion
Power). Similarly, the anode was prepared via painting a catalyst solution on Toray carbon
paper resulting in a catalyst loading of 2 mg cm-2
of Pt/C catalyst (50 % mass on Vulcan carbon,
E-Tek) and 0.1 mg cm-2
of Nafion binder. More details on this catalyst deposition procedure
have been reported previously [33].
6.2.3 Fuel Cell Testing
The fuel cell assembly was tested under alkaline conditions using an anolyte comprised
of fuel + 1 M potassium hydroxide (KOH, Sigma Aldrich), and a catholyte comprised of 1 M
KOH. The KOH electrolyte solutions are not expected to damage the gas diffusion electrodes
via carbonate poisoning because the application of a flowing electrolyte removes carbonates
from the electrolyte system [34].
Four methanol (Sigma Aldrich) concentrations (0.375, 0.75, 1.5, and 3.0 M) were used in
the anolyte. The flow rates of the fuel and electrolyte streams within the fuel cell were regulated
with a syringe pump (Harvard Instruments, PHD 2200). Fuel cell measurements were conducted
at 0.05 V intervals using an in-house built load box. The current was measured once steady-state
97
had been achieved at a given potential. The geometric surface areas used to calculate the
reported current and power densities were equal to the exposed electrode areas (0.0625, 0.125,
0.25, 0.5, and 1 cm2 for the 3-mm wide and 2.25, 4.5, 9, 18, 36-mm long Teflon windows).
After flowing through the reactor, the electrolyte stream exited the fuel cell through a plastic tube
(Cole Palmer, i.d. = 1.57 mm) and was collected in a glass beaker. As shown in previous work,
no significant potential drop occurs between the fuel cell and the collection beaker [35].
Individual electrode potentials were measured using multimeters (Fluke 8 III) relative to an
Ag/AgCl reference electrode (saturated NaCl, BAS, West Lafayette, IN) in the collection beaker
for the fuel stream. Fuel cell measurements were conducted at 0.05 V intervals using an in-
house built load box. The current was measured once steady-state had been reached for a given
potential, which typically took between 3 and 5 minutes. The geometric surface areas used to
calculate the reported current and power densities were equal to the exposed electrode areas
(0.0625, 0.125, 0.25, 0.5, and 1 cm2 for the 3-mm wide and 2.25, 4.5, 9, 18, 36-mm long Teflon
windows). All conversions are calculated based on the current recorded by the potentiostat, and
consequently, do not include Faradaic losses due to fuel crossover in the conversion calculation.
6.2.4 Modeling
Based on the fuel conversion in the presence of various length electrodes in a single
LFFC unit, we modeled the anticipated conversion for a fuel cell stack comprised of a series of
electrodes. Several assumptions went into the development of this model. First, performance
losses associated with fuel crossover were not take into account because the use of a separator
drastically reduces the crossover effect as we previously reported [28]. Second, we assumed
that, for the conditions tested, the current density of the fuel cell can be interpolated in a linear
fashion relative to the inlet fuel concentration (see Appendix A). Third, we assumed that fuel
98
streams between consecutive cells had the opportunity to restore any concentration depletion
layers generated on the electrodes in the previous cell. In other words, a well-mixed fuel stream,
but of lower fuel concentration, will enter each successive cell.
For this model, a full factorial experimental design was developed in which we varied the
methanol concentration of the fuel stream (0.75, 1.5, and 3.0 M), the flow rates of the methanol
and electrolyte streams (0.1, 0.2, 0.4, and 0.8 ml min-1
), and the length of the exposed electrode
(9, 18, and 36 mm). In each experiment, the KOH concentration in the electrolyte and fuel
streams was 1 M. For each condition, a polarization curve was obtained to identify a power
output for a given current density. For each electrode length and flow rate tested, the measured
current was plotted as a function of the three methanol concentration. Then, after interpolation
over the regions 0.75 to 1.5 M and 1.5 to 3.0 M, the slope m of these interpolated lines is
calculated using the following equation:
(6.1)
where i is the current density at respectively the high and low methanol concentration, c, of each
region (so 3.0 and 1.5 M, or 1.5 and 0.75 M). Now equations to calculate an estimate current
density (iest) for any inlet feed concentration (cinlet) in each of the two regions can be obtained:
0.75 < [MeOH] < 1.5 (6.2a)
1.5 < [MeOH] < 3.0 (6.2b)
Using the calculated values for iest, the outlet concentration (coutlet) for an individual cell
can be calculated using:
(6.3)
where F is Faraday’s constant, and Q is the volumetric flow rate of the methanol stream.
These aforementioned equations were used in a repetitive fashion to simulate several
electrodes in series until coutlet, was 0.75 M, which corresponds to a conversion of 75 %.
99
Specifically, in the empirical model, the inlet methanol concentration, cinlet, for the first electrode
was 3 M. The current from the first electrode, iact, was estimated based on calculations using Eq.
(1) and Eq. (2). The outlet concentration, coutlet, for the first electrode was calculated using Eq.
(3). This process was repeated for additional electrodes with the inlet concentrations for the
subsequent electrodes being defined by the outlet concentration of the previous electrode. Once
the final outlet concentration, coutlet, was below 0.75 M, the total electrode surface area was
calculated by multiplying the number of electrodes required for the given conversion with the
surface area of the given electrode (e.g., 16 electrodes × 1.08 cm2 per electrode = 17.3 cm
2).
6.3 Results and Discussion
6.3.1 Influence of Electrode Length on Performance
The influence of electrode length on power and current output was investigated by
Figure 6.1: Schematic of an alkaline, air-breathing, direct methanol, laminar flow fuel cell (LFFC). Two 1-mm thick
PMMA windows provide the flow fields for the electrolyte and fuel streams between the current collectors and the
nanoporous separator.
MeOH
Electrolyte
Air
Track-Etched Membrane
(porous filter to minimize crossover)
Electrolyte
MeOH
Cathode
Anode
Air Air
Methanol to
Reference Electrode
Electrolyte
Cathode
Air Air Air
Electrolyte
Methanol
Anode
Track-Etched Membrane
100
changing the exposed electrode length (2.25, 4.5, 9, and 18 mm) of both the anode and cathode
in a direct methanol LFFC (Fig. 6.1). Peak power outputs of 7.5, 4.3, 3.1, and 1.7 mW at current
outputs of 24, 19.2, 14.4, and 8.1 mA were observed for 18, 9, 4.5, and 2.25-mm length
electrodes, respectively (Fig. 6.2a). These results corresponded to peak power, and max current
densities of 13.9, 16.1, 23.1, and 24.9 mW·cm-2
and 45, 71, 107, and 121 mA·cm-2
, respectively
for 18, 9, 4.5, 2.25-mm length electrodes (Fig. 6.2b). While the power output consistently drops
with electrode length, the first eighth (2.25 mm) and quarter (4.5 mm) of the electrode produces
23 % and 41 % of the total power output obtained when using the full 18-mm electrode length,
respectively. Both the maximum current and power densities at electrode lengths of 4.5 and
2.25 mm are similar to or better than the current and power densities typically observed for
Nafion-based DMFCs and miniaturized conventional DMFCs (100 to 150 mA cm-2
and 15 to
20 mW cm-2
, respectively [2,36-39]).
Fig. 6.2c shows the individual electrode polarization plots for the aforementioned
experiments. In general, both anode and cathode performance per unit area improved with
decreasing electrode length due to the reduced boundary layer effects. The 2.25- and 4.5-mm
length anodes had later onsets of mass transport losses, around 150 mA·cm-2
, as compared to the
18- and 9-mm length anodes, which had onsets of mass transport losses in the range of 50 to
75 mA·cm-2
(Fig. 6.2c). However, little difference was observed between the 4.5- and 2.25-mm
length anodes. This observation suggests that other factors may dominate in the first 4.5 mm of
the electrode length. Specifically, catalyst roughness, typically on the order of 10 to 30 μm [40],
may hinder diffusion more than boundary layer depletion for the first 4.5 mm, given that the
boundary layer thickness is on the order of 15 μm [29]. The roughness within the catalyst layer
may explain why the power density curves for 4.5- and 2.25-mm lengths anodes are similar,
101
whereas simulations predict significantly
higher power density curves for even shorter
electrodes [27,29].
6.3.2 Optimal Aspect Ratio for a Single
Electrode
The experiments presented above show
that shortening electrode length leads to higher
power densities, but the absolute amount of
power generated is reduced, as expected (Fig.
6.2a). This result suggests that using wider
electrodes will be beneficial to achieve both
higher power densities as well as higher
absolute power. Changing the electrode aspect
ratio (i.e., by widening the flow channel),
while keeping the total electrode area and the
volumetric flow rate constant, affects the
development of the boundary layer in two
ways: (1) The lower linear velocity increases
the boundary layer thickness at the end of the
electrode; and (2) shorter electrodes result in a
less developed (thinner) boundary layer.
To systematically study the influence
of electrode aspect ratio, as defined as the ratio
Figure 6.2: (a) Polarization and power curves of an air
breathing LFFC for electrode lengths 18, 9, 4.5, and
2.25 mm. (b) Corresponding polarization and power density
curves and (c) anode and cathode polarization curves
normalized to the electrode area. In all experiments were
performed with [MeOH] = 2 M, [KOH] = 1 M, and a flow
rate for each stream of 0.1 ml min-1
.
(a)
(b)
(c)
102
of the electrode length-to-width, on
both current and power output, we
relied on geometric and dynamic
similarities between the
aforementioned LFFC and various
aspect ratio electrodes and to
empirically model electrodes of
various aspect ratios via experimental
data with the aforementioned LFFC.
So, rather than building several new
cell designs, we equipped this LFFC
with 4.5, 9, and 18 mm electrodes to
experimentally simulate electrode
aspect ratios of 1:2 (4.5 x 9 mm2), 2:1
(9 x 4.5 mm2), and 4:1 (18 x 2.25
mm2). The electrode surface area is
identical for these three cases (0.4
cm2). The flow rate of the electrodes
simulated (0.15 ml min-1
) was
achieved by flowing the fuel streams
at 0.2, 0.1, and 0.05 ml min-1
in the actual experiments performed in cells with 18, 9, and 4.5 mm
electrodes, respectively (Fig. 6.3a), and thus making these experimental conditions equivalent to
the conditions we are actually interested in. The flow channel width was assumed to have a
Figure 6.3: (a) Schematic of the conditions tested and their
correlation to simulated scenarios. (b) Polarization and power
density curves of an air breathing LFFC for electrode lengths 18, 9,
and 4.5 mm operated at fuel and electrolyte flow rates of 0.2, 0.1,
and 0.05 ml min-1
, respectively. In all experiments, [MeOH] = 0.
75 M and [KOH] = 1 M.
Qi=0.05 ml/min
Qi=0.1 ml/min
Qi=0.2 ml/min
1:2
4:1
2:1
Correlation to
Aspect Ratio
Tested
Condition(a)
(b)
l:9 mm / w:3 mm
l:18 mm / w:3 mm
l:4.5 mm / w:3 mml:4.5 mm
w:9 mm
l:9 mm
w:4.5 mm
l:18 mm
w:2.25 mm
103
minimal influence on performance because the flow profile is primarily affected by the distance
between electrodes, which is the critical dimension for the development of the flow profiles,
which allowed us to use a fixed-width channel of 3 mm to simulate electrodes with three
different widths.
Fig. 6.3b shows the normalized cell performance for electrode lengths of 18, 9, and
4.5 mm operated at 0.2, 0.1, and 0.05 ml min-1
to simulate electrode areas of 4.5 x 9 mm2 (1:2), 9
x 4.5 mm2 (2:1), and 18 x 2.25 mm
2 (4:1). Maximum power densities of 24, 15, and 14 mW·cm
-
2 at current densities of 120, 64, and 54 mA·cm
-2 were observed for simulated electrode areas of
4.5 x 9 mm2, 9 x 4.5 mm
2, and 18 x 2.25 mm
2, respectively. The shorter and wider aspect ratios
exhibit much higher current densities and fuel conversion for a given flow rate and electrode
area. Specifically, the electrode performance increased by 70 % with the low aspect ratio (1:2)
configuration compared to a high aspect ratio (4:1) configuration. Most microfluidic LFFCs
reported in the literature [5] use electrodes with high aspect ratios (long and narrow) as they
easily develop uniform flow fields and facilitate high fuel conversion at reasonable flow
velocities. Our data suggests that a low aspect ratio (short and wide) would significantly
improve the performance of a single electrode LFFC. The results obtained here, which show that
shorter and wider electrodes will improve the performance of an air-breathing alkaline fuel cell
configuration, will be applicable to the design of other single cell microfluidic systems for
electrochemical processes, not just fuel cells, with different arrangements and reactants.
6.3.3 Optimal Electrode Length for Multichannel Configurations
Above, we reported how to optimize the performance of a single electrode configuration.
Scaling to achieve the power output needed for specific applications will require the creation of
104
multichannel configurations. Through a combined simulation and experimental approach, we
study how to optimize fuel conversion when using different multi-electrode arrangements.
To model conversion for a series of electrodes, the current density across a single
electrode was measured for various methanol concentrations (0.75, 1.5, and 3.0 M), flow rates
(0.0125, 0.025, 0.05, 0.1, 0.2, 0.4, and 0.8 ml min-1
), and electrode lengths (9, 18, and 36 mm) at
the potential which corresponds to peak power (0.2 V). The main effects and interaction plots
may be found in Appendix A. By interpolation of the experimental data over the concentration
range of 0.75 to 1.5 M and 1.5 to 3.0 M, as explained in Section 2.4 of the Experimental, we can
estimate the current for any inlet concentration (Eq. 2a and 2b), and by using the estimated
current, we can calculate the outlet concentration (Eq. 3). Next, we used the outlet concentration
of the first electrode as the inlet concentration of the second to calculate (again with equations 2
and 3) the outlet concentration of the second electrode, and so on, until the outlet concentration is
only 25 % of the original inlet concentration. So for each of the three electrode lengths, we
determined the number of them needed to accomplish a conversion of 75 %. By adding the area
of the electrodes required for the specified conversion, the total electrode area needed was
calculated for each electrode length and flow rate tested (Table 6.1).
A decrease in the electrode length necessitates more electrodes to accomplish the
aforementioned 75 % conversion. However, as Table 6.1 shows, the corresponding total
electrode surface area decreases. This decrease in normalized electrode area is more pronounced
Table 6.1. Absolute electrode area required for a multiple electrode arrangement for 75 % conversion of a methanol
fuel stream
Electrode
Length (mm) Flow Rate (ml/min)
0.8 0.4 0.2 0.1 0.05 0.025 0.0125
36 228.96 113.4 59.4 31.32 17.28 11.88 8.64
18 184.68 92.34 46.98 23.76 12.42 7.02 4.32
9 187.38 91.53 44.82 23.22 11.61 6.21 3.51
105
at slower flow rates. Specifically, at a flow rate of 0.0125 ml min-1
, the necessary electrode area
for 75 % conversion was reduced by 61 % (3.5 vs. 8.6 cm2) by using thirteen 9-mm electrodes as
compared to eight 36-mm electrodes. In contrast, at a flow rate of 0.8 ml min-1
, the necessary
electrode area was only reduced 18 % (187.4 vs. 229.0 cm2) by using 694 9-mm electrodes as
compared to 212 36-mm electrodes.
Next, we wished to compare these datasets with respect to surface reaction rates to better
enable comparison of the total electrode areas needed to accomplish 75 % conversion at different
flow rates. Figure 4 shows methanol reaction rates normalized to the total electrode surface area
(in mmol/min·mm2) for a 3.0 M methanol fuel stream converted across a series of electrodes
(each 9, 18, or 36-mm long) to 0.75 M (75 % conversion) for each flow rate tested.
Figure 6.4: Reaction rate of methanol within the fuel stack relative to the total electrode area, predicted based on a
model for the use of a fuel cell stack for the conversion of methanol from 3.0 M to 0.75 M with variable length
electrodes (9, 18, and 36 mm) in a stack arrangement.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
0.8 0.4 0.2 0.1 0.05 0.025 0.0125
No
rmal
lize
d r
eac
tio
n r
ate
(m
mo
l/m
in·m
m2)
Flow Rate (ml/min)
9 mm 18 mm 36 mm
106
While the shorter electrodes universally improve the normalized reaction rates, the
improvement is more pronounced for slower flow rates because the boundary layer becomes
thicker. Specifically, at a flow rate of 0.0125 ml min-1
, the resultant normalized reaction rate for
75 % conversion increases by 146 % (1.07 vs. 0.43 mmol/min·mm2) by using 9-mm electrodes
instead of 36-mm electrodes. In contrast, at a higher flow rate of 0.8 ml min-1
, the necessary
electrode area only increases by 22 % (1.28 vs. 1.05 mmol/min·mm2) by using 9-mm electrodes
instead of 36-mm electrodes.
The more pronounced influence of electrode length on conversion at slow flow rates
stems from an increase in reactant depletion due to slower fuel flow rates. This effect can be
explained by applying the Blasius solution for boundary layer thickness (Eq. 4):
(4)
where is the boundary layer thickness as defined by the point where the fluid velocity has
come within 1 % of the ‘free stream’ velocity, , is the kinematic viscosity, and is the
distance down the electrode [40]. This equation shows that the boundary layer becomes thicker
upon decreasing the flow rate. As a result, the cell becomes diffusion limited at slower flow
rates, a limitation that can be overcome by using many short electrodes.
The results here show that using several “short” electrodes in series results in much
higher conversion normalized to the surface area than fewer “longer” electrodes, especially at
slow flow rates. Consequently, applications that require low power for long periods of time, for
which fuel utilization is important, would benefit from stack designs with multiple electrodes. In
contrast, applications requiring large amounts of power for short periods of time need to operate
at higher flow rates and do not benefit as much from series of short electrodes.
107
One practical aspect that we did not discuss above is the need to homogenize the fuel
concentration in the fuel stream before it enters the inlet of the next cell, after emerging from the
outlet of the previous cell with a depleted boundary layer. In the simulations above, we assumed
that a stream of uniform, but lower, fuel concentration would enter each successive cell. To
homogenize the fuel stream between successive cells, one could use one of many microfluidic
mixing methods, e.g., a herringbone mixer [27,41] or multiple inlets (or outlets) along a single
electrode to periodically push away (or remove) the depleted boundary layer.[30] However, the
addition of mixers or multiple inlets (outlets) adds complexity to the system and reduces spatial
density of the stack. Consequently, an optimal design for a given application will need to
balance the desire to minimize electrode length with the need for homogenization of the fuel
stream between successive cells, which adds additional volume and complexity to the stack.
6.4 Conclusions
Here we studied the influence of boundary layer depletion on performance in an alkaline,
air-breathing LFFC as a function of electrode length, electrode arrangement, and flow rates.
Both the current and power densities obtained here for LFFCs are similar to or better than
Nafion-based DMFCs and miniaturized conventional DMFCs [2,36-39]. We confirmed that
higher current densities can be achieved when using shorter electrodes. The power density
increased, respectively, by 65 and 88 % for 4.5-mm and 2.25-mm long electrodes compared to
an 18-mm electrode. While shorter electrodes lead to higher current densities, the absolute
current decreases, suggesting that arranged multiple short electrodes in series can improve
overall performance.
To utilize the increased power densities from shorter electrodes, we explored different
electrode arrangements and aspect ratios. Shorter and wider electrodes exhibit notably higher
108
current densities and conversions for a given flow rate and electrode area. Specifically, a 67 %
increase in performance was observed when switching from a 1:2 (length-to-width) to a 4:1
aspect ratio. So by changing the standard operating aspect ratio, the performance of single
channel electrodes can be improved significantly.
Alternatively, the use of shorter electrodes in series will allow for higher fuel conversion
while maintaining higher current densities. We observed that the improvement in reaction rate
normalized by surface area as a result of shortening the electrode length is much more
pronounced when operating at slow flow rates. Consequently, arrangements comprised of
shorter electrodes will be most suitable for applications that require low power for long periods
of time, for which fuel utilization is important. However, in arrangements comprised of multiple
electrodes, the fuel stream (with depleted boundary layer) must be homogenized between
consecutive cells, which will add additional volume to a stack. The analysis reported here for
LFFCs can also be applied in the development and optimization of other microfluidic
electrochemical reactor designs [42-44].
6.5 References
1. L. Carrette, K. A. Friedrich and U. Stimming, Fuel Cells - Fundamentals and
Applications, ChemPhysChem, 2001, 1 p. 5-39.
2. A. Kundu, J. H. Jang, J. H. Gil, C. R. Jung, H. R. Lee, S. H. Kim, B. Ku and Y. S. Oh,
Micro-fuel cells-Current development and applications, Journal of Power Sources, 2007,
170 (1), p. 67-78.
3. S. Pennathur, J. C. T. Eijkel and A. Van Den Berg, Energy conversion in microsystems:
Is there a role for micro/nanofluidics?, Lab on a Chip - Miniaturisation for Chemistry
and Biology, 2007, 7 (10), p. 1234-1237.
4. C. K. Dyer, Fuel cells for portable applications, Journal of Power Sources, 2002, 106 (1-
2), p. 31-34.
5. E. Kjeang, N. Djilali and D. Sinton, Microfluidic fuel cells: A review, Journal of Power
Sources, 2009, 186 (2), p. 353-369.
6. C. Potera, Beyond batteries: Portable hydrogen fuel cells, Environmental Health
Perspectives, 2007, 115 (1), p. A38-A41.
109
7. P. O. López-Montesinos, N. Yossakda, A. Schmidt, F. R. Brushett, W. E. Pelton and P. J.
A. Kenis, Design, fabrication, and characterization of a planar, silicon-based,
monolithically integrated micro laminar flow fuel cell with a bridge-shaped
microchannel cross-section, Journal of Power Sources, 2011, 196 (10), p. 4638-4645.
8. N. T. Nguyen and S. H. Chan, Micromachined polymer electrolyte membrane and direct
methanol fuel cells - A review, Journal of Micromechanics and Microengineering, 2006,
16 (4), p. R1-R12.
9. A. V. Tripkovic, K. D. Popovic, B. N. Grgur, B. Blizanac, P. N. Ross and N. M.
Markovic, Methanol electrooxidation on supported Pt and PtRu catalysts in acid and
alkaline solutions, Electrochimica Acta, 2002, 47 (22-23), p. 3707-3714.
10. K. F. Blurton and E. McMullin, A comparison of fuel cell performance in acid and
alkaline electrolyte, Energy Conversion, 1969, 9 (4), p. 141-144.
11. J. R. Varcoe and R. C. T. Slade, Prospects for alkaline anion-exchange membranes in
low temperature fuel cells, Fuel Cells, 2005, 5 (2), p. 187-200.
12. M. Schulze and E. Gülzow, Degradation of nickel anodes in alkaline fuel cells, Journal of
Power Sources, 2004, 127 (1-2), p. 252-263.
13. N. Wagner, M. Schulze and E. Gülzow, Long term investigations of silver cathodes for
alkaline fuel cells, Journal of Power Sources, 2004, 127 (1-2), p. 264-272.
14. Fikile R. Brushett, Matthew S. Thorum, Nicholas S. Lioutas, Matthew S. Naughton,
Claire Tornow, Huei-Ru “Molly” Jhong, Andrew A. Gewirth and Paul J. A. Kenis, A
Carbon-Supported Copper Complex of 3,5-Diamino-1,2,4-triazole as a Cathode Catalyst
for Alkaline Fuel Cell Applications, Journal of the American Chemical Society, 2010,
132 (35), p. 12185-12187.
15. C. Coutanceau, L. Demarconnay, C. Lamy and J. M. Léger, Development of
electrocatalysts for solid alkaline fuel cell (SAFC), Journal of Power Sources, 2006, 156
(1 SPEC. ISS.), p. 14-19.
16. E. R. Choban, L. J. Markoski, A. Wieckowski and P. J. A. Kenis, Microfluidic fuel cell
based on laminar flow, Journal of Power Sources, 2004, 128 (1), p. 54-60.
17. E. Kjeang, R. Michel, D. A. Harrington, D. Sinton and N. Djilali, An alkaline
microfluidic fuel cell based on formate and hypochlorite bleach, Electrochimica Acta,
2008, 54 (2), p. 698-705.
18. Ranga S. Jayashree, Seong Kee Yoon, Fikile R. Brushett, Pedro O. Lopez-Montesinos,
Dilip Natarajan, Larry J. Markoski and Paul J. A. Kenis, On the performance of
membraneless laminar flow-based fuel cells, Journal of Power Sources, 2010, 195 (11),
p. 3569-3578.
19. E. Kjeang, R. Michel, D. A. Harrington, N. Djilali and D. Sinton, A microfluidic fuel cell
with flow-through porous electrodes, Journal of the American Chemical Society, 2008,
130 (12), p. 4000-4006.
20. W. Sung and J. W. Choi, A membraneless microscale fuel cell using non-noble catalysts
in alkaline solution, Journal of Power Sources, 2007, 172 (1), p. 198-208.
21. E. Kjeang, J. McKechnie, D. Sinton and N. Djilali, Planar and three-dimensional
microfluidic fuel cell architectures based on graphite rod electrodes, Journal of Power
Sources, 2007, 168 (2), p. 379-390.
110
22. K. S. Salloum, J. R. Hayes, C. A. Friesen and J. D. Posner, Sequential flow membraneless
microfluidic fuel cell with porous electrodes, Journal of Power Sources, 2008, 180 (1), p.
243-252.
23. Kamil S. Salloum and Jonathan D. Posner, A membraneless microfluidic fuel cell stack,
Journal of Power Sources, 2011, 196 (3), p. 1229-1234.
24. J. L. Cohen, D. A. Westly, A. Pechenik and H. D. Abruna, Fabrication and preliminary
testing of a planar membraneless microchannel fuel cell, Journal of Power Sources,
2005, 139 (1-2), p. 96-105.
25. A. Li, S. H. Chan and N. T. Nguyen, A laser-micromachined polymeric membraneless
fuel cell, Journal of Micromechanics and Microengineering, 2007, 17 (6), p. 1107-1113.
26. M. H. Chang, F. Chen and N. S. Fang, Analysis of membraneless fuel cell using laminar
flow in a Y-shaped microchannel, Journal of Power Sources, 2006, 159 (2), p. 810-816.
27. S. K. Yoon, G. W. Fichtl and P. J. A. Kenis, Active control of the depletion boundary
layers in microfluidic electrochemical reactors, Lab on a Chip, 2006, 6 (12), p. 1516-
1524.
28. A. S. Hollinger, R. J. Maloney, R. S. Jayashree, D. Natarajan, L. J. Markoski and P. J. A.
Kenis, Nanoporous separator and low fuel concentration to minimize crossover in direct
methanol laminar flow fuel cells, Journal of Power Sources, 2010, 195 (11), p. 3523-
3528.
29. A. Bazylak, D. Sinton and N. Djilali, Improved fuel utilization in microfluidic fuel cells:
A computational study, Journal of Power Sources, 2005, 143 (1-2), p. 57-66.
30. A. Ebrahimi Khabbazi, A. J. Richards and M. Hoorfar, Numerical study of the effect of
the channel and electrode geometry on the performance of microfluidic fuel cells, Journal
of Power Sources, 2010, 195 (24), p. 8141-8151.
31. R. S. Jayashree, L. Gancs, E. R. Choban, A. Primak, D. Natarajan, L. J. Markoski and P.
J. A. Kenis, Air-breathing laminar flow-based microfluidic fuel cell, Journal of the
American Chemical Society, 2005, 127 (48), p. 16758-16759.
32. D. P. Davies, P. L. Adcock, M. Turpin and S. J. Rowen, Stainless steel as a bipolar plate
material for solid polymer fuel cells, Journal of Power Sources, 2000, 86 (1–2), p. 237-
242.
33. F. R. Brushett, W. P. Zhou, R. S. Jayashree and P. J. A. Kenis, Alkaline Microfluidic
Hydrogen-Oxygen Fuel Cell as a Cathode Characterization Platform, Journal of the
Electrochemical Society, 2009, 156 (5), p. B565-B571.
34. Matt S. Naughton, Fikile R. Brushett and Paul J. A. Kenis, Carbonate resilience of
flowing electrolyte-based alkaline fuel cells, Journal of Power Sources, 2011, 196 (4), p.
1762-1768.
35. E. R. Choban, P. Waszczuk and P. J. A. Kenis, Characterization of limiting factors in
laminar flow-based membraneless microfuel cells, Electrochemical and Solid State
Letters, 2005, 8 (7), p. A348-A352.
36. S. C. Kelley, G. A. Deluga and W. H. Smyrl, Miniature methanol/air polymer electrolyte
fuel cell, Electrochemical and Solid-State Letters, 2000, 3 (9), p. 407-409.
111
37. T. Shimizu, T. Momma, M. Mohamedi, T. Osaka and S. Sarangapani, Design and
fabrication of pumpless small direct methanol fuel cells for portable applications, Journal
of Power Sources, 2004, 137 (2), p. 277-283.
38. T. J. Yen, N. Fang, X. Zhang, G. Q. Lu and C. Y. Wang, A micro methanol fuel cell
operating at near room temperature, Applied Physics Letters, 2003, 83 (19), p. 4056-
4058.
39. V. Baglio, A. Stassi, E. Modica, V. Antonucci, A. S. Aricò, P. Caracino, O. Ballabio, M.
Colombo and E. Kopnin, Performance comparison of portable direct methanol fuel cell
mini-stacks based on a low-cost fluorine-free polymer electrolyte and Nafion membrane,
Electrochimica Acta, 2010, 55 (20), p. 6022-6027.
40. Alex Bauer, Elod L. Gyenge and Colin W. Oloman, Electrodeposition of Pt-Ru
nanoparticles on fibrous carbon substrates in the presence of nonionic surfactant:
Application for methanol oxidation, Electrochimica Acta, 2006, 51 (25), p. 5356-5364.
41. A. D. Stroock, S. K. W. Dertinger, A. Ajdari, I. Mezic, H. A. Stone and G. M.
Whitesides, Chaotic mixer for microchannels, Science, 2002, 295 (5555), p. 647-651.
42. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Microfluidic Reactor for the
Electrochemical Reduction of Carbon Dioxide: The Effect of pH, Electrochemical and
Solid-State Letters, 2010, 13 (9), p. D109-D111.
43. D. T. Whipple and P. J. A. Kenis, Prospects of CO2 Utilization via Direct Heterogeneous
Electrochemical Reduction, Journal of Physical Chemistry Letters, 2010, 1 (24), p. 3451-
3458.
44. Brian A. Rosen, Amin Salehi-Khojin, Michael R. Thorson, Wei Zhu, Devin T. Whipple,
Paul J. A. Kenis and Richard I. Masel, Ionic Liquid–Mediated Selective Conversion of
CO2 to CO at Low Overpotentials, Science, 2011, 334 (6056), p. 643-644.
112
Chapter 7
Summary of Accomplishments and Future Directions
7.1 Summary of Accomplishments
An electrochemical reactor to reduce CO2 can be used as a means of producing chemical
fuels and feedstocks to store electrical energy from renewable sources such as wind and solar in
chemical form. When combined with a “green” renewable source, this technology provides a
means of (a) reducing dependence on foreign oil via enabling penetration of renewable
technologies into the transportation sector and (b) reducing CO2 emissions by moving away from
fossil fuels.
For large scale implementation of processes for the electrochemical reduction of CO2, the
following need to be improved: (a) energetic efficiency, (b) reaction rate, and (c) product
selectivity. Significant research has focused on catalyst development, reactor operation and
electrolyte optimization, primarily for the production of large hydrocarbons (e.g., CH4, C2H4).
Because the electrochemical reaction to produce CO has much higher energetic efficiencies as
compared to formic acid, methane, ethylene, and other C1, C2, and C3 products, we believe an
electrochemical process for the reduction of CO2 to CO will be most favorable for
commercialization. Consequently, we used a microfluidic platform to study the influence of
electrolytes, reactor conditions and catalysts to improve (1) reaction energetic efficiency,
(2) reaction rate, and (3) reaction selectivity for the electrochemical reduction of CO2 to CO.
In Chapter 2, cation size within electrolyte solutions was observed to play a significant
role in the electrochemical reduction of CO2 into CO. Specifically, small cations (e.g., Li+, and
Na+) favor H2 evolution, and, consequently, have a low Faradaic yield for CO production,
113
whereas large cations (e.g., Rb+, Cs
+) increase CO production and suppress H2 evolution,
resulting in a high Faradaic yield for CO, here up to 96 %. We observed the highest partial
current density of 175 mA/cm2 for CO at a cell potential of 3.0 V in the presence of RbOH.
Furthermore, we observed strong trends between anion choice and cell performance. Anions
influence both the anode and cathode polarization and consequently, the partial current density
for CO. Specifically, our data suggests that small anions reduce the reaction overpotential and
that the presence of OH- (high pH) further improves performance.
In Chapter 3, a microfluidic platform was modified to look at the onset potential for CO
production from the reduction of CO2 with Ionic Liquids. With an 18 mole% EMIM BF4
solution, CO production was achieved at a total reactor potential of only 1.5 V, which constitutes
a maximum cathode overpotential of 0.17 V. This low of an onset potential constitutes a drastic
reduction in the onset potential for CO production in an arrangement that achieves excellent
selectivity for CO. However, the individual electrode polarization plots clearly showed that
EMIM BF4 poisons the Pt anode. This anode poisoning was drastically reduced via the
development of a dual electrolyte reactor, with a Nafion cation-exchange membrane. While the
use of a Nafion membrane eliminated anode poisoning, the EMIM BF4 solution appeared to
poison the membrane as was evident by an increased reactor resistance when using the
membrane, which resulted in low current densities. The reactor developed here enabled the
analysis of the overall reactor onset potential, however, at the expense of low current densities
due to high reactor resistances.
In Chapter 4, drastic improvements were achieved regarding the partial current density
for CO with EMIM salts. Specifically, substituting of the BF4 anion for either an OH or Cl anion
enabled partial current densities for CO greater than 60 mA/cm2 at a reactor potential of 3 V.
114
Furthermore, the substitution of BF4 for OH or Cl eliminated anode poisoning, which stemmed
from the BF4 anion, thus enabling the use of a single channel electrolyte stream. The elimination
of both the membrane reduced the reactor resistance from EMIM+ poisoning the Nafion as
discussed in Chapter 3, which led to the drastic improvements in the current densities.
In Chapter 5, amine-based novel organometallic complexes were developed for the
electrochemical reduction of CO2 to CO, which reduce the onset potential by roughly 0.1 V and
more significantly, increased the partial current densities for CO as compared to a more
traditional catalyst, Ag. Specifically, AgDAT, AgPc, and AgPz have comparable or better
current densities with increased selectivity for CO as compared to Ag. Furthermore, when
comparing the performance relative to the silver loading, the amine-based organometallic
catalysts outperform the silver catalyst by more than an order of magnitude. Furthermore, when
using copper-based organometallic catalysts (i.e., CuDAT), the product selectivity changed.
This resultant change in selectivity and current density demonstrates that organometallic
complexes have potential to “tune” catalysts to produce the desired products. Further
exploration of the effect of amine-based catalysis on the electrochemical conversion of CO2 to
CO, or to other products, may lead to further improvements in performance.
Chapter 6 reported on the influence of boundary layer depletion on performance in an
alkaline, air-breathing LFFC as a function of electrode length, electrode arrangement, and flow
rates. Both the current and power densities obtained here for LFFCs are similar to or better than
Nafion-based DMFCs and miniaturized conventional DMFCs [1-5]. We confirmed that higher
current densities can be achieved when using shorter electrodes. The power density increased,
respectively, by 65 and 88 % for 4.5-mm and 2.25-mm long electrodes compared to an 18-mm
electrode. While shorter electrodes lead to higher current densities, the absolute current
115
decreases, suggesting that arranged multiple short electrodes in series can improve overall
performance. To utilize the increased power densities from shorter electrodes, we explored
different electrode arrangements and aspect ratios. Shorter and wider electrodes exhibit notably
higher current densities and conversions for a given flow rate and electrode area. Specifically, a
67 % increase in performance was observed when switching from a 1:2 (length-to-width) to a 4:1
aspect ratio. So, by changing the standard operating aspect ratio, the performance of single
channel electrodes can be improved significantly. Alternatively, the use of shorter electrodes in
series will allow for higher fuel conversion while maintaining higher current densities. We
observed that the improvement in reaction rate normalized by surface area as a result of
shortening the electrode length is much more pronounced when operating at slow flow rates.
Consequently, arrangements comprised of shorter electrodes will be most suitable for
applications that require low power for long periods of time, for which fuel utilization is
important. However, in arrangements comprised of multiple electrodes, the fuel stream (with
depleted boundary layer) must be homogenized between consecutive cells, which will add
additional volume to a stack. The analysis reported here for LFFCs can also be applied in the
development and optimization of other microfluidic electrochemical reactor designs [6-8].
7.2 Future Directions
Numerous areas of opportunity exist within the field of electrochemical reduction of
CO2. The following potential directions of research appear to hold much promise:
The development of a wide range of organometallic catalysts could open the door to
further stabilizing the CO2- intermediate, which has an overpotential of -1.9 V vs. SHE, and
thereby make the electrochemical reduction process much more efficient from an energetic
standpoint. Furthermore, we showed in the case of silver modified organometallic catalysts, that
116
organometallic catalysts can increase both current density and product selectivity. Potentially
more interesting is that organometallic catalysts can change product distributions and onset
potentials. Organometallic complexes may be able to change the intermediate adsorbate-surface
bond strengths, and thus change the catalytic activity of other catalysts. The added stabilizing
effect of organometallic complexes could shift volcano curves from Sachtler-Fahrenfort and
Tranaka-Tamaru plot and thereby, enable less expense or more active catalysts.
Tremendous improvements in catalytic activity can potentially be achieved by improving
catalyst utilization. For common fuel cell catalysts (e.g., Pt), carbon supports have been shown
to drastically improve performance. However, in Chapter 5 a carbon support was shown to
decrease the CO production and favor hydrogen evolution. Other catalysts selective for CO, i.e.,
Au and Zn, should be explored with carbon supports. Furthermore, the use of other supports
such as carbon nanotubes should be explored.
While some long term durability experiments were carried out in this study, further
analysis of catalyst stability is necessary at reasonable overpotentials, with a wide range of
electrolytes and conditions.
A commercial process must consider the cost of CO2 capture in an economic feasibility
analysis. Coupling this work with a capture technology (i.e., Rectisol process) could enable
energy savings. Specifically, by combining this technology with a Rectisol process, a methanol
stream could be used to capture CO2 from an emission source and fed directly into an
electrochemical reactor. Using a liquid with a high solubility for CO2 has the potential to
increase the selectivity for CO by increasing the local concentration of CO2 (the reactant) in the
liquid at the electrode interface.
117
Most of the catalysts used in these studies were lone transition metals (i.e., Cu and Ag)
supported on Nafion. Minimal studies have been carried out regarding the catalytic activity of
metallic alloys for CO2 reduction. In the same fashion that organometallic catalysts improved
the reaction activity, metallic alloys have tremendous potential to improve the overpotential,
reaction rate, and selectivity (the three areas needing improvement).
Co-catalysts have been shown (i.e., EMIM BF4) to reduce the onset potential for CO. A
wider range of co-catalysts and ionic liquids should be explored to influence the reaction onset
potential and current density. Specifically, ionic liquids with high conductivities, diffusivities,
and solubilities for CO2 should be selected to minimize mass transfer and resistance limitations.
Overall, the CO2 reduction process needs to be further improved through catalyst
development, optimization of operating conditions, and developing compatibility with CO2
capture technology. Of all these areas, catalysis has the greatest potential to improve the (1)
onset potential, (2) reaction rate, and (3) selectivity. Consequently, a much wider range of
catalysts and co-catalysts need to be explored for their activity for CO2 reduction.
7.3 References
1. S. C. Kelley, G. A. Deluga and W. H. Smyrl, Miniature methanol/air polymer electrolyte
fuel cell, Electrochemical and Solid-State Letters, 2000, 3 (9), p. 407-409.
2. T. Shimizu, T. Momma, M. Mohamedi, T. Osaka and S. Sarangapani, Design and
fabrication of pumpless small direct methanol fuel cells for portable applications, Journal
of Power Sources, 2004, 137 (2), p. 277-283.
3. T. J. Yen, N. Fang, X. Zhang, G. Q. Lu and C. Y. Wang, A micro methanol fuel cell
operating at near room temperature, Applied Physics Letters, 2003, 83 (19), p. 4056-
4058.
4. A. Kundu, J. H. Jang, J. H. Gil, C. R. Jung, H. R. Lee, S. H. Kim, B. Ku and Y. S. Oh,
Micro-fuel cells-Current development and applications, Journal of Power Sources, 2007,
170 (1), p. 67-78.
5. V. Baglio, A. Stassi, E. Modica, V. Antonucci, A. S. Aricò, P. Caracino, O. Ballabio, M.
Colombo and E. Kopnin, Performance comparison of portable direct methanol fuel cell
mini-stacks based on a low-cost fluorine-free polymer electrolyte and Nafion membrane,
Electrochimica Acta, 2010, 55 (20), p. 6022-6027.
118
6. D. T. Whipple, E. C. Finke and P. J. A. Kenis, Microfluidic Reactor for the
Electrochemical Reduction of Carbon Dioxide: The Effect of pH, Electrochemical and
Solid-State Letters, 2010, 13 (9), p. D109-D111.
7. D. T. Whipple and P. J. A. Kenis, Prospects of CO2 Utilization via Direct Heterogeneous
Electrochemical Reduction, Journal of Physical Chemistry Letters, 2010, 1 (24), p. 3451-
3458.
8. Brian A. Rosen, Amin Salehi-Khojin, Michael R. Thorson, Wei Zhu, Devin T. Whipple,
Paul J. A. Kenis and Richard I. Masel, Ionic Liquid–Mediated Selective Conversion of
CO2 to CO at Low Overpotentials, Science, 2011, 334 (6056), p. 643-644.
119
Appendix A
Supplementary Section to Chapter 6
A.1 Linear Interpolation of Current Density relative to Methanol Concentration
To model conversion for a series of electrodes, it is necessary to estimate the current for
intermediate methanol concentrations not tested. The current output for methanol concentrations
between 0.75 and 3.0 M were estimated using linear interpolation (Eq. 6.1 and Eq. 6.2) as
explained in the Experimental section within Chapter 6.
In Figure A1, we show the measured current output at each of the methanol
concentrations and electrode lengths recorded. Furthermore, Figure A1 shows the actual lines
resulting from the interpolation of experimental data (current) over the two concentration regions
(0.75 to 1.5 M and 1.5 to 3.0 M) for each of the seven flow rates tested (0.0125, 0.025, 0.05, 0.1,
0.2, 0.4, 0.8 ml/min) for 9, 18, and 36-mm electrodes. Table A1 shows the calculated values for
the slope, m, from Eq. 6.1 in the range of 0.0125 to 0.8 ml/min interpolated linearly for 9, 18,
and 36-mm electrodes at a cell potential of 0.2 V between [MeOH] of 3.0 M and 1.5 M as well
as 0.75 M and 1.5 M.
120
Figure A1. Measured current for 9, 18, and 36 mm electrodes at 0.2 V with flow rates between 0.0125 and 0.8
ml/min interpolated linearly between inlet [MeOH] of 3.0 M and 0.75 M.
Table A1. Calculated values for the slope, m, in the interpolation (Eq. 6.1) of the values for the current for 9, 18,
and 36 mm electrode at 0.2 V with flow rates between 0.0125 and 0.8 ml/min interpolated linearly between inlet
[MeOH] of 3.0 and 0.75 M.
Electrode
length (mm) Conc. (M)
Flow rate (ml/min)
0.8 0.4 0.2 0.1 0.05 0.025 0.0125
36 1.5 to 3.0 10.23 6.26 6.71 1.50 0.67 2.67 2.67
0.75 to 1.5 -10.39 -0.36 0.00 2.74 5.33 18.67 18.67
18 1.5 to 3.0 3.57 2.25 3.11 2.53 2.27 1.67 0.94
0.75 to 1.5 14.48 17.79 16.09 8.28 8.92 8.30 11.48
9 1.5 to 3.0 3.00 1.84 1.61 1.20 0.14 0.75 0.60
0.75 to 1.5 -3.78 -0.77 -5.47 0.44 2.13 0.84 2.66
A.2 Main Effects and Interactions from Design of Experiment
Figure A2 shows the main effects plots from the design of experiment discussed in the
modeling section of Chapter 6 based on a 90 % confidence influence. These graphs confirm
many of the trends observed and discussed in the main text of this work. An increase in
methanol concentration, as seen in Figure A2a, increases the current output from the cell.
0
10
20
30
40
50
60
70
80
90
100
0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25
Cu
rre
nt
(mA
)
Methanol Concentration (M)
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0
10
20
30
40
50
60
70
80
90
100
0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25
Cu
rre
nt
(mA
)
Methanol Concentration (M)
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0
10
20
30
40
50
60
70
80
90
100
0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25
Cu
rre
nt
(mA
)
Methanol Concentration (M)
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0
10
20
30
40
50
60
70
80
90
100
0.5 0.75 1 1.25 1.5 1.75 2 2.25 2.5 2.75 3 3.25
Cu
rre
nt
(mA
)
Methanol Concentration (M)
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
0.8 ml/min
0.4 ml/min
0.2 ml/min
0.1 ml/min
0.05 ml/min
0.025 ml/min
0.0125 ml/min
9 mm 18 mm 36 mm
121
Similarly, an increase in electrode length increases the total current output, at a diminishing rate
for larger electrode lengths as seen in Figure A2b. As seen in Figure A2c, the flow rate has a
drastic effect only at very low flow rates. At higher flow rates, the change in flow rate has only
a small influence on the current output.
Figure A3 shows the interaction plots for (a) the effects of methanol concentration for all
eight flow rates tested on output current, (b) the effects of methanol concentration for all three
electrode lengths on output current, and (c) the effects of methanol concentration for all eight
flow rates tested with respect to the current output. The data in Figure A3a, which shows the
43
44
45
46
47
48
49
50
51
52
53
0.5 1 1.5 2 2.5 3
Cu
rre
nt (m
A)
Methanol concentration (M)
0
10
20
30
40
50
60
70
80
5 10 15 20 25 30 35 40
Cu
rre
nt (m
A)
Electrode Length (mm)
30
35
40
45
50
55
0 0.2 0.4 0.6 0.8
Cu
rre
nt (m
A)
Flow Rate (mL min-1)
(a) (b)
Figure A2: Schematic of an alkaline, air-
breathing, direct methanol, laminar flow fuel cell
(LFFC). Two 1-mm thick PMMA windows
provide the flow fields for the electrolyte and fuel
streams between the current collectors and the
nanoporous separator.
(c)
122
effect of methanol concentration for the various flow rates, reveals that methanol concentration
has less of an effect at high flow rates compared to low flow rates. This trend can be explained
by the fact that less depletion is observed at the high flow rates. Figure A3b shows that methanol
concentrations are more important for long electrodes than for short electrodes, because
depletion is more significant with long electrodes.
Comparing all three interaction plots shows that electrode length is the most sensitive factor
(Figure A3c). Especially at low flow rates, the effect of depletion on the electrode becomes
more significant, as is evident from a drastic decrease in the improvement when using longer
electrodes.
123
25
30
35
40
45
50
55
60
0.5 1 1.5 2 2.5 3
Cu
rren
t (m
A)
Methanol Concentration (M)
0.8 ml min-1
0.4 ml min-1
0.2 ml min-1
0.1 ml min-1
0.05 ml min-1
0.025 ml min-1
0.0125 ml min-1
20
30
40
50
60
70
80
0.5 1 1.5 2 2.5 3
Cu
rren
t (m
A)
Methanol Concentration (M)
36 mm
18 mm
9 mm
20
30
40
50
60
70
80
90
5 10 15 20 25 30 35 40
Cu
rren
t (m
A)
Electrode Length (mm)
0.8 ml min-1
0.4 ml min-1
0.2 ml min-1
0.1 ml min-1
0.05 ml min-1
0.025 ml min-1
0.0125 ml min-1
Figure A3. Interaction plots for (a) effects of
methanol concentration for all eight flow rates
tested on output current, (b) effects of methanol
concentration for all three electrode lengths on
output current, and (c) effects of methanol
concentration for all eight flow rates tested on
output current.
(a)
(b) (c)